Simple unipolar driver with current limiter

[DaRy]

Hi there!

I have made a simple unipolar driver for small steppers. I use this to drive steppers recoverd from discarded 5.25" diskette drives. It is based on a voltage regulator LM317, used "backwards" as a current limiter. The current is limited by the shunt resistor (10 ohm here) and the internal referens voltage of the regulator (1.3 volt), this gives a current of 0.13 A. This can be changed by replacing the resistor. A resistor of 1 ohm would give 1.3 A insted. With a little change and some more components it can be made adjustable. But I wanted to keep it simple instead. The LM317 can handle current up to 1.5 A, and a heat dissipation of 20W each. Heatsink is needed!

The TTL inputs is driven directly from a parallel port, data pin 2 and 3. And I've made a small c program (linux) to test the driver.

Comments... suggestions...


Have a good day!

[adamziegler]

Woah.. very nice! You know of anything like LM317 that can handle more Amps? I am also attempting to design a simple unipolar driver, but what you have looks as simple as you can get!

http://www.cnczone.com/forums/showthread.php?t=30670

[DaRy]

To get more amperage change the voltage ragulator. Here are some other types that might work (I have not tested... so... YAOYO):

LM317T: 2.2A, 40V (in the earlier post I was referring to the min 1.5A, not typical)
LM317HVT: 2.2A, 60V
LM338K: 5A, 35V
LM350T: 3A, 36V
LT1083CP: 7.5A, 31V
LM1084IT-ADJ: 5A, 31V

The higher current versions seems rather pricey except LM1084...
There is one more option and that is to use a powertransistor in conjunction with the regulator, but this ofcource adds to the complexity. Note that for some reason using this bridge the coils release voltage spike goes away. I'm not sure why, but I'm unable to messure any spikes at all (therefore I've removed these protective diodes, to add to the simplicity)

I forgot to mention that the driver is a fullstep 2fase driver that and it needs a qudra...um... something input sequence, but only using two signals instead of four.

Happy testing!

[Brandano]

Can this design be adapted to drive each coil in turn? I don't have much experience with electronics, electronics based on discrete components is black magic to me. How is the "matching pair" lm317 triggered?

[DaRy]

Yes you can drive each coil individually (you need halfstep I guess), but this requires a negative voltage... See the last attachment below. Beware, this circuit is untested and may not work, and burn... You have been warned!

I will try to explain the "matching pair" circuit.
See attachment 1. The LM317 acts as a voltage follower. The OUT is 1.2V higher than the ADJ and if you raise the ADJ then OUT will follow. If ADJ is grounded OUT will be 1.2V. The resistor will have 1.2V over its terminals and this will generate a constant current of 0.12A independent of the input voltage to the regulator. It is this current that then drives the stepper coil (not shown).

See attachment 2. Now we have two opposing regulators. Regulator B Has ADJ to ground, and so its OUT pin will set to 1.2V. Assume the current flows through B, and this drive one coil (not shown). Now if we raise the ADJ voltage on A just a tiny bit so its OUT voltage goes to 1.25V. B senses this tiny raise and lower its current to compensate so much that it is completely turned off. And now the current flows through A instead.

Does this explain it?

[pminmo]

You really only need two LM317...LM350. One for each phase in the centertap. Now where in a simple full step of simple half step sequence do you ever have both more than one/half phase energized at the same time. if you want to control the seq via a parallel port then put a simple mosfet switch on each of the 4 coil legs.

[DaRy]

Yes, you can do it with only two regulators... But atleast here in sweden one IRL640 will cost more then twice as much as one LM317 (and the complexity: six power componenets instead of four per motor)...

[pminmo]

Have you tried the circuit? If the coils were purely resistive I would see no reason your circuit wouldn't work, but they are not purely resistive. I would be concerned that when changing states, the field collapse would cause a start of a potentially unstable circuit that could result in a nice oscillator?

[DaRy]

Yes, it works beautifully now here on my desk... tested on a experiment board (see http://www.cnczone.com/forums/attach...4&d=1170699140). I guess it is the regulators job to do just that, supress noise and stabilize.

I was concerned with the coils release spikes, but for no apparant reason there was no spikes. Not a single volt. Can't really grasp why but, well, I wont complain or so. There is however normal back-EMF when the motors run at high speed. So there has to be some voltage space between LM317 max and supplied voltage.

[Brandano]

DaRy, thanks for th explanation, I doubt I'll ever be able to think in anything other than a "digital" mindset, but you made it simple to understand.

[666]

Excellent idea with the 317´s...

If any switching noise is present you couls add 2 100nF caps from Vin to GND and From Vout to GND.

It should kill spikes and noise generated if any present...

Regards from Denmark...

666

[666]

And the caps should also prevent the device from oscillating...