Select Controller and Overload

[96RAF]

@Eric

I have an ESU test rig so hanging the meter onto the exposed motor will be somewhat easier than trying to do it on a loco.

 

@Fishy

I will try an R8249 which as you say cant set Vstart and a Sapphire which can and see what the outcome is. 

...but if a decoder can’t set Vstart then as soon as you move the controller to speed step one or more the decoder will send power to the motor, so there is no way you can have a decoder passing power to a stationary motor. If it is Vstart capable then it holds back until the set value is reached and it sends from there.

I will report back later if i still have my eyebrows.

Rob

[96RAF]

What I did was set up an R8215 into the ESU rig as shown in the picture with a multi-meter set to DC Volts clipped to the motor terminals.

 

Using the Elite I attempt to program the R8215 Vstart to speed step 100 (just under half throttle) and as we know the dumb controller tries to change the CV2 value but it never happens as it is not supported by the decoder, therefore speed step 1 still = speed step 1.

 

I apply throttle and the motor turns immediately speed step 1 is applied and the volts rise to around 0.7v DC across the motor. This proves the motor is not seeing a standing PWM whilst waiting for higher speed steps as it starts immediately.

 

Now I swap out the R8215 with a Sapphire and change CV2 to 100 (just less than half throttle). This time as soon as I apply throttle the motor gets a kick straight to almost half throttle and the volts rise to 5.64vDC shown in the picture (i.e. speed step 1 now = speed step 100).

 

/media/tinymce_upload/5c9641c619cac6fc3d4a9528fd086e79.JPG

 

Now the cruncher. I swap out the Sapphire for a DC blanking plug and even before I apply throttle the motor is live and buzzing but remains stationary whilst seeing full DCC voltage - note the multi-meter range is now on vAC and leds are lit showing functions F0 and F1 are activated. Applying throttle makes no difference as that command is being sent to loco 3.

 

/media/tinymce_upload/4579993d9633f6dc866d7c112a22a240.JPG 

 

This proves in my mind that Vstart setting has no latent standing voltage effect on the motor as it doesn't see that set voltage until the throttle is applied whereupon the decoder passes the voltage to the motor (i.e. Speed step 1 now = speed step 100 in my test example).

 

It also proves conclusively to me that you should not leave any DC loco on a DCC track if you value your motors.You can feel the motor is in distress, even that chunky motor on the test rig.

 

Rob 

[2e0dtoeric]

Thanks for that trial, Rob.

That will settle some 'discussions'!

[Fishmanoz]

Well constructed and described Rob. Just one problem - you’ve tested a motor setup where Vstart = 1. What you need to test my scenario is a motor setup where Vstart is other than 1, let’s say 5, using a decoder with no Vstart setting CV.  Consequently, what will happen when you set the throttle to 1, 2, 3 or 4 speed steps?  We know the motor will not be rotating, it doesn’t until speed step 5, and we also know that full voltage excursion DCC waveform is being applied with an average DC value of  up to around 300mV (128 speed steps, how many speeds steps were set in your rig, given you had 700mV at speed step 1?).

 

Also, you need to get the scope out to find the actual throttle waveform, not just the average DC value. 

 

PS.  Apologies for complicating this by mentioning Vstart, it just seems to have confused people.  I was simply using it as shorthand to, I thought simply describe, a setup where the motor remains stationary at any non-zero speed step and so has DCC motor volts applied but is not rotating.

[96RAF]

Fishy

The Elite and all decoders are all on 128 speed steps.

 

I set the R8215 to Vstart of 1, but the Sapphire was set to 100, so I did test your what-if Scanrio and the voltage passed when the controller moves to speed step one is whatever you have set in Vstart. Imagine a DC rig with an On/Off master switch in series with a rheostat. Set the rheostat to say half way, then switch on the master switch. This is what the decoder does when you alter Vstart to a higher figure.

 

The Vstart value is stored in the decoder so only that decoder sees that effect. I.e. speed step 1 on the controller passes Vstart speed step nn to that locos motor Immediately the controller is moved.

 

Any other decoder sees speed step 1 at the motor as soon as the controller is moved. Any decoder not being Vstart changable also puts speed step 1 on the controller to the motor.

