GRJ's Original DIY Constant Current PAX Car Lighting Module Files

The capacitor is simply a storage device, the larger the capacitance, the more energy it can store.  The intensity of the lighting (and thus the current draw) will determine how long the energy in the cap lasts to keep the lights on.

Truthfully, I really have not had issues with the 470uf value of the cap, even at slow speeds over switches.  If you're seeing issues, I'd be checking the integrety of the pickup rollers and wheel axle wipers that supply the power.  Normally with passenger cars over switches, at least one roller should be on the powered track, and any power interruptions would be very small.  What kind of switches are you seeing this effect on?

The flicker is slight and only over old Gargraves switches. I have been replacing these with Ross and no problem at all with Ross.

Don't have a lot of experience with Gargraves switches, so I can't say what that would be.  My layout is all Ross switches and Ross/Gargraves track.

@gunrunnerjohn posted:

The amount of storage and the current setting directly affect the hold-up time for the lighting.  The dimmer you run them, the longer they'll last during power interruptions.

Here's a 1,000uf cap that  will drop in and double the capacitance.  It is taller than the 470uf cap, but the same lead spacing and diameter.

Rubycon 35PK1000MEFCT810X20

Today, I got around to building a PCB with the 1,000 uf cap. But I had a frustrating afternoon. Could not get it to work and wonder if I messed up somewhere. I know it's a long shot, but can you tell from the photo where I went wrong?

What happens when I input 17.5 - 18 V AC: I get about 6 - 6.5 V DC out, measured at the output terminals. This output won't light my strip of LEDs (3528 ribbon style, rated for 12V DC.) Turning the pot makes no difference. If I apply DC current to the same LED strip directly, ie not through the board, it lights just fine, even at 6 V DC, although it is fairly dim at that voltage. Should I swap out the 1000 uf cap for the 470 uf and see if that works?

(I checked and re-checked everything I could think of for over an hour. Ready to jump off a bridge but my wife poured me a big glass of Malbec and now I am feeling better.)

Bob

Bob; it's doubtful the cap is the problem IMO. Those boards have all worked perfectly for me, even with 1000 uF caps installed. I would bet the problem is the 317T since everything else looks good. From the pix the center pin soldering looks a bit suspicious, though from the bottom it might be just fine. I would recheck that for sure. If that doesn't solve the problem i would cut out the 317 and put in a new one. They are pretty cheap after all. Several months back another gentleman was building a few of these and had some troubles. It turned out the pack of bulk 317's he bought had several defective ones. Seems hard to believe I know, it sure seemed odd to me. Don't think I have ever had a bad one. Worth checking.

Rod

The cap should make no difference in the output current.  That is a LM317 regulator, right?  It looks OK visually, but I can assure you that many people have built the same board and are using them, so something is amiss.

What voltage to you get measuring directly across the capacitor at the two arrows?

@Rod Stewart posted:

Bob; it's doubtful the cap is the problem IMO. Those boards have all worked perfectly for me, even with 1000 uF caps installed. I would bet the problem is the 317T since everything else looks good. From the pix the center pin soldering looks a bit suspicious, though from the bottom it might be just fine. I would recheck that for sure. If that doesn't solve the problem i would cut out the 317 and put in a new one. They are pretty cheap after all. Several months back another gentleman was building a few of these and had some troubles. It turned out the pack of bulk 317's he bought had several defective ones. Seems hard to believe I know, it sure seemed odd to me. Don't think I have ever had a bad one. Worth checking.

Rod

Rod,

Here is the bottom of the PCB. The regulator is from DigiKey, LM317TG. I am a novice at this type of soldering. However, I have built about a dozen of these, using the 470uf and they are all installed and working flawlessly. Nevertheless, I am no expert!

@gunrunnerjohn posted:

The cap should make no difference in the output current.  That is a LM317 regulator, right?  It looks OK visually, but I can assure you that many people have built the same board and are using them, so something is amiss.

