LM3886 Power Transistor help needed

[jimmy74]

Well I must say that these bulbs are becoming like good old mullards, not one shop, supermarker, handy store, electronics store etc etc supply them anymore and fridge bulbs are max 25watts. I managed to find a working 40w bulb in an old box in the garage so I'll use that.
I also found an old variac (iskra brand) from the 60's I think, could this be handy too?

I'll start doing all the testing tomorrow, though I have one good question for you, before Roly posted telling me not to turn it on, I had already done so, do you think I'll need new amp chips? or do you think that they'd be fine?

thanks
J.

[Roly]

A 25W globe is handy too.  I have a box with a collection from about 5 watts up to 100, not to mention a few duplicates just in case.

do you think I'll need new amp chips? or do you think that they'd be fine?

Part of what comes next is answering that question.

The game plan from here is;  clean up the board and check the power supplies without power chips in.

Possibly discover the regulators are in backwards, and equally possibly boofed, sort that out as required.

Once the supplies are confirmed okay, fit one chip amp and see what happens.  If it's boofed then try another, etc.

Once one position has come up okay, try another chip amp in the next.

Keep looping until we have good chips in all positions.

A Variac can be useful, but in this situation you really would need a limiting lamp as well, so perhaps better to leave the Variac as a later project for the moment.

[jimmy74]

Ok I've made the limiting lamp and I've carefully removed all the amp chips. I've done some tests for continuity as you suggested:

Check PINS 1 & 5 = OK ALL CHIPS
Check PINS 3 & 2 = OK U7 ONLY (U8 = Pin 3 to output, Pin 2 to Pin 1) ( U5 = Pin 2 to Pin 1, Pin 3 to ?)
Check PINS 4 & 6 = OK ALL CHIPS
Check PIN 7 = ARE THEY SUPPOSED TO GO DIRECTLY TO GROUND? THEY DON'T
Check PIN 8 = C54 100uf 25v ALL CHIPS
Check PIN 9 = NOT SURE EXACLY WHAT TO CHECK HERE
Check PIN 10= NOT SURE ABOUT THIS ONE EITHER.

And I've also verified that both voltage regulators are in the correct way. What's next?

Also tested the limiting lamp with the amp attached to it, on pins 1 &5 am getting 35vDC and on pins 4&6 am getting -35vDC.

thanks
J.

[J M Fahey]

1)

Pin _x_ OK ALL CHIPS

What does this exactly mean?

2) With the simple JPG schematic posted earlier you already had all the data needed to repair this amp.

3) pull all chipamps, verify that each chipamp leg or pad is connected where it should, through a healthy track.
If any missing, repair it (small piece of wire, whatever)

4) still chipless, turn amp on through bulb limiter, measure voltage to grounf pn each ñleg of each chipamp.
Post it here.
Keep it simple: IC _x_ Pin _y_ = _z_ V (or whatever)

[Roly]

+/-35V looks reasonable.  Please also check that you have +15V and -15V in the preamp area - that's what the regulators do.

[jimmy74]

well I started out today to try and get those voltages written down but something strange was happening, I wasn't getting any voltages anymore, I checked around the whole board and found that the ground wire had snapped off the lug on the mains plug. Fixed that and went ahead to reading voltages to each pin and each chip:

http://www.flickr.com/photos/73835524@N00/11173829406/in/set-72157638111984374

As for the +15/-15v readings there's some strange stuff going on here too, reading off the regulator pins I'm getting:
LM317
Pin 1 adj= -6.5v
Pin 2 out= -5.8v
Pin 3 in= 35.1v

LM337
Pin 1 adj = -32v
Pin 2 in = -36v
Pin 3 out = -34.5v

and here's the other half of the schematic:
http://www.flickr.com/photos/73835524@N00/11174186145/in/set-72157638111984374

[Roly]

LM317
Pin 1 adj= -6.5v
Pin 2 out= -5.8v
Pin 3 in= 35.1v

LM337
Pin 1 adj = -32v
Pin 2 in = -36v
Pin 3 out = -34.5v

hummmm ... well that ain't right.   

{I can't see how you can have a more negative voltage on the LM317 "adj" than on the "out", however...}

You've got -5.8V where you should have +15V, and -34.5V where you should have -15V; so I'd guess that the LM337 is shorted (or the "adj" lead is open).  The LM317 may not be damaged but I can't imagine it's happy with a negative voltage on its output.  At this point I'd replace both, being careful to get them in the right place, the right way around.

I'm seriously annoyed with your "tech", he seems to have made matters quite a bit worse with his blundering around.   :trouble

We will now have to see what other damage may have been caused to the preamp by excessive -ve supply voltage.


I've merged the two half circuits so we can have an overview of what is going on (attached).

The upper LM3886 feeds the high frequency speaker, while the lower two appear to be in a bridge configuration driving the main speaker.  A "bridge" means that each amp is arranged to drive the speaker from each end in opposite directions, so doubling the signal voltage across it.

Most of the op-amps appear to be BA4560 duals and if needs be can be replaced with just about any available dual.

