USB Made Simple - Part 1

Index

Part 1 - Introduction to USB

Part 1
Part 2
Part 3
Part 4
Part 5
Part 6
Part 7

This series of articles on USB is being actively expanded. If you find the information useful,
you may wish to come back to this page in the future to check for newly added parts.

Links

General Introduction
The Universal Serial Bus (USB)
is a specification developed by
Compaq, Intel, Microsoft and
NEC, joined later by HewlettPackard, Lucent and Philips.
These companies formed the
USB Implementers Forum, Inc
as a non-profit corporation to
publish the specifications and
organise further development in
USB.

The aim of the USB-IF was to

find a solution to the mixture of
connection methods to the PC,
in use at the time. We had
serial ports, parallel ports,
keyboard and mouse
connections, joystick ports, midi
ports and so on. And none of
these satisfied the basic
requirements of plug-and-play.
Additionally many of these ports
made use of a limited pool of
PC resources, such as
Hardware Interrupts, and DMA
channels.
[Link] (1 of 6) [13.03.2008. 5:14:49 Peca]

Forward

USB Made Simple - Part 1

So the USB was developed as

a new means to connect a large
number of devices to the PC,
and eventually to replace the
'legacy' ports. It was designed
not to require specific Interrupt
or DMA resources, and also to
be 'hot-pluggable'. It was
important that no special userknowledge would be required to
install a new device, and all
devices would be
distinguishable from all other
devices, such that the correct
driver software was always
automatically used.
It may be apparent that, to
make a system which is so userfriendly is going to mean a lot of
work behind the scenes for the
developer.

Data Speeds
The USB specification defines
three data speeds, shown to the
right. These speeds are the
fundamental clocking rates of
the system, and as such do not
represent possible throughput,
which will always be lower as
the result of the protocol
overheads.
Low Speed

This was intended for cheap,

low data rate devices like mice.
The low speed captive cable is
thinner and more flexible than
that required for full and high
speed.
Full Speed

This was originally specified for

all other devices.

[Link] (2 of 6) [13.03.2008. 5:14:49 Peca]

Name

Speed

Low Speed

1.5 Mbit/s

Full Speed

12 Mbit/s

High Speed

480 Mbit/s

USB Made Simple - Part 1

High Speed

The high speed additions to the

specification were introduced in
USB 2.0 as a response to the
higher speed of Firewire.

Specification
The current specification is
'Universal Serial Bus
Specification, Revision 2'. This
can be obtained free of charge
on the USB-IF website. Please
note that this specification
replaces the earlier 1.0 and 1.1
Specifications, which should no
longer be used. The Revision
2.0 specification covers all three
data speeds, and maintains
backwards compatibility. USB
2.0 does NOT mean High
Speed.
Click here for an overview of
the specification.

Architecture
The USB is based on a socalled 'tiered star topology' in
which there is a single host
controller and up to 127 'slave'
devices. The host controller is
connected to a hub, integrated
within the PC, which allows a
number of attachment points
(often loosely referred to as
ports). A further hub may be
plugged into each of these
attachment points, and so on.
However there are limitations
on this expansion.
As stated above a maximum of
127 devices (including hubs)
may be connected. This is
because the address field in a
packet is 7 bits long, and the
address 0 cannot be used as it

[Link] (3 of 6) [13.03.2008. 5:14:49 Peca]

USB Made Simple - Part 1

has special significance. (In

most systems the bus would be
running out of bandwidth, or
other resources, long before the
127 devices was reached.)
A device can be plugged into a
hub, and that hub can be
plugged into another hub and
so on. However the maximum
number of tiers permitted is six.
The length of any cable is
limited to 5 metres. This
limitation is expressed in the
specification in terms of cable
delays etc, but 5 metres can be
taken as the practical
consequence of the
specification. This means that a
device cannot be further than
30 metres from the PC, and
even to achieve that will involve
5 external hubs, of which at
least 2 will need to be selfpowered.
So the USB is intended as a
bus for devices near to the PC.
For applications requiring
distance from the PC, another
form of connection is needed,
such as Ethernet.

Typical 4-port Hub

[Link] (4 of 6) [13.03.2008. 5:14:49 Peca]

USB Made Simple - Part 1

Types of Host Controller

There are three commonly encountered types of
USB host controller, each with its own history
and characteristics.

Host is Master
All communications on this bus
are initiated by the host.
This means, for example, that
there can be no communication
directly between USB devices.

OHCI (Open Host Controller Interface)

Compaq, Microsoft and National
Semiconductors cooperated to produce this
standard host controller specification for USB
1.0 and USB 1.1. It is a more hardware oriented
version than UHCI. Low speed and full speed.
UHCI (Universal Host Controller Interface)

A device cannot initiate a

transfer, but must wait to be
asked to transfer data by the
host. The only exception to this
is when a device has been put
into 'suspend' (a low power
state) by the host then the
device can signal a 'remote
wakeup'.

On-The-Go
An extension to the USB
specification has been defined,
to allow a device to also
become a limited role host. This
specification is known as OnThe-Go. A later part is planned,
to cover this specication in
detail.

[Link] (5 of 6) [13.03.2008. 5:14:49 Peca]

Intel's more software-oriented version of a

controller for USB 1.0 and USB 1.1. Requires a
license from Intel. Low speed and full speed.
EHCI (Extended Host Controller Interface)
When USB 2.0 appeared with its new high
speed functionality, the USB-IF insisted on
there being a single host controller specification,
to keep device development costs down. The
EHCI handles high speed transfers, and hands
off low and full speed transfers to either OHCI
or UHCI companion controllers.

USB Made Simple - Part 2

Index
Part 1
Part 2
Part 3
Part 4
Part 5
Part 6
Part 7
Links

Back

Part 2 - Electrical

Cables
USB cables have been
designed to ensure correct
connections are always made.
By having different connectors
on host and device, it is
impossible to connect two hosts
or two devices together.
Unfortunately it is possible to
buy non-approved cables and
adapters with illegal
combinations of connector.
These may be useful in certain
development situations, but can
lead the unsuspecting user to
make connections which can
easily damage their equipment.

[Link] (1 of 7) [13.03.2008. 5:15:15 Peca]

Forward

USB Made Simple - Part 2

Cables - Electrical
USB requires a shielded cable
containing 4 wires.
Two of these, D+ and D-, form a
twisted pair responsible for
carrying a differential data
signal, as well as some singleended signal states. (For low
speed the data lines may not be
twisted.)

Makeup of USB Cable

The signals on these two wires

are referenced to the (third)
GND wire.
The fourth wire is called VBUS,
and carries a nominal 5V
supply, which may be used by a
device for power.
'A' Plug, 'B' Plug and 'Mini-B' Plug

Connectors
As stated above, USB uses
different connectors on host
and device to enforce correct
connections.
"A" receptacles point
downstream from a Host or
Hub, while "B" receptacles point
upstream from a USB device or
hub.
Series A plugs mate with A
receptacles, and B plugs mate
with B receptacles.
Standard "A" and Standard
"B" Plug and Receptacle Pin
Assignments
[Link] (2 of 7) [13.03.2008. 5:15:15 Peca]

USB Made Simple - Part 2

Typical
Contact Signal
Cable
Number Name
Colour
1

VBUS Red

D-

White

D+

Green

GND

Black

Shell

Shield

Drain
Wire

A mini-B plug and receptacle

has also been defined as an
alternative to the standard B
connector on handheld and
portable devices. The mini-B
connecter has a fifth pin, named
ID, but it is not connected.
Mini-B Plug and Receptacle
Pin Assignments
Typical
Contact Signal
Cable
Number Name
Colour
1

VBUS Red

D-

White

D+

Green

ID

no
connection

GND

Black

Shell

Shield Drain Wire

Cable Types
The USB specification defines
three forms of cable:
1. A high/full speed
detachable cable with
one end terminated with
an A plug and the other
end with a B or mini-B
plug.
[Link] (3 of 7) [13.03.2008. 5:15:15 Peca]

USB Made Simple - Part 2

2. A captive high/full speed

cable where one end is
either hardwired to the
vendors equipment or
connected via a vendor
specific connector and
the other end is
terminated with an A plug.
3. A low speed version of 2.
The maximum length of a high/
full speed cable is determined
by the attenuation and
propagation delay. But, for a
low speed cable, it is the signal
rise and fall times that
determine the maximum length.
This forces the maximum length
for low speed cable to be
shorter than that for high/full
speed.

Power Distribution
A device (or hub) can only sink
(consume) current from its
upstream port.
A 'self-powered' device is one
which does not draw power
from the bus.
A device which draws its power
from the bus is called a 'buspowered' device. In normal
[Link] (4 of 7) [13.03.2008. 5:15:15 Peca]

USB Made Simple - Part 2

operation, it may draw up to

100mA, or 500mA if permitted
to do so by the host.
A device which has been
'Suspended', as a result of no
bus activity, must reduce its
current consumption to 0.5 mA
or less.

If a device is configured for high power (up

to 500 mA), and has its remote wakeup
feature enabled, it is allowed to draw up to
2.5mA during suspend.

Device Powering
The availability of a 5V supply is
a very attractive feature of USB,
and can simplify the design of a
device considerably. And a
device with a single connection
is also attractive to the user.
However before designing a
bus-powered device it is well to
consider the limitations of this
approach.
The voltage supplied can fall to
4.35V at the device. There can
also be transients on this taking
it 0.4V lower, due to other
devices being plugged in. Your
device needs to cope with these
voltage levels.
The standard unit load available
is 100mA. No device is
permitted to take more than this
before it has been configured
by the host. It must also reduce
its current consumption to
0.5mA whenever it is
'suspended' by a lack of activity
on the bus. Note that this
suspend condition will occur at
least once before the device is
configured.
It should be remember that of
this 0.5mA, the required 1.5k
pullup resistor is already
drawing 0.3mA. This leaves you
a budget of 0.2mA to power the
rest of your device circuitry. If
[Link] (5 of 7) [13.03.2008. 5:15:15 Peca]

"Hot-Pluggable"
To achieve the goal of being able to plug a
device into and out of a running system,
some design rules must be followed. Firstly
it is important to realise that if you pull a
plug out at the far end of the cable from the
device while current is being drawn, then
the cable will develop a potentially large
flyback voltage across your device. The
specification suggests that a minimum
solution to this is to place a capacitance of
at least 1uF across Vbus and GND.
The second thing to consider is that when
you plug your device in, any capacitance
between Vbus and GND will cause a dip in
voltage across the other ports of the hub to
which you are connecting. To limit the
consequences of this (such as crashing
other devices), the specification places a
maximum on the value of capacitance
across Vbus and GND of 10uF.
For the same reason, the hub port supply
must be bypassed with at least 120uF.

USB Made Simple - Part 2

the device contains a microcontroller it will need a sleep

mode which meets this
requirement, but do not forget
that a badly placed resistor can
very easily draw current which
you hadn't expected. Measure
your suspend current with a
meter.
A device may draw up to
500mA after it has been
configured as a high-power
device. Being configured is
dependent on the Hub being
able to supply 500mA, which
implies a self-powered hub. So
there is always a degree of
uncertainty whether more than
100mA will be available. It
would be well to offer the option
of external power via a socket
on such a device.

Self-Powered Devices
When designing a self-powered device,
remember that you must not pull a D+ or Dline above the Vbus voltage supplied. This
means that you must, at the very least,
sense when Vbus is connected.
The D+ or D- resistor should, strictly
speaking, be pulled up to a 3.3V supply
derived from Vbus, or controlled by Vbus in
such a way that the resistor never sources
current to the data line when Vbus is
switched off.
If you pull, say D+, high in the absence of
Vbus then you will risk faulty operation with
On-The-Go hosts. (See later).

Devices requiring more than

500mA are obliged to be selfpowered. The practice of
attempting to draw power from
two adjacent USB ports, using a
modified cable, is not permitted
and can easily damage the
ports.

Summary
We have looked at the electrical
requirements of USB.

