Explain this to me

[tom mann]

I picked up one of the Atlas HO scale HH600s for my Hinshaw Valley.  It ran pretty stiff out of the box and the rear light would flicker while going forward.  It did this for a few weeks with no improvement.  One day, I cleaned the track with a few drops of oil.  Almost instantaneously, the HH ran perfectly.  It was quieter, coasted farther on flywheel power, and the rear light stopped flickering.

So what happened here?

[DKS]

I picked up one of the Atlas HO scale HH600s for my Hinshaw Valley.  It ran pretty stiff out of the box and the rear light would flicker while going forward.  It did this for a few weeks with no improvement.  One day, I cleaned the track with a few drops of oil.  Almost instantaneously, the HH ran perfectly.  It was quieter, coasted farther on flywheel power, and the rear light stopped flickering.

So what happened here?

Dirty track. Try cleaning it with a solvent, alcohol or whatever, and see what the loco does. I'd imagine its performance will be comparably good. Oil is non-conductive, so chances are it helped remove fine particulate matter from the rails.

[tom mann]

But how could dirty track improve the flickering light that shouldn't have been on in the first place? 

[wcfn100]

I seem to vaguely remember this issue and I thought it was the motor.

edit: Here's what I remembered from the A-Board regarding flickering lights.

The flicker is caused by feedback from the motor. It has always been there, but until the use of white and golden-white LEDs, you didn't notice it.

Don't know if it's the same thing.

Jason

[C855B]

Yes - it's the same thing. It happens because of "BEMF", or back-EMF. This is the current produced by the magnetic field (flux) collapsing whenever power is broken to an inductor (coil). Recall that the motor is 3 or 5 coils. The flickering happens because the intermittent wheel contact breaks the connection, and it lights-up the rear headlight because the BEMF polarity is backwards from the direction of travel. You would rarely see this with incandescent bulbs since they needed a couple of milliseconds to "warm up" to full brightness and the effect is very short, while LEDs instantaneously respond to the current.

BEMF is the acronym TCS uses to describe their advanced DCC motor control where they "read" the BEMF to precisely compute motor speed.

Oh - "EMF" is "electromotive force". Has nothing to do with the 1:1 locomotive maker. Nor flux capacitors.

[Zox]

Yes - it's the same thing. It happens because of "BEMF", or back-EMF. This is the current produced by the magnetic field (flux) collapsing whenever power is broken to an inductor (coil).

Makes sense to me. I had a timer circuit that blew itself out because of back EMF. I forgot to put a reversed diode across the relay coil to give the decay current an easy dissipation path, so it dissipated itself back through the timer chip. This is what happens when you let a software guy play with hardware. 

Of course, that was nothing compared to the previous failure, which was caused by a lightning-induced power surge:

[peteski]

Kato (and probably other manufacturers) solves this problem by attaching a capacitor in parallel with the LED. While I haven't measured its capacitance, I suspect it is about 0.1 micro Farads.  The capacitor acts as a shunt for the spikes so there isn't enough voltage for the LED to light up.

[wcfn100]

Kato (and probably other manufacturers) solves this problem by attaching a capacitor in parallel with the LED. While I haven't measured its capacitance, I suspect it is about 0.1 micro Farads.  The capacitor acts as a shunt for the spikes so there isn't enough voltage for the LED to light up.

We did add the capacitor to some of our N scale models, but then it causes a short when you try to run a DCC unit on analog (not a common occurance, but a problem of a different sort).

Jason

[Chris333]

I'm no pro as I rip the lights out of everything, but I ordered LEDs from Richmond Controls.

He sent along some info about wiring either resistors, caps, or diodes across the LED leads to keep them from flickering.

[peteski]

Jason, I don't know all the details about Atlas' capacitors but if they cause a DCC signal short, their value is probably way larger than needed.  A 0.1 micro Farad cap (or whatever value Kato uses on their light boards) should not cause a short with analog locos running under DCC power.

[wcfn100]

I'll make sure to pass that along to Paul next time we hang out.


Jason

[peteski]

I'll make sure to pass that along to Paul next time we hang out.
Jason

The capacitor in question is wired in parallel with the headlight LED (not across the track). There is also a series resistor (of at least few hundred ohms) in the headlight circuit. Even if the LED/cap was totally shorted, because of that resistor, there would not be a short across the track.

[Ian MacMillan]

Here is a page with a photo of the N scale board with the cap.

http://n-scale-dcc.blogspot.com/search/label/Decoder%3A%20TCS%20M1

[TiVoPrince]

Before  
I run over to Fry's I have a question on application of capacitors.  Does the capacitor go across the blue-white pads before reaching the resistor and LED?  Will a capacitor be necessary for each circuit or will one capacitor be needed for each circuit or will one capcitor work for all lighting?  Going to affect my quantity to purchase...

[peteski]

Before  
I run over to Fry's I have a question on application of capacitors.  Does the capacitor go across the blue-white pads before reaching the resistor and LED?  Will a capacitor be necessary for each circuit or will one capacitor be needed for each circuit or will one capcitor work for all lighting?  Going to affect my quantity to purchase...

If your LED is powered through a DCC decoder, you don't need any capacitors across the LED. The decoder's circuitry insures that there are no voltage spikes get into that circuit.  But if you needed the cap, it would be wired in parallel with the LED directly (not before the resistor).

Cap is only needed for locos without decoders.