USB hubs in the lab, any caveats?

Ok, folks, just got a new scope for work. And a new PC-driven receiver
for EMI stuff, other smaller things will follow. All that has USB. I am
ever so grateful that the GPIB era with those dreaded garden hose cables
seems to be over but now comes the question: Are there any special
precautions when using USB hubs with lab equipment?

Some gear like the scope is host-capable, something that isn't normally
the case when you use hubs to connect mice, printers etc. I don't want
to fry anything and also I'd prefer a non-powered hub. Got enough cables
in there as it is and nothing needs much power from the bus anyhow.

BTW, before anyone starts cracking jokes, I made sure all this can run
sans USB when radio silence is needed for EMI work ;-)

 

My experience is that it's pretty much impossible to fry anything with USB;
even cheap implementations do provide current limiting.

A non-powered hub will limit its "downstream" devices to 100mA each, in almost
all cases (I've seen a few "hybrid" design where one port gets more current
and the rest get very little, but it's uncommon); you'll want to check to make
sure your various peripherals will work with that. (Anything with its own
power supply generally will, of course, as will mice, keyboards, joysticks,
and some webcams... anything else you just have to try it.)

I would make an effort to use decent quality cables -- some of the cheap ones
seem to leak noticeably more, and I very much suspect the USB certification
stickers on them are counterfeit.

I was playing with a receiver design last week that didn't want to work with
low input signal levels, and the problem turned out to be that one of our
newer Agilent network analyzers puts out a birdie at exactly 45MHz -- the same
as one of my IFs. @#$#$%(@$

---Joel

 

Thanks, Joel. Ok then. I'll get a hub and just connect it all up. USB is
amazing, no addresses to set or worry about, it just works. Usually. The
only thing that sometimes doesn't are RS232-USB kludges. But to my
surprise the new scope was the 2nd piece of gear that came with a RS232
that wasn't mentioned in the advertising. So if it all goes wrong I know
I can still talk to it the old-fashioned way.

Yes, I buy brands like Belkin or Targus. The good stuff.

Another reason why I shunned the big mfgs for this new scope. My first
encounter of this kind was also an Agilent analyzer. An EMC analyzer no
less and one of the more expensive kind. There was a small stand of
trees on the screen that remained stubborn. I was slowly beginning to
worry whether I still had all my marbles, at least in terms of EMI
fixing. Then I found the source: The screen of that very analyzer ....

The latest one was a TDS220 which sent so much conducted EMI into a new
circuit that I got fooled into believing my PID loop was slightly
unstable. I mean, it could happen. Finally I hooked up a real scope, a
Tek 2465, back from when they built the good stuff. Bingo, the noise was
gone. Turned out it also originated at the screen, probably the
backlight inverter.