While we were tuning our PID control constants, the cRIO stopped working (no power light). We finally figured out the 2A surface mount fuse inside the cRIO is blown. Fortunately, we have a backup cRIO, we quickly replaced it and tried several more runs and the 2nd one is blown too. Did anybody experience this? We couldn’t figure out why the fuses blew. Visually, I don’t see any metal debris in any of the slots. We’ve been driving the robot hard and nothing blew until we did the PID turning constants calibration. In both times, the robot was turning back and forth vigorously (PID ringing). I would think the cRIO can take this kind of motion. I am going to get some replacement fuses tomorrow but I am a little concern what actually happened. Because it could happen again if we don’t figure it out. Did anybody have similar experience and figured out what happened? There are several theories but none make sense. One concern was the back emf from the motors somehow got into the power bus of the cRIO. I can’t believe this because I would imagine the digital side car and the cRIO should have voltage regulators, diodes and capacitors etc to guard against power contamination like this. Another theory was the violent left and right turning motion physically breaks the fuse. Can’t believe that either. Of course the most make sense theory was something shorted while the robot is shaking, but we can’t find anything visually. I will take a much closer look again tomorrow. But I am running out of ideas to look.
I would definitely look into the shaking causing a transient short.
The cRIO is chassis grounded so if it contacts the metal of your frame during violent movements (and you already have one existing connection to your chassis) the situation you described would be the likely result.
I would check ALL your connections and make sure you don’t have a connection between any electrical systems and the robot chassis. This means taking a multimeter and verifying that resistance is greater than 10M ohms between the robot chassis and the following:
- Battery negative (probably easiest to hook to at the PDB)
- Battery positive (probably easiest to hook to at the PDB)
- Each motor lead (you have to test each one on the motor side of the Spike/Jag/Victor)
The last one is the one that doesn’t get tested very often, but can cause the exact same problems as the other two.
I heard this was a known issue from last year, and that they’ll overnight you a new unit, and a label to send the old one back. Can anyone confirm?
Our cRIO is mounted in a wooden box, so it wouldn’t contact our robot chassis. Not that we have any deliberate connection of the cRIO chassis to the battery’s negative terminal, I can understand 2 and 3, but why is it a problem that the cRIO chasis is grounded to the battery’s negative terminal? Wouldn’t the black lead of the 24V supply is exactly the negative terminal of the battery? In fact, I measured the voltage between the negative terminal of the battery and the black lead of the 24V supply which is the ground of the cRIO. The voltage between them is in the mV so it is practically shorted to the battery’s negative terminal. In any case, we do not connect the cRIO chassis to the negative terminal of the battery (at least not deliberately). I will check 2 and 3 tomorrow.
Who should we contact to get the replacement? What was the “known problem”?
Thanks.
Call National Instruments at (866) 511-6285 from 7 a.m. to 7 p.m. Central Time.
Thanks for the info. I found that on the NI forum as well.
With the violent shacking it might also be due to debris that’s worked it’s way inside the cRIO chassis and is being thrown around creating a traveling short circuit.
Did you inspect the inside of the cRIO?
Did you install the KOP gaskets supplied to prevent this, at the beginning of the season?
Our team managed to blow 2 cRIOs as well. For one of our cRIO the problem was actually debris, a small strand of cable got inside through the serial port. The second one is a mystery for us as well, however both seemed to blow when we powered the robot over the bump. So maybe the physically damaging may not be so absurd, but i doubt it as well. However after installing the KOP gakets we haven’t blown a cRIO so maybe we just had stuff get into the cRIO while practicing. However maybe the victor connections were not tight and going over the bump caused the issue. We didnt check our victors as number 3 stated in posts above.
Just to share our story.
Keeping the electronics clean is critical. If you hit a bump and suddenly things stop working, chances are that sheet metal your team member drilled through in the pits left a little extra behind.