Coming up...
Next we will examine the
elements of signalling which are
used by the USB protocol
Copyright 2006 MQP Electronics Ltd

[Link] (6 of 7) [13.03.2008. 5:15:15 Peca]

Forward

USB Made Simple - Part 3

Index

Back

Part 3 - Data Flow

Forward

Part 1
Part 2
Part 3
Part 4
Part 5
Part 6
Part 7
Links

The following discussion on data flow covers full and low speed. High speed signalling will be
covered in a later part.

USB is a Bus
Picture a setup of plugged-in hubs
and devices such as that on the
right. What we need to remember is
that, at any point in time, only the
host OR one device can be
transmitting at a time.
When the host is transmitting a
packet of data, it is sent to every
device connected to an enabled port.
It travels downwards via each hub in
the chain which resynchronises the
data transitions as it relays it. Only
one device, the addressed one,
actually accepts the data. (The
others all receive it but the address is
wrong for them.)
One device at a time is able to
transmit to the host, in response to a
direct request from the host. Each
hub repeats any data it receives from
a lower device in an upward only
direction.
Downstream direction ports are only
enabled once the device connected
to them is addressed, except that
one other port at a time can reset a
device to address 0 and then set its
address to a unique value.

[Link] (1 of 17) [13.03.2008. 5:12:54 Peca]

USB Made Simple - Part 3

Transceivers
At each end of the data link between
host and device is a transceiver
circuit. The transceivers are similar,
differing mainly in the associated
resistors.
A typical upstream end transceiver is
shown to the right with high speed
components omitted for clarity. By
upstream, we mean the end nearer
to the host. The upstream end has
two 15K pull-down resistors.
Each line can be driven low
individually, or a differential data
signal can be applied. The maximum
'high' level is 3.3V.

The equivalent downstream end

transceiver, as found in a device, is
shown to the right.
When receiving, individual receivers
on each line are able to detect single
ended signals, so that the so-called
Single Ended Zero (SE0) condition,
where both lines are low, can be
detected. There is also a differential
receiver for reliable reception of data.

[Link] (2 of 17) [13.03.2008. 5:12:54 Peca]

Upstream End Transceiver

USB Made Simple - Part 3

Not shown in these simplified

drawings is the rise and fall
time control on the differential
transmitters. Low speed
devices need longer rise and
fall times, so a full speed / low
speed hub must be able to
switch between these rise and
fall times.

Downstream End Transceiver (Full Speed)

Speed Identification
At the device end of the link a 1.5
kohm resistor pulls one of the lines
up to a 3.3V supply derived from
VBUS.
This is on D- for a low speed device,
and on D+ for a full speed device.
(A high speed device will initially
present itself as a full speed device
with the pull-up resistor on D+.)
The host can determine the required
speed by observing which line is
pulled high.

[Link] (3 of 17) [13.03.2008. 5:12:54 Peca]

USB Made Simple - Part 3

Line States
Given that there are just 2 data lines
to use, it is surprising just how many
different conditions are signaled
using them:
Detached
When no device is plugged in, the
host will see both data lines low, as
its 15 kohm resistors are pulling each
data line low.
Attached
When the device is plugged in to the
host, the host will see either D+ or Dgo to a '1' level, and will know that a
device has been plugged in.
The '1' level will be on D- for a low
speed device, and D+ for a full (or
high) speed device.
Idle
The state of the data lines when the
pulled up line is high, and the other
line is low, is called the idle state.
This is the state of the lines before
and after a packet is sent.
Bus State
Differential '1'

Levels
D+ high, D- low

Differential '0'

D- high, D+ low

Single Ended Zero (SE0)

D+ and D- low

Single Ended One (SE1)

D+ and D- high

To make it easier to talk about the

states of the data lines, some special
terminology is used. The 'J State' is
the same polarity as the idle state
(the line with the pull-up resistor is
high, and the other line is low), but is
being driven to that state by either
host or device.

Data J State:
Low-speed
Full-speed

Differential '0'
Differential '1'

Data K State:
Low-speed
Full-speed

Differential '1'
Differential '0'

Idle State:
Low-speed
Full-speed

D- high, D+- low

D+ high, D- low

Resume State

Data K state

The K state is just the opposite

polarity to the J state.

Start of Packet (SOP)

Data lines switch from

idle to K state

End of Packet (EOP)

SE0 for 2 bit times

followed by J state for
1 bit time

Disconnect

SE0 for >= 2us

J, K and SEO States

The Single Ended Zero (SE0) is

when both lines are being pulled low.
[Link] (4 of 17) [13.03.2008. 5:12:54 Peca]

USB Made Simple - Part 3

The J and K terms are used because

for Full Speed and Low Speed links
they are actually of opposite polarity.

Connect

Idle for 2.5us

Reset

SE0 for >= 2.5 us

Bus States
This table has been simplified from the original in the USB
specification. Please read the original table for complete
information.

Single Ended One (SE1)

This is the illegal condition where
both lines are high. It should never
occur on a properly functioning link.
Reset
When the host wants to start
communicating with a device it will
start by applying a 'Reset' condition
which sets the device to its default
unconfigured state.
The Reset condition involves the
host pulling down both data lines to
low levels (SE0) for at least 10 ms.
The device may recognise the reset
condition after 2.5 us.
This 'Reset' should not be confused
with a micro-controller power-on type
reset. It is a USB protocol reset to
ensure that the device USB signaling
starts from a known state.

EOP signal
The End of Packet (EOP) is an SE0
state for 2 bit times, followed by a J
state for 1 bit time.

[Link] (5 of 17) [13.03.2008. 5:12:54 Peca]

USB Made Simple - Part 3

Suspend
One of the features of USB which is
an essential part of today's emphasis
of 'green' products is its ability to
power down an unused device. It
does this by suspending the device,
which is achieved by not sending
anything to the device for 3 ms.
Normally a SOF packet (at full
speed) or a Keep Alive signal (at low
speed) is sent by the host every 1
ms, and this is what keeps the
device awake.
A suspended device may draw no
more than 0.5 mA from Vbus.
A suspended device must recognise
the resume signal, and also the reset
signal.
Resume
When the host wants to wake the
device up after a suspend, it does so
by reversing the polarity of the signal
on the data lines for at least 20ms.
The signal is completed with a low
speed end of packet signal.
It is also possible for a device with its
remote wakeup feature set, to initiate
a resume itself. It must have been in
the idle state for at least 5ms, and
must apply the wakeup K condition
for between 1 and 15 ms. The host
takes over the driving of the resume
signal within 1 ms.

Keep Alive Signal

This is represented by a Low speed
EOP. It is sent at least once every
millisecond on a low speed link, in
order to keep the device from
suspending.

[Link] (6 of 17) [13.03.2008. 5:12:54 Peca]

If a device is configured for high power (up

to 500 mA), and has its remote wakeup
feature enabled, it is allowed to draw up to
2.5mA during suspend.

USB Made Simple - Part 3

Packets
The packet could be thought of as
the smallest element of data
transmission. Each packet conveys
an integral number of bytes at the
current transmission rate. Before and
after the packet, the bus is in the idle
state.

Serial Interface Engine (SIE)

You need not be concerned with the

detail of syncs, bit stuffing, and End
Of Packet conditions, unless you are
designing at the silicon level, as the
Serial Interface Engine (SIE) will deal
with the details for you. You should
just be aware that the SIE can
recognise the start and end of a
packet, and that the packet contains
a whole number of bytes.
In spite of this packets often expect
fields of data to cross byte
boundaries. The important rule to
remember is that all usb fields are
transmitted least significant bit
first. So if, for example, a field is
defined by 2 successive bytes, the
first byte will be the least significant,
and the second byte transmitted will
be the most significant.

A packet starts with a sync pattern to

allow the receiver bit clock to
synchronise with the data. It is
followed, by the data bytes of the
packet, and concluded with an End
of Packet (EOP) signal. The data is
actually NRZI encoded, and in order
to ensure sufficiently frequent
transitions, a zero is inserted after 6
successive 1's (this is known as bit
stuffing).

[Link] (7 of 17) [13.03.2008. 5:12:54 Peca]

The complexities and speed of the USB

protocol are such that it is not practical to
expect a general purpose micro-controller to be
able to implement the protocol using an
instruction-driven basis. Dedicated hardware is
required to deal with the time-critical portions of
the specification, and the circuitry grouping
which performs this function is referred to as the
Serial Interface Engine (SIE).

Data Fields are Transmitted Least

Significant Bit First
The first time when you need to know this is
when you are defining 'descriptors' in your
firmware code. Many of these values are word
sized and you need to add the bytes in the low
byte, high byte order.

Idle

SYNC

DATA BYTES

A Single Packet

EOP

Idle

USB Made Simple - Part 3

Before we continue, some

definitions...

Endpoints
Each USB device has a number of
endpoints. Each endpoint is a
source or sink of data. A device can
have up to 16 OUT and 16 IN
endpoints.
OUT always means from host to
device.
IN always means from device to
host.
Endpoint 0 is a special case which
is a combination of endpoint 0 OUT
and endpoint 0 IN, and is used for
controlling the device.

Pipe
A logical data connection between
the host and a particular endpoint,
in which we ignore the lower level
mechanisms for actually achieving
the data transfers.

Transactions
Simple transfers of data called
'Transactions' are built up using
packets.

lsb

msb

PID0 PID1 PID2 PID3 \PID0 \PID1 \PID2 \PID3

Packet Formats
The first byte in every packet is a
Packet Identifier (PID) byte. This
byte needs to be recognised quickly
by the SIE and so is not included in
any CRC checks. It therefore has its
own validity check. The PID itself is 4
[Link] (8 of 17) [13.03.2008. 5:12:54 Peca]

The PID is shown here in the order of

transmission; lsb first.

USB Made Simple - Part 3

bits long, and the 4 bits are repeated

in an complemented form.

Cyclic Redundancy Code (CRC)

A CRC is a value calculated from a number of data
bytes to form a unique value which is transmitted
along with the data bytes, and then used to
validate the correct reception of the data.
USB uses two different CRCs, one 5 bits long
(CRC5) and one 16 bits long (CRC16).
See the USB specification for details of the
algorithms used.

PID Type

Token

There are 17 different PID values

defined. This includes one reserved
value, and one value which has been
used twice with different meanings
for two different situations.
Notice that the first 2 bits of a token
which are transmitted, determine
which of the 4 groups it falls into.
This is why SOF is officially
considered to be a token PID.

Data

Handshake

Special

PID Name

PID<3:0>*

OUT

0001b

IN

1001b

SOF

0101b

SETUP

1101b

DATA0

0011b

DATA1

1011b

DATA2

0111b

MDATA

1111b

ACK

0010b

NAK

1010b

STALL

1110b

NYET

0110b

PRE

1100b

ERR

1100b

SPLIT

1000b

PING

0100b

Reserved

0000b

* Bits are transmitted lsb first

There are four different packet
formats based on which PID the
packet starts with.

[Link] (9 of 17) [13.03.2008. 5:12:54 Peca]

USB Made Simple - Part 3

Token Packet

The token packet contains two addressing

elements:

Sync PID ADDR ENDP CRC5 EOP

8
7 bits
bits

4 bits 5 bits

Used for SETUP, OUT and IN

packets. They are always the first
packet in a transaction, identifying
the targeted endpoint, and the
purpose of the transaction.
The SOF packet is also defined as a
Token packet, but has a slightly
different format and purpose, which
is described below.

Data Packet
Sync PID
8
bits

DATA CRC16 EOP

(01024) 16 bits
x 8 bits

Used for DATA0, DATA1, DATA2

and MDATA packets. If a transaction
has a data stage this is the packet
format used.

Address (7 bits)

This device address can address up to 127

devices. Address 0 is reserved for a device which
has not yet had its address set.
Endpoint number (4 bits)

There can be up to 16 possible endpoints in a

device in each direction. The direction is implicit in
the PID. OUT and SETUP PIDs will refer to the
OUT endpoint, and an IN PID will refer to the IN
endpoint.

