Back EMF issues

[dr pepper]

The correct way to do it would be to Rc snubber it.
The R calc'd in such a way that it pulls the average current the motor pulls if it were connected direct to the supply, and the C calc'd so that it holds enough energy to empty that in the inductance of the motor.
So at switch off the current drops back to zero failry smoothly, instead of making a huge voltage spike.
I have seen and used a common technique and thats solder a 100n across the motor terminals, and from each terminal to the case of the motor.
If its a fairly beefy motor from a drill you might need class x caps to take the beating.

 

[ClydeCrashKop]

In the original plan, my understanding (or misconception) was that a motor would be stronger at slow speeds with pulsed 12 volts at 25% than it would be with a steady 3 volts. That is why I wasn't happy when the scope pattern looked like the voltage was being dragged down. So when scoping the current, it looks like I am still getting strong pulses. That made me happy again. Especially when putting a load on it, the current went up, like it had plenty to spare.
With the diode across the motor terminals, I don't see any destructive spikes. Do I want to smooth it out with RC snubbers or keep the PWM?

My book "The Art of Electronics" only has a few paragraphs on inductors. This experience and visualizing it, helped with understanding it.

 

[MrAl]

Hello,

Normally you use an H bridge to control a motor not a single drive transistor.
There are different reasons for this, like the ability to reverse, braking, faster di/dt and much better efficiency.
Do some tests and/or sims and see how big the difference is. Might surprise you.

 

[ClydeCrashKop]

There will be 2 motors. I planned on using a DPDT relays to swap polarity for direction change. I don't expect to change direction often or have to do it fast.

I hadn't considered braking. I may have to rethink.

 

[ClydeCrashKop]

The correct way to do it would be to Rc snubber it.
The R calc'd in such a way that it pulls the average current the motor pulls if it were connected direct to the supply, and the C calc'd so that it holds enough energy to empty that in the inductance of the motor.
So at switch off the current drops back to zero failry smoothly,

The more I think about that, the more I like it. I will give it a try.

 

[dr pepper]

You know how to calc the resistor, start with say a 22n cap and observere on the 'scope, fiddle with the C till you get voltages well within the capability of your switching devices.
If you go too far and try and get a nice looking squarewave you'll end up wasting lots of power in the R, its a tradeoff to get right.

 

[crutschow]

The more I think about that, the more I like it. I will give it a try.

Adding a snubber will just waste power, that otherwise would power the drill.
I see no advantage to that over just the diode, which wastes a minimal amount of power (you could use a Schottky to further minimize the loss).

 

[MrAl]

There will be 2 motors. I planned on using a DPDT relays to swap polarity for direction change. I don't expect to change direction often or have to do it fast.

I hadn't considered braking. I may have to rethink.

Hi again,

Also, i was assuming PWM, but are you doing that too or just turning on and off once in a while?
If not PWM the efficiency is not as much of an issue usually.

 

[dr pepper]

I mentioned the snubber as I was thinking being a drill that it would reverse.
Or were you thinking a transorb diode or whatever.

 

[ClydeCrashKop]

As of right now, this is the circuit I am using except to double it for the other motor, using the second half of that half-H drive chip.
I am assuming that reverse will be like driving a car. Only used occasionally.
This is for my GPS controlled lawn mower project. GPS instability I thought I would play with it while waiting for the second GPS module to arrive. I got the RF transceivers to send GPS position info between 2 Arduinos and my PC last week.

It is a variation of this circuit that one of our "most helpful" members posted lately.

 

[ClydeCrashKop]

I got a pair of motor>gear reduction>wheel assemblies from one of these that I salvaged.
Power Wheels Smart Drive Mustang Ride On Cars for Kids Battery Powered 12V Toy Review

I probably won't use the cool car noises or 3 preloaded songs.

 

[MrAl]

As of right now, this is the circuit I am using except to double it for the other motor, using the second half of that half-H drive chip.
I am assuming that reverse will be like driving a car. Only used occasionally.
This is for my GPS controlled lawn mower project. I thought I would play with it while waiting for the second GPS module to arrive. I got the RF transceivers to send GPS position info between 2 Arduinos and my PC last week.

View attachment 115524

It is a variation of this circuit that one of our "most helpful" members posted lately.

View attachment 115525

Hi again,

So you are doing PWM, at around 500Hz.
Do you have any spec on the max (and maybe min) duty cycle?
I ask because there could be conditions where the PWM does not work even though the transistor is being turned on and off at the right rate and times. That's because the single diode solution always leads to more limitations in these designs.

 

[ClydeCrashKop]

Because I am running out groups of pins on the Arduino, I am now using 980 Hz PWM. The motor starts turning at 6% at 0.5 amps with or without a load. 50% is 1.1 amps. Full speed is 97% at 2 amps.

The R calc'd in such a way that it pulls the average current the motor pulls if it were connected direct to the supply, and the C calc'd so that it holds enough energy to empty that in the inductance of the motor.

So at 12 volts, 50% duty cycle is 1.1 amps, would R be 10.9 ohms? Or would R be calculated at full speed, 12volts at 2 amps = 6 ohms?

 

[dr pepper]

Probably 10 ohms, and select a cap for acceptable noise/spikes without lots of heat in the R.
However I agree with Cruts, and having seen your schem a diode is probably better, while its probably possible to waste less power with a well designed RC its a whole lot easier just yo bung on a diode.