Gainclone 3886, Vcc- fuse blows

[polo16mi]

Hi there:

Had been a long time since my last post.  We survived Puyehue´s eruption, and my new house building, so here i m again.

Now i return to guitar amp construction hobby, cause i still playing with my ruby version, and last week try to play with other guy´s, equiped them with a Roland 30 cube & Gibson SG  8|, other guy with a Traynor 20 and a Squier, and i with my old Faim and Ruby  :grr.  It is like trying to scream at front of Boca Juniors hooligans. I cant ear my self nothing at all.

So now i working in a Gainclone like http://www.generalguitargadgets.com/projects/13-amps/58-lm3886-amp

I mounted a power supply with a 24/24 6A trafo, rectifiers, 2 4700uF electrolitical per rail, so i have +34.9V,0V,-34.9V.
I put a fuse between each rail and V+ and V- of the gainclone, so goes like this:


---Vcc+ --> Fuse----Vcc+(gainclone)

---Vcc- --> Fuse----Vcc-(gainclone)


I switch on, and fuse at Vcc- blows.

I checked disconecting the gainclone, new fuse, at -34.9V ok. No blows.

I checked the pcb and founded a short between Vcc- and Pin11 (pin 11 says datasheet that not conected, anyway, i removed the short). Regard of this, every looks ok.

The tab of LM3886 are insulated from heatshinker with a mika film and a plastic ring in the screw. I measured continuity between the tap and heatshinker, and read something, and quickly goes infinit.

Can you help with some tips?

[J M Fahey]

Hi Polo16, long time no see.
*If* you used the GGG PCB, I see the negative rail track runs very close to the edge where the LM3886 is mounted to the heatsink. *very* close.
Maybe that track is touching the heatsink.
It will blow the "-" fuse at once.

[polo16mi]

I ripped the track a little from the edge of the pcb.

I´ll buy more fuses and try this at afternoon.

I using 1,5 A fuse for each rail, fast blow. Its ok?

Thanks for you help.

[J M Fahey]

1) build and use a lamp bulb limiter, it will not only save fuses, but let you measure *something*, which otherwise is impossible.
2) turn the amp on without speaker or any other load connected, check that you do not have DC on its output.
Good luck.

[polo16mi]

Now connected a 75W limiter bulb. Using 2A fuse for each rail.

No heatshink attached (just for test ). I did next tests:

1- Power amp stage disconnected from power supply. Turn MAIN switch on. +33.5V, 0, -33.5V at rails. Bulb is OFF.

2- Now connected power stage at supply. Using 8ohm-80W "FAHEY" Speaker. MUTE switch OPEN. Turned MAIN ON. -8,5V at VCC- and 9V at VCC+. Bulb is ON. Turned MAIN off.

3. Now using cheap 8 ohm, W, speaker. Mute switch CLOSE (ON). Turned MAIN ON.  -8,5V at VCC- and 9V at VCC+. Bulb is ON. Turned MAIN off. As result of this C5 (2200uf/63V) get is head a few inflated, and somewhat hot.

No more 2200/63V caps at hand, so i planning to pull them (c5 and c6) off of pcb, and retry. Perhaps c5 in short? It was new brand one. I thinking to pull them out cause i see that them are in parallel with the power supply ones (my supply has 2x4700uf for each rail), so ¿with them are enough filter caps?

I checked the pcb again and nothing seems wrong. At first test yesterday, as i say, PIN11 was wrong joined with VCC - . Datasheet says that PIN 11 is NotConnected. Anyway, the chip can blow for this?

Attached picture from power supply, and back of amp pcb.

[J M Fahey]

1)

turn the amp on without speaker

"without" means NO speaker connected (not even a Fahey  :lmao:)
2) besides that, I think I found the culprit, you mounted C5, the 2200uF cap which filters -35V on the power board itself, upside down.
(You connected it with the polarity inverted).
That´s why you did blow the "-" fuse and now it bulged (inflated) and became hot.
Replace it with a new one, because it´s unreliable by now.
Just for kicks, check the other 2200uF too, maybe you mis-interpreted the markings and *both* are upside down.
3) strictly for testing, to save some time considering you live in an isolated place, replace C5 with *any* 470uF or higher, 35V or higher electrolytic cap you happen to have, and turn amp on again (still no speaker) to check what happens.
If the Chip is fine, you should have around +/-30V on rails, the bulb should blink and then be off, and no DC on speaker output (well, less than 100mV DC).
If not so, order another LM3886.
Good luck.

[polo16mi]

J M;  YOU ARE THE MAN!!!!   :dbtu:

I had readed your bulb and speaker suggestion after do my tests.

I checked again the datasheet and ggg schematics, and you are right!!!

I confused putting the capacitors. Negative pole of electrolytic, always at most negative tension. Using now  2x new 1000uf /63V.

I try replacing speaker of coihue micro (aka ruby with big muff control tone) with lm3886 input and trip like a charm  :dbtu:  :tu:   ( Guitar-> CoihueMicro->Lm3886 input)


Thank you a lot!!!!!

PS: Next step! SansAmps GT2 clone as a preamp!