DATA0 and DATA1 PIDs are used in Low

and Full speed links as part of an errorchecking system. When used, all data
packets on a particular endpoint use an
alternating DATA0 / DATA1 so that the
endpoint knows if a received packet is the
one it is expecting. If it is not it will still
acknowledge (ACK) the packet as it is
correctly received, but will then discard the
data, assuming that it has been re-sent
because the host missed seeing the ACK
the first time it sent the data packet.
DATA2 and MDATA are only used for high
speed links.

ACK

Handshake Packet

Receiver acknowledges receiving error free

packet.
NAK

Sync

PID

EOP

8 bits

Receiving device cannot accept data or

transmitting device cannot send data.
STALL

Used for ACK, NAK, STALL and

NYET packets. This is the packet
format used in the status stage of a
transaction, when required.

Endpoint is halted, or control pipe request is

not supported.
NYET

No response yet from receiver (high speed

[Link] (10 of 17) [13.03.2008. 5:12:54 Peca]

USB Made Simple - Part 3

only)
SOF Packet
Frames
Sync

PID

Frame
CRC5 EOP
No.

8 bits 11 bits 5 bits

The Start of Frame packet is sent

every 1 ms on full speed links. The
frame is used as a time frame in
which to schedule the data transfers
which are required. For example, an
isochronous endpoint will be
assigned one transfer per frame.

Transactions
A successful transaction is a
sequence of three packets which
performs a simple but secure
transfer of data.
For IN and OUT transactions used
for isochronous transfers, there are
only 2 packets; the handshake
packet on the end is omitted. This is
because error-checking is not
required.
There are three types of transaction.
In each of the illustrations below, the
packets from the host are shaded,
and the packets from the device are
not.
OUT Transaction

A successful OUT transaction

comprises two or three sequential
packets. If it were being used in an
Isochronous Transfer there would
not be a handshake packet from the
device.
On a low or full speed link, the PID
shown as DATAx will be either a

[Link] (11 of 17) [13.03.2008. 5:12:54 Peca]

On a low speed link, to preserve bandwidth,

a Keep Alive signal is sent every
millisecond, instead of a Start of Frame
packet. In fact Keep Alives may be sent by a
hub on a low speed link whenever the hub
sees a full speed token packet.
At high speed the 1 ms frame is divided into
8 microframes of 125 us. A SOF is sent at
the start of each of these 8 microframes,
each having the same frame number, which
then increments every 1 ms frame.

USB Made Simple - Part 3

DATA0 or a DATA1. An alternating

DATA0/DATA1 is used as a part of
the error control protocol to (or from)
a particular endpoint.

IN Transaction

A successful IN transaction
comprises two or three sequential
packets. If it were being used in an
Isochronous Transfer there would
not be a handshake packet from the
host.
Here again, the DATAx is either a
DATA0 or a DATA1.
SETUP Transaction

A successful SETUP transaction

comprises three sequential packets.
This is similar to an OUT transaction,
but the data payload is exactly 8
bytes long, and the SETUP PID in
the token packet informs the device
that this is the first transaction in a
Control Transfer (see below).
As will be seen below, the SETUP
transaction always uses a DATA0 to
start the data packet.

Data Flow Type

Description

Control Transfer

Mandatory using Endpoint 0

OUT and Endpoint 0 IN.

Bulk Transfer

Error-free high volume

throughput when bandwidth
available.

Interrupt Transfer

Regular Opportunity for

status updates, etc.
Error-free
Low throughput

Data Flow Types

There are four different ways to
transfer data on a USB bus. Each
has its own purposes and
characteristics. Each one is built up
using one or more transaction type.

[Link] (12 of 17) [13.03.2008. 5:12:54 Peca]

USB Made Simple - Part 3

Isochronous
Transfer

Guaranteed fixed
bandwidth.
Not error-checked.

Bulk Transfers
Bulk transfers are designed to
transfer large amounts of data with
error-free delivery, but with no
guarantee of bandwidth. The host
will schedule bulk transfers after the
other transfer types have been
allocated.

Example Bulk Transfer

If an OUT endpoint is defined as

using Bulk transfers, then the host
will transfer data to it using OUT
transactions.
If an IN endpoint is defined as using
Bulk transfers, then the host will
transfer data from it using IN
transactions.
The max packet size is 8, 16, 32 or
64 at full Speed and 512 for high
speed. Bulk transfers are not allowed
at low speed.
Use Bulk transfers when you have a
lot of data to shift, as fast as
possible, but where you would not
have a large problem if there is a
delay caused by insufficient
bandwidth.
The diagrams to the right illustrate the
possible flow of events in the face of
errors.
Error Control - IN
If the IN token packet is not
recognised, the device will not respond
at all. Otherwise, if it has data to send it
will send it in a DATA0 or DATA1
packet, If it is not ready to send data it
will send a NAK packet. If the endpoint
is currently 'halted' then it will respond
with a STALL packet.

[Link] (13 of 17) [13.03.2008. 5:12:54 Peca]

BULK Transfer Error Control Flow

USB Made Simple - Part 3

In the case of DATA0/1 being sent, the

host will acknowledge with an ACK,
unless the data is not validly received,
in which case it does not send an ACK.
(Note: the host never sends NAK!)
Error Control - OUT
If the OUT token packet is not
recognised, the device will not respond
at all. It will then ignore the DATAx
packet because it does not know that it
has been addressed.
If the OUT token is recognised but the
DATAx packet is not recognised, then
the device will not respond.
If the data is received but the device
can't accept it at this time, it will send a
NAK, and if the endpoint is currently
halted, it will send a STALL.

Example Interrupt Transfer

Interrupt Transfers
Interrupt transfers have nothing to do
with interrupts. The name is chosen
because they are used for the sort of
purpose where an interrupt would
have been used in earlier connection
types.
Interrupt transfers are regularly
scheduled IN or OUT transactions,
although the IN direction is the more
common usage.

[Link] (14 of 17) [13.03.2008. 5:12:54 Peca]

Error Control Flow

USB Made Simple - Part 3

Typically the host will only fetch one

packet, at an interval specified in the
endpoint descriptor (see below). The
host guarantees to perform the IN
transaction at least that often, but it
may actually do it more frequently.
Interrupt packets can have any size
from 1 to 8 bytes at low speed, from
1 to 64 at full speed or up to 1024
bytes at high speed.
Use an interrupt transfer when you
need to be regularly kept up to date
of any changes of status in a device.
Examples of their use are for a
mouse or a keyboard.
Error control is very similar to that for
bulk transfers.

Isochronous Transfers

Example Isochronous Transfer

Isochronous transfers have a

guaranteed bandwidth, but error-free
delivery is not guaranteed.
The main purpose of isochronous
transfers is applications such as
audio data transfer, where it is
important to maintain the data flow,
but not so important if some data
gets missed or corrupted.
An isochronous transfer uses either
an IN transaction or an OUT
transaction depending on the type of
endpoint. The special feature of
these transactions is that there is no
handshake packet at the end.
An isochronous packet may contain
up to 1023 bytes at full speed, or up
to 1024 at high speed. Isochronous
[Link] (15 of 17) [13.03.2008. 5:12:54 Peca]

Error Control Flow

USB Made Simple - Part 3

transfers are not allowed at low

speed.

Example Control Read

Control Transfer
This is a bi-directional transfer which
uses both an IN and an OUT
endpoint. Each control transfer is
made up of from 2 to several
transactions.
It is divided into three stages.
The SETUP stage carries 8 bytes
called the Setup packet. This defines
the request, and specifies whether
and how much data should be
transferred in the DATA stage.
The DATA stage is optional. If
present, it always starts with a
transaction containing a DATA1. The
type of transaction then alternates
between DATA0 and DATA1 until all
the required data has been
transferred.
The STATUS stage is a transaction
containing a zero-length DATA1
packet. If the DATA stage was IN
then the STATUS stage is OUT, and
vice versa.
Control transfers are used for initial
configuration of the device by the
host, using Endpoint 0 OUT and
Endpoint 0 IN, which are reserved for
this purpose. They may be used (on
the same endpoints) after
configuration as part of the devicespecific control protocol, if required.

Error Control Flow

SETUP STAGE

Notice that it is not permitted for a device to

respond to a SETUP with a NAK or a STALL.
DATA STAGE

(same as for bulk transfer)

The max packet size for the data
stage is 8 bytes at low speed, 8, 16,
32 or 64 at full Speed and 64 for high

[Link] (16 of 17) [13.03.2008. 5:12:54 Peca]

STATUS STAGE

USB Made Simple - Part 3

speed.

(same as for bulk transfer)

Summary
We have examined 4 different types
of data transfer, each of which uses
different combinations of packets.
We have seen Control Transfers
which every device uses to
implement a Standard set of
requests. And we have seen three
other data transfer types, which a
device might use depending on its
purpose.

Coming up...
Next we will examine the standard
set of requests which every USB
device has to implement.

Forward

ADVERTISEMENT

Packet-Master USB Bus Analysers from MQP

Electronics

Special Offer

Radically cut development time

Intuitive graphical interface
Detailed timing information
Full analysis of all standard events
Results can be printed
Optional class analysis modules
Click for more information

[Link] (17 of 17) [13.03.2008. 5:12:54 Peca]

USB Made Simple - Part 4

Index
Part 1
Part 2
Part 3
Part 4
Part 5
Part 6
Part 7
Links

Back

Part 4 - Protocol

Controlling a Device
Before we go into detail, we
need to look at how the host
recognises and installs a
device when you plug it in. We
need to do this in general
terms without getting bogged
down with the detail.
When you plug a USB device
in, the host becomes aware
(because of the pullup resistor
on one data line), that a
device has been plugged in.
The host now signals a USB
Reset to the device, in order
that it should start in a known
state at the end of the reset.
In this state the device
responds to the default
address 0. Until the device
has been reset the host
prevents data from being sent
downstream from the port. It
will only reset one device at a
time, so there is no danger of
two devices responding to
address 0.
The host will now send a
request to endpoint 0 of
device address 0 to find out its

[Link] (1 of 17) [13.03.2008. 5:16:28 Peca]

Forward

USB Made Simple - Part 4

maximum packet size. It can

discover this by using the Get
Descriptor (Device) command.
This request is one which the
device must respond to even
on address 0.
Typically (i.e. with Windows)
the host will now reset the
device again. It then sends a
Set Address request, with a
unique address to the device
at address 0. After the request
is completed, the device
assumes the new address.
(And at this point the host is
now free to reset other
recently plugged-in devices.)
Typically the host will now
begin to quiz the device for as
many details as it feels it
needs. Some requests
involved here are:

Get Device Descriptor

Get Configuration
Descriptor
Get String Descriptor

At the moment the device is in

an addressed but
unconfigured state, and is
only allowed to respond to
standard requests.
Once the host feels it has a
clear enough picture of what
the device is, it will load a
suitable device driver.
The device driver will then
select a configuration for the
device, by sending a Set
Configuration request to the
device.
The device is now in the
configured state, and can start
working as the device it was
designed to be. From now on
it may respond to device
[Link] (2 of 17) [13.03.2008. 5:16:28 Peca]

USB Made Simple - Part 4

specific requests, in addition

to the standard requests
which it must continue to
support.
We can now see that there is
a set of requests which a
device must respond to, and
need to look at the detailed
means by which the requests
are conveyed.
We saw in the last chapter
that data is transfered in 4
different types of transfer:

Control Transfers
Interrupt Transfers
Bulk Transfers
Isochronous Transfers

The only transfer type

available before the device
has been configured is the
Control Transfer. The only
endpoint available at this time
is the bidirectional Endpoint 0.

Configurations,
Interfaces, and
Endpoints.
The device contains a number
of descriptors (as shown to
the right) which help to define
what the device is capable of.
We will examine these
descriptors further down the
page. For the moment we
need to have an idea what the
configurations, interfaces and
endpoints are and how they fit
together.
A device can have more than
one configuration, though
only one at a time, and to
change configuration the
whole device would have to
stop functioning. Different
[Link] (3 of 17) [13.03.2008. 5:16:29 Peca]

USB Made Simple - Part 4

configurations might be used,

for example, to specify
different current requirements,
as the current required is
defined in the configuration
descriptor.
However it is not common to
have more than one
configuration. Windows
standard drivers will always
select the first configuration so
there is not a lot of point.
A device can have one or
more interfaces. Each
interface can have a number
of endpoints and represents a
functional unit belonging to a
particular class.
Each endpoint is a source or
sink of data.
For example a VOIP phone
might have one audio class
interface with 2 endpoints for
transferring audio in each
direction, plus a HID interface
with a single IN interrupt
endpoint, for a built in keypad.
It is also possible to have
alternative versions of an
interface, and this is more
common than multiple
configurations. In the VOIP
phone example, the audio
class interface might offer an
alternative with a different
audio rate. It is possible to
switch an interface to an
alternate while the device
remains configured.

