Help with 3V to HVac converter

[Paolo Cravero]

Sirs,

on the subject of power supplies, I need to fix a circuit that drives modern "EL wire" (or "neon wire") that works at about 350Vac from 2 AA 1.5V cells.

The given PSU, of which I have drawn a partial circuit (attached) probably of the HV conversion stopped working after one hot month break and no more than 10 minutes total use. The effect is that luminescent areas start with much weaker intensity and in a couple of seconds they become invisibile. It is like the PSU has lost its ability to deliver enough energy. HV oscillation is present, confirmed on the o'scope.

With a different PSU I can drive luminescent areas individually, so they are confirmed to be working. I need to fix the original circuit because it also makes some light effects with the 12 areas.

I would appreciate any suggestion to what I should check.

While playing with the second PSU I realised that this "EL wire" can be used to model giant "Nixie" tubes without the vacuum/gas requirement. See the second picture with a figure "6" approximately 1 x 0.5 metres made with 3 metres of wire. EL wire costs about 1 USD/metre (without driver).

Help, if possible.

Paolo

[gregebert]

The temperature problem is a good clue; there must be a larger transistor not shown in the schematic that is the actual driver, and I suspect it's overheating. Does it have a heatsink, and does it get warm/hot ?

Does the supply work correctly when illuminating shorter segments of EL wire ?

The transistors shown in the schematic are probably pre-drivers. They should continue to operate correctly even as the EL wire starts to dim. You can confirm this with your scope.

What frequency/voltage are you seeing on the HV ?

[Paolo Cravero]

The temperature problem is a good clue; there must be a larger transistor not shown in the schematic that is the actual driver, and I suspect it's overheating. Does it have a heatsink, and does it get warm/hot ?

One moment. Heat was in the environment while the whole thing was at rest: we were away on holidays and the circuit was without batteries! Nevertheless there is no larger transistor, just one SMD "3D" (BC856? PNP) for each "EL line" (12 in total). Their emitters all tied to ground. The circuit has three groups of 4 electroluminescent areas (that's more a flat panel than wire, but it is all hidden out of sight in the commercial product).

The "common out" is 350 Vpk sine with 400 us period (2500 Hz). Those 12x PNPs collectors pull "cathodes" to ground one at a time. The "common out" doesn't weaken or change frequency.

 

Does the supply work correctly when illuminating shorter segments of EL wire ?

No. And it looks it is getting worse and worse the more I keep it running with my tests!

 

The transistors shown in the schematic are probably pre-drivers. They should continue to operate correctly even as the EL wire starts to dim. You can confirm this with your scope.

Confirmed.

Unfortunately I do not have other HV AC equipment, so I am stuck. All I can do is replace the SS8550 and M28S in case they have failed for some reason, but I am mostly sure that is not the solution.

My daughter won't be happy, neither will I.

Paolo

[gregebert]

Well, the BPC856 datasheet says it's rated for 80V, so if it's running at 350V then it's not suitable for this design.

I'd suggest looking for another PNP that can handle ~400V. Digikey has a 400V device that has an SOT-23 footprint (ZXTP08400BFFTA), so that's worth looking into.

I doubt there's anything wrong with the pre-drivers, but you may want to poke around with your scope to make sure none of those devices get stressed more than 40V across their terminals, particularly collector-emitter. If they are stressed beyond that, then they should be replaced with a more robust device.

Since your device was off during the hot weather, it wouldn't be affected. Storage temp ratings are often 150 C for solid-state devices, so they are about the last thing that would get cooked. The case would melt first.