DISH
10 W
- Joined
- Oct 19, 2009
- Messages
- 94
Thanks for the additional info!
Issues like this can be difficult to analyze as there can be minor wiring differences which can have major effects when it comes to ground loop issues. Your comments suggest that most of your problem was caused by PWM noise on the throttle control signals caused via the ground loop back to the battery.
That would be the worst case situation as the PWM is active to hold off the motor speed and causing all kinds of spikes and noise on the common ground line.
Under those conditions your PEAK current drawn from the battery is still quite high but average current is low and that is what causes the noise to appear on the common ground loop. The throttle signal is not only sensitive to DC level shifts but noise as well and millivolts of level shift or noise riding on the DC throttle signal is enough to cause problems.
A very small amount of noise on the throttle control signal can get magnified as a very small change in throttle control signal causes the PWM signal to change which then gets fed back to the throttle control signal via the common ground loop which then causes more change in the throttle control signal! This creates a feedback condition where things can get really weird very quickly and is why it's so difficult to figure out what is going on.
I used the large 1.0 volt drop reference only to make things easier for discussion. I thought it would be easier to understand what I was talking about if I stuck to whole numbers (like 1 volt) rather than .001 or millivolts. Sorry if that misled the discussion as millivolts of noise or DC level shift can cause problems due to the feedback issue mentioned above.
When the throttle is wide open PWM should be inactive meaning there is no PWM noise on the common ground between the controllers. The test you mentioned (100 A to one motor while other is freewheeling) suggests that most of your problem may be PWM noise on the throttle control signal and the simple RC filter that someone suggested early on should eliminate that, provided of course you do not also have a DC level shift problem going on at the same time you have a PWM noise problem. An RC filter will kill the noise but will not correct any DC level shift issues.
Issues like this can be difficult to analyze as there can be minor wiring differences which can have major effects when it comes to ground loop issues. Your comments suggest that most of your problem was caused by PWM noise on the throttle control signals caused via the ground loop back to the battery.
EXCELLENT observation and very helpful!!!methods said:Your argument is straight forward - not hard to follow at all.
In actuality - the worst cases I get is near freewheel and at very low motor loading - I am talking like 5 - 10A
At those kinds of currents with the size cabling I am running I should not be seeing volts of drop... maybe mV....
That would be the worst case situation as the PWM is active to hold off the motor speed and causing all kinds of spikes and noise on the common ground line.
Under those conditions your PEAK current drawn from the battery is still quite high but average current is low and that is what causes the noise to appear on the common ground loop. The throttle signal is not only sensitive to DC level shifts but noise as well and millivolts of level shift or noise riding on the DC throttle signal is enough to cause problems.
A very small amount of noise on the throttle control signal can get magnified as a very small change in throttle control signal causes the PWM signal to change which then gets fed back to the throttle control signal via the common ground loop which then causes more change in the throttle control signal! This creates a feedback condition where things can get really weird very quickly and is why it's so difficult to figure out what is going on.
I used the large 1.0 volt drop reference only to make things easier for discussion. I thought it would be easier to understand what I was talking about if I stuck to whole numbers (like 1 volt) rather than .001 or millivolts. Sorry if that misled the discussion as millivolts of noise or DC level shift can cause problems due to the feedback issue mentioned above.
Once again a very helpful observation!methods said:The troubling thing is that this is not happening while one motor is pumping 100A and the other is freewheeling...
When the throttle is wide open PWM should be inactive meaning there is no PWM noise on the common ground between the controllers. The test you mentioned (100 A to one motor while other is freewheeling) suggests that most of your problem may be PWM noise on the throttle control signal and the simple RC filter that someone suggested early on should eliminate that, provided of course you do not also have a DC level shift problem going on at the same time you have a PWM noise problem. An RC filter will kill the noise but will not correct any DC level shift issues.
The 5 volt rails should already be pretty well isolated by the internal regulators and filter caps in the controllers. The 5 volt voltage level usually doesn't vary much from controller to controller unless of course there is something wrong with the regulator and in that case I think you would notice it running that particular controller by itself.methods said:If this is the case, anything that can be done to keep the 5V rails equal and the 5V Grounds equal will help.
I suspect that you will find that tying the grounds together with a VERY SHORT piece of HEAVY wire plus putting the cases of the two controllers together so that they are in electrical contact with one another will be a BIG help but you may still need the RC filters on the the throttle control lines. I don't think that tying the 5 volt lines together will help and in fact may be a bad idea as you could end up injecting the common ground loop noise directly into the 5 volt bus. The 5 volt bus should already be pretty well isolated due to the internal regulator and filter caps.methods said:I am going to try bringing out a +5V 12AWG wire and a 12AWG ground tap - as close to the uC as possible.
That's a good start but the wires going to the throttle are too small in diameter to have much effect on eliminating the common ground loop issue back to the battery where large currents are flowing.methods said:In my circuit I am bringing out 5V, GND, and Throttle from each controller and tieing them to the same potential at the Y point where the throttle is.