Attention: electronic smarts needed in this minor customization

Try a filter capacitor first, or adding in a less sensitive diode for each led, or resistance .

The issue is likely the LED's are not receiving enough voltage under actual operation.  A solution that won't break the bank is to add a couple of diode pairs in series with the motor leads after the LED connection, this will force more voltage for the LED's.  Since I'm lazy, I use bridge rectifiers to minimize the wiring.  This arrangement will drop the voltage about 3V to the motors and allow reliable operation of the LED's.

After the above suggestions I added a 1uf 50v NP electrolytic cap across the motor leads - no change. Added another one across the AC input leads to the e-unit - no change. I left those in place and added a pair of bridge rectifiers in series with each other and the motor - the only change this produced was to noticeably increase the voltage needed to run the engine.

I should point out that the flickering occurs very noticeably while the engine is running on some slightly dirty O27 track. On the bench powering with alligator clips, very little to no flickering.

The e-unit is an old one, and it has seen better days - I threw it in really just to see if it still works, which it does. It has some green corrosion on a couple of the discrete diodes, but works reliably in the F-N-R-N cycle and the engine maintains a steady speed around the circle. 

More thoughts??

Well, giving them a little more voltage and adding the capacitor should solve both issues pretty totally.

Here's the "deluxe" version.

The cap is a 47uf non-polarized cap, and the two resistors are just to keep you from affecting the reverse unit with excessive output capacitance, some of them don't like that.

Since the undesired "opposite" LED flickers more strongly with dirty track, this is likely related to intermittent loss of track voltage.  When the current is interrupted to the motor, the inductance (motor windings) generates an inductive "kickback" voltage of the reverse polarity.  This is not to be confused with motor back-emf which is of the same polarity as the drive voltage (hence it would not turn on the opposite LED).

As Adriatic said earlier the idea is to make the LED less sensitive so that this intermittent reverse voltage does NOT light up the reverse/opposite LED.  Putting a bridge-rectifier as GRJ proposes does in fact reduce the amount of reverse voltage to the LED but it does come at a substantial cost.  The bridge must be capable of carrying full motor current so you can't use tiny 1-cent diodes.  But a bigger penalty is you waste a lot of power in the bridge when the objective is to lower the voltage that the LEDs see which is relatively negligible current (milliAmps vs. Amps).

Since you obviously have white LEDs which you probably have a bag of, why not insert a 2nd pair of back-to-back white LEDs.  This will drop the voltage to the LEDs by ~3V.  This means you now need ~6V to drive both the active (desired) and backward (undesired) LED.  The 2nd pair of LEDs can be placed together and serve as an interior light since one or the other will always be on when the motor is active. 

If this does not starve the opposite LED enough, then add non-polarized capacitance.  If the "6V" LEDs are too dim in the desired direction, cut the 470 Ohm resistor value.  If all you have are 470 Ohm resistors, place 2 in parallel to "make" a 235 Ohm resistor.

Stan, while it's true that the bridges carry full motor current, that's probably not a major issue in this case.  He's doing the Lionel Commando Assault train, it's a very small locomotive with a tiny DC motor!

1 amp bridges should be MORE than enough for this set.  I'm confused about the comment about lowering the voltage the LED's see, the series resistor does that.  However, you need enough headroom for the LED to light at all.  Since the white LED's need 3+ VDC, that's perhaps a limitation with that light train.  If the voltage from the reverse board isn't enough to properly light the diodes when the train gets up to the desired speed, the solution is to give more headroom to the LED's.

I'm not seeing where putting two LED's in series is solving this problem, I suspect with 6VDC on that little motor, that locomotive is flying down the tracks, they don't take much to move.

As far as wasting power, you're wasting a watt or so from a train transformer, I don't see that as an issue!

gunrunnerjohn posted:

...

I'm confused about the comment about lowering the voltage the LED's see, the series resistor does that.  However, you need enough headroom for the LED to light at all.  Since the white LED's need 3+ VDC, that's perhaps a limitation with that light train.  If the voltage from the reverse board isn't enough to properly light the diodes when the train gets up to the desired speed, the solution is to give more headroom to the LED's.

I'm not seeing where putting two LED's in series is solving this problem, I suspect with 6VDC on that little motor, that locomotive is flying down the tracks, they don't take much to move.

Note that he has plenty of track voltage to light the LED in the intended direction.  It's the anti-parallel LED in the unintended direction that is flickering albeit less brightly.  The premise is this reverse voltage is coming from the "-L di/dt" kick each time motor current is interrupted by dirty track contact.  This reverse voltage then pulses/flickers the unintended LED.

Apparently the LED in the intended direction is plenty bright.  By making the LEDs less sensitive (requiring more voltage), this might be able to starve the unintended LED of whatever reverse voltage pulse is coming from the motor during each dropout.  The intended LED would also require more voltage to operate but, again, it appears this is not an issue.  