 

A DC blanked loco will see the DCC waveform direct as there is no decoder firewall.

 

There is no way any decoder equipped loco motor sits there feeling voltage until the decoder passes that voltage and then that voltage is imediately whatever is set in Vstart whether 1 or 99 or 255, there is no latent voltage standing on the motor as the decoder acts as the master switch in my analogy.

 

No need for the scope as I know that would show exactly what the meter saw, i.e. no voltage trace until the controller moves off stop, then the PWM would snap to the Vstart voltage spacing. There would be no standing voltage else the motor would move as all the decoder can do is convert the DCC waveform and pass DC to the motor. I can do the scope thing if you want me to.

 

Rob

[Fishmanoz]

Sorry Rob, please lock Vstart out of your mind then  read my PS again.

 

Now go back to believing Vstart.  The reason for having Vstart at all is that in some instances, locos won’t move on speed step 1, they need a higher speed step, maybe only 2, 3, or 4, to overcome the inertia in the loco drive train, load etc.  if this is the case and you fit them with such as an 8249 without a Vstart Setting, or you have one and don’t bother to set it, then if you apply speed step 1 on the throttle, the loco will sit there with the motor unable to turn but with full DCC motor control voltage applied with a mark/space ratio of almost 1.

 

With a loco sitting as described above, is the motor as likely to be damaged as it would be if it were configured DC sitting on a DCC track?  If not, why not, what is the difference?

[96RAF]

Fishy

Lets see if I am clear about the scenario now....

 

We have a loco fitted with an R8249 or similar decoder which does not have the capability of altering  Vstart so it will apply PWM voltage to the motor from speed step 1 as soon as the controller is turned.

 

This particular loco is a bit of a dog and needs a good kick to get it going so when we apply throttle nothing happens until it gets to the higher speed step where this motor will turn and in that period it will see PWM DC voltage from the decoder applied to a stopped motor.

 

This voltage is unidirectional unlike raw DCC therefore applying a steady torque to the armature as if it was held for a stall test, as opposed to vibrating it back and forth. That is the difference, similar to a DC loco using a PWM controller under the same sticky motor scenario. The frequency of the pulse will be the most apparent difference - raw DCC being much higher and alternating positive to negative, than PWM steady positive or negative DC - no doubt someone will put numbers to that.

 

Taking the scenario to an extreme, you select this dog loco and apply a bit of throttle (not enough to get it going), then SWMBO shouts you down for lunch and you leave it with this PWM DC voltage running through the motor. What happens - I guess it gets warm and if it needed a lot of kick to get going it could see a goodly current running through the armature whilst you were feeding your face.

 

(At last he sees the problem cries Fishy)...

 

Whats the answer? If you have a loco that needs a Vstart kick then use a decoder that can change this or it is possible you could leave a loco with the engine running but stalled on load and maybe damage the motor.

 

Rob

[Fishmanoz]

By jove, I think he’s got it!

 

All except one thing - you can’t just write off the non-zero frequency components in the DCC waveform as doing nothing in this scenario and doing something in the raw DCC waveform.  The fact they rise and fall, whether crossing zero or not, will give rise to a greater or lesser torque on the motor and still cause it to vibrate.  That’s why you can hear the DCC waveform in (older) lower frequency decoders, it’s the vibration from these non-zero frequencies. And the thing is these non-zero frequency components are at full DCC PWM motor drive voltage, even if low current.

 

Again I ask - how is this different to raw DCC damaging the motor?

[96RAF]

Some decoders can tune this DC motor PWM frequency e.g. ESU uses CV49 with a choice of 15 or 30 kHz.

 

The only difference is raw DCC is pulsing both pos and neg with a mean value to the motor of zero, whilst DC PWM from the decoder output to the motor is either positive or negative so you don't get the reversal effect. i.e. DCC is push-pulll-push-pull whereas DC PWM is either push-push or pull-pull so DC could be much less damaging due to cyclic fatigue.- e.g. paper clip reversals versus repeated unidirectional bend attempts.

 

You have to rely upon the decoder electronics to damp the raw DCC frequency before it hits the motor and that is more or less effective dependant upon the decoder make and model. I know some are better than others but my reference books are a tad distant at present so I can't quote chapter and verse.

