PWM Frequency Decision

One problem you can run into at high speeds is the time it takes to switch your current on and off. For instance, if it takes 100uS to turn your switch on or off, that limits your frequency to 5000Hz or so. Your switch and motor effectively becomes a low pass filter that won't let your high frequencies get through.

I've done PWM with pics (a 16F model) at much lower frequencies (~100Hz) by doing the PWM in code, rather than using the built in PWM function. If you do it that way, you can make it as slow as you want.

I have been thinking about doing the PWM in code. The only downside is that I would probably need to dedicate a chip to that. Not too big a deal, but I am trying to keep the complexity down as much as possible.

I am looking at the high side driver chips I am using (VN05N chips from ST Microelectronics). The switching time may be the issue. It that is it I will let the forum know.

thanks,

Doug

PWM frequencies are generally kept above about 20kHz to keep any buzzing/whining noise above audio range.

As bhankiii says, a problem can be the switching time. And there's a trade-off here (as ever!). You can use faster transistors/drive them harder - fast switching is good to reduce the power loss in the switching elements, but it's bad because it generates more RFI. It's all a balancing act!

I had typically gone in the other direction. Keeping the pulse train at very low frequency to prevent the whining. It had the added benefit of inspiring fear from the low frequency tones. :-)

I had always worried about inductance effects at the higher frequencies. Not that I had a specific engineering reason for that, just gut instinct. I guess my gut was wrong. :-(

Is there any formula or rule of thumb to use in these situations? Is there any way to calculate the power loss at specific frequency ranges. It seems my high side driver chip is capably of switching fast enough. It claims a 30 uS switching time.

Thanks,

Doug

1/(30uS x2) = 16KHz as the top speed (your minimum pulse is 60uS wide - and that's all rise and fall time).

We use an Intersil HIP4080A full bridge driver that has T_R and T_F < 35nS. I don't know if that part is applicable to you, but it does show what's available out there.

I assume you're using a high side driver because you're using an H-Bridge for forward/reverse. If you don't need both directions you can use a low side N-channel MOSFET switch and drive it with a simple transistor circuit straight from your PIC.

I was actually using the VN05N because it was inexpensive and had a direct logic level input for extremely easy integration. At this point I am considering converting to a MOSFET switch. I only need one direction since it is just managing a spindle motor for a mill.

Do you have a circuit examples or a MOSFET you would recommend. I would actually like to be able to drive the motor at 32v and up to ~3.5A continuous. The motor seems to handle it fine and it would give me a wider speed range.

Once again, thanks for your help.

-Doug

I've used the IRF1010 and the IRF540. Both of these can handle 100V, IIRC. Remember you are switching an inductive load, though and slap a transient suppressor (TVS) across the load to keep the switching spikes down.

You can look at the data sheets and see if they will work for you. International Rectifier has a huge variety of fairly inexpensive MOSFETs.

I drove the PIC output to a 2N2222 transistor with a collector pullup to 12V for the gate voltage. There are logic level gate MOSFETs available as well if you don't have a 10-15V supply handy, as well as gate driver chips that convert your logic supply to a high gate voltage.

The inductance effects aren't too much of a worry for a DC motor (tho' it plays merry hell in a stepper drive system - if the inductance is too high, you can't get the motor current up quickly enough at high step rates, unless you use higher voltages).

There must be a way to work out the actual power loss as a function of switching frequency and turn-on & turn-off time of the switches (tho' I don't know how, off-hand). Sounds a bit complicated to me!

Generally, the higher your switching frequency, the more power loss (because if you're switching more frequently, the time spent switching ON and OFF will be a higher proportion of the total time). When I did some work with home-brew systems, we used to work at the lowest frequency we could that didn't produce a headache. If you get it right, you can keep the kids away! These were stepper drives, so the chopping rate had to be well above the fastest step rate needed (say 4 - 5kHz). I use Parker drives now, which chop at 20kHz.

The motor doesn't seem to be putting out as much torque as I would expect, even at a 98% duty cycle.

The other guys seem to be answering your other questions just fine.
Do you get enough power when it is running of DC? If so then maybe the control circuit is dropping some voltage somewhere. There are plenty of very low R_{DS ON} Mosfets out there.

It seems a bit over complicated to me.
If you want to use a DC motor I'd have thaught a motor with plenty of power, geared well down and simple voltage (or pwm) control would do it, and forget the feed back.

But maybe I'm barking up the wrong tree

Del

Yeah, I always seem to get great advice on this forum. Actually the goal is to be able to control the speed to specific RPMs. Different types of metal being worked require different surface speeds for the bit. Different size bits give you a different surface speeds at different RPMs. Another issue is that while the bit is working the metal it requires additional torque to maintain the proper speed.

My final goal is to have a small controller box that you can enter in the bit diameter and required surface speed. The system would then calculate the proper RPMs and use the PWM system with feedback to maintain the appropriate speed while it is doing its milling operations. There is a keypad for input and a LCD screen for displaying options and spindle status.

I guess I should have given the project details earlier but I was mostly focused on the PWM issues at the moment.

-Doug

You mean miauwing up the wrong neighbours window .

I think he wants the feedback to keep rpm constant under variable loads.

Doug, I think the others have helped you out and what I say may not be any more important... *sniff*

But if its one of those scooter motors, as you say, then they have a large inductance and huge losses at a high frequency - the same as a mains powered transformer would have at several kHz.

Those scooters have a variable pwm drive to alter the power to the moters... have you got the other parts to the scooter?

The throttle is just a potentiometer connected to a little box which houses the ingenious pwm circuit to supply power to the motor.... I've a couple of these someone was throwing away and so I had a look at how they work... the scooter motor is about 100 watts upwards... so at full power you should be seeing a current draw of several amperes, not just a few!!

I suspect the frequency is way too high for these motors, as others have said... why not grab one os these speed controllers and input a variable voltage in place of the throttle to give speed control?

Then all the power electronics are built for you!

John.

Regrettably I don't have any of the other parts since I got the motor off eBay.

Between the inductance and the rise time off the switch I may be losing a fair amount of my power. There is a significant difference between straight DC and the PWM signal at 98% duty cycle. I was thinking of a possible solution using a hand built PWM made from 555 timer ICs. It would be similar to your solution of using the original motor controller. I also liked the idea of just using an extra small PICAXE chip as a dedicated PWM system I could control more precisely. Since I am a better programmer than electrical engineer this idea appeals to me.

Over the weekend I will be able to take some of these ideas and try them out. Hopefully I can get this issue fixed and move on.

Thanks for the help,

Doug

Doug, if you want one of those speed / power controllers, I've still got a few left over together with motors and complete scooters...

A 'friend' dumped 6 of them on my doorstep saying he was taking them to a skip - me being me had them for some strange project that I've forgotten about so they are sitting around doing nowt!!

If you want one just ask...

John.

A 'friend' dumped 6 of them on my doorstep..

Presumably these are the small kids scooters 'Make speedy walk on level ground' not the big sit on jobs?

I've been thinking of an electric scooter for some time... (think Lambretta...not scootalong with foot)...

Yes, they are the 12 volt stand on things that are illegal to use on the pavement or road...

Probably why there are so many being dumped!!

The motors are good though... 100 watt with a power controller (twist grip) and nice poly wheels.... ideal for a project or two...

Motor for robotics / generator? sealed lead acid battery ... and wheels... mmmmmmmmm

John

http://www.nutsvolts.com/ has several circuits in their back issues.

I love Nuts and Volts. This month's electromagnetic coil launcher pistol is very tempting.