DC Motor Control

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

adamasd

Well-known member
Joined
Jun 17, 2004
Messages
472
Location
Duluth MN
I am getting ready to build a coil winder and I am not completely sure about the motor control circiut. My plan is to just make a fixed frequency oscillator with PWM control and run its output into a power MOSFET switch to get the current needed. Is that a sensible way to go for this?

What about oscillator frequency? It seems if it is to slow the motor would get jerky at lower speeds, if it is to fast the control will not be as fine/have the range it should. Is that about right? Is there a good frequency to use as a starting point?

thanks,
adam
 
I read that earlier today, did not find much about the DC motor control, but it was a good read, puts what I am planing to build to shame. Someday I will make something fancier.

If only I had the time.

thanks
 
> Is that a sensible way to go

Sounds over-complicated. How BIG a motor do you need??? Unless you are winding heavy gauges on huge cores, wire tension and breakage is going to keep the horsepower needed down to maybe a dozen watts.

DC supply. Pot across to get 0%-99% of supply voltage. Get a large power Darlington bolted to an old mower blade or hi-fi heatsink, wire as cathode follower, motor as load. Yeah, at least half your electric bill comes off as heat... worry about it when your electric bill spikes.

Watch for a $20 sale on a junky cordless drill. You get E-Z connection chuck and considerable gear-down to sane RPM.
 
www.theopiumdenproductions.com/potPWM4.pdf

This is a professional design I did for lighting/motor control from an isolated line voltage.

rather simple but very robust.

softstart for .75 seconds and speed is adjustable.

R5 adjusts deadtime, R6/c5 adjust frequency. each IGBT(or NFET) handles 50% of the duty.


I don't remember if the finished layout got any type of pulldown resistors on the gates but I don't think so since the device is totempole output. I can check on my prototype but I'd have to dig it out of storage.

You are free to use it for non-commercial applications as I've already been paid and I was allowed to keep the design(very rare these days..).

:thumb:

EDIT: you can run this anywhere from about 10v(without the rectifier and zener regulator) on up to about 190VDC/140VAC(with rectifier and regulator.)
 
-Sounds over-complicated.

Maybe but a dual opamp and a power mosfet is less complicated for me then a power darlington and a lawn mower blade/heatsink. I have boxes of dual opamps and mosfets, I have no power darlingtons or spare lawn mower blades/big heat sinks.

-Watch for a $20 sale on a junky cordless drill. You get E-Z connection chuck and considerable gear-down to sane RPM.

Done it, it is fine for smaller coils and chokes but when you need to do a pickup with 10,000 turns the noise really starts to get to you. Its time to move up to a nice quiet motor and foot control.

-you could buy a working/junk or inexpensive sewing machine. Then, you'd have a foot control too.

That was always my plan, but I can not junk a perfectly good sewing machine and I have yet to find one that was trashed but had a good motor still. So I found a 12VDC motor in the junk box that will work and I have a wah pedal carcass, should do the job.

A schematic Svart sent me shows my idea will work and is on the right track, but that one along with the one he just posted use parts I do not have and can not get locally, I plan on having this made by the weekend.

So I have the basic design sorted out, just need to get an idea on frequency. I have come across everything from 100hz to 20khz mentioned, probably will just have to spend awhile experimenting to see what works best.

thanks
 
The lower frequencies are easier on the gate drivers as they don't have to slew so much cap to get the gate on/off and the motor will give more torque but a lot of motors will whine at audio frequencies which is pretty annoying. Higher freqs are generally more energy conserving and have other assorted reasons for using them but I think you should stick with something around 50khz-100khz.

isn't Digikey right down the street practically?

the sg/uc3525 is generally less than 1$ and the other assorted parts will bring you up to around 8$ if you go with hefty NMOSFETs.

If I had known this the other day I could have dug out some through-hole 3525s for you and mailed them..

dig out any old SMPS you might have laying around and take a peek to see what the PWM ic is.. some of the older ones used this chip a whole lot.

even the 3842 IC can be used for this but being a current mode IC instead of a V mode, it makes life a little more difficult.
 
Have a look at this:
http://www.produktinfo.conrad.com/datenblaetter/175000-199999/196460-as-01-de-Drehzahlsteller_Gleichstrommotoren.pdf

and this:
http://www.produktinfo.conrad.com/datenblaetter/175000-199999/191507-an-01-de-Drehzahlsteller_10A.pdf

analogguru
 
Sure he can use a 555 timer for the PWM, I threw together one based on a 556 timer setup once. I got pretty horrible results actually UNLESS I did a lot of extra work getting things like gate drive transistors(needed to enhance the drive to the mosfet from the 556..) into the mix and getting the output jitter under control, but then again I was working on a commercial design and it HAD to be somewhat bulletproof and perfect.

