Coreless motors and controller

[johnarcher]

I am soon to have a first go with one of the High Level coreless motors (in an Albion Terrier kit just found on ebay). Now I know these motors don't like feedback, and my only controller is a 1980s ECM Compspeed. It is at least the sort with switchable feedback, so would it be OK with feedback off, or would it be better to invest in a modern non-feedback controller? (Gaugemaster Combi?)

[Miss Prism]

It's a bit of an urban myth that coreless motors 'don't like feedback'. It's a question of degree.

 

The ECM Compspeed is a great controller, and will be fine with those motors IMO.

 

 

[johnarcher]

1 hour ago, Miss Prism said:

It's a bit of an urban myth that coreless motors 'don't like feedback'. It's a question of degree.

 

The ECM Compspeed is a great controller, and will be fine with those motors IMO.

 

 

Thank you, the things these electrons get up to are a bit of a mystery to me.

The ECM has two levels of feedback, as well as feedback off, I'll try off and maybe the lower level.

[kevinlms]

Good to read that you can do proper research, rather than just relying on myths.

[AndyID]

21 hours ago, johnarcher said:

I am soon to have a first go with one of the High Level coreless motors (in an Albion Terrier kit just found on ebay). Now I know these motors don't like feedback, and my only controller is a 1980s ECM Compspeed. It is at least the sort with switchable feedback, so would it be OK with feedback off, or would it be better to invest in a modern non-feedback controller? (Gaugemaster Combi?)

 

It depends 😀

 

If the controller outputs pulses rather than smooth DC that can make the motor overheat. This affects coreless motors more than other DC motors because coreless motors have little mass to act as a flywheel. That makes the motor accelerate and decelerate a bit with every pulse and that produces heat in the windings.  The heat can't shunt into a core because there isn't one.

 

If the pulse frequency is high enough the motor will effectively "see" DC and there won't be excessive heating. DCC decoders output high frequency pulses that are compatible with coreless motors.

 

So it depends on the pulse frequency and I don't know what the ECM controller output looks like. I'd suggest you try running the Terrier with the body off so you can monitor the motor temperature at various speeds with and without feedback.

[johnarcher]

11 minutes ago, AndyID said:

 

It depends 😀

 

If the controller outputs pulses rather than smooth DC that can make the motor overheat. This affects coreless motors more than other DC motors because coreless motors have little mass to act as a flywheel. That makes the motor accelerate and decelerate a bit with every pulse and that produces heat in the windings.  The heat can't shunt into a core because there isn't one.

 

If the pulse frequency is high enough the motor will effectively "see" DC and there won't be excessive heating. DCC decoders output high frequency pulses that are compatible with coreless motors.

 

So it depends on the pulse frequency and I don't know what the ECM controller output looks like. I'd suggest you try running the Terrier with the body off so you can monitor the motor temperature at various speeds with and without feedback.

Thanks, I'll check that (when I've built it!)

So there may be pulses even with feedback off?

I did see, looking at stuff about this online last night, that the NLG  Bodmin people used a lot of Portescaps and controlled them with a Compspeed F ( which mine is) with feedback off. Seems encouraging?

[AndyID]

1 minute ago, johnarcher said:

Thanks, I'll check that (when I've built it!)

So there may be pulses even with feedback off?

I did see, looking at stuff about this online last night, that the NLG  Bodmin people used a lot of Portescaps and controlled them with a Compspeed F ( which mine is) with feedback off. Seems encouraging?

 

The only way to know for sure is with an oscilloscope. I have one but the leads are not long enough to reach your controller 😄

 

It could be that the Compspeed outputs DC and when FB is on it periodically interrupts the output current so that it can sense the motor's EMF. That can be a problem with a coreless because it has so little flywheel mass. Anyway it sounds like the Bodmin folks have your answer.

 

BTW, does your Terrier have a flywheel?

