Current reversed on compressor spike controller

[scottish144]

For some odd reason, the spike controller that runs to our compressor continually blows its fuse when we are turning. All connections are sound, so we believe that the current on the spike is being reversed. However, we are unsure how this is happening. Is it possible that the motors could be producing feedback to the spike? They suffer a lot of mechanical stress while turning.

[Mike Betts]

Scott,

Last year's rules allowed you to replace the 20A fuse with a 20A self-resetting circuit breaker because everyone was having this problem.

The issue is that, as the battery voltage goes down, the current to the compressor goes up. The mechanical work that the compressor is attempting to do is independent from the drive system (or anything else) on the robot. If the mechanical power remains the same, the electrical power will remain the same (neglecting efficiency) and the current must go up as the voltage goes down.

Replace the fuse with a circuit breaker and move on...

Mike

[sanddrag]

While your at it, you should verify that your compressor is not running in reverse. Symptoms can include extreme heating (to the touch, it feels very hot after running for a little while) and a prolonged time to reach full pressure in the tanks. Running the compressor in reverse has the potential of drawing more current and it's efficiency goes way down. It could eventually lead to failure.

The reason I mention this is that I have seen it too often.

[mechanicalbrain]

THIS IS A UPDATE. No nothing like that. I'm updating for Scott. We know its not the compressor (I got it running on a 9 volt) and its not the spike, battery, fuse, wire, or pressure sensor. The spike is literally not receiving a signal. In fact nothing connected to the relay is getting a signal. I'm wondering if their is a real easy way to test if the relay works. Chances are its not the code because the compressor stopped working after running fine with no changes to the code. Weird fact: we commented out the pressure sensor code and we got a response from the compressor but all that happened is the fan made a twitch (I'm guessing the motor switched polarity) once every second. I'm completely clueless. I tried switching out every part in the electronics but nothing changed!

[sanddrag]

Try a new pwm cable and a new port on the rc (with code changed for that port). Also, if you can get your hands on an older Spike try that. We had problems with the 2005 ones too (not working, oscillating, etc).

[sciguy125]

Originally Posted by mechanicalbrain

The spike is literally not receiving a signal.

If you haven't already tried another spike, do so. But, we had this problem too. It took us days to figure out. You're probably not putting the PWM connector in right. It might go all the way in, but the pins might not actually be in the connector. Sometimes, we would even get only one or two of the pins in.

Just to make sure though, pull the connector out of the spike. Test it for voltage between ground and the appropriate signal pin. It should be a solid 5V. If you have that right, then I probably found your problem.

Actually, I just reread your post...

Originally Posted by mechanicalbrain

we got a response from the compressor but all that happened is the fan made a twitch (I'm guessing the motor switched polarity) once every second

Do the voltage test still to check your code. But also pay attention to the LED once you plug it back in. Orange is off, red is reverse, green is forward. I also think that the spike makes a clicking noise when it changes direction (I can't remember). You might also want to put in a meter instead of the compressor to make sure that it's getting a solid 12V.

[mechanicalbrain]

Originally Posted by sanddrag

Try a new pwm cable and a new port on the rc (with code changed for that port). Also, if you can get your hands on an older Spike try that. We had problems with the 2005 ones too (not working, oscillating, etc).

We tried 3 diferent spikes (including older ones) , 4 pwm wires, and a number of ports. Im pretty frustrated! The light was a steady orange the entire time. I doupt it could be the code since it gave out on us while we were using it... One thing im going to try is to solder on sewing needles onto a PWM to make sure we get a connection.

[sciguy125]

Originally Posted by mechanicalbrain

One thing im going to try is to solder on sewing needles onto a PWM to make sure we get a connection.

I hope you mean onto a cable and not the controller. I think the pins on the controller are gold plated (corrosion resistant). It would be a shame to get solder on them.

But I seriously think that you should try the meter first (if it's pulsing too fast, you might be better off with a scope). It'll tell you if you're even getting voltage/correct signals. If you're not getting anything at the end of the cable, you should try the pins directly on the controller.

