Help Request: Simple Stepper Hardware with EMC's HAL

[adamziegler]

Hello all,

I have some 8 wire 5.4V 1.5A 0.9 deg/step stepper motors. I think they are in the realm of 100-120 oz/in. For the time, I plan to go unipolar with the steppers. http://www.cnczone.com/forums/showth...t=vexta+8+wire

I have successfully configured EMC's HAL to act as a half stepping 4 phase pulse generator. Pins 1-4 on my parallel port single through an optocoupler. The optocoupler output triggers a combination of 4 transistors... (1 transistor for each phase.)

My power source comes from an old computer. It is a ~150W switching power supply.

My electronics background is "ok" but I have been out of it for a number of years, and have been re-climbing the learning curve.

I need some help picking out an adequate transistor. I think I am looking for a mosfet NPN that can handle at least 1.5A. I managed to parallel a couple (per phase) 800mA NPN's I had laying around for a test... stepper moves fine, but when I stop the stepper the transistors get very hot. (They were only rated at 1.8W a piece) These were not meant to be permanent... just a test.

I have also noticed that I can run my steppers faster at greater 5.4V... is this a bad idea? Do I need to add something else to the circuit to protect the stepper?

Should the power source be modified in any way? Any other suggestions or words of caution? Any other information I can supply?

Thank you in advanced for your help.

[kreutz]

Feeding more than the 5.4 volt will create currents higher than the Imax stated on your datasheets for that motor. In order to do that, without risking damaging your motors, you need to limit the current by either adding current limiting resistors, en series with each phase's central tap (the one connected to VCC), or a chopper circuit.

Look at this thread unipolar motor microstepping there is a lot of information there, look also in the Pminmo's website (http://www.pminmo.com/ ) and Gecko's website http://www.geckodrive.com/

[adamziegler]

Great links! I am familiar with some of those links, but need to study some more.

A couple quick questions. As is now, I have VCC going through the each phase with the central taps on ground. In recent days I have noticed it is usually done the other way around. ( http://www.cs.uiowa.edu/~jones/step/circuits.html ) Is there a particular reason for this?

A chopper seems the best way to go, and the resistor are simple but inefficient. (Does that sound like an accurate assessment?) For now I will probably look at using the resistor method. Since the stepper is rated at 1.5A, and my power source could supply 24V...

24V / 1.5A = 16 Ohms >>> 16 Ohms – {coil resistance} = Resister needed. 36 – 40 watts? Does this look about right?

[kreutz]

I have VCC going through the each phase with the central taps on ground. In recent days I have noticed it is usually done the other way around.

Reason: It is easier to drive a low side N-Channel Mosfet (reference to ground) than a high side Mosfet, where you must get between 12 and 15 volts above VCC in order to drive the gate.

A chopper seems the best way to go, and the resistor are simple but inefficient. (Does that sound like an accurate assessment?)

Yes, it does.

24V / 1.5A = 16 Ohms >>> 16 Ohms – {coil resistance} = Resister needed. 36 – 40 watts? Does this look about right?

You will need (24-5.4) / 1.5 =aprox 12 Ohm, 50 watt resistors (2), your resistors will disipate 1.5 x 1.5 x 12 =aprox 28 Watt.

[adamziegler]

You will need (24-5.4) / 1.5 =aprox 12 Ohm, 50 watt resistors (2), your resistors will disipate 1.5 x 1.5 x 12 =aprox 28 Watt.

For the heck of it, I measured the resistance of one of the stepper coils... 3.6 Ohms. V = IR >>> V = 1.5A * 3.6Ohms = 5.4V (which is the voltage rating of the motor.) Is it always this straight forward?

Ok 12 Ohms and 28 watts follows... for one coil. Now 2 coils at the same time... the 50 watts follows.. but is it still 12 ohms? If so could you attempt to explain why?

[kreutz]

You have 4 coils in two groups, the coils in each group are joined at the center tap where the connect to VCC. You place each resistor between VCC and this tap. Hence 2 resistors.

From each group only one coil (half) will be activated at the time. So the peak current flowing on each resistor will be about 1.5 amps.

Ohm's Law is always true: I * R = V

[kreutz]

50 watt resistors are needed in order to dissipate heat without burning the resistors. You always leave 50% margin on power resistors.

[adamziegler]

How does the attached image look?

[kreutz]

Why do you want to use PnP Transistors in that configuration? What kind of driver circuit will you use for that configuration?. Look at the data-sheets for your transistors, your base-emitter junction will be very stressed when switching inductive loads.

Why don't you use Npn transistors, Collectors connected to the coils, emitters to ground, that way they are easier to drive.

Use anti-parallel fast diodes (at least 3 Amperes) across collectors and emitters for each transistor.

Mount your power transistors to a good heatsink, use thermal pads (mica or thermal) between the metal tabs and the heatsink..

[adamziegler]

Why do you want to use PnP Transistors in that configuration? ... Why don't you use Npn transistors, Collectors connected to the coils, emitters to ground, that way they are easier to drive.

My test circuit is setup to use NPN, but after reading various sources such as the one I posted above ( http://www.cs.uiowa.edu/~jones/step/circuits.html ) it looked as if the norm was to switch on the ground side of the coil. I asked about the reasoning for this in which you replied:

Reason: It is easier to drive a low side N-Channel Mosfet (reference to ground) than a high side Mosfet, where you must get between 12 and 15 volts above VCC in order to drive the gate.

I took that as meaning it was the recommended method. I guess I am a bit confused now.

What kind of driver circuit will you use for that configuration?.... Look at the data-sheets for your transistors, your base-emitter junction will be very stressed when switching inductive loads.

I am useing EMC's hal to supply the pulses. Hal has a pulse generator, and each phase can be assigned to a pin on the parallel port. Please see page 72 of http://linuxcnc.org/docs/HAL_Documentation.pdf (Step Type 9) for the step type I plan to use.

As of right now, I do not have driver specific transistors picked out... but I will look into the recommendations you made. Thank you again for the help!

[kreutz]

Originally Posted by kreutz
Why do you want to use PnP Transistors in that configuration? ... Why don't you use Npn transistors, Collectors connected to the coils, emitters to ground, that way they are easier to drive.

Look at figure 3.7 on the link http://www.cs.uiowa.edu/~jones/step/circuits.html , that is what I meant.

[adamziegler]

Look at figure 3.7 on the link http://www.cs.uiowa.edu/~jones/step/circuits.html , that is what I meant.

Ahh!! Yeah I feel dumb. Thank you for your patience.. I understand now!