Author Topic: Still stumped by USB link  (Read 1258 times)

0 Members and 1 Guest are viewing this topic.

Offline IconicPCBTopic starter

  • Super Contributor
  • ***
  • Posts: 1564
  • Country: au
Still stumped by USB link
« on: December 16, 2019, 12:19:03 am »
Looking into USB2.0 standard the actual USB cable is specified as one twisted pair @Zo= 90 ohms plus two additional non twisted conductors for power.
USB2.0 further specifies the delay ( is it group delay inequality ?) and the maximum cable length of 5 meter.
I need to repair machine vision based on a USB camera with an overall cable length of close to the limiting value.
The cable needs to be ultra flexible as it lives inside a drag chain.
Finding such a cable has proven to be difficult.
I can get ( out of Sydney ) two pair cat5 ethernet cable which would survive millions of flexures inherent in the application however the characteristic impedance is 100 ohms =/-15%.

How critical is the deviation in Z0 form the spec?
What problems might be experienced?
Is it possible to measure/ evaluate cab;e performance using the USB ports on a PC?

But more to the point other than Sydney based LAPP cable is there anyone else who might be able to supply USB cable or similar data cable?
 

Offline SeanB

  • Super Contributor
  • ***
  • Posts: 16385
  • Country: za
Re: Still stumped by USB link
« Reply #1 on: December 16, 2019, 07:33:56 am »
How about some USB extenders to use that CAT5 cables in the loom, solving your problem?

Try this one for example,

https://www.amazon.com/AV-Access-Extender-Operating-Synchronously/dp/B01EV33R8S/ref=sr_1_7?keywords=usb+extender

Or this.

https://www.ebay.com/itm/USB-Over-Ethernet-Extension-RJ45-Cat5e-Cable-LAN-Adapter-Extender-4-Port-HUB-/391745074308

Just the first sort of sane results from searching and looking at images.

« Last Edit: December 16, 2019, 07:35:33 am by SeanB »
 
The following users thanked this post: tooki

Offline ConKbot

  • Super Contributor
  • ***
  • Posts: 1400
Re: Still stumped by USB link
« Reply #2 on: December 16, 2019, 02:50:51 pm »
If the cable is shielded pairs, id definitely try it, with something to act as a repeater on either end (active cable, or hubs, etc) so performance won't vary with different hosts plugged in.

If it's not shielded pairs, find a cable with shielded pairs, as USB2 has more than a little common mode content, so it will be emitting a lot, and be more susceptible to inference. Especially in an area with motors and motor drives going.
 

Offline SiliconWizard

  • Super Contributor
  • ***
  • Posts: 15797
  • Country: fr
Re: Still stumped by USB link
« Reply #3 on: December 16, 2019, 03:45:24 pm »
Beyond the data, you'll have definite problems with VUSB with cables this long, so you'd need to provide the power supply locally (close to the camera) IMO.
 

Offline fcb

  • Super Contributor
  • ***
  • Posts: 2135
  • Country: gb
  • Test instrument designer/G1YWC
    • Electron Plus
Re: Still stumped by USB link
« Reply #4 on: December 17, 2019, 06:25:31 pm »
https://electron.plus Power Analysers, VI Signature Testers, Voltage References, Picoammeters, Curve Tracers.
 

Offline Leo Bodnar

  • Frequent Contributor
  • **
  • Posts: 820
  • Country: gb
Re: Still stumped by USB link
« Reply #5 on: December 17, 2019, 07:40:33 pm »
How critical is the deviation in Z0 form the spec?
We have measured all sorts of dodgy USB cables.
Some rubbish ones were as low as 69Ohms yet still worked - sometimes.
Most brand name cables we have tested have Z0 close to 100Ohms
A lot depends on how well device and host are terminated internally, their transmit levels and receive thresholds.
Common mode impedance is important as well. SE0 state that ends every packet is not a differential transmission.
Leo
« Last Edit: December 17, 2019, 07:50:51 pm by Leo Bodnar »
 

Offline mariush

  • Super Contributor
  • ***
  • Posts: 5170
  • Country: ro
  • .
Re: Still stumped by USB link
« Reply #6 on: December 17, 2019, 08:19:05 pm »
There are some other options, like for examples cables that use optical fiber to transmit the data.

Here's an example: https://www.monoprice.com/product?p_id=16377

It carries usb 2.0 and usb 3.0 over 10 meters (32ft) and the fiber should have no problems with repeated bending.
Note that the fiber itself doesn't carry power over the cable. The cable comes with a power adapter which you'd plug near the machine, and you power the connector by the machine to decode the data coming through the fiber and to transmit data from that end to host computer.
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf