Best Of The TP56/TP70 Kitbash thread N scale

[peteski]

PWM was used in electronic throttles long before DCC came along.  One of the original reasons was claimed to be smoother running than regular DC.  As motors became more efficient, they also became more sensitive to the pulses.

Which of the pre-DCC DC throttles used PWM?  Do you recall any specific examples?
Majority of the so called "transistorized" DC throttles were still using transistors in analog mode.  Some injected pulses  which were either sine-wave or square-wave. But those weren't PWM type of pulses.

[u18b]

For some reason, the shot with the new rear light on did not post.

So here it is.

[Lemosteam]

That frame is so simple it could be printed out of sintered stainless steel by SW.  It could also be extended under the walkway with a hole for tapping and mounting the coupler. Based on the dimensions you mentioned it meets the printing requirements.

[ Guests cannot view attachments ]

[mmagliaro]

If you are really worried about the magnitude of the PWM pulses, two 5.1v zener diodes, back to back, in series with one of the motor leads will work.    In each direction, one of the zeners is fully conducting (drops only 0.7v) and the other will drop the full zener voltage (5.1) so 5.8v are across the two diodes and only 6.2 can reach the motor.  By having two zeners back to back, it works in either polarity.  The other nice thing is that the PWM pulse shape should still be preserved, just with reduced magnitude of the pulses (which is what you want),

*** as long as the zeners have a fast enough switching time to keep up.   I don't know how high a frequency those decoder PWM streams can use.  ***  I can't seem to find any specs on zener switching times, even in their data sheets!  Maybe they are too slow for this.

 You are essentially chopping the top off the pulses.

The only downside is that the engine won't move until the average PWM voltage gets over 6, so the first half of your speedsteps might act like "nothing" - the motor won't turn.  Of course, you can adjust the DCC CV parameters so that speed step 1 is 50% of full, and that would take care of that problem.

[peteski]

If you are really worried about the magnitude of the PWM pulses, two 5.1v zener diodes, back to back, in series with one of the motor leads will work.    In each direction, one of the zeners is fully conducting (drops only 0.7v) and the other will drop the full zener voltage (5.1) so 5.8v are across the two diodes and only 6.2 can reach the motor.  By having two zeners back to back, it works in either polarity.  The other nice thing is that the PWM pulse shape should still be preserved, just with reduced magnitude of the pulses (which is what you want),

*** as long as the zeners have a fast enough switching time to keep up.   I don't know how high a frequency those decoder PWM streams can use.  ***  I can't seem to find any specs on zener switching times, even in their data sheets!  Maybe they are too slow for this.

 You are essentially chopping the top off the pulses.

The only downside is that the engine won't move until the average PWM voltage gets over 6, so the first half of your speedsteps might act like "nothing" - the motor won't turn.  Of course, you can adjust the DCC CV parameters so that speed step 1 is 50% of full, and that would take care of that problem.

Zener diodes are a good idea.

As far as 50% of speed steps being wasted is not accurate. The motor sees an average voltage derived from the width of the full-voltage pulses.  So using PWM at speed step 1, a 12 volt motor will see lets say 0.5V.  If a 6V motor is used with Zener diodes, the pulses will be the same width but voltage will be clamped to 6V.  So the average voltage the 6V motor sees will be about 0.25V (proportionally the same as the 12V motor with no Zener diodes.   So the full speed steps range will be available with either 12V or 6V motor (with Zener diodes).

Ron, Brass is heavier than steel. Your 3D printed frame will be lighter than a same frame made from brass.

[u18b]

After reading the big Digitrax Decoder Manual, I tinkered with CVs 54, 56 and 57.

This little switcher runs almost flawlessly now.   The hesitation is greatly reduced.

With a great running chassis, it's time to turn to detailing.

Nose ladder grabs came from on old JnJ set.  David's holes are 18 inches apart and that's exactly what these grabs were.
The big grab is 15 mil brass wire.




Lift rings are from BLMA set 90.




Uncoupling bar on the pilot is a BLMA set..

Holes are also all drilled for handrails.



Handrails and windshield wipers will come next.