The BA6110 may be a different story.  This is a special type of op-amp called an OTA or VCA - Voltage Controlled Amplifier, and you may have to hunt around a bit for a supplier who stocks it if you need one.  Together with U3B and associated circuitry (D2, D3, D6, &c) this acts as a limiter/automatic gain control to throttle the amp back if over-driven.

First priority now is to get the +/-15V supplies correct, then we see what else might be damaged.

[jimmy74]

well I replaced both of the regulators, triple checking that both are in the right way, and triple checking for continuity to the surround components following the schematic. BTW there is some monkey business in the schematic too C48 and C49 are the other way around and so are C51 and C50, though + and -ve connections are still the same. However powering it on through the limiting lamp I first notices a very high pitch whistle and that the power led doesn't turn green as it had been doing up untill today.
Anyways here are the voltages I'm getting off the regulators:
LM317
Pin 1 adj = 35.4
Pin 2 out = 36
Pin 3 in = 36

LM337
Pin 1 adj = 2.6v
Pin 2 in = -36
Pin 3 out = 1.1

so what went wrong now?  :grr

[Roly]

C48 and C49 are the other way around and so are C51 and C50, though + and -ve connections are still the same

I'm not sure what you mean by this.  If the component numbering has C48/9 in swapped positions that is of no matter.  If you mean these caps are in backwards, +ve and -ve swapped that's very serious.  (C50 and 51 are not polarised).

LM317
Pin 1 adj = 35.4
Pin 2 out = 36
Pin 3 in = 36

LM337
Pin 1 adj = 2.6v
Pin 2 in = -36
Pin 3 out = 1.1

Well we now seem to have an opposite situation where the +ve regulator is shorted and pulling the -ve rail +ve.

You are "off the circuit" here, meaning that whatever problem you have will only be apparent by physical inspection of the board and components, no amount of circuit-gazing will help because there is something about your board that isn't on the circuit (such as a solder bridge between two regulator pins, broken track, dead regulator, one of the associated resistors open circuit, esp R52, etc)

I'd pull the LM317 and see if that restores the -ve rail.

Since the LED is driven by these low voltage rails I'm not surprised that it's behaving strangely.

a very high pitch whistle

You do have all the power chips out, and no speakers connected?

[jimmy74]

Ok I'll try all of this tomorrow, and yes I have pulled all the amp chips out and there's no speaker attached, but I still did hear that very high pitch whistle, it lasted about 5 seconds.

As for the caps, no they're not backwards in terms of polarity, just in terms of component numbering. If you look at the photo of the board showing the two voltage regulators, the far right lug of the LM317 was soldered directly to a hanging track which was shorting with one of the lugs on that small white connector. So this time round I cut off the hanging track and connected the LM317 input lug directly to the +ve side of C46 using a piece of wire. Tomorrow I'll take it out and see if that fixes anything.

http://www.flickr.com/photos/73835524@N00/11085317413/in/set-72157638111984374

What a mess! :duh

[jimmy74]

Ok I pulled out the Lm317, the power led turns green and these are the voltages I'm getting:
LM317
Pin 1 adj = -23.3
Pin 2 out = -22.6
Pin 3 in = 39.1

LM337
Pin 1 adj = -36
Pin 2 in = -38.5
Pin 3 out = -37.1

what now?

[teemuk]

You have reference voltages listed in the schematic. Input pin voltages should be -close- to 40VDC, both negative or positive. These seem to be ok, not exact 40 but close enough.

Output pin voltages are obviously incorrect. They should be about 15 VDC, negative and positive. Likewise "adjustment" pin voltages are incorrect: These should be about 1.25V below output voltages. This voltage potential is derived with the resistive divider from the output that serves as a feedback loop. Consult datasheets for further info.

So, given the readings I would check the condition of those resistive dividers as well as the condition of the regulator ICs.

[jimmy74]

can you point out which ones in the schematic are the resistive dividers?

[teemuk]

The string of two resistors at the output of each regulator, their joining node connecting the "adj" pin of their corresponding regulator, other end connecting the common.

In other words: R54 and R52 for LM317 positive voltage rail regulator, and R55 and R53 for LM337 negative voltage rail regulator.

They are called "dividers" because the voltage at their joining node divides by the ratio of the resistances. This is basically how the voltage adjustment of the regulator works.

I bet you didn't even consult the datasheet like advised. The datasheet would have clearly explained the purpose of these resistors, how they are designed/calculated accordingly and how the regulators ICs work.

www.onsemi.com/pub/Collateral/LM317-D.PDF
www.onsemi.com/pub/Collateral/LM337-D.PDF

[jimmy74]

I've tested the 4 resistors and they all seem ok, 215r and 1.8k. I've followed the tracks making sure there aren't any shorts with other tracks or solder points in that area and all is fine. I've soldered back in the LM317 and taken the voltage readings. They are still all over the place and mostly incorrect. I think there may be another problem, that white 2 pin connector underneath the board is not shown in the schematic, well atleast I can't find it anywhere. One of its pins goes to ground and the other goes to the out pin of the LM317 regulator. Here are the new voltages:

LM317
Pin 1 adj = 34.8
Pin 2 out = 36.4
Pin 3 in = 36.4

LM337
Pin 1 adj = 9.4
Pin 2 in = -36.3
Pin 3 out = 8.1

And the power led remains red.