[Link] (4 of 17) [13.03.2008. 5:16:29 Peca]

USB Made Simple - Part 4

The SETUP Packet

The Standard requests are all
conveyed using control
transfers to endpoint 0.
Remember that a control
transfer starts with a SETUP
transaction which conveys 8
bytes. These 8 bytes define
the request from the host.
The structure of
bmRequestType makes it
easy to use it to switch on
when your firmware is trying
to interpret the setup request.
Essentially, when the SETUP
arrives, you need to branch to
the handler for the particular
request, so for example bits
6:5 allow you to distinguish
the mandatory standard
commands, from any class or
vendor commands you may
have implemeted for you
particular device.

Offset

Field

Size

Value

Description
D7 Data direction
0 - Host-to-device
1 - Device-to-host
D6:5 Type
0 = Standard
1 = Class
2 = Vendor
3 = Reserved
D4:0 Recipient
0 = Device
1 = Interface
2 = Endpoint
3 = Other
4-31 = Reserved

bmRequestType

Bitmap

bRequest

Value

Specific Request

wValue

Value

Use varies according

to request

wIndex

Index or
Offset

Use varies according

to request

wLength

Count

Number of bytes to
transfer if there is a
data stage

The meaning of the 8 bytes of the SETUP transaction

data, which are divided into five named fields.

Switching on bit 7 allows you

to deal with IN and OUT
direction requests in separate
areas of the code.

Here is a table which contains all the standard requests which a host can send. The first 5
columns are the SETUP transaction fields in order, and the last column describes any
accompanying data stage data which will have the length wLength.
bmRequestType

bRequest

wValue

wIndex

wLength

Data

00000000b
00000001b
00000010b

CLEAR_FEATURE
(1)

Feature
Selector

Zero
Interface
Endpoint

Zero

None

10000000b

GET_CONFIGURATION
(8)

Zero

Zero

One

Configuration
Value

10000000b

GET_DESCRIPTOR
(6)

Descriptor
Type (H) and
Descriptor
Index (L)

10000001b

GET_INTERFACE
(10)

Zero

[Link] (5 of 17) [13.03.2008. 5:16:29 Peca]

Zero or
Descriptor
Language
Length
ID
Interface

One

Descriptor
Alternate
Interface

USB Made Simple - Part 4

10000000b
10000001b
10000010b

GET_STATUS
(0)

00000000b

SET_ADDRESS
(5)

Zero

Zero
Interface
Endpoint

Two

Device,
Interface or
Endpoint
Status

Device
Address

Zero

Zero

None

Zero

Zero

None

00000000b

SET_CONFIGURATION Configuration
(9)
Value

00000000b

SET_DESCRIPTOR
(7)

Descriptor
Type (H) and
Descriptor
Index (L)

00000000b
00000001b
00000010b

SET_FEATURE
(3)

Feature
Selector

Zero
Interface
Endpoint

Zero

None

00000001b

SET_INTERFACE
(11)

Alternate
Setting

Interface

Zero

None

10000010b

SYNCH_FRAME
(12)

Zero

Endpoint

Two

Frame
Number

Zero or
Descriptor
Language
Length
ID

Descriptor

GET_DESCRIPTOR
It is probable that this request
(with the descriptor type set to
Device) will be the first that
will be received after USB
reset. The host needs to know
the max packet length in use
by the control endpoint and
this information is available in
the 8th byte of the device
descriptor.
Typically when the host is
Windows, the device will
receive the request with the
required length wLength set to
64. The host will then input 1
packet, and then reset the
device again. Whatever the
value of the max packet
length, the host now has the
value of the 8th byte and
knows what the packet size is
for all future control transfers.
The second reset is probably
to guarantee that the device
does not get confused after
not being allowed to complete
the transmission of all 18
bytes of the device descriptor.
[Link] (6 of 17) [13.03.2008. 5:16:29 Peca]

Descriptor Types

Value

Device

Comments

Configuration

Request for this also

returns OTG,
interface and
endpoint descriptors

String

Qualified by an index
to specify which
string is required

Interface

Not directly
accessible

Endpoint

Not directly
accessible

Device Qualifier

Only for high speed

capable devices

Other Speed
Configuration

Only for high speed

capable devices

Interface Power

Obsolete

On-The-Go (OTG)

Not directly
accessible

Table of wValues use in Get Descriptor requests to

select the required descriptor.

USB Made Simple - Part 4

Device Descriptor
This descriptor will most likely
be the first one fetched by the
host. We should point out
some important features.
bLength and bDescriptorType

All descriptors start with a

single byte specifying the
descriptor's length, and this is
always followed by a single
byte defining the descriptor
type.

Offset

Field

Size

Value

Description

bLength

Number

Size of this descriptor

in bytes

bDescriptorType

Constant

DEVICE descriptor
type (= 1)

bcdUSB

BCD

USB Spec release

number

bDeviceClass

Class

Class code assigned

by USB-IF
00h means each
interface defines its
own class
FFh means vendordefined class
Any other value
must be a class code

bDeviceSubClass

SubClass

SubClass Code
assigned by USB-IF

bDeviceProtocol

Protocol

Protocol Code
assigned by USB-IF

bcdUSB

The only valid version

numbers are 0x0100
(USB1.0), 0x0110 (USB1.1)
and 0x0200 (USB2.0). If you
design a new device it should
be identified as USB2.0
because that is the current
specification.
bDeviceClass,
bDeviceSubClass and
bDeviceProtocol

This triplet of values is used to

describe the class of the
device in various ways as
defined in the various class
specification documents from
the USB-IF.

bMaxPacketSize0

Number

Max packet size for

endpoint 0.
Must be 8, 16, 32 or
64

idVendor

ID

Vendor ID - must be
obtained from USB-IF

10

idProduct

ID

Product ID assigned by the

manufacturer

12

bcdDevice

BCD

Device release
number in binary
coded decimal

Index

Index of string
descriptor describing
manufacturer - set to
0 if no string

Index

Index of string
descriptor describing
product - set to 0 if
no string

14

iManufacturer

idVendor, idProduct and

bcdDevice

The combination of idVendor

and idProduct (also known as
the VID and PID) must be
unique for the device. This
means that the VID you use
must be one issued by the
USB-IF and which you have
the right to use. You can
either buy a VID from the USBIF, or you may be able to
[Link] (7 of 17) [13.03.2008. 5:16:29 Peca]

15

iProduct

16

iSerialNumber

Index

Index of string
descriptor describing
device serial number
- set to 0 if no string

17

bNumConfigurations

Number

Number of possible
configurations

Device Descriptor

USB Made Simple - Part 4

acquire the right to use a VID

from another manufacturer
together with a particular PID
which they have issued to
you. If you use a VID/PID
combination which is already
in use then you will probably
have major problems with
your product in the field.
When are requests valid?

SET_ADDRESS
After the host has determined
the max packet size for
endpoint 0, it is in a position to
begin normal communications
with the device. As mentioned
above, there may be a second
reset from the host. The host
now needs to issue a
SET_ADDRESS request to
the device, so that each
device on the bus has a
unique address to respond to.
SET_ADDRESS is a simple,
outward direction request in a
control transfer with no data
stage. The only useful
information carried in the
SETUP packet is the required
address.

The device can be in one of three states which

determine whether a particular request is valid at
the time.
The states are:
Default
After reset but before receiving Set Address.
In the Default state, the only valid requests are
Get Descriptor, and Set Address.
Addressed
After the device has been assigned an address
via Set Address.
Now the device must recognise the following
additional requests:

When implementing this

request in firmware, you
should note the following. All
other requests must be
actioned before the status
stage in completed. But in the
case of SET_ADDRESS, you
should not change the device
address until after the status
stage. The status stage will
not succeed unless the device
is still responding to address 0
while it is taking place. The
device then has 2ms to get
ready to respond to the new
[Link] (8 of 17) [13.03.2008. 5:16:29 Peca]

Set Configuration
Get Configuration
Set Feature
Clear Feature
Get Status
Set Descriptor (optional)

Configured
After the host has sent Set Configuration with a
non-zero value, to select a configuration. The
device is now operational.
In the Configured state, only Set Address is not
a valid request. Three further requests are
restricted to Configured state only:

Get Interface

USB Made Simple - Part 4

address.

Set Interface
Synch Frame

Note that this was only a brief overview. The

specification gives more detailed information, which
you should read when implementing a USB device.

Other Information
Gathering Commands
The host is likely to start using
the GET_DESCRIPTOR
request mentioned above, to
fetch other information
describing the device. A major
piece of this information is the
configuration descriptor.
Get Descriptor
(Configuration)
The Get Descriptor
(Configuration) warrants
special explanation, because
the request results in not just
a Configuration Descriptor
being returned, but also some
or all of a number of other
descriptors:

Interface Descriptor
Endpoint Descriptor
OTG Descriptor
Class-specific
Descriptors
Vendor-specific
Descriptors

A Get Configuration
Descriptor fetches the
descriptors for just one
configuration depending on
the descriptor index in wValue
of the SETUP packet. Most
devices only have one
configuration, because built-in
Windows drivers always
select the first configuration.
The diagram opposite shows
a typical set of Descriptors
[Link] (9 of 17) [13.03.2008. 5:16:29 Peca]

The actual descriptor which is fetched by a

GET_DESCRIPTOR request is determined
by the high byte of the wValue word in the
SETUP data.
So the request we call here 'Get Descriptor
(Configuration)' is simply a Get Descriptor
request with the high byte of wValue set to 2.

USB Made Simple - Part 4

which is fetched. It starts with

the configuration descriptor,
and the vertical position
shows the correct sequence,
with the interfaces being dealt
with in turn, each one followed
by its own endpoints.
The position of class
descriptors is defined in the
appropriate class
specification, and of course
vendors descriptor positions
would be up to the vendor
concerned.
An OTG descriptor position is
not defined but typically
appears immediately after the
configuration descriptor.

Offset

Configuration Descriptor
The configuration descriptor
format is shown to the right.
The wTotalLegth value is
important because it tells the
host how many bytes are
contained in this descriptor
and all the descriptors which
follow.

[Link] (10 of 17) [13.03.2008. 5:16:29 Peca]

Field

Size

Value

Description

bLength

Size of this descriptor

Number
in bytes

bDescriptorType

Constant

CONFIGURATION
descriptor type (= 2)

wTotalLength

Total number of bytes

in this descriptor and
Number
all the following
descriptors.

bNumInterfaces

Number of interfaces
Number supported by this
configuration

Value used by Set

Configuration to
Number
select this
configuration

bConfigurationValue

iConfiguration

Index

Index of string
descriptor describing
configuration - set to 0
if no string

USB Made Simple - Part 4

bNumInterfaces describes
how many interfaces this
configuration supports.

bmAttributes

bMaxPower

D7: Must be set to 1

D6: Self-powered
Bitmap
D5: Remote Wakeup
D4...D0: Set to 0

mA

Maximum current
drawn by device in
this configuration. In
units of 2mA. So 50
means 100 mA.

Configuration Descriptor

Interface Descriptor
The interface descriptor
format is shown to the right.
bAlternateSetting needs some
explanation. An interface can
have more than one variant,
and these variants can be
switched between, while other
interfaces are still in
operation.
For the first (and default)
alternative bAlternateSetting
is always 0.
To have a second interface
variant, the default interface
descriptor would be followed
by its endpoint descriptors,
which would then be followed
by the alternative interface
descriptor and then its
endpoint descriptors.
bInterfaceClass,
bInterfaceSubClass and
bInterfaceProtocol

Offset

[Link] (11 of 17) [13.03.2008. 5:16:29 Peca]

Size

Value

Description

bLength

Size of this descriptor

Number
in bytes

bDescriptorType

Constant

bInterfaceNumber

Number identifying
Number this interface. Zerobased value.

bAlternateSetting

Value used to select

Number this alternate setting
for this interface.