[johnarcher]

56 minutes ago, AndyID said:

 

The only way to know for sure is with an oscilloscope. I have one but the leads are not long enough to reach your controller 😄

 

It could be that the Compspeed outputs DC and when FB is on it periodically interrupts the output current so that it can sense the motor's EMF. That can be a problem with a coreless because it has so little flywheel mass. Anyway it sounds like the Bodmin folks have your answer.

 

BTW, does your Terrier have a flywheel?

It hasn't got any wheels yet! It won't have though, I plan to use a High Level 1219, which has a single short shaft (on a Roadrunner+).

I just hope I can trust the Bodmin story.

[DCB]

Like everything else it's a matter of degree. If you gear  for a sensible speed revving its head off like 30/40 mph, and drive at scale speed it's really not going to matter. If you run at speeds the real thing could not dream of, like 2mph with a heavy load then it will end in tears, pretty much irrespective of what controller you use. If Morley made sensible single units I would recommend a Morley which gives nice controllable variable voltage, but they don't any more.  Shame.

Mine is charging an Acer Laptop battery at present set to 12 volts.  Earlier I was testing LEDs with it at 2.4 volts.  I have a question on this forum and am still waiting to hear the Gaugemaster can do the same trick, but really a High level. a decent gear ratio and as long as it runs freely there shouldn't be a problem what power supply you use, 3  AAA cells and a on off reverse switch would  probably be adequate for a shunting plank.

[AndyID]

2 hours ago, DCB said:

Like everything else it's a matter of degree. If you gear  for a sensible speed revving its head off like 30/40 mph, and drive at scale speed it's really not going to matter. If you run at speeds the real thing could not dream of, like 2mph with a heavy load then it will end in tears, pretty much irrespective of what controller you use. If Morley made sensible single units I would recommend a Morley which gives nice controllable variable voltage, but they don't any more.  Shame.

Mine is charging an Acer Laptop battery at present set to 12 volts.  Earlier I was testing LEDs with it at 2.4 volts.  I have a question on this forum and am still waiting to hear the Gaugemaster can do the same trick, but really a High level. a decent gear ratio and as long as it runs freely there shouldn't be a problem what power supply you use, 3  AAA cells and a on off reverse switch would  probably be adequate for a shunting plank.

 

You might want to take a look at this.

 

https://dccwiki.com/Coreless_Motor

[Miss Prism]

4 hours ago, johnarcher said:

I just hope I can trust the Bodmin story.

 

I'm no longer involved with the layout, but the Portescap locos are still perfectly happy on the wonderful Compspeed Fs.

 

[AndyID]

31 minutes ago, Miss Prism said:

 

I'm no longer involved with the layout, but the Portescap locos are still perfectly happy on the wonderful Compspeed Fs.

 

 

From what I've been able to find it seems the ECM Compspeed definitely outputs PWM (pulse width modulation) at a frequency of 100 hertz. That is not at all good for coreless motors but it's likely not a problem if the motor has a flywheel.

 

Do they have flywheels on Bodmin?

[Miss Prism]

8 hours ago, AndyID said:

Do they have flywheels on Bodmin?

 

No.

[ejstubbs]

15 hours ago, AndyID said:

The only way to know for sure is with an oscilloscope. I have one but the leads are not long enough to reach your controller 

 

Cheap digital oscilloscopes like this one or this one are more than adequate for this kind of job.  Maybe difficult to justify the cost for a one-off task but whoever turned down the opportunity excuse to have more toys tools?  (And you might be surprised how useful one can be for other things.  I recently used mine to find the "outside foil end" of a cylindrical foil guitar capacitor - and it definitely made a difference installing it with that side connected to ground.)

 

15 hours ago, AndyID said:

It could be that the Compspeed outputs DC and when FB is on it periodically interrupts the output current so that it can sense the motor's EMF.

 

That is how traditional feedback controllers work.  I doubt the Compspeed is any different.

 

15 hours ago, AndyID said:

If the controller outputs pulses rather than smooth DC that can make the motor overheat. This affects coreless motors more than other DC motors because coreless motors have little mass to act as a flywheel.

