A couple more motor control questions - Braking

Whats more important? Saving/recycling the energy using regen, or just having a modulatable brake for the downhill run? Going downhill the motor will be generating, but I think you need a big contactor to change the top of the h-bridge from the battery to the load resistor... then the throttle (pwm duty) is the brake.

First of all, dynamic braking (based on the power generated by the motor) works only when the motor is turning, so you probably still need some sort of mechanical brake to hold the train when it is stopped or nearly stopped, instead of using battery power to hold the motor still.

Controllable dynamic braking with an H bridge involves briefly shorting the motor with the bridge, so that the generated voltage ramps the motor current up to some desired value (the braking torque is proportional to the motor current) with a ramp rate dependent on the generated voltage and the motor inductance. Then open the short circuit and let the motor inductance pump that current into the battery, with a decay rate dependent on the difference between the battery voltage and the generated voltage and the motor inductance. You repeat this cycle fast enough that the current is controlled between a maximum and minimum limit, to produce the desired average braking torque, specified by the position of the brake pedal. To perform this process accurately and smoothly, the motor current needs to be measured during the cycle, so that switching decisions can be made based on its rising and falling through the setpoint limits.

If you add an electrical magnetic mechanical brake, you can bring this into play once the motor is turning so slowly that the desired braking torque (motor current) is not achievable during the shorted part of the cycle in a reasonable time, or if the current falls instead of rising during the short.

Good explanation John! I have never been able to get a clear explanation of how a motor generating somewhat less than the battery voltage could be used for regen until now. Sounds like they use the motor windings and the h-bridge switches like a boost converter!

Exactly so. This method prevents the motor from actually being driven in reverse, but just extracts energy from the generation process. There is nothing technical preventing you from augmenting the current ramp-up part of that cycle by applying battery voltage across the motor, adding to the generated voltage with the bridge, effectively driving it in reverse. But then you have to have controls that treat the motor as a servo, with speed feedback, to prevent the brake pedal from sending the train backwards after it stops.

I'm not really worried about saving energy too much, but if I can manage to implement it, I will. It is more important to have a modulatable brake or "top speed".

Yes, that is of course true. I would be happy with a speed limit for keeping the thing from zooming down the hill, and use a mechanical brake for bringing the train to a stop. I should be able to bring the train to a very slow speed with the throttle, and then I could have an emergency brake" for stopping, and also for obstacles on the track that might require a quick stop. The emergency brake could disconnect the motor when it is engaged.

Ok, so it is sort of like a charge pump? More complicated, of course, but sounds similar. I have been looking for regen braking circuits etc; is there a chip or an example schematic you might know of that I might be able to put to use?

Also, if I were to do it as you've explained, is it likely that I can get significant braking action over a prolonged (30 s) period of time with no danger to the battery? I'm using a 12V car battery; if I were to slam on the pedal with 500lbs on the train going downhill, might I put too much current through the battery? I should go re-read some stuff on SLA batteries (and charging them) to see how much current they can handle. I will have to do quite a bit of testing, I imagine. I could always put in another battery to help out.

I hadn't thought of that sort of brake. I like it! I was thinking only purely mechanical. You're thinking of something like a disc brake with a solenoid or something controlling the pads?

Ok, so I'm going to go look up boost converters, for one.

This method of essentially driving the motor backwards is interesting

- If I had a controller that had a pot where halfway was stop (or, neutral), and one way was forward and one way was backward, I could simply turn the pot the other way to brake the motor. I would still be worried that the fight between the battery and the motor under a high load situation might create fireworks.

(snip)

That speed control should probably be in effect during powered motion, as well as during braking. This involves measuring the speed (usually just by averaging the voltage across the motor) and having an override circuit that takes over and lowers the throttle if the speed is too high, or activates a regenerative dynamic braking once the overridden throttle goes to zero. This takes some sort of control loop to be a smooth and stable operation.

Or add to dynamic braking.

(snip)

I haven't designed this, lately, and would have to look around.

You could safely produce as much braking torque this way ans you could safely produce driving torque. Since all mechanical and electrical losses add to the braking deceleration capability, but subtract from the driving acceleration capability, this method could slow the train more aggressively than the same motor and battery could drive it. Also, there is no limit to how long you can ask the battery to keep the train in motion, but there is a time limit on how long you can decelerate it, so you can break in an overload condition that doesn't last long enough to overheat anything. Also, you can switch in a dump resistor to absorb some of the energy, so the motor and H bridge capacities are really the braking limit, not the battery.

I think the length of the down hill run is a lot bigger limit than the initial stop. Car batteries can deliver and absorb rather large pulses of power, briefly. If the battery can drive the train up the hill, it can almost certainly control its speed going down the same hill.

Reducing the peak stresses on the batteries will definitely extend their life.

Magnetic safety brakes are generally activated with a permanent magnet or spring to break by default, and released with a small electromagnet that cancels the permanent magnet field or compresses the spring. That way, when you turn the power off, the brakes hold the thing still. This is a pain, however, if the battery goes dead and you have to push. So a normally released brake that is activated by an electromagnet might be better, with a mechanical parking type brake in addition.

As I said, the biggest problem with reverse drive to brake is that things don't stop when you reach zero speed, but change direction.

look into the controllers used for electric scooters, they do most of what you want.

Bye. Jasen

Yes, that was my intention. A likely use for the train is to let younger kids operate it, so I would want to be able to set a reasonably slow speed limit, but also have it adjustable for hauling cars for track work etc.

Yes, it seems that a combo dump resistor/regen braking would be the most robust.

Well, on this track, the uphill portion is much more gradual; the same elevation is done in 2 or 3 times the length of track. So, the downhill will be more current. When you say that car batteries can handle large brief pulses of current, do you mean the pulsing during regen braking, or the average pulse over the entire braking period?

Thank you very much for your advice. It has been quite helpful.

All that may be necessary to control the dump resistor is an over voltage sensor across the battery. Any time the battery exceeds, say 15 volts, switch the dump resistor in until the voltage falls to below, say, 14 volts. Most of the time this would result in brief , repetitive periods of having the dump connected.

Both. Once the battery has been used a bit and is no longer fully charged, you can blast rather large pulses of current into it, as long as it doesn't get anywhere close to being full charged, again. There is some internal resistance in the battery that will produce heat, regardless of the direction of the current through it, but as long as the temperature of the electrolyte doesn't overheat, you should not worry too much about braking current if the battery voltage does not get excessive. The H bridge should have considerable low internal series resistance capacitor across it, so that the high frequency part of the chopped wave does not need to go all the way back to the battery. Adding a little inductance between this local energy store and the battery would greatly increase the filter effect, lowering the RMS current through the battery caused by the pulsing. This will lower the battery temperature rise caused by the high amplitude, short duration (millisecond) pulses, whether during driving or breaking.

A battery temperature interlock might be a good idea, though. It wouldn't have to disable the motor, just lower the maximum speed to a very low value, so you can still limp home.