The end result was moving onto a PWM IC. you can purchase the 3525 for as little as .50$USD and save about a dozen parts over a proper 555/556 setup.

My schematic shows a number of things that aren't really needed. the opto was leftover from a part of a feedback setup that allowed me to isolate and drive the unit from logic as well as the switch setup shown. You could easily replace all of that with a momentary switch to ground.

The deadtime resistor(r5) can be taken out and a jumper put in it's place. You could also remove one of the FETs and put both outputs through signal diodes to one fet with a 220R to ground for load. this will cause one fet to run up to 100% duty but you will need to be careful for temp and current load.

all in all, you can cut costs down to around 5$ for a 12v controller.
 
Model railroaders have been using pulse width speed controls for a long, long time. Might want to check some of those sites.
 
-The lower frequencies are easier on the gate....with something around 50khz-100khz.

Hmm, never made an oscillator above the audio range, should be fun.

-isn't Digikey right down the street practically?

Sure, about 250 miles down the road.

-the sg/uc3525 is generally less than 1$ and the other assorted parts will bring you up to around 8$ if you go with hefty NMOSFETs.

Its not really a matter of cost, its more about making use of what I have on hand and learning abit as I go. I will learn more using parts that are not 1 chip solutions. Sure it may be more work, but I used those one chip solutions for years in synths, never learned much from them.

-dig out any old SMPS you might have laying around and take a peek to see what the PWM ic is.. some of the older ones used this chip a whole lot.

I dug about looking for some old switchers I had sitting about but I think they have already been packed away, moving in a few weeks here.

-can you post more info about your design?

My plan was to just build an oscillator around opamps or what ever and use a comparator to get the PWM and drive a big NMOSFET. Have not gotten any further then that yet, still trying to find out what I need.

-Have a look at this:

I developed an aversion to the 555 back when I used to build alot of synth stuff, I just got so sick of using it, its everywhere. But these PDFs still give some good info, to bad my German is not so good these days.

-Model railroaders have been using pulse width speed controls for a long, long time. Might want to check some of those sites.

I will look into that, maybe there are some good sites floating about for this.

All I really need is a something to allow me to vary the speed and give me a smooth enough rotation to keep a constant tension on the wire. I do not need the bulletproof commercial setup, I just need the gets the job done setup.

So are there any pitfalls going with an opamp triangle oscillator (comparator or comparator integrator type) and then going into a comparator to set the PW. Most any decent opamp comparator should be able to drive the MOSFETs gate should it not? Unless there is some major flaw in this idea I am most likely going to just build it tomorrow and tweak it until it works. I am ready to start winding in style.

Thanks for all the responses,

adam
 
Its not really a matter of cost, its more about making use of what I have on hand and learning abit as I go. I will learn more using parts that are not 1 chip solutions. Sure it may be more work, but I used those one chip solutions for years in synths, never learned much from them.

To each his own, I guess it's my aversion to using 20 parts when I can use 1 doing the talking! Personally I like the simple solutions, much easier to master all of the part's "personality". It sounded like you wanted something really easy, which gludged timer IC based PWM is neither easy nor reliable.

-isn't Digikey right down the street practically?

Sure, about 250 miles down the road.

Ok, I only asked because you could probably get your package overnight with ground shipping. Much better than my 1500+ miles away...

So are there any pitfalls going with an opamp triangle oscillator (comparator or comparator integrator type) and then going into a comparator to set the PW. Most any decent opamp comparator should be able to drive the MOSFETs gate should it not? Unless there is some major flaw in this idea I am most likely going to just build it tomorrow and tweak it until it works. I am ready to start winding in style.

The oscillator and comparator are fairly easy. Use good parts around the ramp generators or your ramp might induce a slight jitter, but that might not be of any matter to your needs. Also, like I mentioned earlier, the timer IC was never meant to drive mosfet gates, you might want to use some kind of buffer setup between the IC and the gate. It takes a lot of charge to pump the gate of a higher power mosfet, sometimes up to an amp if you are attempting to slew the gate at really high frequencies. At lower frequencies your dv/dt will suffer and you'll have a much longer transconductance period than you should if your gate drive is slow and/or weak. This will cause your FET to heat up signifigantly and ultimately you will get a slight pulsing of the motor due to the resistive drag during the extended transconductance periods. If for nothing else than your sanity, this is one of the reasons I suggested a part designed for the duty.

There is a lot more to it than just getting a motor to go around and around, especially when your FETs overheat or you reach some crazy oscillation frequency where all your problems come together(the voice of experience..). But again, I speak of high power/high speed dc brushed motors.