 

A key factor is the significantly lower inductance of coreless motors cf cored motors.  That means that, at the kind of PCM frequencies used by the type of feedback controller which uses pulses at around or about mains frequency, a coreless motor's impedance can be 20% or less than that of a cored motor.  Low impedance means that the current that flows during the pulses is higher - and high current causes overheating.  The lack of a heat sink in the form of an iron core just makes the problem worse.

 

The impedance of an inductor is directly proportional to frequency, so at higher frequencies such as the 10-20kHz used by DCC decoders, the impedance of the coreless motor's coils is greater (100 times as much with e.g. 10kHz PCM vs 100Hz PCM) so the current flow - especially the initial peak current during each pulse - is significantly lower and the heating effect is much reduced.

Edited October 21, 2022 by ejstubbs

[PaulG]

Just to add my "twopenneth", I had to change my ECM 100 controllers for Gaugemaster Model 100, when I purchased a Model Rail J70 tram loco, which has I believe a corless motor.

 

Paul

[AndyID]

6 hours ago, ejstubbs said:

 

Cheap digital oscilloscopes like this one or this one are more than adequate for this kind of job.  Maybe difficult to justify the cost for a one-off task but whoever turned down the opportunity excuse to have more toys tools?  (And you might be surprised how useful one can be for other things.  I recently used mine to find the "outside foil end" of a cylindrical foil guitar capacitor - and it definitely made a difference installing it with that side connected to ground.)

 

 

That is how traditional feedback controllers work.  I doubt the Compspeed is any different.

 

 

A key factor is the significantly lower inductance of coreless motors cf cored motors.  That means that, at the kind of PCM frequencies used by the type of feedback controller which uses pulses at around or about mains frequency, a coreless motor's impedance can be 20% or less than that of a cored motor.  Low impedance means that the current that flows during the pulses is higher - and high current causes overheating.  The lack of a heat sink in the form of an iron core just makes the problem worse.

 

The impedance of an inductor is directly proportional to frequency, so at higher frequencies such as the 10-20kHz used by DCC decoders, the impedance of the coreless motor's coils is greater (100 times as much with e.g. 10kHz PCM vs 100Hz PCM) so the current flow - especially the initial peak current during each pulse - is significantly lower and the heating effect is much reduced.

 

Someone out there must have a scope and a Compspeed 😀

 

From what I can find it seems the Compspeed works by modifying rectified DC. If the pulses are always initiated at or close to zero volts the current rise will not be too extreme. It would definitely be a lot less than a system that uses rectified and smoothed DC then chops it.

 

There actually is a way to make a controller with speed feedback that does not pulse the output at all. It works by continuously sensing the motor current and increasing the voltage to compensate for most of the voltage drop across the internal resistance of the motor. As the load on the motor increases the voltage increases to compensate for the increase in current. Works great too!

 

 

 

[johnarcher]

20 hours ago, AndyID said:

 

From what I've been able to find it seems the ECM Compspeed definitely outputs PWM (pulse width modulation) at a frequency of 100 hertz. That is not at all good for coreless motors but it's likely not a problem if the motor has a flywheel.

 

Do they have flywheels on Bodmin?

But would a Compspeed F be outputting that with feedback switched off? If it does what would be the difference between feedback on and off? 

Probably a silly question, but I'm afraid I am sadly ignorant of electronics.

[AndyID]

1 hour ago, johnarcher said:

But would a Compspeed F be outputting that with feedback switched off? If it does what would be the difference between feedback on and off? 

Probably a silly question, but I'm afraid I am sadly ignorant of electronics.

 

No, it's a very good question. My educated (or not so educated 😀) guess is it outputs pulses 100 times per second. The pulses start when the rectified AC is close to zero volts and the control knob widens them as it is turned towards max output. When feedback is turned on it further adjusts the width of the pulses based on the motor's back-emf but in both cases it is supplying pulses to the motor.

 

I think the reason Bodmin operates with FB off is simply because it does not work well with coreless motors due to their very small inertia (flywheel effect). My guess is it would actually work if the motors had flywheels (or very low friction drive-trains - no worm gears).

