Repair Vivitar 3500 speedlight

[Peabody]

I have three Vivitar 3500 flashes dating back to the 1980s.  When you switch them on, you should hear that rising whine of some kind of charge pump charging the main capacitor.  One of my flashes does that.  The second one does that only after I switch it on and off several times, but then it works perfectly after that as long as it stays on.  The third one used to require many more attempts, but now doesn't come up at all - no whine at all,  and it starts to get warm while I'm waiting.

I use these as manual flashes using a radio trigger, and they work fine for my needs.  A replacement wouldn't cost that much, but if I can fix this one, I'd like to do that - particularly since the middle flash apears to be headed in the same direction, and might have the same fix.  I've taken the bad one apart, and it has two boards (see pics below) - one I call the power board down by the battery compartment, and the other I call the flash board which is up near the strobe, and to which the main capacitor is connected.  Of course this is before the days of cheap microcontrollers, and in fact there aren't any chips of any kind.  The boards are single-sided, through hole, all discreet components.

I don't see anything that looks leaky or burned or otherwise out of place.  I've been unable to find anything in the way of a schematic.  Since the history of these things is that they just gradually get more difficult to power up, and ultimately won't at all, I tend to fall back to electrolytic capacitors as the primary suspect.  I was able the remove the main stobe capacitor (380uF, 350V), and it checks out good.  But there are a bunch of little ones that I would need to remove from the circuit and test.  There is also a trim pot of unknown function that might be worth checking.  But otherwise I don't know what to check, or even which board to start with.

Does anyone have any suggestions for how to approach this?  I don't think spending hours tracing out the circuit would help, particularly since I don't even know how it's supposed to work.

[rvalente]

Maybe the capacitor went bad and refuses to charges some times?

When you turn them on what current do they drain?

Have you tried to switch the caps? If the cap is charged, turn it off, wait a little for the cap to loose some charge (lets say 200 of 330V), do they charge easily now?

[Peabody]

I removed all of the electrolytics, one at a time, and  only one was even marginal.  I replaced that one, and put the others back, but the flash still doesn't work.  The chargeup whine doesn't begin, and there is no charge built up on the strobe capacitor.

However, I did find out what is heating up when I turn on the power.  It's two cylindrical things enclosed in a heat sink.  They are three-pin parts which I assume are transistors or thyristors.  They are marked with the Mitsubishi triangle logo and B324ST and 12 on the next line.  I get nothing when I search for that.  And of course I don't know if the problem is actually with these devices or with their driver circuits.  I think these should be part of an oscillator, but it's not oscillating.

Any ideas?

[floobydust]

The two 2SB324 transistors are suspect, they are PNP germanium parts that don't last that long. I would replace them in the oscillator that has troubles starting up. Maybe with TO-126 packaged parts like BD139. The hassle is what can fit and yet have a different pinout. It's silicon (not germanium) so might be a problem with low battery voltages like 3V. This should be 6V?

[Peabody]

The two 2SB324 transistors are suspect, they are PNP germanium parts that don't last that long. I would replace them in the oscillator that has troubles starting up. Maybe with TO-126 packaged parts like BD139. The hassle is what can fit and yet have a different pinout. It's silicon (not germanium) so might be a problem with low battery voltages like 3V. This should be 6V?

The flash operates on two AAs, so 3V, not 6V.  But thanks for the correct part number and manufacturer (Matsushita, not Mitsubishi).

[Peabody]

I've confirmed that the two 2SB324 transistors are wired in parallel, with their respective bases, emitters and collectors connected together.  I've also tested  the two transistors.  One of them has a dead short between emitter and collector regardless of what's happening on the base.  So that explains why it won't oscillate.  The good transistor switches on when the base is grounded (from 3.3V), but even when the base is pulled up to the emitter, I'm seeing about 130mV bleeding through.   That's with a 1K load resistor coming off the collector.