What voltage to you get measuring directly across the capacitor at the two arrows?

John,

The voltage across the cap, with 17.5 V AC in, is 2.3V DC. I have no idea if that is good or bad.

Thanks to both for your prompt replies.

Bob

Bob; that voltage across C1 is way out of whack. It should be about 24VDC with 17.5 VAC supply. The cap may be shorted out or something like that. For the heck of it why not put in one of your 470 uF caps and see what happens?

And your soldering on the 317 looks just fine BTW.

Rod

Bob; is the C1 cap getting hot to the touch? What about the bridge rectifier?

Rod

@Rod Stewart posted:

Bob; is the C1 cap getting hot to the touch? What about the bridge rectifier?

Rod

I don't think any of the components got hot, but I did not specifically check. Will do, but I think I will wait until tomorrow before doing any further investigation. More later.

Thanks again for helping.

Bob

The voltage across the cap should be in the mid 20's, so that's what to look at!

I wonder if the cap is actually defective?  I'd pull it and put one of the 470uf ones in and see if it works.

@Rod Stewart posted:

For the heck of it why not put in one of your 470 uF caps and see what happens?

Rod

@gunrunnerjohn posted:

I wonder if the cap is actually defective?  I'd pull it and put one of the 470uf ones in and see if it works.

Rod and John,

I started from scratch and built a new board with all new components, paying attention to my soldering and using a 1000uF cap again. No luck, didn't work. So I pulled it and put in a 470uF cap and everything works fine!!

Go figure. I have lots of the 470 caps, so I will stick to those for now.

Thanks for helping.

Bob

Obviously, something is going on with the caps, are they 35V or greater rating?  If you got them from my link on Digikey, I'd be surprised if they're bad!

Bob, that's very strange. Glad it's solved though. I wonder if you got a bad batch of 1000UF caps?? If they are Digikey you would have some recourse to get them replaced. If they are offshore, hmmm, good luck!

Rod

I've bought thousands of off-shore components, I rarely ever get a bad one, and I don't recall ever getting a whole bad batch.

@gunrunnerjohn posted:

Obviously, something is going on with the caps, are they 35V or greater rating?  If you got them from my link on Digikey, I'd be surprised if they're bad!

@Rod Stewart posted:

Bob, that's very strange. Glad it's solved though. I wonder if you got a bad batch of 1000UF caps?? If they are Digikey you would have some recourse to get them replaced. If they are offshore, hmmm, good luck!

Rod

They are Digi-Key and yes, John, I used your link. I ordered only 10 of them, so no big expense. Not sure I could provide any evidence to Digi-Key that they are faulty. It is much more likely that I did something incorrectly, twice, but no idea what.

No doubt I will try and use the 1000uF caps in the future but going to give it a break for now. More later if the third attempt succeeds. With your combined experience, Rod and John, that these are reliable devices, I am satisfied that it must be mysterious human (mine) error.

Appreciate your time on this.

Bob

They looked to be installed in the correct orientation, not much else you can do to those caps to cook them in that circuit.  I looked at the picture again, and the only thing that comes to mind is you connected the AC to the wrong connector.  Noting that you used the same style connectors for both AC and DC, if you connected AC to the DC side, you would get some very funny readings!

FWIW, I always use different connectors for power and signals on anything I build, that way it's not possible to connect the power to the wrong place.

Bob; my curiosity is burning; have you tried swapping out one of the 1000 uF caps for a 470 in either of the failed boards??

Rod

@Rod Stewart posted:

Bob; my curiosity is burning; have you tried swapping out one of the 1000 uF caps for a 470 in either of the failed boards??

Rod

Rod, yes - my second build was with the 1000uF cap and when it did not work, I de-soldered it, removed and replaced it with a 400uF cap. The operation was then perfect.

