[raymondg]

Well, good timing then. Looking forward to your progress.

Thanks Again,
Raymond

[Mariss Freimanis]

Don't expect anything real soon. I'm on this until it's done. The goal is closed-loop from zero RPM to 3,000 RPM. Zero motor heating at zero load. Just like you'd expect from a true servodrive.

Mariss

[raymondg]

Mariss,
In post#24, you showed a proto board for this drive. Did you ever get a chance to test it? Also, 3000 RPM seems high for a stepper. Does this mean that this drive can turn a stepper motor faster than a "normal" stepper drive like a G201?

Thanks,
Raymond

[Mariss Freimanis]

I had many chances to test it and I used every one. After that, the board in message #24 was discarded, never to be used again.

There is a little misunderstanding in how all this works. I start with an ideal of what I want to have. Just like you, I dream. I imagine in my mind what I want the ideal servo to do. I hear it, I touch my fingers on it as I load it and in my my mind it responds as I want it to do. I then dream it while I'm asleep and and I fix it in my mind while I'm awake. It's something I do 24 hours a day, 7 days a week.

Once I have the ideal in my head, it sits on the rim of the Grand Canyon while the physical realization sits on the far rim. What yawns in between is the width and breadth of what I don't understand. It separates me from where I am to where I want to be.

I have the patient power of ant to bridge that yawning gap. Every day I work on how I can make what's in my mind to become real. I build a bridge across the canyon one grain of sand at a time.

Like everyone, I become stale. This is hard work and I become tired. I then go off to do something else while my mind subconsciously restores itself and burps up new ideas. I then go back and add new grains of sand to the bridge that will ultimately span the canyon.

This problem is hard. I laid it aside for over six months and tossed off some easy stuff like the G540, redesigned the G320 into the much superior G320X and the the G201 into the G201X. All the while my mind was working out the step servo while it was resting.

It's done resting now and it is percolating with fresh ideas. That's how I know it's good to go. Ideas I hadn't thought of before that cut through previous obstacles. We'll see if they are any good or not.

About circuits and boards: Each gives a little knowledge I didn't have before. Once each gives up every bit of what it has to give and then it has no further value. The knowledge gained gives rise to the next circuit and so on; each is a stepping stone, adds towards the bridge and what lies on the far rim of the canyon.

You really have to see the Grand Canyon. What separates you and the far rim is only a mile or so. Yet you look beneath your feet and the gap between sinks a mile below before it rises again to the other side. The far side is a hopeless entirety away unless you build a bridge or know how to fly.

Mariss

[CJL5585]

Mariss,

I salute you for your endeavor into an unknown and unexplored area of motion control. I also wish you great success in your quest.

When this is completed and successful, You should make technology history, and be added to the list of people who have been responsible for some of the major inventions of our country, just as Eli Whitney, Samuel B. Morse, Thomas Edison and many others.

Thank you for your endeavors and contributions.

Jerry

[raymondg]

Mariss,

An eloquent metaphor of the creative process made all the more poignant for my recent relocation to the Grand Canyon State, and having lost a race to make an inexpensive high powered microstepping drive to the G201 years ago.

I think we are all very happy you do what you do, even when we are still tossing in a few grains of sand, only to look up and see you smiling from the other side.

Raymond

[amplexus]

I like the grand canyon metaphor, the whole dream problem solving thing is so similar to the way I look at problems it is scary including setting things aside for a while learning some stuff and returning to them later. Do you wake up in the middle of the night and have to get up and write things down? Does everyone dream design and code?

Amplexus Ender

[.xXACEXx.]

i have to say,when i first saw the post about this stepservo thingy (yes im electrically illiterate) i was thinking who is this Mariss guy , what he wants to do sounds impossable to me (again im electrically illiterate)

after seeing what he has done with the g250 /g251 thread i have the upmost respect for him (yes you Mariss, stop blushing ) i only wish i had the time to devote my self to becoming as good as he was at age 7 you must have been a super bright kid Mariss .

oh and if ever any of your knowledge leaks out, can you some how e=mail it to me



===just sitting back in the corner in awe, waiting to see what happens next===

[amplexus]

xXACEXx
No need to wait for the knowledge to leak out, Mariss is working overtime to help anyone willing to take the time to learn. Read the cpld tutorial, hell read everything he has posted. If it doesn’t make any sense think about it for a few days and read it again, it will eventually make sense and you will be electronically literate. Learn verilog it is easy and very cool. It doesn’t matter if you are old, even my father is learning verilog, although he has the advantage of more analog experience than I have. Steppers are subtle and amazing devices, especially when run closed loop as a PID servo, actually Mariss hints at losing the integral term to prevent hunting (still thinking about that tidbit)
Amplexus Ender

[DeusExMachina]

Oh this is amazingly cool, basically converting off the shelf stepper motors in to high pole, low speed high torque servos.

I really hope this project gets finished within the next year, I'd love to get those controllers for my next machine build!

[Mariss Freimanis]

I have knuckled down now to pounding out the step servo and I'm not coming up for air until I have it. This work is taking all of my attention to the exclusion of everything else.

