PW Searchlight car improvement?

I have the 3520 Rotating Searchlight Car which is similar to the earlier 6520 searchlight car. It puts out a bright beam above 10 volts. I don't know if the 3520 is inherently brighter but you might look for light bulbs that have a brighter rating at the voltages you unusually apply.

Wow.  Those LEDs look like they will be really bright.  Thanks for the pointer.  I'm going to get some and see if I can work them into the searchlight car.  In my version of the car, the searchlist does not rotate, so I don't think I'll have any wiring problems.  I can probably squeeze a rectifier and resistor underneath the car or maybe even hidden under the plastic generator.

I guess I'll have to wait and try it, but I hope the LEDs isn't so bright as to be a problem with blinding someone as the car travels around the track and the searchlight points in different directions!

I tried the half wave rectification, but the LED flickered without a good size capacitor, and I wasn't sure I could find room under the car for a big capacitor too.

Good point about the shrink tubing on the leads -- I've got the shrink tubing and should have done that.  If I have to fiddle with it again, I'll add it.

I wish I had an oscilliscope.   My track power is a commercial transformer through a TPC 300 set wide open, so the input to the searchlight car should be a good sine wave.  But it would interesting to see the CL6 output. 

This is great i got the new Bethlehem searchlight car for ****.  I was thinking about doing an led in there just for a little project and the possible advantage, like stated, a more "spot light effect".  Thanx

Well, bad news on the LED addition to my searchlight car.  The CL6 constant current chip doesn't like something about the way I'm using it, and after some usage it partially fails such that it is still passing current to the LED, but it doesn't regulate it.   The current becomes somewhat proportional to the input voltage, and at 18v, it's passing 200+ milliamps.  Not unexpectedly the CL6 get *really* hot, and the 100ma LED eventually burns out.

I originally bought two CL6s, and both have suffered the same fate.  For the second one, I used a pretty decent heat sink (complete with heat conductive paste) thinking the problem might have been heat related, but that didn't help.  The second one failed the same way as the first one.

I'm guessing that the problem is the rectified but unfiltered input voltage.   The CL6 is designed for DC input, and perhaps it doesn't like chasing the sine-wave like input voltage.  Or it doesn't like the extra noise from the pickup roller when rolling down the track, with occasionally interruptions on turnouts.

The next time I place a Mouser order, maybe I'll get a couple more CL6s and try a capacitor on the input side to smooth out the input voltage.  In the meantime, I'll probably have to just go back to some resistors that I can get locally.  I run mostly with TMCC so usually have constant track voltage.

I think you probably need a TVS diode on the input track voltage.  My guess is that voltage spikes are killing the CL2.

Did you ever solve the issue of the CL6 going dead? 

You may need either the TVS I mentioned previously, or more likely you need a heatsink.  Remember, the CL6 has to drop the full wave rectified voltage all the way down to 3V for the LED!  So, if you are getting like 25 volts DC after the bridge, you are dropping 22 volts at 100MA, or 2.2 watts!  That's way more than the power handling capability of the CL6 without a heatsink!  A TO-220 package that is not heatsinked typically dissipates around one watt of heat, at a temperature 62.5°C higher than the ambient temperature.

One easy fix is to use a single diode for command operation so that you have much lower DC voltages to start with.

John

Without a filtered DC input,I also wonder if the CL6 is hunting. perhaps the back of the chip could be mounted against some metal part. 2 watts is a lot to shed but in conventional it would be less as input voltage is less..

Dale H

I didn't take notice there was no cap, but I suspect with no filter cap you're working the CL6 pretty hard, that's another issue.  However, for command with a bridge, I think it would be mandatory to have a heatsink on that chip in any case.  Even with just a diode and cap I think you'd need the heatsink.  Running the CL6 at 60°C above ambient is getting close to it's shutdown limit, and that would be if you only dissipated one watt in the device!