@gunrunnerjohn posted:

They looked to be installed in the correct orientation, not much else you can do to those caps to cook them in that circuit.  I looked at the picture again, and the only thing that comes to mind is you connected the AC to the wrong connector.  Noting that you used the same style connectors for both AC and DC, if you connected AC to the DC side, you would get some very funny readings!

FWIW, I always use different connectors for power and signals on anything I build, that way it's not possible to connect the power to the wrong place.

John, good tip on using different connectors. To avoid that blunder of mixing up the AC input with the DC output, I checked and re-checked with the markings on the spare PCB, as seen in the first photo.

Both you guys are awesome in the interest taken in my problem. That's what makes this Forum so valuable!

Bob

I'm stunned that a 1,000uf cap fails and a 470uf cap works, that makes no sense!  I've actually tested that circuit with a 2,200uf cap to see how long it would hold the LED's up, and it didn't change the operation, and it never even occurred to me to think it might!

FWIW, my previous production run of the lighting boards had a 680uf cap, that worked just fine.  That was going to be my standard going forward, but I couldn't get them for the latest production.  The electronic parts shortage is biting everywhere.  I had to settle for a 470uf cap again, the largest value I could get in the form factor for the production board.

Bob, I too am shocked that the cap switch solved the issue. Clearly the smoking gun is the 1000uF caps, for whatever reason. It would be interesting to see what they show on a DMM that has a cap tester.

Rod

@gunrunnerjohn posted:

I'm stunned that a 1,000uf cap fails and a 470uf cap works

@Rod Stewart posted:

Bob, I too am shocked that the cap switch solved the issue. Clearly the smoking gun is the 1000uF cap

I couldn't stop thinking about this. So I built another board but did not solder in the cap. Instead, I inserted first the 470uF and bent the legs to get good contact, then applied power - working. Then I put in the 1000uF cap and bent the legs and it worked too!!

So you are both right - nothing wrong with the caps. I must have erred somewhere, twice, when soldering the components on both the first 2 boards. Reminds me of the saying, "I learn from my mistakes and can repeat them exactly."

I will test all the caps that I have left on this temporary board and maybe try to test the first caps that got blamed for the failure. I nipped off the legs to de-solder them but will cobble something up.

And the decay in the lighting after power off is noticeably longer with the 1000uF caps. Happy camper here. Thanks to both.

Bob

Well, that restores my faith in electronics, I couldn't figure out how the cap caused this!

Whew; now I'll be able to sleep! Thank you Bob.

Rod

@Bob "O" posted:

And the decay in the lighting after power off is noticeably longer with the 1000uF caps. Happy camper here. Thanks to both.

Obviously that would be the case when everything is working.  FWIW, I found that even my original 330uf cap for my lighting modules was sufficient to kill almost any flicker.  I increased it to 470, and briefly to 680 for the production runs.  With the component shortages, I can no longer get the 680 caps, so I'm back to the 470.

Quick question on this subject.  If I have a strip of led lights, and I just hook it up right to a 9 volt dc battery, will it lit up without any other circuitry?  I have a Halloween display item I would like to light up but not much time to build anything fancy electronically, but I could certainly find time to quick wire up a battery and a switch.

jhz; yes they will light up on 9vdc as long as they are 12vdc led strips. The design voltage should be stamped right on the strip. They will be at maybe 2/3 of full intensity. Nothing to lose, give it a try.

Rod

I hasten to add, with the ampacity of the typical 9V Alkaline being 400 MAH, if you run have 40ma of draw, they'll only last 7-8 hours.   A simple variable DC supply like this 120VAC->9VDC Module at Digikey for $4.48 would be a much better choice.

Datasheet for LD05-23B09R2

@gunrunnerjohn posted:

I hasten to add, with the ampacity of the typical 9V Alkaline being 400 MAH, if you run have 40ma of draw, they'll only last 7-8 hours.   A simple variable DC supply like this 120VAC->9VDC Module at Digikey for $4.48 would be a much better choice.

