Stepping motor controller - is this chip suitable?

Why do you find that staggering? The chip uses switch mode current control and there should be no other losses than the switching and RdsON losses. There are SOT23 FETs rated at 5A.

[...]

Start off by reading the classic stepper tutorial:

--
Stef    (remove caps, dashes and .invalid from e-mail address to reply by mail)

Too much is not enough.

[snip]

Steppers usually win on grunt, servos on speed. That stepper is

26mNM, versus the brushless DC-servo's 11.8mNM. Maybe you're thinking about the torque of the servo motor /after/ gearing down?

-- Cheers, James Arthur

snipped-for-privacy@yahoo.com wrote

Actually, on a closer look, the stepper does about 10mNm and the brushless is similar. I was looking at the stall torque on the brushless one.

Stef wrote

Yes, I have seen them, but would *you* design a product which had to be highly reliable, over a wide temperature range, which used SOT23 FETs carrying 5A?

With a stepper motor driver one might need to adjust things like the current and various pulse timings, and if anything goes slightly out of perfection, a device like that will go up in a puff of smoke in about 1 second.

The one thing I have sussed out when looking at these stepper controllers is that most new ones work in "current mode" and one chooses a stepper motor which is specced at a small fraction of the supply voltage e.g. a 3V motor with a 30V supply. The resulting performance is much better, over simply switching a constant voltage onto the motor coils which is what all the older stepper chips do.

That was not the point, I just mentioned them because of your amazement by the rated current of a certain size chip. Just compare the rated currents.

But to answer your question: Probably not. I'd feel more comforable if there was at least a 2:1 safety margin. But I would have to know all operating conditions and read the entire datasheet for a definitive answer.

But you seem to be on the virge of dicarding a 10:1 overrated chip just because of it's physical size, that does not make sense to me.

Study the specs, and make sure it is overrated by a marging you are comfortable with under all expected conditions. And if the chip still fits the application, just be happy with the small size. :-)

In chips like these (IIRC you were considering the A3987), current is set by a resistor, switch timing is not adjustable and it has built in protections. Adding external components increases the possible points of failure and you must make sure your upper and lower switch times don't cause overlap. And if you use standard mosfets for your external switches, you lose the benefit of thermal shutdown.

These current mode drivers have been around for a while now, so I would not call them 'new'. ;-) But you can still drive your 12V stepper from 12V with one of these. Have you read the Jones tutorial? I think it has a section on the pro's and cons of current drivers (it has been a while since I read it myself).

--
Stef    (remove caps, dashes and .invalid from e-mail address to reply by mail)

Life is like a bowl of soup with hairs floating on it.  You have to
eat it nevertheless.
		-- Flaubert

The FET is the thing that's going to fail -- not the controller. Spend your money there :>

The exact time required is a bit over 0.83 ohnoseconds.

With full H-bridge drives, you have to be careful never to turn on the wrong pair of switches at the same time -- either intentionally or unintentionally. I.e., making sure an "up" switch turns off before turning *on* the "down" switch on the same side of the bridge (which would short the supply through these two switches).

Your controller needs to have enough dead time built into it to guarantee this is true IN ALL CASES.

With more compliance, you can accelerate "harder" and usually reach higher speeds for a given torque (or, higher torques at a given speed). Remember, you are fighting the back EMF from the motor so having more voltage makes it easier to "force" current into the motor's winding(s).

The ugly secret of using steppers in place of servos is that steppers have resonance points where the motor can easily stall. You either avoid operating at those points, live with GREATLY reduced torque at those points, *or* change the resonance of the motor by tinkering with the mechanical load.

The other little ugly secret is that you can't just "apply a current" (i.e., apply a set of stepping pulses) like you can with a servo and magically expeect it to work. You have to determine the maximum speed that you want to operate at and how quickly you need to attain that speed. Then, you have to characterize the motor/load to see what sort of acceleration profile is attainable. If you can get to your desired operating speed in "one step", life is good. :> Otherwise, you have to ramp up the pulse rate to the controller at the rate prescribed by the acceleration profile. Similarly, ramp down when you want to stop the motor.

IF YOU PUSH THE MOTOR TOO HARD AND IT MISSES A STEP, you are screwed -- unless you have some other feedback from the shaft to detect the missed step. I.e., you don't *know* that the motor has just stalled and that your "step pulses" are to no avail.

Much of this is affected by how you step the motor: full steps, half steps or microsteps. Microstepping is nicest as it tends to give you more control over the shaft (even if you never stop the motor "between steps"). But, it is more expensive and more difficult to implement. Full steps are much easier, by comparison, but you lose some instantaneous control, motion is coarser, etc.

I had a good reference on all this some years ago. I will have to see if I can locate it. Or, at least some of the calculations done to characterize the control systems...