Power Supply Failure during Testing

[dksoba]

Hello,

I've recently made a motor control board. Rather simple, it has an H-Bridge run with 4 N-channel MOSFETs and half bridge drivers from fairchild, microcontroller from TI. Anyways, I was testing the H-Bridge circuit with a motor (after doing a LOT of tests with both high, and then lower value resistors, to make sure my programming was right) and my power supply failed. The power supply is a 24V Meanwell 5A supply (Desktop power supply, like one from a laptop). I was supplying the H-Bridge's DC bus with this power supply (microcontroller was on a different power supply, but common ground). The DC bus for the H-Bridge was supposed to have a 3300uF 50V capacitor on it, but I hadn't soldered it on because part of the circuit that's near that large capacitor hadn't yet been soldered on or tested. I have a feeling that the long power lines and the lack of a DC bus capacitor could have, over a few hours of testing, damaged something in the Meanwell 24V supply. I'm just curious if anyone has any idea what happened so I can prevent future failures. If the lack of a DC bus capacitor is to blame, then I probably won't run into that problem again, otherwise I need to take precautions to prevent this in the future.

Thanks,
Matt

 

[mtwieg]

Not having a bus capacitor will definitely make things more dangerous. When the failure happened, were you simply running the motor at a steady speed and load? Were you controlling it with feedback? I've seen motor drivers fail sometimes when one tries to reverse the direction of the motor quickly. What happens is the motor tries to dump its energy back into the DC bus, and if there isn't enough capacitance to absorb that energy, the bus voltage can rise and damage the supply or driver.

 

[dksoba]

Yes, there is feedback control. The failure didn't happen right away. It sort of happened over time. The 24V Meanwell supply was making a humming noise for awhile when the motor was under load (I'm modulating PWM at 20khz). Then it started intermittently working. Finally it gave up completely. I took apart the Meanwell supply but didn't find any broken caps or anything else that looked suspicious.

The feedback modes I was using during the period of time which the failure occured was a simple step response (aka command 45 degree position change...pause... repeat) and steady state (hold a particular angle).

How should I go about sizing my DC bus caps? I have 3 motors sharing the bus (only 1 present during testing though). Each has a 3300 MFD 50V dc bus cap. Possibly I need more? Also, the caps are specified on mouser as General Purpose electrolytic (See UVR1H332MRDANA Nichicon Condensateurs électrolytiques aluminium), and the ESR is not specified. I think I need to have a specified ripple current rating and ESR. Each motor should draw no more than 5 amps continuous, 10 amps peak (for a few seconds).

Thanks,
Matt

 

[mtwieg]

Yes, there is feedback control. The failure didn't happen right away. It sort of happened over time. The 24V Meanwell supply was making a humming noise for awhile when the motor was under load (I'm modulating PWM at 20khz). Then it started intermittently working. Finally it gave up completely. I took apart the Meanwell supply but didn't find any broken caps or anything else that looked suspicious.

The feedback modes I was using during the period of time which the failure occured was a simple step response (aka command 45 degree position change...pause... repeat) and steady state (hold a particular angle).

Okay so there are two obvious candidates for the failure. One is the issue I described where braking/reversing the motor causes power to try and flow into the supply. Another is the ripple current from the motor screwing up the supply (I've seen supplies do awful, crazy things when loading them with lots of ripple). Bus capacitors will help with both. I'm assuming that the overall power rating of the supply wasn't an issue (right?).

How should I go about sizing my DC bus caps? I have 3 motors sharing the bus (only 1 present during testing though). Each has a 3300 MFD 50V dc bus cap. Possibly I need more? Also, the caps are specified on mouser as General Purpose electrolytic (See UVR1H332MRDANA Nichicon Condensateurs électrolytiques aluminium), and the ESR is not specified. I think I need to have a specified ripple current rating and ESR. Each motor should draw no more than 5 amps continuous, 10 amps peak (for a few seconds).

ESR probably isn't what you need to look at. Ripple current rating is more important (and caps with higher ripple current will usually have lower ESR). Don't worry about "general purpose" labels. Just stick to good brand names and look at their ratings. So you need to measure the ripple current of the motors at your given bus voltage and pwm frequency. If you are running multiple motors off the same bus, you can keep overall ripple down by using PWM signals that are out of phase (like for two motors, use 180 difference, for three motors, use 120, etc). This will cause the ripple currents to partially cancel and lower the ripple seen by the caps.

If you still have problems with the supply after that, then there are a few things to try. First you could put a choke on the output of the supply, in order to smooth out the ripple it sees. Second, you could put a blocking diode on the supply output, preventing current from flowing back in from the DC bus. Last, you could try dampening your feedback loop so that the peak torque on the motors is less.