I left-off on this project about a year ago because I had run out of ideas. I did something easy instead like the G250 and the G540. When I say easy I mean there was nothing technical that was uncertain about both projects. Those designs were followed by the tougher to design G320X and G201X that will make an appearance next month.

I'm back on the step servo because ideas have returned. The long down-time allowed things to percolate in the back of my mind.

I start bread-boarding tomorrow. The circuit design and the Verilog code is as far as it can be taken right now without getting Mother Nature's approval. If she smiles, this important section will work; if she doesn't, it means there is more to learn and understand.:-)

If she does smile, what I hope to have when done is an open-loop torque-mode (transconductance) amplifier for a step motor. Its only input will be a torque command, it will have a near 0-degree phase shift and a motor that only draws power as a function of load. Closing a PID loop on such an amplifier is a forgone conclusion.

Mariss

[CJL5585]

Mariss,

Hope you are taking lots of detailed notes for the patent application; Then, you may already have a really good patent attorney working on it.

Again, I hope Mother Nature smiles upon your work. It would be an entirely new direction for the stepping motor drive industry.

Jerry

[amplexus]

I'm not asking for anything that would interfere with your patent but I don't understand how a stepper is any different than any other brushless DC motor. I can see that the typical number of windings is different and the stepper obviously has more poles but other than this they seem fundamentally the same. There is a lot of prior art in the field of bldc motor drives.
Amplexus Ender

[Larken]

If she does smile, what I hope to have when done is an open-loop torque-mode (transconductance) amplifier for a step motor. Its only input will be a torque command, it will have a near 0-degree phase shift and a motor that only draws power as a function of load. Closing a PID loop on such an amplifier is a forgone conclusion.

Cool, but I think (know) there are a lot of closed-loop steppers systems in the works right now. Its the obvious next step. But it nothing new and has been done in the past.
I think Its best done with dynamic sine generation and not fixed lookup table like a microstep drive. I agree, its more like a BLDC motor

Larry K

[Mariss Freimanis]

Larken,

What will be new is price. Current step motor servo drives are in the $1,000 to $1,500 range while we are shooting for $150 to $200. Some of the $1K drives performance is unsatisfactory (Vexta Alpha drive).

Used in a servo, you are looking at microstepping from the wrong perspective, dynamic (?) or otherwise. Look at it instead as sinusoidal commutation; it is T sine theta where T is the torque command. The microstep resolution angle "theta" is not particularly important so long as it is in excess of the encoder resolution.

Ampulex Ender,

If that's strictly true then you should be able to run a BLDC motor as an open-loop step motor.:-)

Mariss

[amplexus]

I think you could run at least some bldc motors as steppers but the resolution would be very low, I saw someone running a car alternator as a step motor. I think they added their own hall sensors. Alternatively you might be able to sense back emf zero crossings.
Amplexus Ender

[.xXACEXx.]

im trying to wrap my feeble mind around this,and i wondering if im any where near the right track.....

stepper motors make a pulse when they are turned the faster the more voltage....no science here.just my own "tongue to the wire" method dont laugh it is the fastest way i know to check 9 volt batteries also ,but a stepper motor will put a 9 volt battery to SHAME it is so hot , not so bad when spun slow but really bad when spun fast

now the type im talking about is the ones found inside a dot matrix printer ,rather small yes but hang with me for a second... as this is what i stumbled on ....

it a stepper motor makes voltage, and that voltage could be fed back into a sensor,...could that be used as its own sort of " encoder" ?? that has a definite step per rev and using the sensor to "read" this back fed voltage, would it be able to be used as a encoder?? am i any where near being onto something here??

oh well, just an idea i thought i'd throw out there, catch the idea if you want it ..keep it, throw it away,or maybe explain why (in many ways perhaps) i am WAY off track just trying to keep the "grey matter" functioning

[Mariss Freimanis]

Two problems with using voltage as feedback:

1) The winding is being driven by the drive. The drive's output masks any voltage generated by the motor.

2) The motor voltage is proportional to motor speed. It goes to zero as motor speed goes to zero.

Mariss

[Mariss Freimanis]

There is another method that shows motor position without the use of an encoder and it works all the way down to zero RPM. Motor inductance is a function of motor position between ordinal steps; inductance is greatest at a 1.8 degree lag, least at 0 degree lag. This inductance modulation isn't very large, less than 10%.

Inductance is relatively easy to measure on a PWM powered motor by using the voltage developed across the current sense resistors. This voltage will have a slope proportional to V / L. I have used a mirror C driven by a synchronously switched current source to demodulate the value of L from this signal.

When all is said and done, this method needs careful tuning to be effective and it's relatively complex when compared to the sure-fire simplicity of an encoder. That is why I abandoned it years ago even though it's kind of neat.

Mariss

[.xXACEXx.]

darn ,

and i thought i was on to something.lol

oh well , was worth mentioning i guess, as im sure i wasnt the only one who had thought of it , thanks for the input