[mechanicalbrain]

Originally Posted by sciguy125

I hope you mean onto a cable and not the controller. I think the pins on the controller are gold plated (corrosion resistant). It would be a shame to get solder on them.

But I seriously think that you should try the meter first (if it's pulsing too fast, you might be better off with a scope). It'll tell you if you're even getting voltage/correct signals. If you're not getting anything at the end of the cable, you should try the pins directly on the controller.

No deffinetely the wire. I need to check the multimeter for a signal though because i honestly can't remember what we got.

[Alan Anderson]

Originally Posted by mechanicalbrain

Weird fact: we commented out the pressure sensor code and we got a response from the compressor but all that happened is the fan made a twitch (I'm guessing the motor switched polarity) once every second.

What fan?

If you're using the standard programming, the compressor is supposed to be connected between the Spike "FWD" output and ground. Is that how you have it?

[dez250]

To state the obvious and make sure your using the right components below is a Victor Speed controller which has a fan on it:


and this is a Spike Relay Controller, which has NO fan attached:



One test you may want to do is to use a multimeter and check the continuity of the single wires of your PWM cables to make sure they aren't damaged. Also physically inspect the pins on the PWM cables to make sure they are not bent or loose. If you still haven't figured out the problem, drop me a pm or im and ill try to help you out more in depth.

[mechanicalbrain]

Yes we used a spike! It would be.... interesting to try it with a speedcontroller. I was talking about the fan inside the compressor. I assure you we tested every piece betwean the control board and the compressor and every one works! We just can't get the spike to get a signal from the boards relays (Tried multiple (MANY multiple) pwm cables). Its unlikely its the code because it stoped running while we were using it. We have to ship the robot to Capitol Clash tomarrow. Any help will be really usefull.

[Mike Betts]

Scott, Oz, et al,

You started off with a problem where you blew a fuse. Now it appears that nothing works at all. There is also confusion in your posts concerning Spikes versus Victors, LED indications, Polarity and (perhaps) PWM versus Relay outputs.

I'm going to ask you to take a step back and compose a description of your problem with some diagrams and/or pictures. Perhaps a code fragment as well.

You may have fallen into the trap of "breaking" several other things while trying to troubleshoot the fuse problem. As such, you may now have a multiple problem problem which is complicating your life.

The act of stepping back and documenting what you have can be very beneficial in these circumstances.

Regards,

Mike

[Gdeaver]

Does the compressor run if powered by 12 volts directly? Maybe your compressor is gone. If it runs check the amp draw.

[Al Skierkiewicz]

A lot of info is passing through this thread and we do want you running for the weekend. If what follows doesn't get you help, ask for some of the teams to assist when you get to the Clash.
As to the first post, tripping the fuse while turning is a good indicator of the start current demands of the compressor. Although it is rated for 10 amps that is "run" current, start current is more than twice that amount and under certain conditions we have measured four times that amount. (For a very brief time during start.) When turning, currents are very high for the drive motors. Tank style steering (where no wheels turn but motion is accomplished by speed differences between the two sides) are the one of the highest current demand drives you can use. (Belt style tank being the worst.) As the battery tries to deliver power during these high demands, the terminal voltage can fall below the point at which the compressor continues running. So the effect is turn, compressor shutdown, compressor restart, turn, shutdown, restart, etc. Each restart of the compressor is accompanied by a high current pulse that heats up the fuse and failure is the result. Replace the fuse and continue beating up the spike and failure will result there as well. So that in a nutshell is the normal failure in this application. However, you may have a different problem altogether since you have replaced the spike and wiring multiple times. I have to ask if the location is very tight and other wiring had to be disturbed to make the changes. There is a distinct possibility that 1) the spike was rotated when replaced, 2) the wiring to the spike was exchanged with another component near it, (another relay or controller), 3) the compressor has shorted, or 4)in replacing the fuse and/or spike, the wiring from the pressure switch was broken and your software is no longer receiving a command to start the compressor. High on my list are #4 and #1. A simple check of the switch wiring and connection back to RC is very easy to do. Checking on the orientation of the spike is also very easy to do. My rule of thumb when troulbeshooting a working device is "go where the last person was" which will be the spike replacement.