Number of endpoints
used by this interface.
Number
Doesn't include
control endpoint 0.
Class code assigned
by USB-IF
00h is a reserved
value
FFh means vendordefined class
Any other value must
be a class code

bNumEndpoints

INTERFACE
descriptor type (= 4)

bInterfaceClass

Class

bInterfaceSubClass

SubClass

SubClass Code
assigned by USB-IF

bInterfaceProtocol

Protocol

Protocol Code
assigned by USB-IF

Index

Index of string
descriptor describing
interface - set to 0 if
no string

By defining the class,

subclass and protocol in the
interface, it is possible to have
interfaces with different
classes in the same device.
This is referred to as a
composite device.

Field

iInterface

Interface Descriptor

USB Made Simple - Part 4

Offset

Field

Size

Value

Description

bLength

Size of this descriptor in

Number
bytes

bDescriptorType

Constant

ENDPOINT descriptor
type (= 5)
The address of this
endpoint within the
device.

bEndpointAddress

D7: Direction
Endpoint 0 = OUT, 1 = IN
D6-D4: Set to 0
D3-D0: Endpoint
number

Endpoint Descriptor
The endpoint descriptor
format is shown to the right.

bmAttributes

wMaxPacketSize

bInterval

D1:0 Transfer Type

00 = Control
01 = Isochronous
10 = Bulk
11 = Interrupt
The following only
apply to isochronous
endpoints. Else set to 0.
D3:2 Synchronisation
Type
00 = No
Synchronisation
01 = Asynchronous
10 = Adaptive
11 = Synchronous
D5:4 Usage Type
00 = Data endpoint
01 = Feedback
endpoint
10 = Implicit feedback
Data endpoint
11 = Reserved
D7:6 Reserved
Set to 0

Bitmap

Maximum packet size

this endpoint can send
Number or receive when this
configuration is
selected

Interval for polling

endpoint for data
transfers. Expressed in
Number
frames (ms) for low/full
speed or microframes
(125us) for high speed.

Endpoint Descriptor

[Link] (12 of 17) [13.03.2008. 5:16:29 Peca]

USB Made Simple - Part 4

Get Descriptor (String)

There are several strings
which a host may request.
The strings defined in the
device descriptor are:

Manufacturer String
Product String
Serial Number String

These strings are optional. If

not supported, the
corresponding index in the
device descriptor will be 0.
Otherwise the host may use
the specified index in a Get
Descriptor (String) request to
fetch the descriptor.
Get Descriptor (String), with a
descriptor index of 0 in the low
byte of wValue, is used to
fetch a special string language
descriptor. This contains a
series of 2-byte sized
language specifiers. In theory,
if the language of your choice
is supported in this list, you
can use the index to this
language ID to access the
string descriptors in this
language by specifying this in
wIndex of the Get Descriptor
(String) request. In practise,
with Windows, you will have
difficulties if you do not ensure
that the first language
specified is English (US).

[Link] (13 of 17) [13.03.2008. 5:16:29 Peca]

Offset

Field

Size

Value

Description

bLength

Number

Size of this descriptor in

bytes

bDescriptorType

Constant

STRING descriptor type

(= 3)

wLANGID[0]

Number

LANGID Code 0

...

...

...

...

...

2+
x*2

wLANGID[x]

Number

LANGID Code x

String Descriptor Zero

(Specifies supported string languages)

Offset

Field

Size

Value

Description

bLength

Size of this descriptor in

Number
bytes

bDescriptorType

Constant

STRING descriptor type

(= 3)

bString

Number

UNICODE encoded
string

String Descriptor

USB Made Simple - Part 4

SET_CONFIGURATION
When the host has got all the
information it requires it loads
a driver for the device based
on the VID/PID combination in
the device descriptor, or on
the standard class defined
there or in an interface
descriptor.
The driver may also ask for
the same or different
information using Get
Descriptor requests.
Eventually it will decide to
configure the device using the
SET_CONFIGURATION
request. Usually ( when there
is one configuration) the Set
Configuration request will
have wValue set to 1, which
will select the first
configuration.
Set Configuration can also be
used, with wValue set to 0, to
deconfigure the device.

A Configured Device
Once a device has been configured, it is
allowed to respond to other transfer types
than Control transfers.
As we have seen, the other transfer types are

Interrupt Transfers
Bulk Transfers
Isochronous Transfers

As a result of the information in the

descriptors, the host will now know what
particular transfers on which particular
endpoints the device is prepared to support.
There may now also be new class or vendorspecific requests which may now be
supported on the control endpoint in addition
to the standard requests.
It is all these additional transfers which
perform the functionality that the device was
designed for.

GET_CONFIGURATION
This request compliments Set
Configuration, and simply
allows the host to determine
which configuration it
previously set.
SET_FEATURE
CLEAR_FEATURE
This pair of requests is used
to control a small number of
on-off features on a device, an
interface or an endpoint.
A device has 5 possible
features, an endpoint has one,
and an interface actually has
none at all.
[Link] (14 of 17) [13.03.2008. 5:16:29 Peca]

Feature Selector

Recipient

Value

ENDPOINT_HALT

Endpoint

DEVICE_REMOTE_WAKEUP

Device

USB Made Simple - Part 4

The greyed out features

shown in the table only apply
to OTG devices.

TEST_MODE

Device

B_HNP_ENABLE

Device

A_HNP_SUPPORT

Device

A_ALT_HNP_SUPPORT

Device

3
4
5

ENDPOINT_HALT

Setting this feature will cause

an endpoint to STALL any IN
or OUT transactions.

Table of wValues used in Set Feature and Clear

Feature requests.

DEVICE_REMOTE_WAKEUP

Setting this feature allows a

device which is then
suspended to use resume
signalling to gain the host's
attention.
Status Bit

Purpose

Comment

D0

Self Powered

Set to 1 by the device

when it is self-powered

D1

Remote
Wakeup

Set to 1 if the device has

been enabled to signal
remote wakeup.

D2 - D15

reserved

Must be set to 0

GET_STATUS
This request is used to fetch
status bits from a device, an
interface or an endpoint. In
each case the request fetches
16 bits (2 bytes). The tables to
the right show the status bits
which are currently
implemented.
Note that Remote Wakeup
and Halt status bits can both
be controlled by the host
using [Link] Feature
requests, but the Selfpowered bit is only controlled
by the device.

Device Status Bits

Status Bit
D0 - D15

Purpose
reserved

Comment
Must be set to 0

Interface Status Bits

Status Bit

Purpose

D0

Halt

D1 - D15

reserved

Comment
Set to 1 when endpoint is
halted
Must be set to 0

Endpoint Status Bits

[Link] (15 of 17) [13.03.2008. 5:16:29 Peca]

USB Made Simple - Part 4

SET_INTERFACE
GET_INTERFACE
Once a device has been
configured the host may use
Set Interface to select an
alternative interface to a
particular default interface. It
can use the Get Interface to
determine which interface
alternative it previous set for a
particular interface.
SYNCH_FRAME
This is used with some
isochronous transfer where
the transfer size varies with
the frame. See USB 2.0
specification for more details.
SET_DESCRIPTOR
This Standard request is
optional and not often used. It
allows the host to specify a
new set of values for a given
descriptor. It is hard to
imagine when this might be of
value.

Summary
We have looked at the set of
standard requests which a
device must support to
become operational.

Coming up...
Next we will examine the
complete enumeration and
start of operation of a specific
device.
Copyright 2006 MQP Electronics Ltd

[Link] (16 of 17) [13.03.2008. 5:16:29 Peca]

Forward

USB Made Simple - Part 5

Index
Part 1
Part 2
Part 3
Part 4
Part 5
Part 6
Part 7
Links

Back

Part 5 - Example Device

Typical Human Interface Device (HID)

A Mouse
We are going to look at a typical enumeration
and subsequent operation of one of the simplest
USB devices around; the mouse. Below you will
see the output of a hardware bus analyser which
is capturing all the USB traffic involved when a
mouse is plugged in. Let us emphasise straight
away that this is just a typical sequence, which
can vary widely depending on the host, which in
this case was Windows XP.

For the display shown below, the analyser has been set to display only Bus States,
and the highest level Transfers. Each Control Transfer shown is made up of a
series of transactions, each of which is made up of a series of packets, as we shall
see later.
The capture begins 1.9 seconds before the mouse is plugged in. The analyser
indicates that the device has been plugged in, and that it a low speed device
(because the pull-up resistor is on D-). After 3 ms, because the host is not yet
allowed to send any data, the device is SUSPENDED because of not seeing any
activity on the data lines. Shortly afterwards, the host issues a RESET which in this
case lasted 31 ms.
It then performs a Get Descriptor request (to the default address 0), to discover the
maximum packet size defined for the control endpoint. Having got this information,
[Link] (1 of 13) [13.03.2008. 5:17:34 Peca]

Forward

USB Made Simple - Part 5

it resets the device again, and then sends a Set Address request, setting device
address to 1 in this example.

All the requests from now on are sent to device address 1. The host has now
requested the device descriptor. (Click here to see analysis of the device descriptor
in a separate window). It has also requested the first nine bytes of the configuration
descriptor collection. Remember that when the host requests the configuration
descriptor, it will also get, following it, the interface and endpoint descriptors and
maybe others. But the host doesn't know the total length of this collection. It does,
however, know that the configuration descriptor itself is 9 bytes long, and that this
descriptor contains a value for the total length of the descriptor collection. So it
starts by requesting just the configuration descriptor using Get Configuration
Descriptor with a required length of 9. (Click here to see an analysis of the
configuration descriptor collection in a separate window.)
Additionally the host has requested String Descriptor 0, for the list of supported
string languages, and descriptor index 2, which in this case has been assigned to
the product description string.

[Link] (2 of 13) [13.03.2008. 5:17:34 Peca]

USB Made Simple - Part 5

After this it can be seen that the host has asked for much of the information again.
This can occur for various reasons, such as the different drivers in the stack each
asking the same questions for their own purposes.
The host has then sent the Set Configuration request, and from that point in time,
the device is 'configured' and is able to perform its purpose in life. The host is now
able to send the HID class request 'Set Idle', to tell the device only to respond to an
interrupt IN transaction if a new event occurs. (In any case when an IN request is
sent and there is no change to report, the device will respond with a NAK packet.
The analyser has been set not to display these 'NAKed' transactions so we will not
see them here) It then also requests the HID class report descriptor, which in this
case informs the appropriate driver to expect to receive a 4 byte report of mouse
events on its interrupt IN endpoint.

At this point the driver starts sending interrupt IN requests, and when any event is
available to be reported the interrupt data transfer succeeds and 4 bytes of data
are transferred.
You may notice that there was a nearly 3 second gap before the mouse was first
moved. In the meantime there were still regular interrupt IN transactions, which
were NAKed by the device, as it had nothing to report. To avoid the display being
swamped by these NAKed transactions, the analyser has been set not to display
them.

[Link] (3 of 13) [13.03.2008. 5:17:34 Peca]

USB Made Simple - Part 5

We are now going to examine one of these control transfers in more detail. By clicking a
button on the analyser we can reveal the transactions which make up a particular control
transfer. We will select a Get Device Descriptor control transfer to examine.

As you can see the Get Device Descriptor is made up, in this case, of five transactions.
The first transaction (SETUP) comprises the setup stage.
The next three (IN) transactions represent the data stage, during which the 18 bytes of
the descriptor are transferred back to the host.
The final (OUT) transaction is the status stage, to acknowledge successful completion.

[Link] (4 of 13) [13.03.2008. 5:17:34 Peca]

USB Made Simple - Part 5

Each of these transactions is made up of packets. By selecting some of these

transactions to expand, we can see the details of the packets. Notice that we have
chosen to expand only one of the three IN transactions, to keep the picture smaller.
Notice how, for example the SETUP transaction is made up of three packets. You
should be able to recognise the component parts of each packet by now.

