[mardus]

Nobody talks about microstepping by unipolar stepper drivers. Why? I'm not count on controllers with microchips (SLA ....etc...). I am interested on discrette one.
Comments ?

[kreutz]

Most of the appeal and popularity of unipolar steppers is the simplicity. You can micro-step with a discrete unipolar driver, it will just make it more complex. Just add a micro-controller, couple of D/A converters (or filtered pwm outputs from the micro-controller), sensor resistors, two comparators and a chopper circuit. Plus some programming and, voilà!, micro-stepping on unipolar!

Now, after all of this, the cost is approaching a bipolar micro-stepper driver, so, most people will upgrade to bipolar micro-stepping instead, since it offers more torque from unipolar/bipolar compatible motors.

As a learning tool the design is a good exercise.

[mardus]

Thank you for fast response kreutz.
I avoid to build bipolar driver, because I am afraid of possibility to make a smoke, due to difficulty to drive upper mosfets, as to match right dead-times and 8 mosfets instead 4 and....
I have no experience about driving steppers and I want to try something simple for the beginning.
Could you tell me something about current relations between two phase
windings in case of unipolar microstepping?

Thank's again! I hope that you understand my pure English.

[kreutz]

Don't be afraid of smoke a circuit, Do you have any idea of how many blown components and fuses took us to get a little knowledge? It does not matter how many books you read (reading is very important before doing anything, you have to have the theoretical part as clear as possible), but it is the practice what glues the knowledge into your brain. Even with today's simulation tools there is nothing like a prototype to tell you if a design is working or not.

I don't know about your electronics background, so I've got you a little tutorial (that might be a little basic), let me know. I will help you as my time allows. Don't be afraid to ask either. There is nothing difficult about electronics. Like everything in life, you have to give it a try and learn by trial and errors.

Here is the link: http://www.stepperworld.com/Tutorials/UniTutor.htm

About current relationship it is the same as bipolar micro-stepping, current in the two phases is sinusoidal (or as sinusoidal as possible, depending on how many micro-steps/step), and there is a 90 degrees phase angle relationship between them (sin/cos)

[kreutz]

Driving upper side mosfets is not a problem if you have access to the right ICs. There are half-bridge as well as full bridge mosfet driver ICs available. There are circuit configurations that won't need those IC drivers, it all depends on what voltage and currents you are aiming to.

[kreutz]

More stepper driving information

[mardus]

I just trying to find some driver ICs and Now I'm reading.... Then I'll be back.
Tnx
Tnx

[bunalmis]

Nobody talks about microstepping by unipolar stepper drivers. Why? I'm not count on controllers with microchips (SLA ....etc...). I am interested on discrette one.
Comments ?

Unipolar microstepping driver more complicated then bipolar.

Minumum six power transistor and complex driving signals need.

Therefore not common.

(Problem is storaged energy at the leakage inductances.)

[kreutz]

Mardus;

If you are looking for IC solution to the unipolar microstepper take a look at this: http://www.allegromicro.com/sf/97060/

If you still want to build your own, using discrete components, then I can help you.

[mardus]

Bunalmis;
Where did you find six power transistors for unipolar microstepping? I never saw such design. I think that is enough four transistors (mosfets). Maybe I'm wrong, because I am new in this, but...
Yes, I suppose you mean on windings that's not in charge in some sequences. What's happening in that coils? There are some interactions between coils in charge and free coils..... Through armature....???
Kreutz;
I like to build using discrete components; it's good to learn something. With IC is easy, but I will learn almost nothing. And you know what? I am not in America or Germany or.... Here is very hard to find some special electronic components yet. Goes better, but steel ....learn with discrete buildings.

If you willing to help me and you have enough free time I'll be appreciative.

Regards
[email protected]

[bunalmis]

Search feed forward topology. (This therm for switch mode supply)

Also this address may help to you.

http://www.freepatentsonline.com/5015937.pdf

Unipolar step motor same as push-pull transformator. All transformator has leakage inductance. If you drive it by soft signal no problem (forexample old sound amplifier) but if you want to drive with square wave signals problem begins.

