Driving an old stepper motor: recommendations?

[John_BS]

Some years ago I rescued an old disc-drive from the era when there were 5 eight-inch platters employed to store 20Mb of data! It was, of course, beautifully engineered, and the head-block was driven by a stepper motor via a stainless pre-tensioned steel band wrapped around a drum on its shaft. I can't find any data on the motor (Sanyo Denki 103G775- ) but it appears to have four identical windings, each c. 3 ohms + 4.4mH / 1A and a step-size of 1.8 degrees. I'm thinking of building a kind of wave-winder, and I wondered if anyone had experimented with driving a motor of this type? I don't need oodles of torque, nor speed.

Cheers

John

[jay_oldstuff]

I'm afraid I can't help with your problem.

It's a pity you don’t have the driver board from the drive, with that you could work out what was originally driving the stepper and try to find a modern equivalent.

As an aside I too am working on a basic coil winder based on stepper motors. I’ve started with the chassis from a small dot-matrix printer where the track feed and print head are controlled by two stepper motors, I’m considering using an arduino for control. I've managed to salvage the stepper driver ICs from the printer.

Jay

[julie_m]

If there are four independent coils, not connected to each other at all, that is a good sign as it means you can use a simple "four-phase unipolar" drive circuit. You just need a suitably-rated power transistor or FET to drive each phase from a microcontroller or simple logic circuit.

You can actually work out the order of the phases by trial and error. If you energise phase 1 followed by phase 2, the spindle will move one step clockwise. If you energise phase 1 followed by phase 4, the spindle will move one step anticlockwise. If you energise phase 1 followed by phase 3, the spindle will move two steps in a random direction.

For maximum torque, rather than just energising one phase at a time, energise phases 1 and 2 together, then 2 and 3 together, then 3 and 4 together, then 4 and 1 together to go clockwise; or 1 and 4, 4 and 3, 3 and 2, 2 and 1 to go anticlockwise. For half-steps, energise two phases together, then just one, then 2 together and so on.

[Ed_Dinning]

Hi John, plenty of circuits about for this use and several have been published in the likes of Elektor over the years.
Usual trick is to fit power resistor of 1 or 2 ohms in the feed to the coils and then step in sequence. You can try it as a demonstration with a rotary switch.
Typical motor like this is 200 steps/ rev, but it is best to ramp the speed up to maximum or it may miss steps. It is also possible to 1/2. 1/4 and micro step these motors by energising more than one coilset at a time for smooth drive
We used this type very successfully with my son's Meccano many years ago, much better than the supplied motors, variable speed and can be locked at full torque by leaving a coilset energised.

Ed

[kellys_eye]

8-wire steppers can be configured to ALL possible stepper modes (with the resulting pro's and con's for speed, torque etc) but they are by far the most versatile devices as far as stepper motors are concerned. Once you figure out the winding sequence then you can use them as any version you want and use the appropriate driving electronics.

Use a proprietory stepper motor driver IC (Allegro do a great range) as they include all the circuitry for current management, step, direction and even microstepping if you need it.

Such devices are very cheap and require very few external components to make an interface to this type of motor.

Complete A39** devices on fabricated boards are available requiring only the step and direction signals for full control.

[John_BS]

Thanks for the feedback everyone.

I can find a huge number of near-identical Chinese stepper-drives on ebay, but none using the Allegro chip-set. Most seem to just have an H bridge with no integrated half-step etc. ( e.g. L298N).

John

[julie_m]

If you're going to use a microcontroller to control your stepper motors, you don't need any fancy driver ICs. It's easy to do the driving in software. You just need one GPIO pin per phase. Use logic-level power MOSFETs to do the switching, and don't forget that the anti-kickback diodes must be able to handle the same amount of current as the winding.

I may be wrong but I thought that a MOSFET didn't need an 'anti kick back' diode, but the avalanche rating coped with it.

[electroanorak]

I have a few of these, and used one a while back. I cannot remember the connections, but I did have one running and driven with BUZ11 MOSFETS.

Daniel.

[kellys_eye]

Use "allegro stepper" as the search word in eBay and you'll find plenty of them.

[McMurdo]

If you're thinking of using the motor to generate the waving action of the winder you'll need a ramp generator. Motors of this size dont like being instantly started and stopped or reversed. They will ideally need to be operated from a trapezoidal frequency profile (ie ramp up to speed and ramp down to stop). Otherwise the motor may randomly lock-up and just buzz. If you're doing the control via some sort of micro then you'll programme that into the action of course!

[John_BS]

Thanks for all the advice; I've ordered an Allegro board and I think I've now understood how micro-stepping operates!

I'm now pondering how a wave-wound coil should be specified: the engineers who designed the machines to wind self-suppporting cotton bobbins obvioulsy knew a thing or two..... on a reel 70mm wide there are 5 turns for one double traverse across and back to the starting point, with an OD of 45mm. This gives a "crossover angle" between layers of about 22 degrees?

I'm assuming that with relativley slippery cooper wire, it won't be possible to lay the wire down with an agressive slew angle. I wonder of this is why many coils were wound with rayon or silk covered wire?

John

[emeritus]

The construction technique of VMOSFETS means that an integral antiparallel diode exists between the source and drain. VMOSFETS therefore need no additional diode to cope with reverse voltages. Other types of FET may well need an additional diode.

[Radio Wrangler]

The antiparallel diode in VMOS is the right way round to stop the drain/source pair going reverse biased, but it won't do anything to stop the drain voltage going excessively high from the turn-off transient of an inductive load connected from drain to supply.

In fast circuits (high freq switchers) the body diode proves to be much slower than the MOSFET, and if it ever gets turned on, things expire. You'll sometimes see schottky diodes across drain/source to do the job without risking body diode conduction.

A diode across the inductive load is fine at low speeds, but it acts to maintain the inductor current for a time after switch-off, so fancy schemes like PWM have to get creative and you'll find things like energy-recovery snubbers, where diodes clamp the transient to a high rail... like a boost rail in many ways.

David