Since the flash operates on two AAs, I'm guessing the fact that these are germanium transistors, with only perhaps -0.3V EB voltage needed to begin turning on, is important, and that if a silicon PNP would have worked, they would have used one.  And I would think using a P-channel mosfet would be even worse in this regard than a silicon PNP.  At the same time, germanium transistors are very expensive, and apparently don't last long, so it would be nice to convert this over to silicon if I can. I guess the first attempt would be to use a single silicon PNP and just see if it works.  Does that sound reasonable?

[floobydust]

2SB324 is 32V 1A each, and -S suffix is highest beta rank 150-273 but surely much less at full current.
I would put in a TO-220 PNP transistor, in place of the two germanium parts. So the tab is the heatsink, as TO-126 parts you have to add one which is a hassle. The part would be overkill but should fit. The DC-DC converter can draw 1-2A for a few seconds when recharging the main capacitor. I would use KSB1366G, KSA1010Y have higher gain 100-320 or D45H11 (80) and bend the B-C-E pinout to fit pcb holes. I'm not sure how a TIP32C would do, likely would work.

edit: from the curves 2SB324 hFE stays good from 300mA to 1A, it doesn't droop like silicon. Might need higher gain parts but hard to know the circuit.

[Peabody]

Thanks very much.  I'm going to try to get a better feel for the circuit, and will post what I find.  The main question I have is whether the difference in the base voltage drop with silicon will result in the big flash cap charging to a different voltage.  I don't mind if it takes a little longer to charge, but I want it to end up at the same place.  The flash is powered by two NiMH AAs, so about 2.4V, and I need to figure out whether a silicon transistor will still get to that 300V, or whatever.

Also, I posted about this on a Usenet newsgroup, and a guy responded who has the 2SB324 "in stock".  So depending on the price, that may end up being an option.

Finally, it may just be time to replace these ancient Vivitars, as they fail, with Neewer flashes at about $34 each.  But if I can get the Vivitars working reliably for not much money, I'd be happy to keep using them.  Anyway, I'll post more about the circuit, and you can see what silicon part you think might work.

Edit:  I should also say that I wonder if the dual 2SB324s are there because both are needed for the current, or because there would still be one left when one fails, assuming it fails open.  Mine failed short, which is no good.

[Peabody]

I've traced out the charging circuit.  Attached is the schematic.  I won't pretend I understand how it works or why it oscillates.  Shariar of The Signal Path did a three-part video on how these things work, but it's still a bit obscure.

Anyway, the problem is it won't oscillate, and the transistors heat up.  Looking at the drawing, R1 is going to turn on the transistors, but if the circuit doesn't oscillate, current will continue to flow through the transformer primary, which is why the transistors heat up.

I've removed the transistors, and one has a dead short on EC.  I've also tested all the electrolytics, and only C1 was bad, which I replaced.  But I'm having a difficult time finding good 2SB324 germanium transistors.  From what I read, they don't age well, and of course by now they are all quite old.

So I would like to look at switching to a silicon PNP.  But since this flash operates on two NiMH AAs (2.4V), I wonder if that's going to work.  I also wonder about needing to change the value of R1 to bias a silicon transistor properly, but don't really know how deal with that issue.  Can anyone point me to an analysis or discussion of making this switch.  I think there's room for a single TO220, but is still has to oscillate.

[NiHaoMike]

This would probably be helpful:
http://repairfaq.org/sam/strbfaq.htm
A silicon transistor can be made to work, although R1 and R2 may need to be increased to account for the higher gain.

[Peabody]

I ended up replacing the shorted 2SB324 transistor, and also replaced the two electrolytic capacitors, and it's working fine now.  The next time one of these flashes fails, I may try to switch over to a silicon transistor, but for now I'm gonna leave well enough alone.  But I am going to replace the capacitors in the other two flashes.  After 35 years, it won't be too soon to replace them, and maybe that will help prevent the transistors from blowing out.