I am not familiar with this engine so don't know its intended operating voltage range.  But even if it's a small engine and small motor there are obviously small DC motors that nominally operate at 12V DC; plus, I don't think E-units operate down at 3V AC or whatever?  

So I think the found the manual for the train set which only had a line-drawing of the transformer.  Then I found an image of a transformer for sale on eBay that looks just like it.  0 - 18V DC.   Albeit only 5.5 VA.

Hmmm... I was hoping to finish this today, but short of soldering forty seven 1uf NP caps together in parallel (John's suggestion), I'll guess it will be a couple of days until I can get a single 47uf cap Seems like I read somewhere that you can put two polarized 47uf caps back to back to make one 47uf NP cap - is that correct? Seems like there would be an issue doing that...

I can try Stan's suggestion later this evening cuz I DO have a bag o' LEDS! I think I'll temporarily turn my little switcher into a test bed to make these changes easier.

As an aside, I did just try another e-unit and got exactly the same results, so that doesn't seem to be the root cause...

GeoPeg posted:

After the above suggestions I added a 1uf 50v NP electrolytic cap across the motor leads - no change. Added another one across the AC input leads to the e-unit - no change. 

Did you try adding the 1uF NP across just the LED pair? 

If adding a 2nd pair of LEDs improves matters but not enough, add a 1uF NP across the LEDs as shown in the diagram.

stan2004 posted:

GeoPeg posted:

After the above suggestions I added a 1uf 50v NP electrolytic cap across the motor leads - no change. Added another one across the AC input leads to the e-unit - no change. 

Did you try adding the 1uF NP across just the LED pair? 

If adding a 2nd pair of LEDs improves matters but not enough, add a 1uF NP across the LEDs as shown in the diagram.

 

No, I added the cap across the motor leads and then added another one across the AC input - have not tried across the LEDS yet, but will in a short bit. I had a 9 year old come up to me about 20 min ago and beg me to play some Minecraft with him. I only have this short intermission to respond  Now....back to killing Slime!!!

I think perhaps I misread the original problem, I was thinking one direction was flickering when it was supposed to be on.  Perhaps ignore my suggestions and stick with the guy with better reading comprehension, that would be Stan!   I really need to slow down and read!

Stan - I tried your solution - I first replaced the LEDs with some 2mm yellow, just to get away from that particular model of LED, thinking maybe it was too sensitive to back currents. I then installed only two LEDs, to duplicate my first test - same result. Then I added the 2nd back-to-back pair of LEDs (now 4 total) in series with the first pair - same result. Finally I added the highest value NP cap I had laying around, a 3.3uf 50v NP. I mistakenly installed it across the DC output/motor instead of the way you had shown - same result. When I spotted my error, I resoldered the cap, one lead to the top wire (as you have drawn), the other to the junction of the resistor and the bottom pair of LEDs - worked like a charm!!

Never one to Rube Goldberg an installation, I thought I would enclose a pic of the professional wiring I did on this engine. All possible exposed wires have been carefully isolated by proper bending to avoid other wires, bare metal, and bouncing on the little circular track of warped O27. Painters masking tape was applied where needed. No LEDs were harmed in making this test!

Hey, it worked!!

I then removed the 2nd pair of LEDs and rewired the cap to be directly across the first pair - that also worked great. revised schematic below:

Finally, I removed the 3.3uf cap and replaced with a 1uf cap (got plenty o' those!) but that wasn't quite enough, some flickering started to reappear. So your theories were correct, and the problem has been solved. 

I'd like to thank Stan, John and Adriatic for diving in with yet another helping hand - Awesome job!

George

GeoPeg posted:

Hmmm... I was hoping to finish this today, ...

And so you did - be careful what you wish for! 

I'm impressed you had NP caps lying around.

stan2004 posted:

GeoPeg posted:

Hmmm... I was hoping to finish this today, ...

I'm impressed you had NP caps lying around.

If you could have seen me as a kid, riding my bike to the local landfill (we called 'em dumps back in the day!) and hauling home an entire TV chassis balancing on my handlebars and basket - just so I could strip it for parts to further my electronic tinkering … well, you would understand why I might have some odd (and old) parts lying about!  I also haunted an electronic surplus (mostly parts) store in Minneapolis during those years.

Can't even begin to imagine where I got those NP caps from, but I will be ordering fresh ones before I seal things up!

George, most of the thanks goes to Stan, he's better at reading comprehension than I am.

gunrunnerjohn posted:

George, most of the thanks goes to Stan, he's better at reading comprehension than I am.

You're preaching to the choir on that point. I re-read everything I write several times before hitting the "send" button, and find all kinds of typos, punctuation, grammar and other errors. Days later, when I read what I originally wrote, there have been times I couldn't even understand what I was trying to say! Just part of MY human experience, I reckon 

Completed repair/upgrade to the little Commando engine. I ended up adding two 3.3uf NP caps in parallel to eliminate the flickering - works flawlessly as you can see here - just too impatient to wait for new caps from China. Thanks again!