Controller Behaviour Under Different Load Conditions

[NickM]

Hi Everyone.

This topic is a spin-off from a discussion about weight location for most realistic running - "Weight Location for Most Realistic Motion".

I proposed a way of determining that absolute maximum loco weight, based on making it slightly less than that required to stall the motor if the loco was held against buffers. This was in the context of an experiment to explore the impact of loco weight and weight position on motion, and should not be taken as a practical solution for establish loco weight

There was some discussion about what happens when a motor stalls, and how controllers, DC or DCC might behave under that condition.

This fired up my curiosity (too easy to do!), so I did some crude tests using a Gaugemaster Model D.

I have taken a few measurements, with the motor removed from the gearbox.

These are values measured with the controller set to MAX power. The measurements were made with a digital multi-meter set on the DC range.

Motor unloaded:

Voltage = 16.7 V
Current = 56 mA
Power = 0.935 W

Loaded (applying light friction to worm gear using fingers and a soft cloth):

Voltage = 15.3 V
Current = 120 mA
Power = 1.836 W

Stalled (locking the worm gear with fingers and soft cloth):

Voltage - 10 V
Current = 620 mA
Power = 6.2 W

When stalled, the motor gets hot quite quickly. I certainly would not be happy leaving it in that condition for more than a few seconds.

The current drawn under stall condition was not sufficient to trip the controller. According to the spec the controller can deliver up to 1 A per track - this suggest the motor resistance was sufficiently high to prevent the controller form tripping.

The 'loaded' case is obviously very unscientific - but it does give an indication of power consumption under 'normal' operation.

Such testing is crude at best, since it assumes that the signal us either pure DC or if there is an AC component present, the contribution of the AC component is DC equivalent.

So - I put an scope on the signal to see what was really happening - results to follow in the next post.

Nick

[NickM]

Oscilloscope plots for Gaugemaster D controller under different load conditions.

Controller was set to MAX for each test.

No load:

Scope No Load.jpg

Normal load:

Scope Under Load.jpg

Motor stalled:

Scope Stalled.jpg

In all the above plots, the 0 V point is on the second dotted line form the bottom, indicated by the 1-> to the left of the screen.

The signal supplied to the motor consists of a DC component with an AC component added. The AC component is at approx 100 Hz, so we can assume it is derived from the mains supply using full-wave rectification. We need to be careful using the term 'AC' since at no point does the signal ever go negative, we are simply adding an AC signal to a DC one.

The 'mean' DC measurement is the average value of the DC + AC components - it is not precisely accurate because the AC component is kind-of a sine wave (not a very good one). For a more precise value we would need to add the DC value to the true-RMS value of the AC component - but it's close enough, life it too short!

The short 'spikes' in the waveform are the back-emf pulses from the motor. Some controllers uses these to monitor the motor speed - I'm not sure if the Gaugemaster does though. There is potentially other noise, presumably generated by the brushes / commutator.

The peak value (ignoring the back-emf spikes) is approximately 20 V.

As load is applied and more current is drawn, the DC component reduces and the AC component becomes more apparent. The effect is very slight until a significant load is applied. The average voltage drops slightly, but not by much. The peak value drops very slightly.

At the point the motor stalls, the DC component has disappeared altogether and we are left with the AC component. The peak value has dropped to around 18 V. Since the motor is no longer rotating, the back-emf spikes (and other noise) have also disappeared.

The average voltage is 10.7 V. However this is not an accurate DC equivalent, since the waveform is roughly a sine wave. The RMS voltage, which represents the DC equivalent, in terms of heating power supplied to the motor is much closer to 12V.

All very interesting - but what does this tell us?

- Under no-load conditions, average voltage is approx 16 V and the peak is 20 V
- The signal applied to the motor is a combination of DC and AC, the DC ensures the signal is always positive with respect to 0 V
- As load is applied the AC component increasingly takes over, and the average voltage drops
- At stall, the controller is supplying a full 12 V to the motor, in terms of DC equivalent heating power - we are probably pushing close to 8 W into the motor at this stage
- This motor cannot draw sufficient current to trigger the short-circuit trip when stalled

Obviously, all of the above only applies to this controller / motor combination.

Nick

[NickM]

I just realised I misspelled the title - how do I change that, or is it only something a moderator can do?

[Hardwicke]

I just realised I misspelled the title - how do I change that, or is it only something a moderator can do?

Go into edit mode (pen like image) and the title will appear with a white background. You can then edit away.

[Tim V]

As above but in the first post the OP made.

[Winander]

I've altered the titles for you, any new posts acquire the title of the initial post at the time.

Interesting findings, I think you said it was a Mashima motor but you didn't know which. Measure the length and width across the flats assuming it is that type. The width is the first two model numbers and length the last two, so a 1224 is 12mm across the flats and 24mm long. IIRC!

[NickM]

Thanks - much appreciated

Yes - it is a Mashima motor, based on dimension it is a 1432. It has dual shafts and they are 2mm dia.

Nick

[davebradwell]

Nothing surprising here. I suggest the spikes are noise from the brushes - to measure back-emf the control system has to turn off the applied power and look at the voltage generated. Those familiar with setting Zimo decoders will know as the length of the off period and its interval are adjustable.

I believe the drop in volts is due to lack of metal in the transformer, very common in attempts to keep cost down. Spent time testing transformers in my early years in employment and we used to specify all this stuff. Our regular supplier was consistent but various attempts to use cheaper suppliers all failed due to this droop at full load.

Can't find my motor data sheet but someone will come up with a max current and yes, motor will get hot. I've never heard of any protection for stalled motor being used in our hobby. My NCE system displays current consumption and it's usually around 0.3A, a bit more when we get 4 locos running.

Impressed by access to scope!

DaveB

[martin goodall]

Although I don't have practical experience of this process, I seem to recall that it used to be recommended that the all-up weight of a locomotive should not be so great that it would stall when buffered up to a stop block on half power.

I cannot quote a source for this, and I am certainly not aware of any scientific basis for this assertion. It is just an idea I heard being put forward some years ago. If this proves to be correct, it should presumably prevent overheating of the motor by being stalled on full power.

Those with the necessary technical knowledge and reliable measuring equipment should be able to prove or disprove this theory.

[NickM]

Thanks, Dave.

I used to have a lot more specialised test gear when I was more into building audio electronics, but I have cut it right back recently. I kept the scope, it's quite an old Tektronix digital model, takes up very little space so it stayed.

Scope gets very little use though - nice to have an excuse to fire it up. 99% of the time a multi-meter is all I need.

Nick

[davebradwell]

I've a suspicion that your controller is based on a series long ago in Railway Modeller called "A transistor controller". The 100Hz you are seeing was bled through with a resistor to help in starting sticky motors. You'll need something significantly better to achieve your stated ambition so suggest leaving room for the decoder!

I still think you're worried unduly about this stalling business. I've always stuffed as much lead as possible in a loco and never burnt out a motor. My reasons have already been given but essentially someone is going to turn it off - it doesn't need to be foolproof. The key is to use a motor of an appropriate size for the loco.

DaveB