[u18b]

If you are really worried about the magnitude of the PWM pulses, two 5.1v zener diodes, back to back, in series with one of the motor leads will work.    In each direction, one of the zeners is fully conducting (drops only 0.7v) and the other will drop the full zener voltage (5.1) so 5.8v are across the two diodes and only 6.2 can reach the motor.  By having two zeners back to back, it works in either polarity.  The other nice thing is that the PWM pulse shape should still be preserved, just with reduced magnitude of the pulses (which is what you want),

*** as long as the zeners have a fast enough switching time to keep up.   I don't know how high a frequency those decoder PWM streams can use.  ***  I can't seem to find any specs on zener switching times, even in their data sheets!  Maybe they are too slow for this.

 You are essentially chopping the top off the pulses.

The only downside is that the engine won't move until the average PWM voltage gets over 6, so the first half of your speedsteps might act like "nothing" - the motor won't turn.  Of course, you can adjust the DCC CV parameters so that speed step 1 is 50% of full, and that would take care of that problem.

Thanks for the suggestion.  I'll come back to if if needed.

I'm not worried about the pulses.

I bought about 6 more of these motors from China just to have on hand.
They arrived yesterday.
So I'm not going to worry about this if and until the motor burns up.

So far everything looks good.

[u18b]

That frame is so simple it could be printed out of sintered stainless steel by SW.  It could also be extended under the walkway with a hole for tapping and mounting the coupler. Based on the dimensions you mentioned it meets the printing requirements.

(Attachment Link)

John,
Thanks for that drawing.

BTW- the base of the coupler is not at the same height as this frame.
It would have to sit lower.
David has a receiving pad for the coupler built into the shell and it is just right for a 1015.

So I don't think that extra would be needed.

[nkalanaga]

Peteski:  I don't know if it was used in commercial throttles.  I can remember articles on building throttles that used it, usually with an SCR for control.  Apparently, today, SCRs are obsolete for this purpose, and more complicated circuits that can be fed DC are used.

[peteski]

Peteski:  I don't know if it was used in commercial throttles.  I can remember articles on building throttles that used it, usually with an SCR for control.  Apparently, today, SCRs are obsolete for this purpose, and more complicated circuits that can be fed DC are used.

SCRs, due to their electrical characteristics, are usually used with either AC or unfiltered DC from a rectifier (which is powered by a standard 60 Hz power transformer).  The resulting waveform would not look quite like today's square wave PWM. But still, I can see using an SCR for a PWM circuit (that is how household dimmer switches work too).  Big difference is the pulse frequency. The SCR circuit would have most likely produced  60 or 120 Hz pulses, while today's PWM circuits in DCC decoders use much higher frequency pulses.

[Lemosteam]

John,
Thanks for that drawing.

BTW- the base of the coupler is not at the same height as this frame.
It would have to sit lower.
David has a receiving pad for the coupler built into the shell and it is just right for a 1015.

So I don't think that extra would be needed.

Was thinking for extra mass.  The coupler height could be added, subtracted from the tab.  Eh, just an idea.

[coosvalley]

Question for you shapeways guys...I see that Shapeways offers printing in several varieties of brass, would that work for the printed frame, and be heavy like the custom brass frame made by u18b?

[peteski]

Question for you shapeways guys...I see that Shapeways offers printing in several varieties of brass, would that work for the printed frame, and be heavy like the custom brass frame made by u18b?

I was also contemplating the same thing, but it is way too pricey for me.  While there are several finish options, the way I understand this (as explained on their website) is that they print the item in wax then do a brass lost-wax casting. The finish depends on how much extra polishing and buffing they do. That is why is so expensive.  Steel on the other hand is directly printed as a 3D object.

[coosvalley]

I see.

Thread drift: Any idea of the resolution for that process?..Could it be useable, to say, make a relatively simple cast detail part?

[peteski]

I see.

Thread drift: Any idea of the resolution for that process?..Could it be useable, to say, make a relatively simple cast detail part?

The production details and all the specs are at https://www.shapeways.com/materials/brass
That's where I got my info.