0
\$\begingroup\$

In my design I have a national instruments chassis (cDAQ 9132) inside of a closed rack fixture. I want to be able to access the usb port on the NI chassis from a front panel on the external of the rack. My thoughts were to just connect a male-to-male usb cable from the NI chassis into the front panel via a panel mount usb connector.

My question is, will the usb extender and connector cause signal degradation? The total length would only be a few feet, much less then the 5 meter maximum.

\$\endgroup\$
  • 1
    \$\begingroup\$ There are no connectors which do not degrade a signal ! Question is, how much degradation can you allow. USB 2.0 is a digital interface and it only works at certain speeds. So you won't get 400 Mbps instead of the standard 480 Mbps. If you harm the signal too much, you might get 12 Mbps (that's the next step below 480 Mbps) or (more likely) you will get nothing. Just use good quality connectors and it will work fine. \$\endgroup\$ – Bimpelrekkie Mar 9 '17 at 21:07
  • \$\begingroup\$ @FakeMoustache: wasn't USB speed "hardcoded" by resistors of the interface and thus there is no downgrade if the wanted link speed is unavailable? \$\endgroup\$ – PlasmaHH Mar 9 '17 at 21:21
  • \$\begingroup\$ @PlasmaHH I don't know about those resistors. But I do know that the lower speeds must be supported for building up the link and backwards compatibility. \$\endgroup\$ – Bimpelrekkie Mar 9 '17 at 21:28
  • \$\begingroup\$ @FakeMoustache The lower speeds are indeed supported for devices, but for ones with those speeds. But if you have a HS device and HS host, there is no fall-back mechanism, see my answer below. The USB 3.0, however, does have some partial fall-back possibility: if SS fails to detect normal SS termination, it will time-out, and HS port will take over. However, if the SS link is present and passes the training, but has massive SS error due to poor signal integrity, it is up to software to drop SS attempts and resort to HS mode. I believe this is not mastered yet, which causes SS port hangups. \$\endgroup\$ – Ale..chenski Mar 9 '17 at 22:04
1
\$\begingroup\$

From my experience, USB extender cables work, with caveats.

  • Power

A device I have draws 460mA out of the allowed 500mA. It will only work with good quality cables (ie, with enough copper in them), no extenders, and only some ports on the PC (those on the mobo, not those on the front, which go through more cables). Contact and wire resistance matters...

Less power-hungry devices will cause less problems.

  • Data

USB Bulk mode retransmits packets when an error is detected.

If your device uses bulk mode, a little more bit errors will reduce throughput (slightly). If there are enough bit errors to make enumeration fail, though, it won't work at all.

I have used extenders with high speed devices without problems.

When you think about it, the USB connectors on the front of your PC go through an extender cable... and they're connected to the mobo with 0.1" headers, which I honestly doubt respect the USB impedance spec...

Why not try? I'm pretty sure it will work, as long as all connectors are USB certified.

\$\endgroup\$
  • \$\begingroup\$ While all connector parts might be certified (in clean manufacturer's test benches), but the quality of cable soldering cannot be certified, because there is no certification for illegal cable extenders. The quality of solder jobs varies wildly, and cable impedance can be 70 Ohms or 110 Ohms as well, causing massive reflections and channel degradation. Unless you have a special fixtures and high-bandwidth scope to capture signal eye diagrams, it is difficult to determine. \$\endgroup\$ – Ale..chenski Mar 9 '17 at 21:53
1
\$\begingroup\$

You are essentially asking if you can use a USB 2.0 extender cable in your setup. The answer is "it depends".

First, formally the USB specifications disallow cable extenders, for the reason that every extra cable imperfection (two extra connectors and two cable fan-outs at solder joints inside cable molds) usually degrade the high-speed signals beyond acceptable. Quality of extenders, however, vary greatly, so your mileage can vary.

Then, it depends whether you/your users/customers are going to plug USB devices with extra long 5-m cables. Then your extra extender will likely kill the USB functionality in high-speed mode. You will have flaky behavior, with unpleasant experience, data corruption, etc. Contrary to the belief in USB data recovery mechanism, it works only in a narrow region of degraded signals. If the channel gets degraded a bit deeper (due contact wear, contamination, different mechanical tension), the channel will just drop off, reconnect back, and cause a lot of troubles.

However, if you plan to use only LS/FS USB devices (mice/keyboards/sound) the extender will be fine.

The hopes that USB port will fall-back into FS mode are also unfounded. The reason is that the FS-HS negotiations are coming at a relatively low signal speed (~20kHz) where the signal degradation is negligible, so the initial connection will be always negotiated as "HS mode", and the link will always start in HS at 480 Mbps. Then the channel will see massive transfer errors, and host will try to recover the port by resetting/re-enumerating it, and process will fall into the same loop.

\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.