[Link] (5 of 13) [13.03.2008. 5:17:34 Peca]

USB Made Simple - Part 5

A good analyser will be able to show you the data from the descriptor in an analysed
form, to save you the trouble. (Click here to see analysis of the device descriptor in a
separate window).

Device Descriptor Analysis

In a similar way, we might have selected the configuration descriptor set to display in
analysed form. The interface descriptor tells us, and the host, that it is a HID class
device. The bInterfaceSubClass is usually 0 in HID class devices except for a mouse or
a keyboard which meet the simplified protocol requirements for being operated by the
BIOS code, before the usual USB drivers have been loaded. In this case we notice that
the device will work with the boot interface and (from bInterfaceProtocol) that it is a
mouse. The usual HID driver will learn about this in another way; by parsing the HID
report descriptor.

[Link] (6 of 13) [13.03.2008. 5:17:34 Peca]

USB Made Simple - Part 5

We notice that this device has a single Interrupt IN endpoint in addition to the default
control endpoint, and that it is set to be interrogated once every 10 ms and expects the
host to read 4 bytes each time. (Click here to see a fuller analysis of the configuration
descriptor collection in a separate window.)
As with any HID device the descriptor following the interface descriptor is the HID
descriptor whose main job is to tell the host where to find the HID Report Descriptor.
This is the means by which the device can specify what it is and the detailed content of
reports it may send and/or receive.

[Link] (7 of 13) [13.03.2008. 5:17:34 Peca]

USB Made Simple - Part 5

[Link] (8 of 13) [13.03.2008. 5:17:34 Peca]

USB Made Simple - Part 5

We have learned that when the device has been configured, the host will start an IN
interrupt (on endpoint 1 IN) to read a report or reports up to 4 bytes long at intervals of
10 ms (or possibly less, typically it would be 8 ms with Windows).

So once the mouse is configured, the picture above represents the communication
channels possible in the mouse. It still has the bi-directional control endpoint, and can
receive the six HID class requests shown in the picture, and it has the interrupt IN
endpoint, for sending its reports of mouse events. Typically, and in our example here,
the only class request out of the 6 defined HID requests which is used, is the Set Idle
request.
We can now select the Get HID Report descriptor to analyse. The display below shows
the contents and significance of the HID report descriptor, which, using a form of coding
specified in the HID specification, defines one or more reports which are the means of
transferring information to or from a HID device. The parsed form of the report desriptor
is shown below. Parsing a HID report is a fairly complex operation, so the analyser has
helped out by displaying the defined reports, or in this case the one report defined. It is
a single Input report, with 5 buttons, and X and Y movement, and a wheel movement,
which make up a total of 4 bytes to match the maximum size of the interrupt endpoint.

[Link] (9 of 13) [13.03.2008. 5:17:34 Peca]

USB Made Simple - Part 5

[Link] (10 of 13) [13.03.2008. 5:17:34 Peca]

USB Made Simple - Part 5

To complete the picture we now examine the content of one of the interrupt transfers
(events 4103 to 4105 in the first sequence we looked at). The bytes transferred were 00
01 00 00. If we select the Interrupt Transfer to analyse we will see how the meaning
matches up with the parsed report descriptor analysis.

We can tell from this that the mouse moved 1 unit in the X direction.

What have we not examined?

We have not seen any Keep Alive signals in our displays. There is at least one Keep
Alive signal every frame, or every 1 ms. This makes for a very cluttered display, even
when they are grouped together by the analyser. So we have set the analyser not to
display the Keep Alive pulses for the sake of this discussion.
In the same way, the interrupt IN transfers have only been shown when successful.
Every 8 ms there was an attempt to perform an IN transaction by the host, and most of
these were NAKed by the device. The analyser was set not to shown these NAKed
transactions.
Here is a section of the capture containing these elements with the analyser set to
reveal them. In addition, one of the NAKed interrupts has been expanded so you can
see what it is built up from.

[Link] (11 of 13) [13.03.2008. 5:17:34 Peca]

USB Made Simple - Part 5

Summary
We have examined the actual transfers involved
during enumeration and operation of a real
device, a mouse.

Coming up...
Next we will look at High Speed USB which uses
a data rate of 480 Mb/s on the same cables as
Full Speed USB.
Copyright 2006-2007 MQP Electronics Ltd

[Link] (12 of 13) [13.03.2008. 5:17:34 Peca]

Forward

USB Made Simple - Part 6

Index
Part 1
Part 2
Part 3
Part 4
Part 5
Part 6
Part 7
Links

Part 6 - High Speed Basics

Back

Introduction to High Speed

USB
As mentioned before, the high speed
additions to the specification were
introduced in USB 2.0 as a response to
the higher speed of Firewire.
As High Speed was added as an
afterthought, and had to maintain
compatibilty without compromising
performance, we have left the description
of High Speed until we had covered the
basics of the original specification.

Subjects covered in this part...

Data Transmission
Packet Sync
End of Packet
Compatibility
Negotiating High Speed
Frames and Microframes

[Link] (1 of 9) [13.03.2008. 5:18:04 Peca]

New Packet Identifiers

High Speed Hubs with Full and
Low Speed Devices
Reset
Suspend
Resume
Detecting a Device Unplug

Forward

USB Made Simple - Part 6

Data Transmission
The data rate achieved by High Speed is 480
Mb/s. This needs to be transmitted down cables
which were originally specified for a 12 Mb/s
transmission rate,
To achieve this, when the link is conveying high
speed data, each end of D+ and each end of Dis terminated with a 45 Ohm resistance to
ground.

Data is sent by steering a current of 17.78 mA

(derived from the positive supply) into either the
D+ or the D- line. This results in a voltage of
400mV on the line being fed with current. The
differential state of the line is detected at the
receiving end by a differential receiver. This
arrangement is able to reliably receive data sent
at 480 Mb/s.
In fact the 45 Ohm resistors are provided by the
Full Speed / Low Speed driver, at each end of
the link, applying a Single Ended Zero. The FS/
LS driver is designed to provide as accurate a
termination resistance as possible. By switching
off the high speed transceiver current source,
the line conditions are as defined for full speed /
low speed.

Basics of High Speed Transmission

In addition to the differential receiver, there is

also a 'transmission envelope detector' and a
'differential envelope detector'.
The transmission envelope detector produces a
'squelch' signal if there is less than 100uV
between the data lines, which means that there
is no data being received.
The differential envelope detector detects if the
far end has been unplugged, as the differential
voltage will double to about 800 mV if the far end
terminating resistors are not present.
(Further down the page you will see how this is
used by the host to detect the unplugging of a
high speed device.)

[Link] (2 of 9) [13.03.2008. 5:18:04 Peca]

Further Detail in the Specification

Fig 7-1 in Chapter 7 of the USB
Specification V2.0 shows a complete
suggested circuit for a low / full / high
speed transceiver.

USB Made Simple - Part 6

Packet Sync
Just prior to the
packet sync, both
data lines are low.
The sync is sent
using the NRZI
sequence
KJKJKJKJKJKJKJKJ
KJKJKJKJKJKJKJKK.

High Speed Synchronisation Pattern

A hub may drop up

to 4 bits from the
sync pattern. After 5
hubs the sync field of
a packet may be only
12 bits long.

End of Packet
On a low or full speed link, a brief Single Ended
Zero (SE0) state is used to indicate End Of
Packet (EOP), and idle is indicated by a J
condition.

At high speed, the EOP is indicated, by

deliberately sending a byte which contains a bitstuffing error; '01111111'. This applies at the end
of all packets except SOF.

On a high speed link, the idle state is effectively

a SE0, so that state is not available to indicate
EOP, and a different method for indicating the
end of packet is used. During normal data
transmission there can not be a run of more than
6 1's in a row, because a 0 is automatically
inserted (and will be removed on reception). This
guarantees that there will be sufficient transitions
in the NRZI encoded data stream to allow clock
recovery.

Each high speed SOF packet is terminated with

5 NRZI bytes containing bit-stuffing errors:
01111111 - 11111111 -11111111 -11111111 11111111. This pattern allows the 'disconnection
envelope detector' to detect a rise in data
amplitude above 625 mV, in the event that the
device, along with its termination resistors, has
been unplugged.

Compatibility
Care has been taken to provide as much compatibity
between high speed and full / low speed host, and high
speed and full / low speed devices. USB is a plug-andplay system and must not confuse the user.
So a low or full speed device will always work with a
high speed capable host.

[Link] (3 of 9) [13.03.2008. 5:18:04 Peca]

A high speed device will always work with

a full / low speed host, at least to the extent
of communicating its identity and
capabilities to the host (which it can do at
full speed). The host will then be in a
position to report to the user if they have a
device which relies on high speed
bandwidth to provide any functionality.

USB Made Simple - Part 6

Negotiating High Speed

To maintain the required compatibility, a high
speed device will always present itself initially
as a Full Speed device (by a 1.5K pullup
resistor on D+).

The negotiation for High Speed takes place

during the Reset, which is, as we remember,
the first thing a host must do to a device before
attempting data communication.
The high speed detection handshake is initiated
by the device.
The hub will respond to it, if it is high speed
capable.

What the device does

Hub Response

The device leaves its D+ 1.5K pullup resistor

connected, and does not terminate the lines with
45 Ohm resistors as it would for high speed. But
it drives high speed current (17.78mA) into the
D- line for at least a millisecond. Now, remember
that the hub is applying a reset condition to the
lines, so effectively is already terminated as for
high speed data. As only one end of the link is
terminated, the hub will see about 800 mV on
D-. This condition is called a K-chirp.

If the hub is high speed capable then it will

monitor the K-chirp from the device until it sees
it completing. It must, within 100us, send a
series of K-J chirp pairs to the device. This
means that it will inject 17.78 mA alternately into
the D- and the D+ lines. Each of these chirps
lasts around 50us, and there are no gaps
between them. The device has to see at least 3
chirp pairs before assuming that the hub is high
speed capable.

A full / low speed hub will pay no attention to this

condition, but a high speed hub will detect it
using its differential receiver and the absence of
a squelch signal.

Switching to High Speed

If the hub does not respond, then the rest of the

reset, and subsequent data transmissions will
take place as is normal for a full speed device.

[Link] (4 of 9) [13.03.2008. 5:18:04 Peca]

At this point the device disconnects its 1.5K

pullup resistor, applies the 45 Ohm high speed
terminations (using its full speed data driver in
SE0 mode), and is thus in a state to perform
high speed data transmission and reception.
The hub will continue to send chirp pairs up until
100 - 500 us before the end of reset, and the
device will monitor these chirps. At the point in
time when the device termination is applied, the
amplitude of the chirp signals, viewed on an
oscilloscope would be seen to halve in
amplitude from 800mV to 400mV.

USB Made Simple - Part 6

Frames and Microframes

The 1 ms frame rate in full speed / low speed
USB, is used for a number of purposes, such
as scheduling access to the bus, and as a
timing reference for interrupt and isochronous
transfers.
For high speed, a higher frame rate was
deemed appropriate, while still maintaining a
relationship with the existing 1 kHz rate.
To this end, high speed uses the 'Microframe'
which is 125ms long (8 Microframes per
millisecond). The correspondence with the
1ms frame numbering is maintained in the
high speed SOF packets by repeating each
frame number in 8 successive Microframes.

Packet Length
The maximum length of packets was
increased for high speed, see table to right.

Packets per (Micro)frame

At high speed it is possible to specify up to 3
isochronous or interrupt transfers per
microframe, rather than the 1 transfer per
frame of full speed; giving a maximum
possible isochronous or interrupt transfer rate
of 192 Mb/s.

