Is our cRIO toast?

Team Jagbots, 2929, could use some help in troubleshooting our cRIO.

In our regional this last weekend (Cascade in Seattle) we made it to the second round of quarterfinals without a hitch but lost communications during it. We took a voltometer to the robot and found that it was getting power all the way up to the cRIO, but something internal must have been causing a problem. The cRIO’s power button did not light.

We confirmed this in several ways:
-Flipped the safety mode dipswitch on the cRIO
-Checked if the cRIO received 24v power (checked for current above and below connector)
-Vacummed metal filings and checked for shorts

So before we have an expert crack open our cRIO, what exactly should we test to make sure our cRIO can be repaired, has anyone had a problem like this?

We had that issue and we just opened our cRIO up and blew out all of the swarf inside of it.

And then to prevent this from happening later we installed the gasket kit that came with it inside of the case and on the connection pins.

I’m with Issac.
If it’s an 8-slot cRIO, then I’d clean it out first.
There’s a very real possibility that metal dust & debris has worked it’s way inside the cRIO and is causing intermittent shorts. You fix that by carefully opening and following directions to clean out the cRIO case. They are pretty tough birds and usually don’t suffer lasting damage from that.

If that doesn’t help then it might still be under warranty (I think three years from being shipped from NI). In any case NI support is only a toll-free call away.
(866) 275-6964 from 8 a.m. to 6 p.m. Central Time, Monday-Friday

I would contact National Instruments directly and ask them what to do next. I believe there is a fuse inside the CRIO and there is a chance that your problem is as simple as replacing a blown fuse. However NI may ask that you do not open the CRIO yourself. After all the CRIO is a sophisticated piece of industrial equipment. In the past the NI folks have been quite helpful to FIRST teams so if it were me I would call first and do what they recommend.

When we killed off all of our cRIOs for some unexplained reason, we had a mentor and his friend break them open and check out the problem. If you are running inefficient code, you can make your cRIO draw enough current to cut out a fuse inside of it (at least, this was our problem. Seeing as there are many intricate parts inside of every cRIO chassis, your mileage may vary). We had the problem fixed on both the hardware and software ends, and we put in a far more cost-effective in-line fuse to the cRIO that will kill the power before the juice gets to the cRIO.

I would say crack it open if the NI Troubleshooting still leaves you with a brick.

Definitely call NI first. They were extremely helpful when ours had the exact same symptom 2 years ago. They ended up replacing the fuse for free (not easily replaced BTW) and getting it back to us in under a week.

Could you elaborate on this? I’m baffled as to how inefficient code would result in blowing a fuse (on the cRIO).

It was some muttering from the programmers about how the draw from jaguars while the robot was driving probably cause the cRIO to go out of control.

Seeing as I just wire the board together for them with some juniors and check over the boards (the mentor and I never found a problem with it), that’s what they went for. However, it could have just been any one bad component causing some vicious chain reaction.

I do remember a hastened backup plan to do a demonstration at a public event and a lot of salty words more vividly than the root of the problem we had to solve… oh well :confused:

By the way, the fuse being mentioned here is an important feature to protect the cRIO from faulty wiring. It typically opens due to high current making its way through the cRIO’s power return. This can happen in several ways, some involving current flowing through the robot frame.

I would also say to crack the case and vacuum; in many cases that will solve your problem. Then you should get a null modem cable and see if it outputs anything. Just hook up with a console (9600,8,n,1) and look to see if it boots at all. Even if it doesn’t output anything intelligent, note if it outputs anything. If it doesn’t output anything at all, I would say it’s the fuse, at which point I would see if it is under warranty as those are a real pain to replace.

first off, by cracking, he doesn’t mean cause a crack in the case (making sure that nobody cough software cough] gets that idea) he means make a small opening at a connection point (take the screws out)

Second: they already vacuumed it.

I assumed that no one would try to actually crack a metal case…

I go by that you can never vacuum enough, and there is always a possibility that something didn’t get all the way clean. I would say that if it doesn’t output anything on console, its a fuse.

I’d say that if the cRIO’s power LED doesn’t light when 24 volts is applied to the proper pins, it’s not likely to be sending anything out the console port.

I don’t think it is possible for the code to force excess current to go through the cRIO. The PD block controls all of that…