LM3886 Power Transistor help needed

[jimmy74]

Hi folks I'm trying to figure out if this power transistor can have Pin 1 completely left out of the circuit? That is... not soldered to anything? what I have here is an Italian made active stage speaker, it has 3 x LM3886's, 1 x LM317 and 1 x LM337. The LM3886's are all have a different soldering layout. It was severely hacked by the previous "tech", one of the LM3886's had been pulled out by force and he pulled all the paths along with the pins and did a fabulous job of trying to solder everything back in. I've got myself a set of new LM3886's, new LM337 and a new LM317. I have already installed and cleaned up the hacked LM3886, it was even missing one of the pins, I have also installed the LM317 and LM337 and am currently on LM3886 number 2 which has also been messed with but not so badly. The only solder point missing is on pin 1. Holding the circuit board up against the light I can see where the solder point used to be but it doesn't seem to be pathed to anything, just a solder point.

Do you think that this is normal? what would this pin usually be tied to? And yes I have asked for the schematic directly from the techs of the company that makes these speakers, but no luck whatsoever, even if the speaker is no longer made.

Any help at all would be great
thanks
J.

[Roly]

Hi jimmy74, Welcome!


To avoid confusion the LM3886 is a "chip amp", not a "transistor".

According to the data sheet this chip has two positive power supply pins, 1 and 5.  I've just quickly scanned the data sheet and it seems to be mute on this topic, but if pin 1 isn't connected but pin 5 is in your amp, then this should be sufficient.  This sort of thing is normally done to allow board designers some flexibility in their layouts.

It was severely hacked by the previous "tech", one of the LM3886's had been pulled out by force and he pulled all the paths along with the pins

Ouch!  This is somebody to avoid in the future.

Do you think that this is normal? what would this pin usually be tied to?

Quite possibly.  Supply positive like pin 5.  Compare the two layouts - what do they do with pin 1 on the other one?

In most cases amps built with chip amps will simply use the suggested circuit straight off the data sheet (Fig 2, p5).  As you have two of them it's a fair bet (if it's mono and not stereo) that it uses the bridge circuit with one chip amp driving each end of the speaker (with opposite signals).

HTH

[jimmy74]

There are 3 of these chip amps:

LM3886 1 - pin 1 is tied to pin 3
LM3886 2 - pin 1 is connected to nothing (pin 5 however is soldered to something)
LM3886 3 - pins 1 and 5 are both connected separately to the board

I haven't tried turning it on yet, but before any replacements the amp was giving me very loud noise and thumps through the speaker and going to protection or peak mode. And also they had used heatsink washers that were too short allowing for the mounting screw to short the amp chip to the heatsink.

I just wanted to be sure before plugging it in.

thanks
J.

[jimmy74]

I tried turning it on, and after a couple of seconds the same thing happens as before, loud noise and the peak light tuns on. Maybe it's the caps that need to be replaces, 2 x 6800uf 63v and while I'm at it I'll change the bridge rectifier out too, what the correct replacement for a PBL403?

thanks
J.

[Enzo]

You are just guessing.  Throwing parts at the amplifier is not the way to fix it.

For example, the LM317 and LM337 are voltage regulator ICs.  What led you to replace them?

[jimmy74]

Firstly the way they were soldered in, as I said the previous tech did a fantastic job of ripping out the parts and hacking the paths, the VR's were just hanging on by a blot of solder used to connect the 3 pins to what was left of the signal paths and there was some monkey business with these too. I suspect that they were (and I copied that with the new parts) put in the wrong way around but without a schematic this would be hard to verify.

what do you think?

[tonyharker]

First of all get yourself a schematic.  What amplifier is it - someone here may be able to help you with it.  If you dont know what you are doing then you are just going to destroy more chips. 

[jimmy74]

I have been insisting on this since this active speaker was brought to me, but the distributors and techs of this brand are not willing at all to let the schematic out of their bonds for some unknown top secret reason. This brand is DB and the model is OPERA LYRIC 215. It has been discontinued so I really don't understand why they aren't willing to let the schematic go.

[Roly]

First of all get yourself a schematic.

He's already tried that - see OP.  They're being anally retentive. (ps see footnote and attach)

LM3886 1 - pin 1 is tied to pin 3

(assuming that you are counting pins from the right end - you have downloaded the data sheet, haven't you?) this is seriously WRONG.  Pin 1 is the positive supply.  Pin 3 is the output - tie the supply to the output and say bye bye to your chip.

Disconnect all the speakers from the amps (if you value them).

I tried turning it on

DON'T power it up without a current limiting lamp (unless you actually like replacing brand new semiconductors and getting nowhere fast).  Before doing anything else make one of these up because you already need it.

they had used heatsink washers that were too short allowing for the mounting screw to short the amp chip to the heatsink.

Eh???  Assuming that it once worked properly, this is more likely to be down to your "tech".  Mounting screws in this case normally have a special insulating washer on the mounting screw.

Maybe it's the ...

As @Enzo said, you're guessing.  This is also known as the "blunderbuss method" and it does not work.

We use a method: First we diagnose, next identify the faulty part(s), finally we replace those, then happily play guitar into the wee small hours.  This is the only method that actually works.

