Signal waveform from a stepper motor

[TOTA42]

I was wondering if somebody could take a crack on why I am getting this bug. I am doing an inverted pendulum for my senior level design, and we decided to make it somewhat harder (yeah, we're not the brightest EEs), by making it an arm-driven one rather than using a cart. We're using a Vexta stepper motor, which can provide a lot of torque, the only problem is that there isn't many IC drivers that supply enough current. So we had to resort to creating our own by building Dual H-bridges out of N-channel and P-channel MOSFETs. When we hook the motor to the driver, we get a messy square wave coming out of 3 of the MOSFET pairs (which is normal because of the motor's inductance), but for some odd reason, a nice clean square wave comes out of 1 of the pairs. We changed the MOSFETs, diodes, checked the Source to Drain resistance of each MOSFET, and no matter what, 1 pair always gives a clean square wave. Any ideas here?

 

[TuxDave]

Schematic please. I'm a little lost. So you built four drivers and tried each one on the motor and only one of the four gives you a clean square wave into the motor?

 

[TOTA42]

Sorry if it seems confusing. No, I didn't make 4 drivers. To create one driver, I used 4 pairs of N-Channel and P-Channel MOSFETs for a total of 8 transistors. Each pair to drive each end of the coil, A, and B. Here's the Schematic which drives one coil of the motor. So basically I have two of these scehmatics so that I can drive both coils. The weird thing is that one of these four pairs, (N-channel and P-channel MOSFET pair) gives a clean square wave, which should not happen because of the inductance of the motor. The diodes are in there so that it will contain the back EMF caused by the motor when current changes direction. The really odd thing is that when you stick a probe to the drain of one pair of the MOSFET, you get a clean square wave. But if you put it on the other pair of the MOSFET on the same schematic, you get a messy square wave. This shouldn't happen, because they are essentially connected together by the motor coil, and one end should be the same as the other end, but inverted. If this doesn't make sense, let me know, I'm pretty tired from working on my project and I'm probably delirious.

 

[PottedMeat]

ah EE464?
- its been a while since ive looked at h-bridges
couple of questions -

is it always the same P-N side thats giving the clean square wave?

what nodes are you putting the probe across to get this result?

are the gates all being driven with basically a square wave?

have you tried putting a resistor as the load and seeing if you get the same effect as the motor?

oh and any particular reason you're using a bipolar stepper instead of a unipolar?

 

[TOTA42]

Yesiree, 464, gotta love it.

1. Yes it's always the same side/corner, however it's still the same even after we resolder the wiring, change the MOSFETs, diodes, and even changing where that particular node connects to. Its a little ah heck. We even asked Dr. Bostick and Grady to take a peek at it and they couldn't really figure it out why.

2. I am sticking the probe at the Drain of the two MOSFETs.

3. Yes, the gates are being driven by a nice square wave, it's not directly connected to the 6812, there's an inverter in the intermediate stage. Yes, the inverter also operates correctly according to the oscilloscope.

4. Tried adding a resistor, gave me all square waves for all 4 pairs of MOSFETs, which is to be expected. Current also goes through, thought it was a false contact on some of the wiring at first.

5. Well, at the moment we're using bipolar because we wanted the torque to be able to swing the pendulum upright before we start balancing it inverted. But we got to the point now where we might have to switch to unipolar mode, because of speed issues now. I looked at a couple of schematics and it seems pretty easy to switch to unipolar. Just connect the center taps to Vss and change the timing diagram a little.

 

[blahblah99]

Is your stepper motor functioning correctly? You get a square wave probably because the mosfet isn't being inductively loaded down, as in the other case and hence, the clean square wave.

 

[TOTA42]

So far the stepper motor looks like it functions correctly. We pulsed the motor at really low speeds to see the actual steps and we also made a quick program to see if there were 200 steps in a revolution for a 1.8 degree step. As for your question for the one mosfet not being inductively loaded, if you look at the schematic, that clean square wave shouldn't happen. Let's say the drain on the left side of the schematic shows the clean square wave and the right side doesn't. How is this possible? Shouldn't the drain on the right side be the very same as the left side but inverted because it's reference node is different? If I was running in unipolar mode, I might understand why one pair may not be inductively loaded, but I am running it at bipolar-series mode.

 

[blahblah99]

Originally posted by: TOTA42
So far the stepper motor looks like it functions correctly. We pulsed the motor at really low speeds to see the actual steps and we also made a quick program to see if there were 200 steps in a revolution for a 1.8 degree step. As for your question for the one mosfet not being inductively loaded, if you look at the schematic, that clean square wave shouldn't happen. Let's say the drain on the left side of the schematic shows the clean square wave and the right side doesn't. How is this possible? Shouldn't the drain on the right side be the very same as the left side but inverted because it's reference node is different? If I was running in unipolar mode, I might understand why one pair may not be inductively loaded, but I am running it at bipolar-series mode.

Yes, which leaves me to believe that the problem has to do more with physical hardware than design.

 

[Armitage]

OT, but can anybody give me (a non-EE) a good place to learn about controlling stepper mpotors with a PC?
Thanks