MTH Rail King Halloween Caboose Electric Short

Thanks for posting this.

I'm worried on the diagnosis of the root problem.

Here's my theory of what "went down"- the LED string lights are using enamel coated magnet wire and held to the body of the car using tiny metal wire staples. That said, a more direct cause would be anywhere the string LED wires can contact the metal frame of the caboose- often tucked in under the shell.

Ask yourself- why did the rectifier get hot? The answer is- if the load shorts (example a staple in the plastic body shell only shorts the LED string to itself) but worse, a string to body frame short is several times the current and voltage instantly heating that rectifier to the blow it sky high point.

In other words- the design flaw is NOT the regulator, not the wiring of the truck power wires. No, the flaw is- the string lights are fragile and nothing more than a super thin layer of enamel is between that wire and several places of metal contact. Be it the wire staples holding it to the shell that rub and eventually wear through the micro thin coating on the wire, or the folded and tucked places where the wire is sandwiched between the edge of the metal truck frame (at AC common potential) and the plastic shell, VS the string wires at a different DC potential.

The "fix" I use on all my LED string light based cars and items is to use tiny drops of UV resin cured glue to prevent movent and thus rubbing at any of the staple points. I then also apply heatshrink tubing over anywhere the string lights tuck under the shell as an extra layer of insulation over the wires to prevent any chance of just such a string LED to frame short.

Again, a drop of resin glue over each staple joint, let it soak in, then cure with UV, prevents the rubbing style short at that location over time and usage. But the far more damaging short is the string wires to frame ground short that just blows that poor regulator sky high.

@jjwyatt posted:

1.) Take the initiative to take your caboose apart prior to running to ensure the board is not bend downwards onto the metal frame and/or wires are not also compressed between the metal frame and the board.

Again, please don't take what I have said to invalidate your point. Yes, you should inspect for obvious pinched wires and other flaws. I agree, wires jammed under the board was not good. That said, the failure appeared to be more the rectifier got smoking hot, and melted into the wire as an after effect.

Having done more repairs on this type of string LED car than I can count, the more common and sinister fault is that super thin enamel coated magnet wire the LED string is made of, that is intentionally for cosmetic reasons, folded around the edge of the plastic shell and then often sandwiched between the metal frame which is at AC common potential (same as outside rail/wheels). Again, the typical rub through at a staple joint will cause flickering of the LEDs but most times is not as brute force damaging as a frame to LED string short. Obviously any short is bad, but there is bad (staple short), and then whole new level of bad (frame short).

Some cars, like the Christmas LED string boxcars- this LED string wiring starts inside the shell at one end, edges past the metal frame in the corner, up and long the corner of the shell, then down the roof edge then the worst of the worst is down the corner at the car end and the cut off tail of the wire is tucked up again- another potential frame short. On the boxcar, that's 4 potential frame to wire short- along with each of the staples being another short point that can cause flickering or color- change- blocking one color but allowing others (again, multicolor string LED specific fault).

@Vernon Barry posted:

Again, please don't take what I have said to invalidate your point. Yes, you should inspect for obvious pinched wires and other flaws. I agree, wires jammed under the board was not good. That said, the failure appeared to be more the rectifier got smoking hot, and melted into the wire as an after effect.

Having done more repairs on this type of string LED car than I can count, the more common and sinister fault is that super thin enamel coated magnet wire the LED string is made of, that is intentionally for cosmetic reasons, folded around the edge of the plastic shell and then often sandwiched between the metal frame which is at AC common potential (same as outside rail/wheels). Again, the typical rub through at a staple joint will cause flickering of the LEDs but most times is not as brute force damaging as a frame to LED string short. Obviously any short is bad, but there is bad (staple short), and then whole new level of bad (frame short).

