Hi,

Some of you have already helped me in my other thread about building a layout and for the resons of a defect in one of the wagons in my starter set I happened to measure the voltage that get's to the tracks from the Märklin speed controller.

I'm really newbie in electronics, but I thought pulse width modulation gives full output voltage and the position of the speed knob just changes the pulse length, right?

However my multimeter shows that voltage goes from 0 to about 10 when the speed knob is rotated...

So what am I getting wrong, or is the Märklin speed controller not a PWM device at all?

Best,
-Eino

Are we talking about the recent speed controller that comes with a wall wart? Or the older all-in-one units?

According to this post on Stummi's forum, the Märklin 67013 does not feature PWM:
https://www.stummiforum....in-Trafo.html#msg1397389

So it seems that PWM does not really bring much to the table.

Makes me wonder, would it be a good idea to just have 10V DC power unit (old mobile phone charger perhaps) and then have several potentiometers in parallel in order to have cheap and easy speed controllers?

-Eino

Not entirely true, a switch mode power supply does use pulse width modulation to regulate the output voltage. As the input voltage changes over the allowed +/-10% tolerance the switching regulator will adjust its pulse with to keep the output voltage constant - and the power supplies that will handle 94-260V input (such as the currently supplied power supplies for the cs3) will be capable of adjusting their pulse width over a much wider range.

But as you state, these should not be confused with a PWM speed controller for a model railway.

What comes out of a marklin controller to the track is not pulse width modulated (not PWM).

Yes, Pulse Width Modulation adjust how much of the duty-cycle of the waveform is in the high position (+Vcc), vs how much is in the low (0v) position.

A multi-meter is trying to get you a momentary voltage, so it is giving you an average.
The frequency of the duty-cycle is likely higher than the multimeter can detect as distinct, that is how one wants it (higher frequency to react quickly)... so just how a light looks 'dimmer' to your eyes when the PWM wave is at 50% or 20% duty-cycle, rather than flashing off and on (which is what a camera might see), the voltage shown by a multimeter is an 'average' over some time.

An oscilloscope will give you a visual representation of the voltage.

In that context, "noise" refers to radio interference, not audible noise.
Modern decoders often operate at 20 to 40 kHz, so beyond the audible frequencies.
I think early c90 decoders were around 90 Hz - much slower, too slow for coreless motors, but no unpleasant audible noise either.

Radio interference suppression.
The motor works without them, but the decoder may work better as the effects of sparks occurring at the commutator are reduced.

Correct.



There are two types of noise being referred to here. The first is audible noise. Think in terms of a transformer hum when plugged into the mains. The continuously changing magnetic field makes the laminations physically move creating the hum you hear. Similarly the laminations in an iron core motor will physically vibrate at the PWM frequency and make an audible noise. It won't be a pure hum like a transformer, as it will have many harmonics due to the shape and width of the PWM pulses.

There is also a caution to note here with some motors, specifically 'coreless' motors that do not have an iron core on the rotor, but rather the windings are would on a former and then impregnated with a lacquer to hold them in shape. when mounted in the motor assembly the windings rotate around a fixed magnet. The problem that arises is the windings have very low mass, and if driven at a low frequency PWM will physically vibrate and eventually shake themselves apart destroying the motor. The solution is to have the PWM frequency set very high, above the mechanical resonant frequency of the windings. This frequency is often 20kHz or higher, sometimes called 'supersonic drive' in MRR decoder literature or specifications. However driving an iron core motor at this frequency will cause the motor to get quite hot due to a problem known as 'eddy current loss' in the iron core, so these need to be driven at a much lower frequency. The actual frequency is dependant on the quality of the iron used in the core material. So there is no one ideal fix for this problem.

The second type of noise is radio frequency interference due to the motor brushes arcing as they move between commutator segments. Historically this is what caused tear lines across old analogue TV screens and hash noise in radios. This gets quite bad with MRR layouts because the track acts as an aerial to radiate the RF noise from the motor. So the solution is to put the inductors in series with the motor leads, and typically a capacitor across the motor terminals, to filter the RF noise and so stop the RF noise from reaching the track and radiating out into the airwaves.

I know that I am getting rather off topic here, but just wanted to add that I use strings of anti-parallel diodes to regulate the power to the tracks. (and even to limit the top-speed of individual locos), on my analogue-running layout.

The diodes can stand quite a lot. I once had a short circuit from a derailed and decoupled wagon in a hidden area; Eventually the soldered diode-pair melted the solder and disconnected. The diodes survived!
((Yes, it was a weak solder-point, and the tension was only around 12Vac from a 30VA light blue M-trafo. -Yet it saved both track and wagon, and didn´t start a fire. ))