Cleaning and applying the gasket is a great warranty against this risk - as is keeping your electronics clean, keeping a separate electronics box, and using a shop vac when drilling. At a recent regional I visited, only 15% of teams had applied the “easy” part of the gasket on the outside of the chassis. Less than that had actually applied the internal gasket as well. This gasket has been well tested and dramatically decreases the likelihood of contamination, and the cleaning will remove anything that may have accumulated over time.
In the meantime, to get you back up and running, contact NI through ni.com/first
+Chris
[Did you install the KOP gaskets supplied to prevent this, at the beginning of the season?
Wow, that’s what those were for. I have been warning people about chips for a while but I missed the gaskets. While we have not had any of these issues I will definately try to add at least the easy ones before our next event at Boston.
Thanks for the tip.
Just an update: We soldered a temporary external fuse to the cRIO and was able to make it run the entire day without it blowing again. We did find two loose wires and re-secured them. That may have something to do with the issue. But we are happy that it did not happen again.
Mikets,
I have inspected several robots that thought they had insulated the Crio in wood only to find that the mounting screws went through the wood into the frame. The camera, KOP sensors and custom circuits are all sources of potential connection to chassis. When the chassis finds a short to a motor controller output, several volts will be developed across the the wire feeding power to the Crio. Shorts can come from pinched wires, shorted motor (internal short to case) and moving components. Robots that kick and quickly move over the bump (the Duke’s of Hazard move as it was termed over the past weekend) cause things to move much more than you think.
Great info. I am curious. Can you educate me on why it is important to insulate the cRIO chassis from other things? I have checked the cRIO motherboard and found that its ground is common to its chassis and the negative terminal of the 24V supply. When the robot is switched on, the negative terminal of the 24V supply is essentially the negative terminal of the battery. Therefore, when the robot is running, all grounds are common to the negative terminal of the battery. If anything shorted to the chassis and thus the ground, the corresponding fuses will blow (e.g. if the motor controller is shorted to ground, the fuse that supplies that controller will blow). Therefore, I would think the only way the fuse of the cRIO will blow is when something in the cRIO is shorted including the I/O modules and the digital side car. In our case, I suspect one of the loose wire had shorted one of the pins in the digital side car. But I don’t understand how it will blow the cRIO fuse if its chassis is connected to the robot chassis and something was shorting to the robot chassis.
Mikets,
One needs to look to the real world properties of the components involved. Mostly to the resistive nature of the wiring. There is no ground nor is there anything that acts like ground on the robot and no current passing through the frame as there might be in vehicle electrical systems. The only point at which all is common is the negative lead of the PD. The chassis of the Crio is tied to this common through a #18 wire of unknown length. The camera likewise is tied through a smaller wire also of unknown length and that is tied to the metallic parts of the camera case. Should a motor lead or power lead for a motor controller also be tied to the frame, a voltage is developed across the #18 wire of the Crio between the Crio chassis and the PD and some current will flow in this path. So here is a few possibilities if the Crio is not insulated from frame.
- The return lead of a motor or speed controller input is tied to the frame. The return to the PD has a bad crimp or loose screw. Some of the motor current will attempt to flow through the Crio.
- One lead of a speed controller is tied to frame either due to pinched wire or motor short. In one direction the frame is connected to the negative lead of the PD and only motor noise and some motor current flows on the frame. The Crio will then have noise on the chassis. In the other commanded direction, the frame is now tied to the positive lead of the battery and is subject to the PWM output of the controller pulsing to 12 volts. That raises the chassis of the Crio several volts depending on the current flow and the resulting voltage drop across the Crio power wiring. It may be enough to interrupt Crio operation, communication or if long enough it may even cause a reboot.
There are other possibilities as well, from disrupting the PD power supplies to faults on the side cars. Each one will be slightly different. It is like a road map. There are many ways to get some place, some of them are faster than others but all eventually lead back to the same destination. It is best to close all the side roads, dirt roads, paths, trails and walkways and just leave the expressway.