One simple way you can select the four high voltage transistor and snubber circuits. Other methode is 6 transistor.

[James Newton]

Unipolar drives require only 4 drive transistors. Bipolar require 8. I'm not sure where bunalmis is getting 6, but I would be happy to be shown an example.

The Linistepper is an Unipolar, Microstepping, Open source, Current regulated Linear controller with 4 Power Transistors.
http://www.piclist.com/techref/io/st...step/index.htm

With a really nice kit of all parts, or just the PCB's available.

You haven't said how big your motors are, but if they are less than around 1.5 amps per phase and 12 volts or so, then it will work just fine for you.

[kreutz]

Bunalmis;

There's more than one way to skin a cat. We are not trying to make the most technologically advanced or the most efficient unipolar micro-stepper. Just another working, reliable, and cost effective unipolar micro-stepping controller.

I read your link. It is a clever but a little more complicated way to do the same we are proposing. James Newton utilizes a similar approach on the linistepper, only changing the current controlling and sensing mechanism, which in J.N.'s case is also simpler.

So far leakage inductance has not stopped anyone to make unipolar steppers work, and if it is of any comfort, we don't want to push square current waves on the coils, only a quasi_sinusoidal chopped wave (almost like an push pull type linear audio transformer).

Keep coming to this thread, because you are going to see it working.

BTW, James Newton: A few days ago I P.M. you a suggestion, and you did not answer back.

[kreutz]

mardus;

Lets begin to work!

What parts do you have access to?
Do you have access to: Opto-insulators, Atmel AVR micro-controllers, ttl logic gates, power mosfets or bipolar transistors, Proms or EEproms, and 1% tolerance resistors?.

I need your motor's specifications. What is your electronics background? Have you assembled any electronic kit before? Where are you located? Any big Electronic Components distributors nearby? (or PC junk-yard?)

[bunalmis]

I dont discuss the linistepper. This is an analog and small size driver. (Flea power)

But my suggestion for high power applications. Some small problems enlarge at the high current.

Six switch method is good for unipolar motors but eating yogurt of each brave is different the others.

[kreutz]

bunalmis;

I agree, linear current control, linistepper style, is better used for low current motors. The limitation is in the current regulation vs temperature and thermal runaway possibility at high currents. Some improvements are still possible using that line.

Anyways, your search for a better solution for high current motors is probably the right way to go. I find it interesting to explore.

Now my question: When somebody else buys a high current unipolar motor is because he/she wants high torque, isn't it? Then, why not to buy/build a bipolar driver and same rate bipolar motor? Do you think it is practical (as a business like decision: cost vs performance, return of inversion time, etc.) to dedicate resources to follow that line? If you think Yes, then, by all means, follow it.

As a learning experience is think it is going to be a good one, even if no business/money is involved. Doing it for the sake of knowledge.

[mardus]

Kreutz;
Lets !
Accessing parts:
- Opto insulators ............ Yes ( like 4n25, 36... CNY74 (4xoptoInsul.)....
- Atmel AVR.......... I am not familiar with it. I can work with PIC family MCs.
- TTL logic gates.............. Yes , no problem.
- Power mosfets ............... Yes, like IRFZ44, IRFZ48, IRF9540....
- Proms, EEPROMs ............ Yes, but have no programmer for PROMs, except for some serial EEPROMs (PIC programmer can do it)
- Resistors ........................Yes, no problem to find
- I can't find any mosfet driver ICs here, like IR2184, IR2104, TC4467,68....

- Motors:
I am waiting for (purchased) , and have to prepare some driver for Chinese type stepper 4.74A, 0.9 Ohm, 10.8mH, NEMA32.
I like to be bipolar, but how to drive mosfets efficiently without mosfet drivers? It's seamed to me that unipolar is easy with less troubles. Now, reading this posts, I am in doubt. I like to try both, just curious!


