DC Motor Control - Heat Issue

I found this company:

It seems that their product RL1256-8-470 is the proper item for this purpose. It doesn't list weight, but it does list size (about

2.5" cube, actually cylinder).

I am trying to confirm that adding an inductor of this type would correct my heating problem? Is it normal that I am receiving this heating issue?

I would design an LC filter instead of simply trying to block with an inductor. You may well find that a multi-section LC will let you get away with less overall inductance and smaller volume.

I would predict it. Unless the laminations in your motor are paper thin, the eddy current induced in them will be hellacious at your PWM carrier frequency.

--
--Larry Brasfield
email: donotspam_larry_brasfield@hotmail.com
Above views may belong only to me.

All,

I am new to PWM signals and DC motor control, so I am bascially learning as I go.

Mark, you mention anti-parallel diodes. Would these be located on the motor driver? I did not design the driver, so I have to assume that the company would have taken this into account since they chose a fixed PWM frequency output?

Larry, could you please provide me some further information on the LC filter? This sounds like a great alternative if it takes up less space, and mainly, less weight. I am not familiar with designing LC filters for this application, so any information on this would be great.

Also, Larry, you mention that you would predict the heat problem based on the eddy current induced. What exactly is eddy current? You indicate that the frequency seems to be the cause of the high eddy current, would a higher or lower PWM frequency help this, and do you suggest changing motor drivers rather than trying an LC filter or choke? What frequency would be optimal for this configuration, and why?

Thanks so much for bearing with me, and for your quick and helpful responses! Keep em comin!

Sincerely, Andrew Czop

If the L/R time constant of the motor is the same or less than the switching frequency then excessive heat is to be expected. A DC motor running at constant velocity and driven by a DC voltage (or current) is fairly efficient, with much of the electric power into the motor going out as mechanical power. The mechanism for this is that the spinning motor develops a back EMF (EMF is the technical term for voltage), and the armature current equals the supply voltage minus the back emf, all divided by the armature resistance.

When you run the motor off of a PWM supply and the motor inductance doesn't smooth out the current then you get I-squared-R losses from the AC current, which can be pretty significant.

OTOH, it is sometimes advantageous to intentionally PWM a motor:

when you do so, however, you must take the extra power dissipation into account or you'll burn up your motor.

All the other suggestions you've gotten about LC filters are relevant and should be useful. Expect a 2.5" cubical inductor to weigh about as much as a 2.5" cube of iron.

--

Tim Wescott
Wescott Design Services
http://www.wescottdesign.com

....

I don't have further information other than basic AC circuit theory and knowledge that inductor sizes are related to the amount of energy they can store. To see why a multistage filter would be more volume efficient, consider this sequence:

1. Start with series inductor large enough to bring the AC PWM voltage imposed across the moter to Epwmax. Call its inductance 'L'.
2. Split the inductor into 2 series inductors, sized L/2.
3. Put a shunt capacitor at the new junction, with a value C several times greater than C = 1/(L*(f*2*pi)^2)
4. Observe marked reduction in AC across motor.
5. Reduce L accordingly to obtain original Epwmax.

It is a name for current that circulates within the iron laminations in response to a changing B field. The reason laminations are used in iron core AC devices is to reduce eddy currents to acceptable levels.

The effect of eddy current is two-fold. It heats the material in which it flows. It causes a corresponding current in the the winding which induced the B field. You can think of eddy current, crudely, as flowing in a shorted turn that acts like an unintended secondary of a transformer, where the excitation winding is the primary. Less crudely, eddy current can be analyzed as arising in a composite of secondary windings formed by thin (or infinitesimal) layers within the lamination.

I cannot address optimal frequency. Your 7+ KHz seems low, and I wonder if it cannot be increased. You will still need a low-pass filter, but it can be smaller as you increase the PWM frequency. At some point in that progression, losses in the switching elements will outweigh the diminishing other return from increased frequency. I don't know how to express that optmization process.

As for changing your driver, I would sure have a go at a workable filter first. I presume this is not a high-volume, cost sensitive application, so, from an engineering cost viewpoint, the filter is going to be your cheapest solution.

