Searching \ for '[EE] USB routing and termination' in subject line. ()
Make payments with PayPal - it's fast, free and secure! Help us get a faster server
FAQ page: www.piclist.com/techref/index.htm?key=usb+routing+termination
Search entire site for: 'USB routing and termination'.

Exact match. Not showing close matches.
PICList Thread
'[EE] USB routing and termination'
2007\01\31@121726 by alan smith

picon face
I was given a project that was shelved some time ago, and the engineer that was working it is long gone.  No PIC, has a PSoC on it, being used as a USB bridge among other things.  It was routed as just signals, across the board, so no wonder it doesnt work. I assume that it should be routed as diff pair for one thing, but anyone know off the top of thier head what the route length restrictions are?  Also, the design has a 120ohm transformer in line (assume for EMI?) as well as a 27.4ohm resistor in series with the D+/D- lines.  All the designs I've seen just come from the chip direct to to the header for the USB connector.  Anyone experianced with putting USB down on a board have a comment?  It is designed to be plugged into just a USB port on a PC

---------------------------------
No need to miss a message. Get email on-the-go
with Yahoo! Mail for Mobile. Get started.

2007\01\31@123654 by David VanHorn

picon face
Parallel tracks, and ideally wth ground right next to them, and below them,
as short as possible.

I wouldn't go as far as to twist them for equal length, but I suppose you
could.

2007\01\31@133009 by Vasile Surducan

face picon face
You didn't say if it's a full speed, high speed or low speed USB.
Differential tracks are important only at high speed, maybe if are
very long routes on full speed too. I'm talking about routing on least
two layer PCB with ground plane, not wiring with wire wrapping or
soldering wires (even those are ok for short connections).  I'm
wondering how many has measured 480Mbit/sec on an USB connection ?

Vasile

On 1/31/07, alan smith <spam_OUTmicro_eng2TakeThisOuTspamyahoo.com> wrote:
> I was given a project that was shelved some time ago, and the engineer that was working it is long gone.  No PIC, has a PSoC on it, being used as a USB bridge among other things.  It was routed as just signals, across the board, so no wonder it doesnt work. I assume that it should be routed as diff pair for one thing, but anyone know off the top of thier head what the route length restrictions are?  Also, the design has a 120ohm transformer in line (assume for EMI?) as well as a 27.4ohm resistor in series with the D+/D- lines.  All the designs I've seen just come from the chip direct to to the header for the USB connector.  Anyone experianced with putting USB down on a board have a comment?  It is designed to be plugged into just a USB port on a PC
>
> ---------------------------------
> No need to miss a message. Get email on-the-go
> with Yahoo! Mail for Mobile. Get started.
> -

2007\01\31@133234 by Herbert Graf

flavicon
face
On Wed, 2007-01-31 at 09:17 -0800, alan smith wrote:
> I was given a project that was shelved some time ago, and the engineer that was working it is long gone.  No PIC, has a PSoC on it, being used as a USB bridge among other things.  It was routed as just signals, across the board, so no wonder it doesnt work. I assume that it should be routed as diff pair for one thing, but anyone know off the top of thier head what the route length restrictions are?  Also, the design has a 120ohm transformer in line (assume for EMI?) as well as a 27.4ohm resistor in series with the D+/D- lines.  All the designs I've seen just come from the chip direct to to the header for the USB connector.  Anyone experianced with putting USB down on a board have a comment?  It is designed to be plugged into just a USB port on a PC

I can't comment on the trace lengths, but on a working board I've done
the lines are routed differentially, and they are terminated with 27R
resistors (both lines) at the chip.

Thanks, TTYL

2007\01\31@133555 by alan smith

picon face
That was my thinking as well....what about the transformer and series resistors? Series resistors are typically in for impedance matching but in this case....routing was so bad that it wouldnt have fixed anything at all.

David VanHorn <.....dvanhornKILLspamspam@spam@microbrix.com> wrote:  Parallel tracks, and ideally wth ground right next to them, and below them,
as short as possible.

I wouldn't go as far as to twist them for equal length, but I suppose you
could.

2007\01\31@135035 by Richard Prosser

picon face
I don't thisnk a transformer is going to work as some of the control
signals are not differential.

RP

On 01/02/07, alan smith <micro_eng2spamKILLspamyahoo.com> wrote:
> That was my thinking as well....what about the transformer and series resistors? Series resistors are typically in for impedance matching but in this case....routing was so bad that it wouldnt have fixed anything at all.
>
> David VanHorn <.....dvanhornKILLspamspam.....microbrix.com> wrote:  Parallel tracks, and ideally wth ground right next to them, and below them,
> as short as possible.
>
> I wouldn't go as far as to twist them for equal length, but I suppose you
> could.
> -

2007\01\31@135335 by David VanHorn

picon face
On 1/31/07, alan smith <EraseMEmicro_eng2spam_OUTspamTakeThisOuTyahoo.com> wrote:
>
> That was my thinking as well....what about the transformer and series
> resistors? Series resistors are typically in for impedance matching but in
> this case....routing was so bad that it wouldnt have fixed anything at all.


I'm not sure what the transformer is doing, the designs I've dealt with
didn't use transformers.  The series termination prevents reflections on the
lines, but of course you can still hose that up if you route it badly enough
:)

2007\01\31@141435 by peter green

flavicon
face


> -----Original Message-----
> From: piclist-bouncesspamspam_OUTmit.edu [@spam@piclist-bouncesKILLspamspammit.edu]On Behalf
> Of Herbert Graf
> Sent: 31 January 2007 18:33
> To: Microcontroller discussion list - Public.
> Subject: Re: [EE] USB routing and termination
>
>
> On Wed, 2007-01-31 at 09:17 -0800, alan smith wrote:
> > I was given a project that was shelved some time ago, and the
> engineer that was working it is long gone.  No PIC, has a PSoC on
> it, being used as a USB bridge among other things.  It was routed
> as just signals, across the board, so no wonder it doesnt work. I
> assume that it should be routed as diff pair for one thing, but
> anyone know off the top of thier head what the route length
> restrictions are?  Also, the design has a 120ohm transformer in
> line (assume for EMI?) as well as a 27.4ohm resistor in series
> with the D+/D- lines.  All the designs I've seen just come from
> the chip direct to to the header for the USB connector.
It does seem a bit weired especially the transformer

what does the datasheet for the chip in question say?

is this low speed? full speed? high speed?


2007\01\31@162853 by alan smith

picon face
I haven't been able to gleam from the Cypress website...in fact, the reference schematics do not show any termination resistors at all.
 
 This will be running low speed.....its feeding into a FPGA thats going to be the bottleneck on the system.

peter green <KILLspamplugwashKILLspamspamP10Link.net> wrote:
 

> {Original Message removed}

More... (looser matching)
- Last day of these posts
- In 2007 , 2008 only
- Today
- New search...