Datasheet for LD05-23B09R2

Thanks John,  thanks @Rod Stewart.  This is for a display that happens 1 night year for about 3 hours.  2/3 intensity for 8 hours is perfect!

OK, so I bought this strip of led lights years ago and never did anything with them so I am a novice here.   The end of the string has leads on it. Do I just cut off what I don't want?  I assume those pads are for soldering new leads onto subsequent sections,  is that correct?

OK you can cut on the red line below at any place you see a pair of oblong copper pads. Each 2" section of the strip contains 3 leds and one resistor (the black thing) in a small series circuit. You have to cut only at the pads, in 2" increments.

So lets say you want 10 sections, 30 leds; just cut at the 10th pair of pads. Each section of 3 leds will draw about 3-4 ma of power, so judge accordingly. And yes you can solder wires onto those pads. Observe polarity when hooking up the battery.

Rod

So I have to cut right thru the pads?

That's the idea. Later versions of led strips have included cut lines shown right on the tape. Just use scissors.

Rod

@Rod Stewart posted:

I took the liberty of moving grj's original files for this project over to a new topic just to make them easier to find. The original topic started in late 2016 and is some 7 pages long; these files are found on page 5.

The only thing I have added to this is my user notes in Word format as an attachment. The gerber zip file is also attached below. I want to emphasize that grj is the sole creator of this circuit and these original files!

That you for your help & information however I downloaded your notes & printed out & then downloaded the 7 pages of Gerber files & tried to print but no program in my new computer would allow me to open & printed these pages. Wondering what prgram you use to open & print these files. Your help is greatly appreciated

You don't print the Gerber files, they get send in that ZIP file to the PCB manufacturer of your choice to make the PC boards.  JLCPCB is one of the places I use, they generally do a good job.

Another option is OshPark domestically. Their price per board is higher, but if you only need a few that would work. I beleive their price is about $1.25 a board in batches of 3. Again as grj says you upload the gerber zip file to their site and go from there.

If you only need 10 or 20 boards I can mail them to you inexpensively. Shoot me an email if interested. I can also help with placing a board order for you for delivery to your address if needed.

Rod

I hope I don't cause any ill feelings but I was looking at a low cost alternative for installing LED strip lighting and came across this video on YouTube:

How To Add LEDs To Lionel MPC Passenger Cars

His "design" uses just capacitor and voltage regulator. I'd appreciate any thoughts on this approach.

Jon

@ffjon posted:

I hope I don't cause any ill feelings but I was looking at a low cost alternative for installing LED strip lighting and came across this video on YouTube:

How To Add LEDs To Lionel MPC Passenger Cars

His "design" uses just capacitor and voltage regulator. I'd appreciate any thoughts on this approach.

No hard feelings here, you're free to use whatever technique you find useful.

A couple of points that you don't get with this method.  The first is intensity adjustment, and there is also the lack of DCS compatibility.  Those cars with the huge cap right after a full-wave bridge will hammer the MTH DCS track signal.

@ffjon posted:

I hope I don't cause any ill feelings but I was looking at a low cost alternative for installing LED strip lighting and came across this video on YouTube:

How To Add LEDs To Lionel MPC Passenger Cars

His "design" uses just capacitor and voltage regulator. I'd appreciate any thoughts on this approach.

Jon

Very interesting but I like the idea of the adjustable pots on gunrunner Johns site, but that was very good to know. Thank you for your help.

@ffjon posted:

...

His "design" uses just capacitor and voltage regulator. I'd appreciate any thoughts on this approach.

By my reckoning, your linked youtube video shows "just" a capacitor and a bridge-rectifier (not a voltage regulator).

To each his own but I think the photo (clipped from your video link) is RIDiCULOUSLY bright!  GRJ's module allows brightness control...amongst many other benefits.

Full disclosure, I have absolutely no business relationship to GRJ's goings ons and arguably we disagree here on OGR more than we agree!   Also, you get support from fellow OGR participants with his module which is absolutely priceless IMHO.