filtering PWM output of analog controllers

[hans-diesel]

Hello all,

 

Most modern controllers use PWM control with a frequency around 20 kHz. Meaning that the output looks like very fast switching between 0 Volt and 12 Volt (or a little bit more).

Isn't it better to have a clean DC at rails?

I am thinking about filtering the output. A resistor in series with the output. And a capacitor parallel to the rails. The resistor should have a value around 10 Ohm. For n gauge models the current is low enough to keep the voltage drop low enough. And 10 Ohm is high enough for limiting the current to the capacitor while charging. Yes, it should be a power resistor. The capacitor should be bipolar.

 

What do I forget?

 

Hans

[kvp]

Standard model drivetrains have high starting friction and classic 3 or 5 pole dc motors also have the magnetic resistance of the permanent magnets. Together this gives very bad running characteristics at low speed when run from flat dc. Pwm technology was developed to have full torque at low speeds by having full voltage on the motors. High speed pwm was invented to smooth out the noise and vibration of low speed pwm. Tomix added the constant lighting feature on top of pwm. It's just a small percentage of stationary power to keep the lights on at full brightness regardless of speed. Most modern led lights are cl compatible.

 

Filtering the pwm wave with a series resistor parallel capacitor net has the following effects:

-disables constant lighting and anti flicker buffers in the cars

-disables slow speed smooth running

-causes a voltage drop due to the resistor acting as a voltage divider together with the motor

-generates a nonlinear speed curve for the motors, so you have to go fast if you want stable speeds

 

If you really want a flat DC signal, just use a Kato (or Lego 9V, Maerklin Z, Minitrix N) throttle as they have transistor controlled flat outputs. But at least they are linear.

[hans-diesel]

ok, thanks.

 

I thought that high frequency PWM didn't help with having full torque at low speeds. I thought that frequencies around some tens till maximum 200 Hz of PWM only help. Higher frequencies (e.g. 20 kHz) are just averaged (I thought).

 

I must think about that nonlinear speed curve. I think that I understand it. Thanks for bringing that under my attention.

 

Hans

[Kiha66]

Adding on to what KVP said, many (most/all?) transistor throttles won't output smooth DC either, leaving in the peaks from the AC sine wave to help with torque at low speed.  

[hans-diesel]

indeed, some German "DC" controllers rectify just 1 half of the sine wave, just to have higher peaks at 50 Hz instead of smaller peaks at 100 Hz.

 

Back to the filtering of high frequency PWM: I think that the idea has more drawbacks than advantages: skip and next.

 

Thanks and bye, Hans