3-phase stepper driver inquiry

[extrapilot]

Hello

I have a low current/volt (3amp/12V max/phase) application that requires 4 3-phase stepper motors.

It seems that integrated 3-phase drivers are about as common as hen's teeth. The one I can find is the SLA7611M from Sanken, but that part is rare in the US (apparently there are 30 units available in the US...) and lead times are horrific. Pricing is good on qty 144, but the lack of access is concerning- and Im worried they may discontinue the part due to low demand.

I suppose I can just make my own via an 80-pin PIC and a charge pump, along with some FETs and a little phase coding. But, I was wondering if anyone has implemented a BLDC controller for 3-phase stepper applications; I think if you just managed the hall effect feedback sequencing, it should work. Im not sure if that saves me any pins or coding, but it definately saves on part cost, PCB real estate, and complexity on the current limiting, etc.

Any thoughts are appreciated-
Regards,
Rob

[AbSat]

I build 3 phase drivers with the HA13532 controller from Hitachi and the SLA6023 driver. But both saved from dismantled kopiers.
stragenmitsuko programmed a sequencer for a 3 phase star motor in two 16V8 GAL's and I put them before the SLA6023 and that is working well as long as you do not need microstepping. Only full and halfstep can be done but with my motors halfstep is 1200 steps/revolution and I am running my router with the HA13532 controllers in halfsetp too.
Look at Designing a 3ph stepper driver for the thread stagenmtsuko started about designing a 3 phase stepper driver.
In stead of the SLA6023 driver a set of bipolar darlington transistors can be used of course.
I am now building a more or less universal driver with three SLA4391 drivers I have. I will have 3 full bridges and 6 halfbridges then to test with 3 phase star or delta, 3 phase bipolar and 5 phase configurations. I tried to use mosfet drivers, but that stranded at the complexity of driving the high side mosfets.
If the bipolar drivers are running well I will start again with mosfet driverstages.

AbSat.

[extrapilot]

Absat-

Yep, I saw that thread on a hybrid config. I can do that no problem with a high pin count uC and some half bridges- programming is simple for it. Problem for me is real estate, since I need 4 of these on a small board. Maybe if I used an integrated power stage that would save some space, vs 24 fets and some support circuitry for current limiting and high-side drive.

But really, the subject of my inquiry is- can a BLDC motor driver be used for a 3-phase step driver (full step only) if the hall effect inputs are sequenced? Or more to the point, has anyone tried this and had any luck?

Best,
Rob

[AbSat]

But really, the subject of my inquiry is- can a BLDC motor driver be used for a 3-phase step driver (full step only) if the hall effect inputs are sequenced? Or more to the point, has anyone tried this and had any luck?

Best,
Rob

Nice thought, never came to my mind.
I have no experience in this field but I took a look in the datasheets of some BLDC driver chips and the following questions popped up in my mind.

What is the advantage of using a brushless motor driver ? Only the availability of one?
You still need a sequencer in front of it, in fact the BLDC driver would be only the driver and, if usable, the current limiter. As far as I can see this type of drivers use pure PWM for current control and a current limiter too avoid damage when the motor stalls. In stepper controllers a chopper is used for current control. The whole circuitry about the frequency generator, crystal oscillator and PLL cannot be used for driving stepper motors.
Maybe a nice Mosfet powerstage would be worth using such a device.

AbSat.

[extrapilot]

Well, the thought was that these have internal half bridges, internal charge pumps, dead time logic, and some current limiting, all in a small and inexpensive package. Yea, still requires a uC for base sequencing, but it seems feasible and wondered if anyone had tried (and failed, or had success). I think microstepping is the major limitation here, but in my application it is not required.

I will probably just do it with discretes, since Sanken seems to be very difficult to work with (as a supplier).

R