[johnarcher]

12 minutes ago, AndyID said:

 

No, it's a very good question. My educated (or not so educated 😀) guess is it outputs pulses 100 times per second. The pulses start when the rectified AC is close to zero volts and the control knob widens them as it is turned towards max output. When feedback is turned on it further adjusts the width of the pulses based on the motor's back-emf but in both cases it is supplying pulses to the motor.

 

I think the reason Bodmin operates with FB off is simply because it does not work well with coreless motors due to their very small inertia (flywheel effect). My guess is it would actually work if the motors had flywheels (or very low friction drive-trains - no worm gears).

Sorry, maybe it's because it's late or maybe I'm just dense regarding this subject, but you seem to be saying that there are pulses whether feedback is on or off (just varied according to back emf if it's on). In that case though why do the Bodmin Compspeeds work better with the Portescaps with feedback off, if it's still outputting pulses anyway?

[AndyID]

2 minutes ago, johnarcher said:

Sorry, maybe it's because it's late or maybe I'm just dense regarding this subject, but you seem to be saying that there are pulses whether feedback is on or off (just varied according to back emf if it's on). In that case though why do the Bodmin Compspeeds work better with the Portescaps with feedback off, if it's still outputting pulses anyway?

 

It's because coreless motors slow down as soon as the current is interrupted and that confuses the feedback sensing circuit. The core in conventional motors acts as a flywheel so they don't slow down so quickly.

 

DC motors act as generators when they are turning and the electronics senses the voltage (which is proportional to the speed of rotation) when the current is interrupted but if the motor slows down too quickly the electronics over-compensates.

[Miss Prism]

Iirc, we didn't/don't turn the feedback off on the Compspeeds - we turn it down. This topic is all about pulse shape, and the degree of back emf being allowed to set the outgoing voltage level.

 

In my view, all motors, coreless or otherwise, need to be regulated. The term 'throttle' is very accurate. Try controlling a coreless without controlling its bemf.

 

 

[AndyID]

9 hours ago, Miss Prism said:

 

In my view, all motors, coreless or otherwise, need to be regulated. The term 'throttle' is very accurate. Try controlling a coreless without controlling its bemf.

 

 

A coreless motor can be controlled perfectly well with smooth, regulated DC and no feedback of any sort. Feedback has the advantage that it automatically compensates for variations in load due to curves, gradients etc so that the motor maintains approximately constant speed.

 

A "perfect" DC motor (cored or coreless) would have no internal resistance and it's EMF (meaning its speed) would perfectly correspond with the applied DC voltage. No feedback would be required.

 

But all motors do have some resistance which is why they slow down as the load increases. That causes an increase in current draw which results in an increase in the voltage drop across the internal resistance which is effectively in series with the motor.

[kevinlms]

4 hours ago, AndyID said:

 

A coreless motor can be controlled perfectly well with smooth, regulated DC and no feedback of any sort. Feedback has the advantage that it automatically compensates for variations in load due to curves, gradients etc so that the motor maintains approximately constant speed.

 

A "perfect" DC motor (cored or coreless) would have no internal resistance and it's EMF (meaning its speed) would perfectly correspond with the applied DC voltage. No feedback would be required.

 

But all motors do have some resistance which is why they slow down as the load increases. That causes an increase in current draw which results in an increase in the voltage drop across the internal resistance which is effectively in series with the motor.

Your description also explains how varying the resistance to control the speed of trains, generally works so poorly.

[Izzy]

I’m sure I’ve read somewhere that people have added a small resistor (no idea what value) into the motor feed to ‘tame’ the spiky performance at low rpm that can happen. I’ve just stuck to emitter/follower type DC controllers like the original Gaugemaster 100. 

[roythebus1]

With the very short bursts of power any model loco will run would overheating be a problem, especially on a shunting plank or BLT. It may be a problem on a big roundy roundy with continuous running for long periods. Why worry?