Max Packet Size

Transfer
Type

LS

FS

HS

Control

8, 16, 32,
64

64

Bulk

8, 16, 32,
64

512

up to 8

up to 64

up to 1024

up to 1023

up to1024

Interrupt
Isochronous

New Packet Identifiers

Some new PIDs were added for high speed,
partly to overcome some inefficiencies which
were recognised in the full speed protocol, and
partly to support new isochronous transfer
features, and the new requirement for 'split
transactions' (more about this below).
The identifiers which are designed to overcome
some inefficiencies and improve bandwidth
usage at high speed are:

PID Type

Token

Data

NYET
PING
Handshake

Identifiers which allow for control over multiple

isochronous packets per microframe are:

DATA2
MDATA

[Link] (5 of 9) [13.03.2008. 5:18:04 Peca]

PID Name

PID<3:0>*

OUT

0001b

IN

1001b

SOF

0101b

SETUP

1101b

DATA0

0011b

DATA1

1011b

DATA2

0111b

MDATA

1111b

ACK

0010b

NAK

1010b

STALL

1110b

NYET

0110b

PRE

1100b

ERR

1100b

USB Made Simple - Part 6

Special

The identifiers added to assist with split transfers

are:

SPLIT
ERR

SPLIT

1000b

PING

0100b

Reserved

0000b

* Bits are transmitted lsb first

High Speed Hubs with Full and

Low Speed Devices
In the original USB, there was a built-in
inefficiency, in that the whole bus was held up,
waiting for low speed transactions to take place.
Having gone to the trouble of increasing the data
rate to 480 Mb/s, it would have been wasteful if
this situation was perpetuated, so a different
approach was taken.
All communication to a USB V2.00 hub takes
place at high speed, even when it contains traffic
for low or full speed devices. In a tree of high
speed hubs the packets are tranferred down the
tree at high speed as far as the hub, to whose
port a low or full speed device is connected.

The hub in question assigns special control

circuitry within itself, to take over the role of
communicating with the low or full speed
segment of the bus; initiating the transactions,
getting the response back from the device, and
finally communicating the result back to the host
at high speed.
The mechanism needed to deal with this, without
holding up the high speed segments of the bus,
involves splitting every low or full speed
transaction into 2 stages; the request from the
host, and the eventual response from the device.
The host communicates its requirements with the
high speed hub using a new packet with a SPLIT
identifier. This packet can define a Start Split, or
a Complete Split transaction.
The actual sequence of packets in these two
types of transaction is very dependent on the
transfer type and direction.

This diagram illustrates how full speed and low speed traffic is
kept separate from high speed traffic. The blue lines carry only
high speed traffic, and the red lines only full or low speed. Any
traffic directed at full or load speed devices passes through the
high speed section, as high speed split transactions.

[Link] (6 of 9) [13.03.2008. 5:18:04 Peca]

USB Made Simple - Part 6

Reset
The host will maintain its
SE0, but not send any data,
when it wants to Reset the
device. The device will
initially see a SE0 (with no
data activity) and will not be
able to distinguish this
condition from a Suspend.
After, at the latest, 3.125 ms
of this condition the device
must revert to full speed
termination itself, and then
test whether it sees SE0 or
Idle. If it sees SE0 then it
knows it is being reset, and
will procede with the chirp
handshake described above.

[Link] (7 of 9) [13.03.2008. 5:18:04 Peca]

USB Made Simple - Part 6

Suspend
A high speed host suspends
a device by reverting to a Full
Speed idle state. Again the
device will initially see a SE0
(with no data activity) and will
not be able to distinguish this
condition from a Reset.
After, at the latest, 3.125 ms
of this condition the device
must revert to Full Speed
termination itself, and then
test whether it sees SE0 or
Idle. If it detects idle it must
assume that it it being
suspended, and must go to
its lower power suspended
mode.
Note that both ends of the
link must remember that they
were in high speed mode, so
that when Resume takes
place, no high speed
handshake is required.

Resume
As for full / low speed, the
Resume is signalled by a K
state for 20 ms. When the
link was previously in high
speed mode, the resume is
completed by a transition
back to SE0 at the end of the
resume, and both host and
device must be in high speed
terminated mode within 2 low
speed bit times.

Detecting a Device Unplug

See the description of the EOP signal above.

[Link] (8 of 9) [13.03.2008. 5:18:04 Peca]

USB Made Simple - Part 7

Index
Part 1
Part 2
Part 3
Part 4
Part 5
Part 6
Part 7
Links

Back

Part 7 - High Speed Transactions

High Speed USB Transactions

We have looked at the mechanisms for
communicating at 480 Mb/s. We now examine
the packet formats in more depth, and then see
how these are used to provide the various
transactions: Control, Interrupt, Bulk and
Isochronous.

[Link] (1 of 19) [13.03.2008. 5:18:42 Peca]

Forward

USB Made Simple - Part 7

Subjects covered in this part...

The New Packets

Token Packet
Data Packet
Handshake Packet
SOF Packet
Split Packet

High Speed Bulk Transactions

Ping Protocol
High Speed Isochronous Transfers
High Speed Interrupt Transactions
High Speed Control Transfers

Packet
Identifier

DATA2
MDATA
NYET
ERR
SPLIT
PING

The table on the right gives brief details of the

purposes of the six new packet types.
[Link] (2 of 19) [13.03.2008. 5:18:42 Peca]

Usage
This data packet token has been added as part of a
system for controlling multiple isochronous IN
packets during one microframe at high speed.

The New Packets

The packet types which have been added for
high speed are:

Split Transactions
Split Bulk OUT Transaction
Split Bulk IN Transaction
Periodic Split Transactions
Split Interrupt OUT Transaction
Split Interrupt IN Transaction
Split Isochronous OUT Transaction
Split Isochronous IN Transaction

DATA2

For each isochronous IN packet requested, the

suffix of the DATAx PID represents the remaining
number of packets to be transferred during the
current micro-frame.
This data packet token has been added as part of a
system for controlling multiple isochronous OUT
packets during one microframe at high speed.

MDATA

All but the last packet sent during a microframe use

the MDATA PID. The last packet sent uses DATA0,
DATA1 or DATA2 depending on whether one, two
or three packets were sent.

USB Made Simple - Part 7

NYET handshake packets are used in 2 different

high speed situations.

NYET

One use is when a hub wishes to respond to a

'complete split' transaction to say that it is has not
yet been completed.
The other use is during the high speed control or
bulk OUT PING protocol. It means that the
endpoint has accepted the data but is not yet ready
for further data.

ERR

Used during the high speed split protocol by a hub,

to indicate that there was an error on the full or low
speed bus.
This packet introduces a Start Split transaction or a
Complete Split transaction.

SPLIT

PING

Split transactions are used at high speed to

communicate to a hub, the details of a low or full
speed transaction which it is expected to handle,
and to get back the results.
A PING packet is used on high speed Control and
Bulk OUT endpoints. They may be sent by the host
to establish whether the endpoint is ready to accept
a DATA0 or DATA1 packet, and will result in an
ACK or a NAK packet from the device.
This is an efficiency improvement, as at full or low
speed, the data packet has to be sent in full, and
only then can the endpoint respond with a NAK.

For the sake of completeness we will now look at

the packet formats for all the available packets
type, including the ones already covered for low
and full speed.
There are five different packet formats based on
which PID the packet starts with.
[Link] (3 of 19) [13.03.2008. 5:18:42 Peca]

USB Made Simple - Part 7

Token Packet
Sync

The token packet contains two addressing elements:

PID

ADDR

ENDP

CRC5

8 bits

7 bits

4 bits

5 bits

EOP

Used for SETUP, OUT, IN and PING packets.

They are always the first packet in a transaction,
identifying the targeted endpoint, and the
purpose of the transaction.
(The SOF packet is also defined as a Token
packet, but has a slightly different format and
purpose, which is described below.)

Data Packet
Sync

PID

DATA

CRC16

8 bits

(0-1024)
x 8 bits

16 bits

EOP

Used for DATA0, DATA1, DATA2 and MDATA

packets. If a transaction has a data stage this is
the packet format used.

[Link] (4 of 19) [13.03.2008. 5:18:42 Peca]

Address (7 bits)

This device address can address up to 127 devices. Address 0 is

reserved for a device which has not yet had its address set.
Endpoint number (4 bits)

There can be up to 16 possible endpoints in a device in each

direction. The direction is implicit in the PID. OUT, SETUP and
PING PIDs will refer to the OUT endpoint, and an IN PID will refer
to the IN endpoint.

DATA0 and DATA1 PIDs are used in Low and Full speed
links as part of an error-checking system. When used, all
data packets on a particular endpoint use an alternating
DATA0 / DATA1 so that the endpoint knows if a received
packet is the one it is expecting. If it is not it will still
acknowledge (ACK) the packet as it is correctly received,
but will then discard the data, assuming that it has been resent because the host missed seeing the ACK the first
time it sent the data packet.
DATA2 (along with DATA1 and DATA0) is used in High
Speed links as part of a system for controlling multiple
isochronous IN packets during one microframe at high
speed.
MDATA (along with DATA2, DATA1 and DATA0) is used
in High Speed links as part of a system for controlling
multiple isochronous OUT packets during one microframe
at high speed.

USB Made Simple - Part 7

ACK

Receiver acknowledges receiving error free packet.

Handshake Packet

NAK
Sync

PID

EOP

Receiving device cannot accept data or transmitting

device cannot send data.

8 bits

STALL

Used for ACK, NAK, STALL and NYET packets.

This is the packet format used in the status stage
of a transaction, when required.

Endpoint is halted, or control pipe request is not supported.

NYET

No response yet from receiver (high speed only).

SOF Packet
Frames and Microframes
Sync

PID

Frame
No.

CRC5

8 bits

11 bits

5 bits

EOP

The Start of Frame packet is sent every 1 ms on

full speed links. The frame is used as a time
frame in which to schedule the data transfers
which are required. For example, an isochronous
endpoint will be assigned one transfer per frame.

[Link] (5 of 19) [13.03.2008. 5:18:42 Peca]

On a low speed link, to preserve bandwidth, a Keep Alive

signal is sent every millisecond, instead of a Start of
Frame packet. In fact Keep Alives may be sent by a hub
on a low speed link whenever the hub sees a full speed
token packet.
At high speed the 1 ms frame is divided into 8
microframes of 125 us. A SOF is sent at the start of each
of these 8 microframes, each having the same frame
number, which then increments every 1 ms frame.

USB Made Simple - Part 7

Control, Interrupt or Bulk Endpoints

SC

Start / Complete

0 = Start, 1 = Complete

Speed

0 = Full, 1 = Low

not used

Endpoint Type

00 = Control
01 = Isochronous
10 = Bulk
11 = Interrupt

ET
Split Packet

Sync PID

Hub
SC Port S
Addr

E ET CRC5 EOP

8
7 1 7
1 1 2
bits bits bit bits bit bit bits

5
bits

Isochronous Endpoints
SC

Start / Complete

0 = Start, 1 = Complete
00 = HS data is middle of
FS data payload

The SPLIT packet is the first packet in either a

Start Split transaction or a Complete Split
transaction, sent to a high speed hub when it is
handling a low or full speed device.

01 = HS data is end of FS
data payload
S and E

Start and End

10 = HS data is start of FS
data payload
11 = HS data is all of FS
data payload

ET

[Link] (6 of 19) [13.03.2008. 5:18:42 Peca]

Endpoint Type

00 = Control
01 = Isochronous
10 = Bulk
11 = Interrupt

USB Made Simple - Part 7

High Speed Bulk Transactions

These work much like full speed bulk
transactions. The differences are:

The maximum packet size can only be

specifed as 512 bytes.
A more efficient method of doing OUT
transfers has been introduced, involving
the use of PING transactions, and the new
NYET handshake packet.

