Safely probing USB devices

[blurpy]

I've got a TL866II universal programmer that has a fault on ground pin 20. I do not know what the fault is, but I'm thinking about doing some probing with the scope to see what happens during the failing self test of the device. In order to start a self test, it needs to be connected to a computer via USB.

There is  a risk of me probing something wrong, or shorting something. So I'd like to setup the test as safely as possible. I want to try to avoid destroying the device under test, the oscilloscope, the computer, and (definitely) me.

I've read tips about using a powered USB hub in order to protect the computer. Currently looking at maybe buying a TP-Link UH720. It boasts:

a sophisticated circuit design with multiple protections for your devices against over-heating, over-current, over-voltage and short circuit.

Sounds pretty good. Yet, it says in the manual that it works even without external power. Actually all the powered USB 3 hubs I found says that. It makes me wonder how protected the power lines between the hub and the computer really is?

Are there any other measures I should take?
I've read about USB isolators. Some say to always use them, other say they are dangerous.
I've read about differential probes, but it's not completely clear to me if there will be any advantage since it's low voltage (on the DUT, at least).
I've read about probing when connecting the USB device to a laptop on batteries. I feel there is conflicting advice about this as well.

Appreciate any advice.

[chriva]

If it were my computer I'd just go with it. Most if not all have polyfuses and current measurement protection.
Stopped counting the times I've had that annoying over current warning thrown in my face by windows and have yet to kill anything

[madires]

I think you are talking about two connected issues. The first one is about protecting the computer's USB port from a short. If you are concerned about that you could power the programmer by a current limiting lab PSU. The other issue is about earth/PE. The ground of the output of the computer's PSU is connected to earth/PE. At the same time the scope's ground is also connected to earth/PE. If you would connect the probe's ground clip to some power rail of the programmer you would create a short. To prevent that you could use an USB isolator, a battery powered laptop or differential probes.

[blurpy]

I think you are talking about two connected issues. The first one is about protecting the computer's USB port from a short. If you are concerned about that you could power the programmer by a current limiting lab PSU.

I can power it with a lab PSU, but then I won't be able to test the functionality. It's completely computer operated, and the power is only through the USB connection to the computer. I can probably power a USB hub through the lab PSU though, and limit the current it can draw that way. If it's actually using only the power from the external power source when that's connected, which I was confused about.

The other issue is about earth/PE. The ground of the output of the computer's PSU is connected to earth/PE. At the same time the scope's ground is also connected to earth/PE. If you would connect the probe's ground clip to some power rail of the programmer you would create a short. To prevent that you could use an USB isolator, a battery powered laptop or differential probes.

OK, so would it be correct to say that I can "float" the device in this case to avoid ground loops without much risk because it's low voltage? And if it was mains voltage, I could risk making metal parts "live"? Just trying to understand the reason some say it's dangerous to isolate devices from ground.

In the case of differential probes, people seem to recommend Micsig DP10013. They are 50x probes. Would I be able to see a 5 volt signal, or would it disappear in noise?

[madires]

I can power it with a lab PSU, but then I won't be able to test the functionality. It's completely computer operated, and the power is only through the USB connection to the computer. I can probably power a USB hub through the lab PSU though, and limit the current it can draw that way. If it's actually using only the power from the external power source when that's connected, which I was confused about.

Why? Simply cut the 5V wire of the USB cable.

OK, so would it be correct to say that I can "float" the device in this case to avoid ground loops without much risk because it's low voltage? And if it was mains voltage, I could risk making metal parts "live"? Just trying to understand the reason some say it's dangerous to isolate devices from ground.

I assume you mean earth/PE. There are pros and cons for isolating a DUT, for example powering it by an isolation transformer. You'll find several threads about that topic if you're interested in the details.

In the case of differential probes, people seem to recommend Micsig DP10013. They are 50x probes. Would I be able to see a 5 volt signal, or would it disappear in noise?

That ratio isn't ideal for low voltages but it should be good enough for seeing what's going on. BTW, you can also use two channels to create a poor man's version of differential probes (ch1 + ch2 inverted, remove ground clips). The standard 10x probes are be better for low voltages than a 50x DP10013.

[blurpy]

Why? Simply cut the 5V wire of the USB cable.

Good point, I hadn't thought about that.

I assume you mean earth/PE. There are pros and cons for isolating a DUT, for example powering it by an isolation transformer. You'll find several threads about that topic if you're interested in the details.

 

That ratio isn't ideal for low voltages but it should be good enough for seeing what's going on. BTW, you can also use two channels to create a poor man's version of differential probes (ch1 + ch2 inverted, remove ground clips). The standard 10x probes are be better for low voltages than a 50x DP10013.

Thanks for the tip, and advice in general!