 

Rob 

[Fishmanoz]

The decoder electronics doesn’t damp the DCC waveform Rob, it produces the waveform. Damping is due to the highly inductive nature of the motor load.  It’s the motor that averages out to the DC value, although not completely else you wouldn’t be able to hear it or feel the motor vibrating.

 

I‘ll try answering my own question.  Raw DCC from the track will have a higher voltage excursion and may have up to 4 Amps of current available in the case of Hornby controllers. Decoder motor output has lesser voltage excursion (at least due to losses in the decoder, if not more) and at most can suppy only 1 Amp.  In either case if the motor is stationary or vibrating at DCC waveform frequency, one and possibly 2 poles of the armature may be subject to this power source with no cooling available as when rotating. The former is clearly on occasion just outside the capabilities of the armature coils to dissipate the power.  The latter isn’t. 

[96RAF]

Fishy

Damping was a poor choice of words by me, when what I meant was the decoder processes the RAW DCC signal into a regulated and controlled voltage to send to the motor Upon command.

 

I suppose we need a motor specialist to fully answer your question with authority.

Rob

[2e0dtoeric]

Raw DCC to a dc motor that is stopped, is cycling the pseudo ac at full power - 16v average, and as stated up to 4 amps, with no 'off' period except the minor one where the waveform crosses the zero-volts.

Pulsed dc from the decoder (or a pulse-width modulated controller) is still at full volts to the stopped winding, but the power pulse is very brief, with a comparatively long off period that allows heat to dissipate, before the next pulse arrives. As the decoder (or PWM) is 'ramped up, the pulses get longer in duration, and the 'off' gets shorter, until the motor spins up at the point it 'de-sticks'.

[Fishmanoz]

Thought you might have had something there 2e0, but on reflection not.  To turn the motor forward, you need an average positive voltage, and for reverse, an average negative.  This means the motor drive waveform must have both positive and negative excursions just like the raw DCC.  Zero volts will occur at equal mark-space ratio, positive volts where the positive pulses are wider than the negative, negative vice versa. Low speed steps will be almost equal mark-space ratio.

 

agsin, as far as I can see, there isn’t much difference between raw DCC and motor drive, as far as a stationary motor is concerned.  My conclusion - raw DCC just enough to blow the motor on occasion, not all the time or instantaneously. Motor drive, just ok.

 

i think we’ve just about done this to death.

[Fireman_Ian]

@ RAF96

 

@ Fishmanoz

 

You guys have now fried my brain  😬. And all this time I thought all I had to do was put it on the tracks, turn the little knob and off it would go.

Ian.

[96RAF]

The problem isn't new Ian so all we are doing in this thread is reinventing the wheel for those who never even knew it existed.

 

Frequency of the PWM seems to be to main killer as described here

...and modern decoders can work upside the human hearing frequency so that gets rid of one indicator.

 

This guy is an interesting read also.

 

Rob

[2e0dtoeric]

@Fishy - isn't that what I put?

The output from the decoder, at dialled zero speed, is zero pulses to the motor.

At dialled speed 1 the motor 'sees' brief pulses of either pos or neg, but not both!

So if the motor 'sticktion' is such that it won't turn until dialled speed 3, the brief pulse of power has a comparatively long 'off' period before the next pulse. So if it is left at dial 2, the motor doesn't overheat in the short term.

Once dialled speed 3 kicks the motor over to the next pole, the loco moves off.

-

Anyway, as suggested, let's leave it at that.

AND DON'T PUT DC LOCOS ON A DCC LAYOUT!

[Fishmanoz]

Sorry to continue 2e0 but no, what I’ve said is different.  The DCC motor drive goes equally positive and negative.  If it was at stop and sending an output (we don’t believe it is), the positive pulses and the negative would be of equal duration (or equal width) giving an average value of zero.  At speed step 1 going forward, the positive pulses are slightly longer than the negative giving an average value of less than + 1V.  The opposite applies in reverse.  The waveform you are describing, very long at one excursion and very short in the other, is what you get when it’s going flat out.  Your error is that the waveform goes between zero and max volts, in fact it swings at max volts positive and negative, zero is in the middle.