ground isolation on VRM and PCM

Yesterday I noticed that the ground is isolated between the input and output terminal of both the VRM and PCM. This means you cannot power a sensor from the VRM, and expect to reliably read the input on the RoboRio. They are on completely isolated grounds. Does anybody know why that is (i.e., was this done on purpose or was it an oversight)?

***Omar Zrien
AKA: Omar
FRC #3539
Team Role: Mentor

Join Date: Sep 2006
Rookie Year: 2003
Location: Sterling Heights, MI
Posts: 135

The VRM is not isolated. If you measure the voltage difference between the VRM ground and the Rio ground you should find it to be very close to zero .***

Thanks for the reply. Measuring the voltage difference won’t actually tell you if the grounds are connected or not - only that the grounds are at about the same voltage reference. Depending on the multimeter that you use, you may actually be shorting the ground together with the probes - which would show not voltage differential. I’m actually looking for a resistance measurement between the input ground connection and any/all output ground connections. Mine are reading in the MOhm range…


This is correct.

What type of multimeter do you have that does this?

Do you mean megOhm (1,000,000 Ohm) or milliohm (0.001 Ohm) when you write “MOhm”? I can check on our parts when I go to our build site this evening.

I’m just referring to a simple multimeter design that would effectively place a resistor in parallel with your measurement points. A floating ground would then “ground itself” to the other connected terminal. It isn’t a real short.

Mega Ohm

*Si prefixes

Seeing you mentioned sensor power, I am guessing you are having issues getting sensors to work on the roborio, so may want to give this post a read. . .

The input impedance of a good quality DVM is greater than 10 megOhm. This would not be considered a short. An inexpensive DVM, like the ones from Harbor Freight, may have lower input impedance but it will still not short two circuits together.

If you are really seeing megaOhms between the input an output grounds, it would indicate that the VRM and PCM are isolating.

With your meter set to read resistance, short the two leads together. Verify that you able to get a reading that is less than a few Ohms.

You cannot measure the DC resistance between the input ground and output ground, and expect a useful measurement, it will be high-z. The reverse battery protection will not allow this. This is true for both PCM and VRM.

The input ground and output ground of the PCM and VRM are not isolated.

Is there a problem that’s preventing your robot from functioning?

It’s just something I noticed while trying to connect a sensor. The sensor was being powered by the VRM, and read by the RS-232 port on the RoboRio (which ultimately wasn’t working). That is what led me to wonder if perhaps the isolated ground reference would be at fault. Out of curiosity - why is there high impedance between the input and output grounds on the PCM and VRM? Is there a technical reason to do this?

When a VRM/PCM is unpowered or reverse battery’ed it disconnects the ground. Its basic reverse battery protection. When you power a VRM/PCM properly the input and out ground is connected.

This is unlikely your problem. If it was then manually commonizing the Rio and VRM ground would fix it.

Can you oscilloscope the rx pin on the Rio serial port? See if there is actual bytes on the wire. What’s the actual symptom? No data bytes or corrupted data bytes?

Is there a way to see the circuit that disconnects the ground?

Once I ran power and ground for sensor off the RoboRio, everything worked fine. This was a floating reference issue.

No voltmeter (DVM or Analog) will short if its used to measure voltage between two points (by plugging the probes into right sockets on meter or turning the knob/dial to V, depending on the meter). Its the ammeter that is used to measure current in a circuit has very very low resistance and will short circuit when used to measure the potential difference between two points.

phestnes already clarified what he meant:

…and what he said above can happen with a low ohms-per-volt inexpensive analog voltmeter.

Once I ran power and ground for sensor off the RoboRio, everything worked fine. This was a floating reference issue.

This does not prove that there is a floating reference issue.

As I said earlier, the way to test that is by reproducing the problem with the VRM, then manually commonizing the Rio and VRM ground to see if it fixes the symptom. If adding a common ground wire between the VRM and RIO fixes the symptom, then it could be a ground offset. In which case remove the commonizing ground wire and measure the voltage difference to see how much the offset is.

Since neither the sensor or the VRM is drawing considerable current, AND since they are both powered by the same battery, AND since the VRM is not isolated, I really doubt there is a ground reference offset.

The problem is likely something else, something different about the two supplies or they way they are wired. What sensor are you using? What power rail are you using (12V or 5V)? What are the voltage requirements of your sensor? How much current does your sensor draw?

I’m glad the sensor is working now, but I’m a little worried later in the season it will “suddenly” start giving you trouble again because the root-cause was never determined.