Some cars, like the Christmas LED string boxcars- this LED string wiring starts inside the shell at one end, edges past the metal frame in the corner, up and long the corner of the shell, then down the roof edge then the worst of the worst is down the corner at the car end and the cut off tail of the wire is tucked up again- another potential frame short. On the boxcar, that's 4 potential frame to wire short- along with each of the staples being another short point that can cause flickering or color- change- blocking one color but allowing others (again, multicolor string LED specific fault).

Appreciate the response as I may have missed that.   I just assumed the bottom of the board being in contact with the metal frame and the damaged wire that fed from the truck was the cause.   I am not an expert but assumed that ultimately I thought the bottom of the board in contact with the metal frame was the issue.   did test the LED wires separately and was unable to detect any issue as noted. Think previously above I did not articulate that as well with the pinched wire.  That was due to the board in contact with frame.      

FWIW, the basic formula for almost all of these string LED cars is a 6V regulator, and typical 150 Ohm resistor in series to limit the current.

I did manage to find your board AG6200007 https://www.mthpartsandsales.com

Interestingly enough, this weekend at a trainshow I had a failure of the same switching regulator in an MTH lighted Christmas car with the snowmen. I also believe I know what caused it, a user was borrowing my demonstration track (2 30 inch sections of straight fastrack) while testing a known faulted pullmor motor based older Lionel engine with a flaky E-unit. They kept cycling the E unit over and over using the direction button on the Z controller and Z500 brick powering the track- all while the lighted car was also on the same track. So I know with some certainty decent voltage spikes were being created at the high inductance of the motor and active mechanical E-unit were being rapidly energized and circuit opened.

It blew the tiny C7WH1 (labeled C4T19 on the new board) SOT23-6 switching regulator. Some searching I found a Chinese data sheet says that's a 1.5A capable 36V max input.

That seems logical, a voltage spike might not kill the 35V capacitor right away, but a tiny switching regulator might easily be taken out by a voltage spike.

Just saying-TVS might not be a bad idea on your track or somewhere in your power system.

Also, reviewing this, there are very minor differences between the original regulator in the beginning of the topic an the replacements. I have a theory on that. The original unit in that Halloween one has what appears to be an additional diode and the bridge rectifier is what cooked. I have no way of knowing without that exact version of the board in front of me but I suspect that could be a TVS diode just after the bridge rectifier on the DC side of the circuit. If a similar voltage spike then blows the TVS shorted, then sure, the square bridge rectifier now is slammed with near short circuit current and quickly overheats and fails. That said, in theory- the regulator was saved????

What I'm getting at is the replacement part number AG6200007 https://www.mthpartsandsales.com

The one that failed in the Christmas car and the replacement part does not have that same diode in near the middle of the board. It just has the bridge rectifier and then the SOT23-6 regulator and a few SMD capacitors and resistors.

One of those catch22 design situations. They put in a TVS but after the rectifier and that kills the rectifier in failure but saves the regulator- but still a bad result.

Alternatively, putting a bi-directional TVS in front of the entire circuit on the AC side might be good, but a failure of the TVS would then cook the wiring and pickups.

I'm just sharing thoughts after seeing 2 failures, and what appears to be minor design changes and trying to give an understanding of what I think caused the failure (high voltage spikes higher than 36V) created by something else on the track likely with some inductance.

EDIT- also, those same spikes are not good for anything else on the track with electronics either.............. including my Z controller when this failure happened in my case.

Thank you Mr. Barry for the deep dive on holiday LED lighted cars.

I just fixed a half a dozen of these cars at MTH.  The strings are glued on and I use a gelled super glue and an accelerator to speed up installation.  This new run of LEDs is a single string of all 4 colors, where previous versions were two strings of dual LED braided together.  The area where the LEDs pass over the frame is notched and can have a sharp edge on it.  I file it down a use a bit of tape over the notch where the LEDs pass through.  
The bridge rectifier that failed may have been caused by sloppy soldering on the PCB where the legs of the voltage regulator bridged together (I had one at MTH that had the same failure.).