- Now, for practicing I have on site:
1. "Astrosyn" Nema32 , 9.5V/phase , 0.6A(?)...6wires
2. "Vexta" 4V, 0.95A................................6 wires.......... 1.8deg.... little, but very strong, exelent!
3. "Minebea" 24V, 0.2A/phase................... 5 wires.......... 1.8deg
4. "Minebea" 3.5V, 1.1A/phase ................. 4 wires ..........1.8deg........ bipolar
5. "Berger-Lahr " 12v/ phase ................... 8 wires ......... ??? deg


- Background ? ...... Well........ "When I was younger , younger than today" ...... I was familiar with practical electronic. Yes, I built many projects, mostly digital and then there was a gap of about ten years ( done nothing about electronics). Last year I was born again and I am starting again. It's great hobby to forget other troubles and to make you happy.
I knew to use and programming in assembler some old processors, like 8085 and Z80. I remember yet, but have no condition....
Now, I am able to write assembler programs for PIC16 Microcontrollers . Only assembler, I don’t like High-level languages, but I can use on beginner level Pascal and Basic ....
I know what I was, but I want to see what I am now, can I build something useful.....

[kreutz]

Mardus;

Here is the plan;

The design process begins by gathering information on prior art, So you don't lose time reinventing the wheel. Here the wheel is already invented, so we are going to take the best of already successful designs, and do the same using discrete components.

Download the data-sheets for the L297 and the L6506, read them. Since this is a learning exercise we are not going to use them now, I just want you to see how they work so what we are going to do later makes sense to you.

As you have working experience with PIC assembly it is good that you see how the normal full-step sequence is used to drive the power stage. We are going to use the same sequence to drive the coils in the right moment.

Also, I will leave you the programming part, so think in terms of doing the same as the L297 chip by using a PIC (only the translator part of the L297). Do a flowchart of sequential programming steps and P.M. me. We don't want to overload the thread with our brain-storming sessions.

We are going to build the chopper controller in hardware by using comparators, Lm555, TTL gates and R-C delays. Then after we have this part working on the proto-board we are going to add micro-stepping in software using two PWM channels and R-C filtering as reference voltage for the chopper.

I also have a PIC programmer at home. You get to choose the micro-controller, the pin assignment, inputs and outputs, just tell me why you did so, in order to be in sync. I will also test your firmware. Lets begin with a 4 MHz clock, We can always upgrade later if we need to. Send me the proposed schematics for your PIC including power supply filtering, I will ask for explanations on your choices so the process is going to be like a real project.

Do you have a scope? if not, I will have to make prototype too, so you can see the waveforms.

Lets use the "Vexta" 4V, 0.95A" as test bench until you get the big one. You are going to need a 24 to 48 volt (minimum) 10 A power supply.

If you are going to make a prototype for the micro-controller part lets make a separate board for the power stage, so at a later time if you want to make a bipolar driver you can still use the same micro-controller board and even the same firmware with just a couple of improvements. After everything is working we will design a proper PCB for the complete drivers.

[pastera]

I have a bunch of PIC18F4433's in the QFN package if you want a donation.
Ordered 26 when I wanted 6 - I should pay more attention, but at <<$3 each not a problem.

I am working on a home project - PIC based servo controller with EPP (enhanced parallel port) or set/direction interface. Currently looking at 10 amp 80 volt SOIC based fets with footprints for TO-220 or TO-247 on the PWB. Adding a second H-bridge or reconfiguring for unipolar is not an issue. High side current sense (discrete) for added safety.

Aaron

[kreutz]

Pastera;

We could use a couple of them, I will let you know when we are ready.
Thank you for your offer. If you need anything just ask.

Are you looking for the Mosfet drivers or just the power FETs?, I've got a few IRFP260NPBF 200v 50A TO-247AC (4 of them) that I can donate to your project. I was planning to work an a servo driver using AVR, it was placed on hold until my next vacation.