Question: Starting a stepper driver?

[chuckknigh]

OK, I've been thinking about the simplest possible driver circuit, thinking about the open source project. That would be a 4 transistor (per axis) driver, using wave control under TurboCNC. A few transistors, some diodes, and some resistors wired up to the parallel port should suffice, though I welcome other input.

My reason for asking is, I'm working on a <$20 "proof of concept" model with a single axis. A CNC Primer unit, if you will...which can function as a Z axis when completed. I intend for it to be an "I can actually do this!" type of project...simple, cheap, and easily understandable.

But, the more I think about it, the worse it gets. (Of course, I have a tendency to overanalyze everything)

OK...we have 4 coils...numbered 0, 1, 2, and 3. The motor is "stopped" at a random coil, let's say coil 2, and power is removed / computer is rebooted. That means the next step would require a pulse on coil 1 or 3, depending on the direction...that would be the next step.

But, since there's no feedback as to the position of the rotor, how does TCNC know where to start the pulses when it's first started? It could be, depending on the orientation, as much as 2 steps off at the start.

Extending this logic, it would seem this problem would exist in all stepper driver circuits. Any time power is removed, do the chips "zero" themselves back to the initial step? Seems like that's how it would work.

So, how is this situation handled? I really can't figure it out. Help?

-- Chuck Knight

[berin]

The easyest way to solve the problem is to goto home then go back to where you where last. But this only works if you know where you where in the middle of a running cycle when using some kind of a drawing/cutting setup.
As for the easy single axis driver, you need only 4 darlington transisters which have built in diodes, You DONT have to have 4 resisters for the stepping phase controll when run directly from the parallel port but it is reccomended. As for setting up a home switch thats easy enuff.

[chuckknigh]

The resistors I mentioned were pull-up resistors, so that the signal would be sufficient to trigger the transistor switching.

Are these not necessary with this setup? I'm assuming a "not quite" compliant port...they all seem to be not quite compliant.

-- Chuck Knight

[ESjaavik]

When the drive starts up, it does this in one known state. It energizes the windings with this state and the motor assumes the nearest position that corresponds to this state. So it may move up to half a step in an arbitrary direction when powered up. In fact you can strike "may". Usually you can hear the motor move as you turn on the drive. From there on, the driver just sequences around left or right and assumes the motor follows it. And as mentioned, the homing routine makes it possible to synchronize to the actual position of the machine.

There are tons of copies of "your circuit" already on the net. Go Google. What is your task is to sort out those that does not work. No diodes to kill the spikes from the motor is the most common. If you try that, make sure your transistors can be changed in a few seconds, and buy a lot of them.

[chuckknigh]

I did google for it -- without the right search terms, they're a little hard to find. I found one in an old robotics book, though...

Regardless, that's what I was aiming for. Something which, despite buying parts at Radio Shack, could be built for $5 or so. Even on a breadboard, if desired.

It doesn't have to work beautifully -- just work as a proof of concept model for a single axis, to prove to people that they can actually do this CNC thing, without having to pay $100+ for Geckos, before anything even moves.

I'm not an electronics guy. If I can build one of these (and ever find enough continuous spare time to focus) then anyone can.

-- Chuck Knight

[pminmo]

I thought I had submitted my answer but I don't see it. Anyway it the motor goes to the state the drive resets to. Think unipolar all four swtich lines open. Motor won't move. If it resets and one of those lines is switch on, the motor will go to that position, direction is only predictable if it the motor in the power off state is positioned at an adjacent coil.

Phil

[Konstantin]

What about that Luberth wave drive.
Only one IC used. LNxxxx dont recall.
The simplest you can get.


Konstantin.

[pminmo]

ULN2803 a darlington transistor driver set with built in clamp resistors. They can be run directly off the parallel port. BUT only sink 500ma, and as I recall there is a watage and sum total limit on the combination of the drivers.

[arvidb]

I built a universal I/O interface for the parallel port some 7 years ago. It used the ULN2803 and I ran steppers with it (the kind you can find in 5.25" disk drives). I did however not know about the clamping diodes, so for some mysterious reason I had to change them every now and then... It was breadboarded, BTW, and ran from a DOS program I wrote in Pascal.

But it worked!

Arvid

[berin]

you know now that I think about it, there is one other way to get back to the very possion you left from if the computer should shut down. Say you use a pic chip like the 16f867 or there abouts. It could be the step to phase converter as well as a number of other things, and a half/full step converter. but it could also remember where it is, just by counting. a quick home and it could be right back where it started. OR it could remember what winding was energised LAST and start that one up again so the motor wont move anyware!

The nice thing about that is that it could also read posision encoders.

speak of has any one created a ONE chip, 3-4 axis all in one step to phase converter? using some kind of a single pic chip?

[berin]

Originally Posted by pminmo ULN2803 a darlington transistor driver set with built in clamp resistors. They can be run directly off the parallel port. BUT only sink 500ma, and as I recall there is a watage and sum total limit on the combination of the drivers.

You can run thoughs in pairs and get 1000ma, good to drive only one stepper motor per uln2803 but its more powerful and doesnt get as hot. then again you can always use a 100ohm dropping resister and remove even more heat.

[sandos]

I have done something along the lines of the topic:

http://sandos.ath.cx:8080/~sandos/pics/GBA_stepper/

Breadborded four power-transistors and controlling them from a GBA (this should readily be substitutable by the parallell port of any PC, the GBA EXT-port gives out 3.3V though) and four protective diodes that Im not sure are fast enough for the task.

The circuit isnt very complicated, although it did take a few hours to get it all right. (I did one phase at a time, also Im buying COLOURED wire next time). The purpose was mainly to experiment with how strong that little stepper was, conclusion: supplying 11V to a stepper rated at 14V and energising one or two phases doesnt get you much torque. I could vary the speed by pressing keys on the GBA.