Randomly replacing components is not only expensive and fruitless, it also tends to hide the original cause of the problem.  I could cite you many examples of having an amp up on the bench that had been given a dose of the blunderbuss method, and you typically spend something like 90% of the repair time undoing the previous "repair" work to dig back to the original fault which then takes 10% of the time to fix.  This is the reason that some techs simply refuse to touch something that has already been got at by the owner.  It looks like this is already what your "tech" has done, and dropped you right in it.

Download the chip data sheet.  Are you counting the pins from the right end?  Look at the application circuit Fig2/p5 and compare with what you've got.  Take some clear crisp pictures (borrow a camera if you have to) of the innards in question and the general arrangement, and post here.

Remove all three new chips.  Clean up the connection area to be sure no tracks are shorted together.

Power up via the limiting lamp and check that you have something close to the correct supply voltages, +/-40V and +/-15V.

After you post a pic of the mangled track side of the board where the chip amps mount we will try, one at a time, to sort out the correct chip wiring.


And well lookie what I found at http://www.scribd.com/doc/101010269/OPERA-212-LYRIC-Schema-Elettrico , attached.  (anyone got a subscription so we can get the whole thing?)

[jimmy74]

Just signed up and downloaded the whole schematic, payed 9$ for it, but that's fair enough instead of going on in the dark!!

Here are some photos of what's under the board:
http://www.flickr.com/photos/73835524@N00/sets/72157638111984374/

Looking at the top of the chip(where I can see LM3886T written) and counting the pins from right to left (if this is the correct way to do it)  pins 1 and 3 are tied together on U7. Actually this looked quite normal to me as there is a path going from one pin to the other. Looking at the schematic and judging from what you logically say, pins 1 and 3 should not be tied together!! Factory monkey business?

thanks
J.

[jimmy74]

Photo showing pins 1 & 3 unsoldered on U7, there is clearly a track between them:

http://www.flickr.com/photos/73835524@N00/11085319253/in/set-72157638111984374/
I wrote pin 5 on the photo but it's pin 3

Photo showing U8 with pin 1 floating:
http://www.flickr.com/photos/73835524@N00/11085206006/in/set-72157638111984374

Photo showing U9 + U10 + U5:
http://www.flickr.com/photos/73835524@N00/11085317413/in/set-72157638111984374

U9 + U10 have their heat plates not facing the main heatsink which is on the top of all 3 photos.

thanks
J.

[Roly]

Well first up you need to post the circuit here so we are all on the same page.

Factory monkey business?

Do you think dB produce amps that don't work?  And nobody notices?  No, I don't either.  The only monkey here is the "tech" between what was, and what now is.  Whatever, how we got here isn't as important as finding a way back.

Before anything else - a limiting lamp.

Pin 1 is indicated by a dimple on the front of the IC.

From the attached you can see that pins 1 and 5 (V+) are indeed connected.

If you follow each of the traces back from the IC pins then 1+5 should go back (along the edge of the board) to the positive end of one of the main filter caps.

The output pin 3 runs through pin 2 (No Connection) to a speaker output (2nd track from board edge).  Check?

Pin 4 runs through pin 6 (N/C) to the negative end of the other mains filter cap, 3rd track in.  Check?

Pin 7 should find its way back to ground/common.  Check?

Pin 8 Mute connects to the other Mutes and to a 100uF slow start de-thump circuit.  Check?

Pin 9 is feedback (indirectly) from the output.  Check?

Pin 10 is signal input from the preamp output.  Check?

As I said, remove all three IC's, clean off the excess solder, do the test above.  Check?

Report back.

If the supplies are okay, then we will put one IC back in place (and find out if it's boofed).

My circuit fragment doesn't show U9 and U10 which I assume are regulators.  If these are intended to mount on the heatsink (are they?) but they don't have their metal plates on that side, then they are in the wrong way around, and that won't be any good for their health either.  I'll assume that these are the regulators for the +/-15V supply, so you also need to check that the right regulator is in the right position.

I suspect from your comments that the +/-15V won't be okay, and we will need to sort those power supplies before the output stage.

A more general pic showing the board top and how it presents to the heatsink would be helpful.

[Enzo]

and counting the pins from right to left (if this is the correct way to do it)  pins 1 and 3 are tied together on U7.

Umm...   we generally count left to right.  Check this data sheet for the LM3886, scroll down to the picture of the IC with pins numbered.

[jimmy74]

Yes for some reason I missunderstood Roly's statement - "assuming that you are counting pins from the right end". Anyways I am building up the light bulb limiter, do I need 100 watts? Can it be one of the modern energy saver bulbs or does it have be the good old type? I could find any old type ones at the local handy store!!

thanks

[Roly]

"Correct" end.

You need to start with a low power globe, 40W at most.

It must be a filament lamp, not a CFL (or LED), but mains QI downlights are suitable.  Try to find the section that deals with globes for special uses such as ovens and fridges; they may be odd-looking or have a strange base, but these will be available for a while yet as electronic CFL's and LED's aren't suited to these hot and cold services.  Most lighting joints still seem to stock a range of filament globes.