--
--Larry Brasfield
email: donotspam_larry_brasfield@hotmail.com
Above views may belong only to me.

the

that

fixed

LC

LC

based

you

brakes

required.

to

and

Low inductance motors such as pancake motors have this problem. If you don't get the frequency up or add an inductor you're going to be getting tons of I^2R heating in the controller and the coils.

More than just being able to handle the current, you have to make sure the inductor doesn't saturate at that current. That means volume of core material.

Check data sheets, but it's not going to be an SMT part, and it's probably going to require more than one screw to hold it down. ;-) A Hammond 195C30 1mH 30A choke weighs 2.5kg.

Best regards, Spehro Pefhany

--
"it's the network..."                          "The Journey is the reward"
speff@interlog.com             Info for manufacturers: http://www.trexon.com
Embedded software/hardware/analog  Info for designers:  http://www.speff.com

This is just my two cents... but are the anti-parallel diodes across the motor fast enough for your PWM signal? If they are not fast enough to "catch" the reverse EMF, then....

eddy currents are the magnetic field that is polarized in such a way so that when, you pass an object over its surface a drag/pull will take place., just think of an alternator in a car when a load is placed on it. you get a magnetic drag.. normally DC output is applied at analog levels because it has to be rather constant with in reason. PWM may not be such a good idea directly into the clutches or brakes due to wide PW at low freq's could cause mechanical chatter and ware, but using a switching output to keep the PW narrow into a high freq transformer then into a bridge rectifier filtered with CAPS could work. the idea is to generate a fixed pulse width but vary the number of pulses/s so in other words the duty cycle would look like 5% on the scope at low output and at max it would be like 50% duty cycle. something simply like a One Shot (narrow pulse time) being triggered by a Astable 555 that varies its freq depending on the output required. this can give you a nice smooth DC out. and you don't have to worry to much about your drive components because they would be in saturation and thus less heating problems.

Yes. There has to be somewhere for the inductive current to go when a PWM switch turns off. If the drive is unidirectional (one switch) then there should be a fast (possibly schottky) diode across the inductive load (motor in this case) that is reverse biased when the switch is closed, but forward biased to carry the load current when the switch opens and the inductance generates a reverse voltage to keep the current going. If you have a bi-directional (H bridge) 4 switch drive, the diode connections get a bit more complicated, but the function is the same, to provide a path for the inductive current when the switches stop applying drive voltage to the motor. The inductance then approximately averaged the applied voltage and the near zero diode drop voltage during each cycle to simulate a variable DC drive voltage.

Do you have a schematic for the drive?

Anytime the magnetic flux passing through a conductive material (or through a hole in a conductive material), there is a voltage generated around that changing flux that drives current to circle the flux. The faster the flux changes and the lower the resistance of the material, the larger the current. That current is called eddy current.

The inductor lowers the rate of change of flux in the conductive materials in the flux paths. Sometimes all you need is to remove the harmonics of the PWM frequency to get a significant reduction in total eddy current.

--
John Popelish

All,

I do have a schematic of the current driver. The details of this 4 switch H-Bridge type driver are located here, as well as two links (at the bottom) to the two pages of the schematic:

The application of this design actually is intended for low level production at some point, and cost (and availability) is a driving factor.

The most recent problem I am seeing concerning a choke is that they are expensive (and large/heavy) for high current applications. Since I am running 6A per motor, for a 12A continuous output (per driver), 20A (maximum allowable) peak, the DC chokes appear to be in the $60 range, which is more than half the cost of the driver itself. I would need two of these chokes, and four smaller ones for four other motor outputs, costing roughly $30 each, or 1/3 the cost of the driver each.

I think, with these costs, I may need to just investigate new drivers. Finding inductors able to handle these high currents doesn't seem to be the most cost effective solution after all.

Also, I am getting mixed input as to whether or not to use a higher PWM frequency or a lower one. It seems that I should be using a higher frequency, but some of you have indicated a lower frequency? This is a little confusing.

Does anyone have any reccomendations of companies offering inductors in the 470uH range that can handle roughly 15A+ continuous current? Renco has these parts, however, they are $60+. I am looking for a cheap alternative, in order to test a DC choke/LC filter. In the meantime, I will look into some new, higher frequency drivers. Thanks for all your help guys.

