Closed loop motor control

[richard-g8jvm]

Hi

just building up another controller as my old thing makes the motors sing a bit.

I'm playing with variable pulse height and width

 

with PWM for closed loop control do you increase the pulse width with an increase in the back emf as the motor works harder  ?

 

 

It would make sense to do so, but increasing the motor speed also increases the back emf.

My layout is not level so being able to set a motor speed and maintain it would be useful.

 

Thanks

Richard

[stewartingram]

I have used an old Wireless World design since the early 70s. This was based on the then new idea of (Black & Decker) drill speed control, measuring the back emf to change the waveform supplied on a pwm type controller. Works excellently for me, I've built quite a few of them over the years for others too.

Basically with a dc controller, the 12v (a nominal figure in reality) is applied to the rails to give full speed. If this is pulsed ON-OFF-ON-OFF in equal amounts, it is the equivalent of an average 6v, ie 1/2 speed. So in a PWM design, the pulse width ON-OFF ratio is varied to give a variable average voltage and thus control the speed. The big advantage is that in trying to start the motor turning, 12v (in pulses) is available to kick the motor into life. A typical motor, already running, might keep running (on pure dc) down to say 1.5v. However to start it going again, maybe 3v would be needed; not ideal for smooth starts and low speed running!

The W.W. controller not only varies the PWM ratio, but also "shapes" the pulse by measuring the back emf from the motor, giving finer control.

It may be old but it works. If you want more info PM me.

 

Stewart

[richard-g8jvm]

Thanks PM sent

[CourthsVeil]

... My layout is not level so being able to set a motor speed and maintain it would be useful.

 

Thanks

Richard

 

Hi Richard,

pardon the question, maybe silly… Still: WHY would it be "useful" to maintain speed when the loco is going uphill?

Regards

 Armin

[richard-g8jvm]

No  not a silly question,

I'm more concerned that it doesn't go too fast down hill, I have one decline which goes in to a tunnel and a bend inside, too fast into that and it will derail.

And to gain access to that section means lifting another above it out.

 

Richard

[CourthsVeil]

Well, a different approach to model railroading then…

I personally feel satisfied by exerting control over the trains by the throttle – means that I would not let the train run uncontrolled, but rather manually change the speed as the situation requires.

That's the reason why I like most the two controllers mentioned in this and the following post. The first copes exactly with the situation described by stewardingram above (2nd paragraph), whilst the second uses back-EMF.

Maybe these both look too simple for an advanced electrician, but they fulfil my needs completely.

 Armin

[richard-g8jvm]

I don't intend running uncontrolled. there's nothing wrong with a belt and braces approach..

There are always things to cause distractions.

One thing which doesn't help with many of the designs published is they rely on using rectification without any smoothing.

Now if you decide to take advantage of the very low cost laptop replacement power supplies, you supply is DC with very little ripple.

The result is , you have to provide a 100Hz clock, and some designers also take advantage of the rounded pulse shape. So to use those designs

you end up building a 100Hz clock a monostable to get the approximate wave form timings , then through a low pass filter to remove the odd harmonics to get

back to a sinusoidal wave wave shape.

That also means your controllers will work as designed off a battery.

I'm trying to stay away from PICs or any other processor controlled devices.

.

The ww design is interesting as there are not any difficult or expensive components to source.

If you control with a 50% or less duty waveform, there is a lot of potential to what you can do with the emf generated by the motor when no power is supplied.

This is an interesting feature for me

I also want to really suppress the armature discharge pulse, the true back emf caused by the disconnection of power as the motor is run.

You can clamp this down to 600mV with silicon diodes, better to use Schottky devices as they clamp at 200mV, but I reason this can be  fully absorbed by an active device.

That will be one less source of RFI to deal with.

Also important if using switching power supplies as inductive loads can cause problems with them.

I'm not going down the DCC route as my layout is point to point single track,

 

Richard