the current and voltage aren't linearly related in time
greg - Philadelphia & Reading / Reading
gregcinductance causes the current to increase more slowly than the voltage. I think the shorter pulse period prevents the current from reaching its maximum possible value, hence the average current is lower at higher frequency and the is less heat ~= R(Iavg)^2
Here's the thing: the voltage is what the PWM is 'controlling', as it determines the PM motor rpm. The current is the dependent variable, drawn proportional to the load, and the aggregate 'amperage' measured as the sum of the electron transfer across all the pulses in unit time will be comparable in both fast and slow modulation. (Note that we are not discussing the speed effect of 'torque ripple' on, say, the observed jerkiness of the locomotive at low speed.)
The extension of this to what the OP was having suggested to him is another matter, one on which I'm less confident: what is the net effect of substituting a special kind of tuned LC circuit (motor together with that fairly substantial electrolytic) without changing how the SCR does PWM. I don't think any slower charge or discharge characteristics due to electrolytic construction are going to matter much at the switching speed involved in his 'legacy' throttle project...
inductance causes the current to increase more slowly than the voltage. I think the shorter pulse period prevents the current from reaching it's maximum possible value, hence the average current is lower at higher frequencey resulting in less heat ~= R(Iavg)^2
gregcare you saying the average current is less with an equal duty-cycle PWM at a higher frequency because there is less time for the current to maximize due to the shorter pulse period and inductance that resists its change?
What is 'less' for the very fast modulation is the delta-I involved in each of the pulses, and that is what is resisted/supported by the energy storage in the 'inductor'. You get an inherently smoothed current because even with substantial modulation dropping the effective voltage, the rate of change between any two pulses is buffered. Remember it is the sum of the I2R losses that generates the 'heat' you don't want, so minimizing the 'high' points of the delta-I swing is important...
OvermodThe effect of the pulses is to increase and decrease current through the windings, and the changes are influenced by the inductance.
are you saying the average current is less with an equal duty-cycle PWM at a higher frequency because there is less time for the current to maximize due to the shorter pulse period and inductance that resists its change?
Thanks gregc and Overmod - in the light of what you are saying, and Randy too, I'd be getting into things a bit above my pay grade. So, I think I'll retire the unfinished SCR throttle and return to base.
The underlying issue is that he Marklin throttle is (by their own admission) a Toy Transformer. It delivers 10v DC which is good for Z - but the control is flaky, to be kind. Any thoughts on an improvement?
"There are always alternatives, Captain" - Spock.
My understanding is that the issue is related to torque ripple when the motor is turning relatively slowly compared to its design rpm and PWM is relatively low-frequency (e.g. via GTO in powerline SCR @ 120Hz.) The effect of the pulses is to increase and decrease current through the windings, and the changes are influenced by the inductance. This would be magnified in tiny motors with equivalent turns in finer wire.
A larger capacitor would create a tuned LRC circuit to smooth some of the effect of the ripple, but a much better solution is to increase the pulse frequency and use the 'energy storage' in the inductance to reduce ripple more directly. Much as I hate to say it, because I suspect the equipment you have may not be able to be rebuilt cost-effectively to work at the best range of higher frequency/potential interference.
der5997So my question would be, would a larger electrolytic capacitor, so wired, solve the Z motor overheating issue?
i was told several decades ago that PWM causes motors to overheat. at the time, i think PWM operated at low frequencies and the current during the On cycle was relatively high. I believe PWM is operating at higher frequencies (~200kHz) and overheating is less of a problem because it averages out better.
i've built a simple PWM throttle that plugs into the jack the NCE PowerCab plugs into to test DC locos before adding a decoder. you could do similar using whatever power supply you currently use
Thank you Randy, just the info I was after. Pity, but back in the project box for this, I guess. But, before I do that one last kick of the can.
Thorne's article also observes that (following directly from the quote I gave about the pulses maybe overheating the tiny Z motors) "...and the AC power supply can be any 12 volt to 20 volt transformer capable of supplying the current required by your trains. For N, wire a 200 - 1000 micro farrad electrolytic apacitor with pin + to pin# 9 and - to pin# 8 (referencing the ciruit schematic) and add this capacitor for HO if your motors shows any sign of overheating."
So my question would be, would a larger electrolytic capacitor, so wired, solve the Z motor overheating issue?
Thanks, John.
Thank you Kevin, and happily the top voltage of this SCR throttle is 10 volts, which is not why I chose to built it back when, but is useful now for t his Z motor application. Thanks also for the truck info, I did not know that, and now I do!
In electronics it's a Silcon Controlled Rectifier. That doesn;t mean that it's controlled by silicon, it means that it's a controlled rectifier made from silicon (as opposed to say a germanium rectifier or a selenium rectifier (commonly found in old train power packs - they're the ones that give off the horrid and very noxious fumes when the rectifier burns out). What happens is that in one direction they don;t pass current, just like any diode, but in the other direction, they don't automatically pass current, there is a gate that controls forward current flow. By turning it on early, you get a longer time, more area under the curve, more apparent DC voltage. Turn it on late in the cycle, and you have a short on time duration, smaller area under the curve, lower DC voltage.
The pulses used in the Thorne throttle are low frequency AND at a fairly high amplitude - well over 12 volts. It's easy to just use the 60Hz AC line, reduced with a transformer - it reverses 60 times a second and handles shutting off the SCR. The combination of low pulse frequency and high amplitude introduces more heating int he motor, the larger the motor, the more weight of metal there is to act as a heat sink, but small Z scale motors aren't meant for even 12 volts fully filtered DC, they were made for 8 volts. (pretty sure it was 8, not 10 - or maybe someone back in the early days made a Z scale transistor throttle and used 8 volts because 8 volts still made them run faster than any real loco, and 10 made them rocket ships - like early N scale with top speeds on 12 volts of 300+ smph)
Do not use an SCR throttle like that with a loco like the new Kato P42 with the truck mounted coreless motors, either - it will burn those up pretty quick as well.
--Randy
Modeling the Reading Railroad in the 1950's
Visit my web site at www.readingeastpenn.com for construction updates, DCC Info, and more.
This is my best guess:
Marklin Z scale operates at 10 volts, not 12-15 like other scales. The maximum voltage of the SCR throttle might exceed the voltage rating.
It sounds like this might be a Square Wave or Pulse Width Modulated throttle, which is not directly realted to SCR, which is just the rectifier. Solid State throttles were kind of new in the 1980s.
Now SCR commonly stands for Selective Catalytic Reduction, an emission control for Heavy Duty Trucks. I had to do a double-take on your post.
On to people who know more...
-Kevin
Living the dream.
Back inthe 1980s I built most of a SCR Throttle from Peter J. Thorne's "34 New Electronic Projects for Model Railroaders". This was intended for use on a HO layout, so overheating of loco motors from the SCR pulses was not an issue. The project was not completed but I kept the components I built.
However, I now have a Nn3 layout running Marklin Z scale locomotive chassis - and Mr. Thorne's text concerning power supplies for the SCR throttle reads "The SCR throttle can be used for any scale except Z (the pulses may overheat these tiny motors)..."
Here's my question. The Thorne article is from the early 1980s. Has Marlin's motor technology advanced in the interveening years suffiently to where the pulses of a SCR throttle are not longer an overheating issue?
To complete the throttle I need to purchase some items, one of which, a 24v latching relay, may set me back $50 Cdn. I don't want to buy that if the overheating thing is still a deal-breaker.
Many thanks in anticipation of your expertise.