The Multi Amp aka Dynalo Mk2

[Pars]

Question time: One of the people I built a mini dynalo for sent it back to me for repair. They apparently had something stacked on top of it with it on for a day. When they turned it on the next day, no workee. When I got it, the first thing I noticed as only 2 of the 4 pcb leds were on, and sure enough, the positive reg (amb) had bit it. I went ahead and removed that and ordered another reg. While waiting for that, I powered it using my bench supply and didn't find anything wrong that I could find. DC offset was fine, bias was fine. Did a diode test on all of the output transistors with no problem. When the replacement reg got here, I went ahead and put it in after seeing no issues on the bench supply, and it turned on fine.

This weekend, I was going to get it finished up to return the amp, and went ahead and let it run for an hour or so with the heatsinks on, but much of the case missing. When I went back down to check on it, the offset was around -900mV and then I noticed that 2 of the leds were off. Positive reg bit it again.

I went down tonight and measured the +/- rail currents at +288mA and -278mA after running for quite awhile on the bench supply. These regs are supposed to handle up to 1A each, and have thermal protection. I took a heat gun and noticed the current was going up pretty fast when I was heating the output section; up to over 350mA after a short time and climbing. I did not wait to see how far it might go. I have no idea how hot one of these could get with something sitting on top of it, but I'm guessing quite hot. I know the regs current capability will vary with heat and how much voltage they are dropping, which in the case of these is 4V(24V in, 20V out).

I'm injecting the bench power on the electrolytic 100uf caps. There are (4) 100uf tantalum polymer caps following the electros, which may not like heat much? I don't know.

Amb's regs are up to $30 ea. with $10 shipping, so I don't want to keep burning these up. Just looking for ideas on where to look next?

Edited March 10 by Pars

[ang728]

did the old regs still working individually after removed from mini dynalo?

4V difference looks fine even a bit too much for amb σ78/79 , I would check the DC voltage ,also check output waveform on scope to see if it's oscillating when symptom occurs

[Pars]

The original reg did not work using my bench supply. The new reg (the now dead one) worked fine off my bench supply, but I haven't removed it yet since it went bad. The data sheet notes that these are supposed to go into protection when stressed. I presume that once they go into protection, they are supposed to come out of it once things settle down without the user having to do anything.

I've never been around when the bad stuff has been happening. This time, I left the amp running on the bench (no signal), and when I returned, it had failed.

Do Amb's regs have a reputation for oscillating? In this amp, the electolytic caps are Elna Silmic IIs, which I presume are not low ESR (the owner mandated these). I'll poke around Amb's forum and see if anyone noted oscillation.

[ang728]

16 hours ago, Pars said:

The original reg did not work using my bench supply. The new reg (the now dead one) worked fine off my bench supply, but I haven't removed it yet since it went bad. The data sheet notes that these are supposed to go into protection when stressed. I presume that once they go into protection, they are supposed to come out of it once things settle down without the user having to do anything.

I've never been around when the bad stuff has been happening. This time, I left the amp running on the bench (no signal), and when I returned, it had failed.

Do Amb's regs have a reputation for oscillating? In this amp, the electolytic caps are Elna Silmic IIs, which I presume are not low ESR (the owner mandated these). I'll poke around Amb's forum and see if anyone noted oscillation.

Hope that new dead ones is still salvageable

TI TPSA7A4701 would oscillate under some extreme circumstance but most likely that's not the case especially you have electrolytic caps installed. 

I would go back to experiment with thermal like force air flow on regs to see if the problem is solve then thinkering with passive setup.

[Pars]

I removed the positive reg and discovered it still worked on my bench supply. Soldering the pins onto these is getting harder for me, and I think one of the solder joints on this one was bad and disconnected when I had the amp running the other day. I think I may have lifted a pad, so will probably order another reg and figure out a better way to hold the reg and the pin(s) when soldering. These things are small.

I hooked my amp up to the bench supply just to confirm the currents I was seeing and it was in range. Mine was actually higher at 330/320mA +/-, but mine is built using MPSW thru hole transistors with creative lead bending. 

[sbelyo]

let me know if you need anything.  I may have the parts if anything is unavailable.  I squirelled away some of these devices.  I have 1 TKD board and 3 Alpha's if anyone needs a board.  I might be able to part with a TKD pot if needed

[Pars]

I know there isn't much interest in these amps anymore, but wanted to post my findings regarding thermals. I spent around 3 hours yesterday testing the one I have in for repair to evaluate whether these run cooler with no heatsink, heatsink only, or heatsink with a spreader bar to tie it to the top cover. @johnwmclean had at one point suggested that he thought these might run cooler without. I forget which user suggested and implemented a spreader bar, which for them resulted in lower temps.

I used a thermocouple with the covers on to do this test. For the 1st case (no HS), I put the thermocouple on top of the upper right transistor in the L- channel. For the other two tests, the thermocouple was touching the collector tab. I tried to wait long enough for things to stabilize completely, which never really happened. Even after an hour or more, the temp would go up 0.1 ºC every minute or so. Here are the results:

bare                       91.8       (shortest test, around 45 minutes)
HS only                 92.6       (approx 1 hour +)
HS w/spreader     87.1        (1 hr 20 min)

all temps in  ºC. Ambient temp 22.1 ºC.

It appears that w/spreader is the winner. That temp was going up slower than the other two, so I would guess that the margin is a bit larger. i made the spreader with a couple of pieces of 1/8" and 1/16" thick aluminum x 1" stock.

[Pars]

Also, I have been thinking about doing a revised layout for this amp, but adding some things such as Kevin's protector board, and possibly adapting the circuit used by nwavguy in the o2 for battery management to delay startup until the 2 Meanwell bricks had both come up, since this is what is causing the large DC offsets at startup (and possibly at shutdown as well). I know the protect board would take care of the startup issue so might be overkill for the additional circuit. The board would have to be larger in size to accommodate these changes, which should also have some benefits thermally. Large heatsinks for both the Amb regs and the onboard regs would also be a good idea.

This is just at concept phase right now, and I'm not really interested in on for myself, but might do it anyway. If so, I will ask Kevin for permission to use his protect board circuitry for this.

[kevin gilmore]

knock your socks off.