Messed around with it enough, and finally got it to work. But it pulses at about 214 beats per minute.
I've got it set to full stepping (MS1 & MS2 pulled low, MS3 floating (it has an internal pulldown))

My electronics are a bit of a mishmash.
I've got an arduino diecimila
I've got a PWM driver V. 1.1 from the rrrf, modded to work at 24V instead of 12V
I've got a Tempurature sensor V1.1 board (rrrf, again)
I've got some scavenged stepper drivers, full stepping @ 24V, 3 channels, which is working fine.
I've got a Mean Well SP-150-24 power supply. (24V)
I've got a breadboarded pololu stepper driver, which is pulsing at 214 PBM
I've got a mares nest of wiring, recently brought under a little better control by moving the power supply to the bottom and using a few zip ties.
The 24V power comes from the SP-150-24, and the 5V power comes from USB.

The pololu works, and I've gotten the attached stepper to spin. Looks fine, except it turns off for an instant about 214 times per minute, and about 1 time in twenty (of those 214 times) it turns backwards instead. There seems to be no pattern to when it decides to turn the wrong way.

The pulsing goes away if I unplug the USB, of course, as there is no 5V to power the stepper driver. I do have an rrrf DC motor driver that has a voltage regulator on it, I could try to salvage that and stick it in the breadboard to supply 5V from there instead of from the laptop. Hrm.. Tried that, that thing got hot quick, and the pulsing problem was still there. Experiment terminated about 5 seconds in due to heat, which is before the atmel chip gets out of the bootloader.

I've no idea what could be causing this. The atmel chip isn't reseting, so it's not a systemic problem.

Ground loop somewhere?

The ground pin is busted off my laptop's power supply, that may have something to do with it? I could try running the laptop off battery, see if that changes things. Nope, same.

The pulse occurred without the USB even hooked up when I tried to power the arduino off 5V generated by that regulator from the DC motor driver board.

Perhaps it's time to map out all the ground wires.

Edited 1 time(s). Last edit at 10/26/2010 10:38PM by jgilmore.

---

--
I'm building it with Baling Wire

probably the chip's thermal protection. they need to be heatsinked. I cut up a video card cooler, others have successfully used those little stick-on ram sinks. The wrong direction turning will be from when it turns off just long enough that when it turns back on, the coil outputs are set so the motor goes the other way.

Heatsinking the chip should eliminate both problems. Make sure you don't short out the Vmotor capacitor! It's higher than the chip on mine.

---

-----------------------------------------------
Wooden Mendel
Teacup Firmware

Maybe my understanding of the thermal behavior of these chips is flawed. It doesn't seem likely to me that it's a thermal problem because:

It starts right away, even from a cold start it behaves the same as a few minutes in.
The board is still cool to the touch. (reportedly they normally run hot enough to burn when touched.)

Ok, check that. You're right. These chips apparently heat up VERY quickly, and are immediately hot enough to burn. No, I didn't actually burn myself, but even the back of the board is burning hot to the touch in less than a second from turning the power on.

Thank you, I hadn't thought to give it a touch check.

Now I need to redo that section of the board, mount it the other way up, etc. Maybe I can attach a heat sink to the back of the board?

==edit==

No, I was not able to even significantly change the period of the halting by pressing a chunk of aluminum to the back of the board.

Quite naturally when I went to mount the driver on my breadboard, I mounted it chips down, so I could see the silkscreened labels for the pins. This means I'll have to completely redo it (and resolder in new headers) to get a heatsink touching the actual chip. I think I'll go with a hacked-off chunk of aluminum touching the chip, and an aluminum heatsink pressing that chunk down onto the chip.

Edited 1 time(s). Last edit at 10/26/2010 10:37PM by jgilmore.

---

--
I'm building it with Baling Wire

Did you try adjusting the potentiometer? It contols the current limit for the stepper driver.

johnnyr Wrote:
-------------------------------------------------------
> Did you try adjusting the potentiometer? It
> contols the current limit for the stepper driver.

Well, yes. But given that the potentiometer was under the board, and I've got an older, slower-starting bootloader on my diecimila, and the whole thing needs to be powered off to unplug anything, I didn't go through very many iterations before giving up.

Now I've taken it off, pulled the plastic rails off, removed the pins one by one, and reassembled the whole thing with the chip on the top this time. Rearranged the breadboard for the new configuration, and now it doesn't work at all. > No chip warming up, no action on the motor.

VDD is connected correctly, VMOT is as well. The motor pins each have .7V on them. !reset and !sleep are tied together, !enable isn't tied to anything (has an internal pulldown anyway) MS1-3 are pulled down. Everything I can think to check checks out, but it's not doing anything. The DIR pin is high, and the STEP pin shows .5 volts, which probably means it's getting pulsed correctly. I hope I didn't fry the chip or anything. But I'm not seeing/testing any cold joints, and as far as I know how to test it's all hooked up correctly. So why doesn't it move?

---

--
I'm building it with Baling Wire