What's new
What's new

OT? DC motor question

metalmagpie

Titanium
Joined
May 22, 2006
Location
Seattle
If I connect a variac to a rectifier and feed the +/- pins to a DC motor, I can smoothly control the speed. If I wanted to bother to design one, I could do a PWM solution too - low duty cycle for low speed.

I'm talking permanent magnet motor here, not field wound.

How can I control such a motor to run at low rpms but still have high torque?

metalmagpie
 
Yes, but only 6% speed regulation, you will see it slow down under load...I use this on a jig bore and a big mill, and drill press,s A wound field with comp works the best...Phil
 
An ideal DC motor will deliver full torque from stall on up as long as the current is kept constant. Of course YMMV, but that is the general answer. A motor with zero resistance fed from a power supply that has no impedance will be automatically almost constant speed. Since the motor is not quite ideal, winding resistance and all that, it will change speed under load. That is what the IR COMP setting on motor controllers is for. It measures the current drawn and adds voltage to compensate. Similarly, a series field increases field strength with a current increase. Short of adding a little motor to adjust the Variac, I don't know of a way to do it without adding some electronics. I would use a magamp, but that is just me.

Bill
 
As 9100 states torque should not be a problem even at zero RPM on a DC perm mag. The curve starts out at max and you have more torque at low RPMs.
Accurate speed control with varying load requires some sort of feedback loop and as usual 9100 is on the money.
Current is mainly torque and voltage is speed but it's not that simple if working a motor hard.
Knowing the application may help many to add useful information.
Bob
 
One caveat. In normal use the voltage is only applied to each commutator bar for a small fraction of a revolution so the armature wires can be much smaller than would be required to carry the current continuously. Stalled or turning very slowly will overheat armature wires. You can generate rated torque when stalled, but only for a very short time.

Bill
 
Isn't that just called stall torque or stall current which is miles below peak torque but a major concern in designing machines that have to carry a gravity load or say a servo driven bolt fastener?
Most DC PM motors I know will take stall torque forever with little air flow around them. Two words current limit, almost all dc servo amps have a peak and much lower time limited current built in.
The amp knows that it may need to kick the motor in the ass on acc/dec but also knows not to do it for very long.
Design anywhere near peak torque and it had better be moving and not at this power level for very long. 2-3 seconds OK, past that....I smell something funny.

Perhaps I'm wrong and admit to learning to pay attention to the different power curves by oops.
Sometimes the winding goes, sometimes you melt the brush holders. (it worked for the first year, 8760 hours seems reasonable...:dopeslap:)
Done right and you get 10 times this or more on a DC brush motor. Bearings and bushings die first.
Bob
 
One caveat. In normal use the voltage is only applied to each commutator bar for a small fraction of a revolution so the armature wires can be much smaller than would be required to carry the current continuously. Stalled or turning very slowly will overheat armature wires. You can generate rated torque when stalled, but only for a very short time.

Bill

An excellent point. And turning slowly, airflow does not help much.
 
You can design a motor to run under almost any conditions but you have to carry through the whole system. In the 80s I was getting GE DC motors with their armatures intact but the leads from the brush holders to terminals melted. Some local people were developing an electric car. These motors can stand a huge overload for a short time, which makes the short 0-60 times Tesla posts possible, but you have to match up the thermal masses of the various components. In this case the leads were the fastest heating weak link.

GE Diesel electric locomotives have an ammeter marked off in green, yellow and red arcs. The yellow arcs are marked in times it can be held at a given current.

Bill
 
I'm not developing an electric car or any system like that at all. I'm trying to make a slow speed lap that won't slow down at the touch of a finger.

I got a different PM motor today. This one has a disc attached to the back of the shaft. That disc has holes around its periphery and there is a sensor and wiring harness. I'm guessing it's a Hall effect sensor. So clearly there is a pulse train coming out. How can I take advantage of this?

metalmagpie
 
The problem that you're running into is the variac. Everyone here's talking about DC motors being driven either with straight DC or with a DC motor drive.

Your variac has a pwm rate of 60hz. A proper DC motor drive has a pwm rate of somewhere between 8000 and 100,000 Hz. That's important. The reason it is important is because motors have a property called inductance. Inductance is to current like capacitance is to voltage: it smooths everything out. Because torque is proportional to current, inductance means you'll have a smooth torque even though you're turning the power from the drive on and off. Well, sort of. At a fast PWM rate, the motor inductance is large relative to the chopping speed, and it acts like it is being fed the average DC voltage. But all the way down at 60Hz, there isn't enough inductance to keep the current going through the long time between pulses. So your motor current is dropping to zero and staying there for part of your pulse. As you turn the variac down, your current (torque) is zero for a longer portion of the cycle, so your average torque goes down, despite what a DC motor driven from a DC source would do.

Short answer: buy any sort of a proper DC motor driver that runs at a proper DC motor driver pwm rate in the kilohertz and you'll get better results. Anything from a $10 mystery parts amazon special to a $1000 servo amp. (Obviously you'll get better better results with the better driver)
 
The other problem with slowing down is also due to the variac.

When you adjust the variac for the speed you want, that is with minimum current (minimum torque, no load).

