Lionel GP7 LED upgrade: check my circuits

can we see a pic of this geep?

Is this an early GP7 with the early TMCC electronics?

As I recall, the lights where controlled by triacs, and the triac was fired for only half of the time, and the it was random whether you are going to get a positive or negative pulse. Then there was another issues of might needing a non-polarized cap to make sure the triac really fired. John Z. knew the details.

Not sure about what age this is.  No paper work.  It has a R2LC board, train sounds, and working couplers.  Rock Island 1274.  Only one motor, which makes it a pretty bad puller.  I got it as part of a large layout package that I bought.  Pics show it working on the bench.

I don't try to run four white LED's in series, as for TMCC, the output of the lighting outputs is a half-wave of the track power.  So, you end up with about 12-13 volts peak voltage, not much more than the operating voltage of the four series LED's.

I'm not what the eight LED's in the middle are, there seems to be four more than you are talking about.  The way the red/green ones are wired, they'll only ever see one polarity in command mode, so you'll only ever get one state.  Also, the 2k resistors at either end limit the current for those LED's to a fraction of what you should be using, small wonder they're not brighter.

I'd suggest a few changes.

Since the smoke output when programmed for a cab light reverse in polarity for direction changes, it would be a bunch easier to put the red/green markers onto that output, then you'll get automatic direction changes for free.  You can wire all of them in series with a single resistor for current limiting, I suggest around 470 ohms.  If you wire them in the proper polarity, you'll have green going forward in front, and red going backwards.  The rear will be opposite.

Wire the front and back number boards with diode isolation to both the front and rear outputs.  This gives you number boards lit all the time the lights are on, but you can still turn them off with AUX2.

Wire the two headlights for the front and rear respectively to the proper outputs with a 470 ohm series resistor.

I used Digi-Key's Schemeit circuit software to make the drawing.  There is no "bicolor LED" symbol so each pairing on the diagram is one LED, one side representing red and other green.  Since they are wired in series, only 4 will ever light up.  I used the 2k resistors as when I used 470's, it would work on the bench fine.  But once the throttle was turned up, the side that was supposed to be off would start to glow.  The 2k's stopped that. 

I'd prefer to leave the cab light set as 'strobe'.  It's hokey, but kids love it. 

The numberboard idea is brilliant.  I will try that.

Why not run four in series?  I've seen reference to six being the magic number.  If it's a question of brightness, I'm fine with the output that I'm currently getting. 

Remember that this setup (seems) to be working fine.  I've tested it for about a half an hour on the bench with no problems that I can perceive.  My concern is that what I am proposing will not damage any other components down the road like the triacs that trigger the lights.

You won't damage anything with the LED's.  The reason I suggest fewer in series is simply because for a lower voltage in command mode, say 15-16 volts, the half-wave signal may not light the four in series.  I haven't tried it, I prefer to have margins for operation when I do something like this.

I also add diodes to all my LED installations for TMCC as LED's don't tolerate reverse voltage all that well.  Obviously, the bi-color ones don't need them, but the others do.

I'm curious how you got the positive side separated with the lighting outputs.  In command mode, the TMCC lighting outputs are negative in respect to frame ground, so you have to have the lights wired that way.  For conventional mode, the triacs pass the full wave power so you won't notice if things are backwards until you run in command mode, then none of the headlights or number boards are going to light.

The way Lionel and Atlas do some of these is with communication lines between the front and rear light boards.  They also use parallel headlight LED with each having it's own limiting resistor.  That is another way to get around the amount of LEDs you want.

As an example, you power your headlights as you are doing (you state they work).  The front headlight input can also feed the front LED Green return to ground.  When it goes to reverse the green goes out and then via feed wire to the rear light setup your rear headlight return to ground, also feeds the red portion of the front markers.  Now the front led will turn red and the rear fed from the rear light circuit will be green, and vice versa with a feed for Rear red to the front lights. 

G

When I began, I had the unit on the bench in command mode.  With the original bulbs in place, I would measure the voltage to each with a meter while lit.  I'd then mark the + side.  I wired up the headlights and numberboards and everything worked fine.  When I began to attempt to add the red/green is when I measured the negative from the front and the negative from the rear and found that it was passing around 10v and flipping negative and positive as I hit the direction button.  That is when I got the idea for the two lead bicolors.  I've never played with the guts of this engine and don't think the original owner did either.  One thing that puzzled me, was that during the initial wiring, I could not ground any of the bulbs to the chassis or frame or even the outside rails.  I thought that ground went through everything.  But I guess that is for the AC current only.  I'm going to continue to play around and try out some of both GGG's and gunrunnerjohn's suggestions and I'll post as I get results.

The other thought is what type of bicolor leds did you get, common cathode or common anode?  It could be why you have an issue with the markers.

I usually trace out the light board to figure out what the Manuf was doing and go from there.  Seems like TMCC based platforms use parallel wired LEDs with individual resistors, while MTH is primarily Series wired for the PS-2 engines.   Some early MTH stuff used parallel wired LEDs. G

If they're two-lead bi-color LED's, there is no "common anode" or "common cathode" designation.

Yep, missed that he had 2 lead LEDs.  G