Sincerely, Andrew Czop

John,

The frequency is very low, it is only up to 7.8kHz, not even quite

10kHz. Any suggestions for something to fit these specs?

Thanks, Andrew Cz> I read in sci.electronics.design that Andrew

wrote

in

Renco

John,

The frequency is always 7.8kHz.

The voltage range is roughly 0-26V.

As far as DC resistance, the lower the better. This would be a DC Motor choke inductor, to be used on the PWM output of the motor driver. The PWM output is driving two DC motors with resistance 0.312ohms each, and inductance of 80uH each. I'm not sure what resistance will start to mess up the motor?

By "design" you mean research available components? I can do the research, I have looked any many companies, but I think that I am having two problems: I'm not sure EXACTLY what I need, and from what I think that I need, the current ratings on almost every inductor I have found is much too low. A suggestion on companies to look into would be helpful.

A selection from a big company like digikey or allied would be great, if possible. Something like this:

on digikey seems alright, for, say, two 1mH 10A rated inductors in parallel, but the weight at 1lb. each is a bit heavy for this application. Also, this is not a stock part.

Any thoughts?

Thanks aga> I read in sci.electronics.design that Andrew

wrote

at

resistance

so

John,

Thanks alot for your time and help!

Unfortunately, a custom part for this application wouldn't go over well with the project. The project calls for widely available parts. Custom parts in this application are frowned upon, as they are no longer available commercially-off-the-shelf. This is reason for using the motor driver that we are, as opposed to designing one ourselves. It is becoming more clear, however, that it may be necessary to design a custom h-bridge for this application!

Thanks again for your time, but I really don't think a custom inductor would suit our needs.

Sincerely, Andrew Czop

Mark,

I thought the large capacitor (2200uF) was catching some of that voltage shot back at the driver when it switched off?

How would I test for built-in recovery diodes?

What would be a proper "big and fast" diode to use in this application, and how exactly would it be inserted? Sorry if this is a really simple question, but I'm not familiar with this.

Thanks, Andrew Czop

Mark,

Also, sorry, the MOSFET pdf is here:

It indicates a maximum diode recovery dv/dt of 5.3 V/ns. Hope this helps.

--Andrew Czop

Hmmm, I might have asked this before but do those MOSFETs have built-in recovery diodes? The drawing doesn't include them. In fact, maybe you should test them to make sure the recovery diodes are still working properly. If there are no diodes across the motors or drivers, each PWM pulse is going to charge up the motor's parasitic inductance and be reflected back at the driver as a large reverse-voltage spike when the driver shuts off, every 140uS or so. Maybe try some big and fast diodes across the motors' terminals themselves (in reverse of course.)

For the hardcore experimenter, maybe the primary coil of a microwave oven transformer (MOT) could be used as an inductor, or rewinding such a core. But as far as I understand it, the laminated silicon-steel is not going to work well at high frequencies, but if you have one laying around, why not try it? It could be a big waste of time, but if you're out of ideas and don't know what else to try...

It's too bad you can't buy ferrite in all kinds of large sizes, like say a 4" diameter toroid. Then again, winding that behemoth would be a bitch! :)

I read in sci.electronics.design that Andrew wrote (in ) about 'DC Motor Control - Heat Issue', on Tue, 18 Jan 2005:

470 uH and 15A could be much cheaper if the inductance has to be effective only at fairly low frequencies (say up to 10 kHz or so).

--
Regards, John Woodgate, OOO - Own Opinions Only. 
The good news is that nothing is compulsory.
The bad news is that everything is prohibited.
http://www.jmwa.demon.co.uk Also see http://www.isce.org.uk

I read in sci.electronics.design that Andrew wrote (in ) about 'DC Motor Control - Heat Issue', on Tue, 18 Jan 2005:

What is the lowest frequency the inductor will see, and what voltage at that frequency could appear across it? What is the maximum DC resistance you can accept? To design the inductor, those figures are necessary, so I've asked for them. However, I doubt that I will have time to design the part for you within the next 7 days, so maybe someone else will volunteer.

--
Regards, John Woodgate, OOO - Own Opinions Only. 
The good news is that nothing is compulsory.
The bad news is that everything is prohibited.
http://www.jmwa.demon.co.uk Also see http://www.isce.org.uk