When you apply a load by trying to lap something, the motor HAS TO slow down somewhat because it has to draw more current to supply the added power. Slowing down reduces the back EMF (voltage) that is opposing current flow, and so allows more current.

When the motor draws more current, the voltage drop due to the variac (which has resistance, etc) increases, lowering the voltage to the motor somewhat. When the voltage is lower, it takes even MORE current for the same power, so the net slowdown is larger than you might expect.

To get around that, a PWM type controller regulates output voltage, which reduces the effect of more current flow.

Some PWM controllers have boost functions which compensate for added voltage drops at slower speeds (lower voltages).

Other PWM controllers can use speed feedback to adjust voltage and current so that the motor speed change is made very small at any speed.

All these effects are made much worse when the motor is made to slow down a LOT, which may be typical of lapping as opposed to straight grinding.
 
Since we don't have any information re size or power, other than it is low powered, it is difficult to make definite recommendations so here are some general ones. Gear the motor down, perhaps with a timing belt. 4 or 5:1 would be easy and could make a big difference. For a power supply, look for a regulated one, preferably with a sense connection that reads the voltage at the motor terminals, eliminating lead voltage drop. These supplies show up in surplus stores for low prices.

If that isn't good enough, I can send you a circuit that compares the pulses from the encoder to a fixed oscillator and adjusts the voltage to keep them in synch.

Bill
 
The problem that you're running into is the variac.

The output of a variac is sinusoidal, not a pulse train whether width modulated or not.

The output of a rectifier powered by a variac is a familiar sinusoid with every other pulse being mirrored so that only positive voltage results.

I see low power at low speeds whether I'm running a variac/rectifier or whether I have the motor connected to a DC power supply.

Unless I'm badly misunderstanding PWM has nothing whatever to do with this.

metalmagpie
 
Since we don't have any information re size or power, other than it is low powered, it is difficult to make definite recommendations so here are some general ones. Gear the motor down, perhaps with a timing belt. 4 or 5:1 would be easy and could make a big difference. For a power supply, look for a regulated one, preferably with a sense connection that reads the voltage at the motor terminals, eliminating lead voltage drop. These supplies show up in surplus stores for low prices.

If that isn't good enough, I can send you a circuit that compares the pulses from the encoder to a fixed oscillator and adjusts the voltage to keep them in synch.

Bill

I'm experimenting with treadmill motors. The ones I've found seem to require peak current of between 12 and 20 DC amps. I would happily list the horsepower ratings on the labels but I believe these data are more marketing than scientific.

My goal is to machine the flywheel contained on all treadmill motors to a flat disk, charge that flat cast iron disk with diamond paste and then sharpen tungsten carbide scraper inserts. The preeminent sharpening machine used for this is the Accu-Finish made by Glendo. This product has been end-of-lifed. That's why I'm trying to find an easy-to-make inexpensive replacement that has no real compromise i.e. it grinds just as well as the $1200 Accu-Finish machines do.

The Accu-Finish also has the advantage of being light. I don't expect to make a lapper that is as lightweight.

Where I live there are no surplus stores any more. Rents in the greater Seattle area are now sky high, so we're stuck with art galleries and nail salons. But I'll keep my eyes open for a DC power supply that has a remote sensing capability. My current DC supply is made by Volteq and can supply 30VDC at 30ADC, in either voltage source or current source mode. But no remote sensing.

I would love to see your circuit, Bill. As long as the "fixed oscillator" can be varied. I really want to try a range of rpms which you might say goes from low speed to lower speeds.
Nothing over 500 rpm. (The Accu-Finish runs at 300 rpm and has a 5" diameter wheel.)

metalmagpie
 
The other problem with slowing down is also due to the variac.

When you adjust the variac for the speed you want, that is with minimum current (minimum torque, no load).

When you apply a load by trying to lap something, the motor HAS TO slow down somewhat because it has to draw more current to supply the added power. Slowing down reduces the back EMF (voltage) that is opposing current flow, and so allows more current.

When the motor draws more current, the voltage drop due to the variac (which has resistance, etc) increases, lowering the voltage to the motor somewhat. When the voltage is lower, it takes even MORE current for the same power, so the net slowdown is larger than you might expect.


To get around that, a PWM type controller regulates output voltage, which reduces the effect of more current flow.

Some PWM controllers have boost functions which compensate for added voltage drops at slower speeds (lower voltages).

Other PWM controllers can use speed feedback to adjust voltage and current so that the motor speed change is made very small at any speed.

All these effects are made much worse when the motor is made to slow down a LOT, which may be typical of lapping as opposed to straight grinding.

These are the important parts. You need a motor controller that detects the added load and boosts the voltage.

Something like this may do it; it will depend on how much current you actually need.
 
I once made a controller like that for a dental handpiece used in polishing teeth. Carefully tweeked, I could make it speed up as the load increased. The idea was that it could be turning slowly and the hygienist could control the speed by pressure on a tooth being polished. We had one test it and she immediately turned it to max speed and that was that.

Working in that realm is difficult, anyway, because the speed will oscillate, varying up and down.

Bill
 








 
Back
Top