Ping Protocol
Before sending an OUT transaction to a bulk
endpoint, the host controller may send a PING
packet. The response to the PING will be ACK or
NAK. ACK means that the endpoint is ready to
accept an OUT transaction of maximum packet
size for the endpoint. NAK means it is not. The
host may continue to send PINGs after a NAK, or
it may choose to wait for a number of
microframes before re-trying.
On high speed Bulk OUT endpoints, the endpoint
descriptor value bInterval is required to specify
the NAK rate of the endpoint. The specification is
misleading when it defines this value. The value
represents the number of microframes which the
host would have to wait, after receiving a NAK in
response to a PING, before a further PING is
guaranteed to elicit an ACK response. 0 means
the endpoint never NAKs.
A further new feature is that, after a successful
OUT transaction, the endpoint may respond with
ACK to indicate that it is already prepared to
accept a further packet, or NYET to indicate that
it received the data correctly, but is not yet ready
to accept further data.
[Link] (7 of 19) [13.03.2008. 5:18:42 Peca]

BULK Transfer Error Control Flow

USB Made Simple - Part 7

High Speed Isochronous

Transfers
High speed isochronous transfers have an
extended control system to allow up to 3
isochronous transactions per microframe,
allowing a data rate of up to 192 Mb/s. (The
specification refers to an isochronous endpoint
with more than 1 packet per microframe as a
'High Bandwidth' endpoint.)
As isochronous transactions do not include a
[Link] (8 of 19) [13.03.2008. 5:18:42 Peca]

USB Made Simple - Part 7

acknowledgement (handshake) packet, a system

was specified which allows the recipient to be
aware, if a data packet is lost, which one it was*.

Isochronous IN Protocol

Isochronous IN
When requesting IN transactions, the device
packages them in DATA2, DATA1 or DATA0
packets, depending on how many packets per
microframe are specified, and which one it is.
See diagram on the right.

Isochronous OUT Protocol

Isochronous OUT
When sending OUT transactions, the host
packages them either in MDATA or DATA0
DATA1 or DATA2 packets, as shown in the
diagram on the right

*In fact this does not apply to Isochronous OUT

transfers with three transactions per microframe
because the first two packets use MDATA as a
PID.

[Link] (9 of 19) [13.03.2008. 5:18:42 Peca]

USB Made Simple - Part 7

Interrupt Transfer Error Control Flow

High Speed Interrupt

Transactions
These work much as for full and low speed,
except that the endpoint descriptor may specify
up to 3 interrupt transactions per microframe,
allowing a data rate of up to 192 Mb/s. (The
specification refers to an interrupt endpoint with
more than 1 packet per microframe as a 'High
Bandwidth' endpoint.)
The protocol is the same as for full and low
speed, in that DATA0 and DATA1 packets are
alternated.
If more than one transaction per frame is
specified and a transaction is NAKed, then the
host should attempt no more transactions on that
endpoint within the same microframe.

Error Control Flow

SETUP STAGE

High Speed Control Transfers

High speed control transfers are the same as full
and low speed transfers, with the following
exceptions:

For high speed control endpoints, the

[Link] (10 of 19) [13.03.2008. 5:18:42 Peca]

USB Made Simple - Part 7

maximum packet size for the data stage is

fixed at 64 bytes.
During a data phase, using an OUT PID,
the transaction follows the rules above for
a Bulk OUT transaction, including the
PING protocol.

Notice that it is not permitted for a device to respond to a SETUP

with a NAK, STALL or NYET.
DATA STAGE

(same as for high speed bulk transfer)

STATUS STAGE

(same as for bulk transfer)

Transaction Translator

Split Transactions

This is a much simplified diagram of the transaction translator

within a high speed hub.

Transaction Translator
Split transactions form a high speed only protocol
between host controllers and high speed hubs,
which are handling full or low speed traffic. This
is the means by which the full or low speed traffic
is prevented from degrading the high speed bus
performance. Each high speed hub is required to
have a 'Transaction Translator' which handles
the full or low speed transactions to particular
ports.

Packet Sequences

Example Split Transaction Sequence

To overcome the difference in speeds, each full

This illustrates a typical split transaction sequence; in this case a

[Link] (11 of 19) [13.03.2008. 5:18:42 Peca]

USB Made Simple - Part 7

or low speed transaction is split into two parts on

the high speed bus; a Start Split Transaction and
a Complete Split Transaction. In between, the
Transaction Translator independently handles
the full or low speed transaction.

split interrupt IN.

The illustration to the right shows how the host

defines the required full or low speed transaction,
allows the hub to deal with the transaction at that
speed and then checks back later for the result.
The actual packet sequence in a split transaction
varies, depending on the type of transaction
involved, and will be described in more detail, in
the following paragraphs.
This will be a brief summary of what to expect,
however the full USB specification contains
masses of detail which would be needed to
actually implement a host controller or hub.

Split Bulk OUT Transaction

For a split bulk OUT transaction, the host sends
a Start Split packet specifying the hub address
and port number, and speed (full), with the
'endpoint type' set to 'bulk'.
Split Bulk OUT Transaction
It then sends an OUT token packet to identify the
full speed transaction type, device and endpoint
address, followed by the data packet for onward
transmission to the target endpoint.
If the Transaction Translator receives these
packets correctly and has an available buffer, it
will respond with an ACK packet, otherwise it will
[Link] (12 of 19) [13.03.2008. 5:18:42 Peca]

USB Made Simple - Part 7

respond with a NAK packet.

At this point the Transaction Translator proceeds
to send this transaction to the specified full speed
endpoint, and to get a response, which it stores,
ready to pass back to the host during a Complete
Split transaction.
When the host controller decides it is
appropriate, it checks for the result by issuing a
Complete Split transaction. The Complete Split
packet and the OUT token packet are the same
as for the Start Split transaction, except for the
Start/Complete bit of the packet. This allows the
hub to verify which transaction it is to present the
results of.
If the full speed transaction is not yet completed,
the hub will respond with a NYET packet and the
host will attempt the Compete Split transaction
sometime later. If it is complete, then a packet
containing the actual result (ACK, NAK or
STALL) is sent to the host.

Split Bulk IN Transaction

Split Bulk IN Transaction

A split bulk IN transaction is similar to a split bulk
OUT, except that the DATA0/1 packet occurs
during the Complete Split stage because it is now
part of the response. If data was returned then
the data packet appears in lieu of a handshake
packet.
The actual ACK of the full speed transaction was
[Link] (13 of 19) [13.03.2008. 5:18:42 Peca]

USB Made Simple - Part 7

sent by the Transaction Translator to the device

on receiving the DATA0 or DATA1 packet from
the device, and the host will keep track of which
data it has received. If it fails to receive a valid
packet it will retry the Complete Split Transaction.

Periodic Split Transactions

Interrupt and Isochronous
Split Interrupt and Isochronous transactions
need special treatment because of the
bandwidth guarantee which they offer. Every
full / low speed device using these endpoint
types, which is added to the bus, is given a
guaranteed allocation on the full / low speed bus,
which also has to be conveyed on the high
speed bus. The mechanism is complex and
described in full in chapter 11 of the USB
specification.
The result of this mechanism is that periodic
[Link] (14 of 19) [13.03.2008. 5:18:42 Peca]

So a longer transaction, which can be up to 1023

bytes long for an isochronous endpoint, will be
transferred in multiple Start Split transactions for
OUT transfers, or multiple Complete Split
transactions for IN transfers.
Furthermore, on periodic IN transactions, data
received by the hub from the device within a
given microframe is sent in response to a
Complete Split, even though it represents only
part of the full or low speed data packet. A full
speed interrupt IN packet of, say, 64 bytes,
which spans two microframes, will be conveyed
in up to two Complete Splits. The data packet in
the first part will use an MDATA PID to indicate
that it is not complete.

USB Made Simple - Part 7

transactions on the full or low speed bus are

transferred in packets of data of at most 188
bytes.

Isochronous IN transactions are similar, but can

span up to 6 microframes.

Split Interrupt OUT Transaction

Split Interrupt OUT Transaction

For an interrupt OUT transaction the maximum
data size is 64 bytes, so these are always
transferred within a single Start Split transaction.
There is no need for a handshake packet from
the hub, as the interrupt will not be retried in the
same frame, if missed.

As with a Bulk transaction, the Complete Split will

elicit a handshake packet from the device if it
responded with one itself, or with ERR if there
was an error on the full or low speed bus, or
NYET if the Transaction Translator has not
completed its task.

[Link] (15 of 19) [13.03.2008. 5:18:42 Peca]

USB Made Simple - Part 7

Split Interrupt IN Transaction

Split Interrupt IN Transaction

For a split interrupt IN, we issue a Start Split
(without expecting a handshake from the hub).
The Complete Split may result in the complete
data packet (DATA0 / DATA1) being returned,
but it is also possible that during the microframe,
only part of the data was so far collected, in
which case an MDATA PID is used to indicate
that a further Complete Split will be necessary in
the next microframe.
Handshake packets NAK or STALL will indicate
that the device responded with this packet. ERR
indicates that there was an error on the full or low
speed bus. NYET indicates that the Transaction
Translator has not completed its task.

[Link] (16 of 19) [13.03.2008. 5:18:42 Peca]

USB Made Simple - Part 7

Split Isochronous OUT

Transaction

Split Isochronous OUT Transaction

On the face of it, this is the simplest split

transaction. There is no handshake from the hub,
and no Complete Split, as the delivery of
isochronous transactions is not checked.
It is complicated by the fact that single full speed
isochronous transactions are divided up into
separate Start Split transactions of at most 188
bytes each, sent one per microframe.
The Start Split packet has its Start and End bits
set to identify which part of the payload is being
transferred.

Split Isochronous IN Transaction

Split Isochronous IN Transaction

For a split isochronous IN, we issue a Start Split
(without expecting a handshake from the hub).
The Complete Split may result in the complete
data packet (DATA0) being returned, but it is also
possible that during the microframe, only part of
the data was so far collected, in which case an
MDATA PID is used to indicate that a further
Complete Split will be necessary in the next
microframe.
ERR indicates that there was an error on the full
or low speed bus. NYET indicates that the
Transaction Translator has not completed its task.
[Link] (17 of 19) [13.03.2008. 5:18:42 Peca]

USB Made Simple - Part 7

Summary
We have examined how each type of data
transfer works at high speed.

Coming up...
Forward
...to be announced...
Copyright 2006-2007 MQP Electronics Ltd

[Link] (18 of 19) [13.03.2008. 5:18:42 Peca]

USB Made Simple - Books and Links

Index
Part 1
Part 2
Part 3
Part 4
Part 5
Part 6
Part 7
Links

USB - Books and Links

USB Complete
(Third Edition)
by Jan Axelson
ISBN: 1931448027
If you buy just one
book on USB then
choose this one.

USB Mass
Storage:
Designing and
Programming
Devices and
Embedded
Hosts
by Jan Axelson
ISBN: 1931448043
Another great book
from the same
author

[Link] (1 of 4) [13.03.2008. 5:19:25 Peca]

USB Made Simple - Books and Links

USB
Implementers
Forum
This is the official
site of the
developers of
USB. All the
important
specifications can
be found on this
site. This is also
the place to
purchase your own
VID. The
developers forum
here is a good
place to get your
questions
answered. (But try
to read up on the
basics first.)

Microchip
Forum
If you want to
develop a USB
device using a PIC
micro-controller
you should join this
forum for
interesting
discussions.

MQP
Electronics
Manufacturer of
the Packet-Master
Series of USB Bus
Analysers.

[Link] (2 of 4) [13.03.2008. 5:19:25 Peca]

USB Made Simple - Books and Links

Lakeview
Research
A mine of
information on all
aspects of USB
development. Well
worth a visit.

The Video
Class
Specification
UVC V1.0
Important Note:
The link above
allows you to
download an
obsolete version
(V1.0) of the USB
Video Class
specification
released in
September 2003.
For the latest
specification
please visit the
official USB
Implementers
Forum site. The
reason we are
hosting, for
reference only, this
obsolete version
here is that the
USB-IF site only
has the new
specification and it
appears that the
Windows
implementation is
based on this older
one. Our reading
of the licence
conditions in the
document lead us
to believe that
distribution of the

[Link] (3 of 4) [13.03.2008. 5:19:25 Peca]

USB Made Simple - Books and Links

document is
acceptable and it
has been put here
in good faith, but
we are fully
prepared to
remove this link if
any connected
party has an
objection.

UVC V1.0a
As above but
V1.0a (2004
version with FAQ
Rev 1.0c).

[Link] (4 of 4) [13.03.2008. 5:19:25 Peca]