Some of the others around here work with the BIG boys, multiphase ac, etc.. I'm sure they will definately tell you the same as I.

Good luck, and have fun though!
 
I once designed a dual axis PID servo loop motor control using 2 LM629/LM628 (if I remember correctly) . The LM628
takes 15 control lines from a micro controller to set its parameters. Very precise speed, acceleration/deceleration and so on.

Perhaps way overkill for your application but it worked very well for the telescope I was controlling. Full RA and DEC control both at siderial, lunar rates and it could be set for independent rates to track asteroids, space junk and anything else.

With the right motors you could make that thing slew.

regards, Ethan
 
-The oscillator and comparator are fairly easy. Use good parts around the ramp generators or your ramp might induce a slight jitter, but that might not be of any matter to your needs.

Thats good to know, you were starting to make me think that there were some special hard to meet needs in the oscillator/comparator area with all your talk of fancy PWM chips.

-you might want to use some kind of buffer setup between the IC and the gate. It takes a lot of charge to pump the gate of a higher power mosfet, sometimes up to an amp if you are attempting to slew the gate at really high frequencies.

I am only using a 2 Amp motor, so no need for some massive high power MOSFET here. I think the comparators output should have no problems driving the gate, if not I will add a buffer. This is not some power transformer winding monster, mostly coils and pickups, and maybe some audio transformers down the road, I do not need all that much power.

-At lower frequencies your dv/dt will suffer and you'll have a much longer transconductance period than you should if your gate drive is slow and/or weak. This will cause your FET to heat up signifigantly and ultimately you will get a slight pulsing of the motor due to the resistive drag during the extended transconductance periods.

I will watch for that, thanks for explaining that out.

-There is a lot more to it than just getting a motor to go around and around, especially when your FETs overheat or you reach some crazy oscillation frequency where all your problems come together

And I will troubleshoot it and figure out how to make it work right, I got two whole free days to sort it out, I am fairly confident I will succeed in this. Most of the mechanical work is done, just need to pick up a few more parts to finish it off. So most of the next two days will be dedicated to the electronics and winding random things when it is done.

-With the right motors you could make that thing slew.

With the right motors you could make it do a 25second mile.

I always wanted to build a motorized computer controlled telescope, not really sure why, but it would be exciting to watch it track objects.
 
remember that with inductive loads such as motors, you WILL get inductive flyback. use a flywheel diode on the mosfet lest you create a Fire Emitting Transistor (thanks to Bcarso for that one). And remember that you will also have an associated voltage spike so choose a mosfet at least 2x the max voltage you will be switching and mosfet about 10 times the current rating of the motor for safety and it will live a long life. I suggest a FRED diode for the flywheel.
 
Thanks for everyones help on this, I got it up and running today and already started winding failure after failure (see my other post). It is working quite well other then I think I might go to a larger MOSFET, even with the heat sink it gets very hot. It survived a few hours of near continuous running though, but for the cost it seems worthwhile to upgrade.
 
That heat is likely a large dv/dt in the gate drive pulse which causes the extended transconductance period I mentioned.

I assume that you are using a comparator as/with a Schmitt trigger?

The other explanation would be a very high RDSon (resistance through the device when ON) which would act like a low value resistor, enough to heat up somewhat.

Couple these two problems together and your FET will be HOT.

The last motor controller I whipped into shape used alternating FETs as low side switches, switching around 150vdc @ 5A continuous. Not a lot of power by today's standards but nothing to sneeze at either. The FETs ran cool enough to touch on a very small heatsink.

Actually now that I think about it, are you using a flywheel diode for the motor?

EDIT:

something I found some time ago that is pretty good:

http://www.nxp.com/acrobat_download/applicationnotes/APPCHP3.pdf
 
-That heat is likely a large dv/dt in the gate drive pulse which causes the extended transconductance period I mentioned.

That is what I was thinking at first, but when I came back an reread your post you also mentioned that that would cause a slight pulsing in the motor. The motor has no noticeable pulsing, as far as my fingers can tell it turns very smoothly.

-I assume that you are using a comparator as/with a Schmitt trigger?

correct.

-The other explanation would be a very high RDSon (resistance through the device when ON) which would act like a low value resistor, enough to heat up somewhat.

Which I have no control over other then replacing the part with something with a lower RDSon correct?

-Actually now that I think about it, are you using a flywheel diode for the motor?

I am.

-something I found some time ago that is pretty good:

Looks great! I will read through this tonight when I get home.

My MOSFET is no where near the 10 times expected current you suggested, I used what I had on hand. I figured this to be the cause of the heat. It seems to be ok with it so far, has about 5 hours of use on it so far and has not blown up. I plan on ordering something that is abit higher powered with my next parts order.
 

Latest posts

Back
Top