Acoustic 370 grounding issue, ch2 buzz

[cbarrow7625]

That makes sense for S101. AGC, huh? I just assumed it was protection. I guess, in a way, it is.

If the DC voltages check out then the next thing to do is figure out what frequency the oscillation is at & where it appears in the circuit. Everywhere? Just at the input? Just affecting the AGC? Does the oscillating frequency appear on the output?

I know you said that there was some "hair" on the rails. HAve you tried parallelling another PS cap to try to eliminate the "hairs"? Are the filter caps in good shape? Have you looked at the DC rails using the AC input on the O'scope? Whadaya see?

[PRNDL]

BAD NEWS

I slipped when measuring 41 volts at the collector of Q303 and hit the ground bar (it's right next to it). Smoke and a burnt fuse. It smokes out the bottom of the board near Q303 with a new fuse, but I can shut it off before the fuse blows.

Now I'll have to figure out what's blown and replace it before I can get back to diagnosing.

BTW, Q316 had incorrect DC voltages. 5 at the collector instead of 10.4 and .24 at the base instead of 0.3. C315 had 5 instead of 9.8 Q316 is the transistor for the AGC.

[cbarrow7625]

Yikes! I hate it when that happens.

Sounds like either Q316 is partially turned on (i.e. bad) or one of the associated passive components is bad. Either R330 has drifted way up in value or R334 or C315 are partially shorting to ground. Makes me suspect C315 or R330 as the most likely culprits. Check them while you deal with the other blowed up stuff.

0.24 at the base sound close enough to me to be OK.

[PRNDL]

R327 is smoking, which means the output transistors are drawing too much current.

I'm wondering what to do ... this amp has already been to several techs without success and they're not that expensive used. The problem is that I can't turn the amp on for more than a second or two before it smokes and blows a fuse.

I tested Q303, 304 and 316 and they seem fine.

Learning by experience can be difficult!!

[cbarrow7625]

If that's the case, I would guess you don't have just 160mVat the junction of R325/326. It's probably more like 40 Volts!

One of those transistors must be bad. Definitely pull & check Q312-314. If you don't find it there, check Q310 & 311as well.

[PRNDL]

I think it's Q310, although it's pretty hard to tell when it's in the circuit, and it's mate (Q305) is NPN instead of PNP. (The power transistors Q312-314 and their driver Q311 tested fine).

Would a test be to take it out off the board and power up to see if there's no smoke?

I could also take out the Q311 to 314 pretty easily.

BTW Are those transistors glued to the square metal case / heat sink?
I'm having difficulty getting Q310 out, although sometimes it's hard to get all 3 pins hot at the same time.

[PRNDL]

Now we wait for UPS delivery - I ordered fuses and two transistors.

It's nice to work on cool amps like this.

[Enzo]

Get all three pins hot at the same time? The way to remove a transistor is to remove the solder, then lift it out. I use a desoldering station, but I have for years used other methods in the field.

Many guys like solder wick or solder braid - same thing - lay it on the spot of solder and then the iron top of it. It absorbs the solder leaving a clean spot.


I don't like braid, but I do like a sucker. These are cheap, but I hate them more than braid.


I like the kind you cock then hit the trigger button.


Trying to melt all the legs of a transistor at once is doomed to fail. Without the heat sink, one can bend back the transostor extracting one leg at time, but not on the heat sink like that.


Q305, Q310 are RCA40409, 40410. The heat sink is PART OF THE TRANSISTOR. If you screw around with the NTE cross, it will not have the heat sink and will burn up pretty soon. MAgi parts sells real 40409s, and I am sure others do too. it is conceivable a TO220 with a finned heat sink screwed on might sub for it, but I use the real part.


Do we know C313 is OK? If it is open, the amp will be unstable.

If the outputs draw too much current sitting there, then I suspect the bias circuits, so look into Q303.304 and all those resistors around them.

[cbarrow7625]

Would a test be to take it out off the board and power up to see if there's no smoke?

NO! Don't do that. You would wreck the biasing & take out a whole bunch of other parts in the meantime.

If you have a meter with a transistor tester on it you could use that to check it. To check it long-hand you need to basically check it for internal shorts. From the emitter to the base, check it like you would a diode (for Q310, theemitter is the anode of the diode & the base is the cathode). You should get about a .6V drop across it in one direction & nothing across it in the other direction. Just like a dioide, if you get a reading in both direction it is bad. Then measure for any leakage between the base-collector (you should measure "overscale", or extremely high resistance) and between the emmitter-collector (same deal). If you measure any resistance between these points, there is a problem with the transistor.

[PRNDL]

The way to remove a transistor is to remove the solder, then lift it out.

I was using solder braid, but didn't know the heat sink was part of the transistor.

MAgi parts sells real 40409s, and I am sure others do too.

Can you tell me the web site (I couldn't find it).
Thanks!!

Do we know C313 is OK? If it is open, the amp will be unstable.

I'm not certain how to test a coupling cap with a DMM.

[Enzo]

Pull the cap. Now set your meter to diode test and hold the probes on the cap leads. After a second or two at most, the cap wil have charged up - you see the ohms display sweep up them max out. Now remove the probes. The cap chould now be charged up to a volt or so by the meter diode test. flip the meter to DC volts and see how long the cap remains charged. The meter will discharge it slowly. COMpare to a good 0.1uf cap from your drawer. if it won't charge up, then it might be open.

Sorry, typo, Magic Parts sells 40409 and 40410. www.magicparts.com wholesale only though
A quick google shows www.vibroworld.com has them for $15.

These guys ask $11 for them http://www.soundchamberrepairs.com

In fact just google RCA 40409 and a whole world of places come up

Google is usually my first place to look when my notmal suppliers don';t have something.

If you didn't realize the heat sink was a part of things, then I guess it WAS hard to get them out. Unsolder the two pins onthe heat sink.

[PRNDL]

It's not smoking!!

I replaced Q310, Q304 and Q316 and pulled two bad power transistors.
I plan on buying a real 40410 and C313 (I used a .047? 600V cap I had on hand).

I think the original problem was Q316, which is part of the AGC circuit. It may have failed while I was measuring voltages and took out some other transistors in the process.

The next step is to see what the scope shows for output, but I need to take a break.
That was hard work!!

Is it best to replace all of the output transistors, or just the bad ones?

[PRNDL]

Something is still wrong - there's no output.

OTOH, the input signal reaches the power board just fine, and there's no smoke, which is nice.

Here's some DC measurements.

DC+ = 85
DC2 (output) = 85 (should be 45)

Q301/2 C =13
Q301 B = 12
Q302 B = 23 (should be 11)

Q316 C = 6.4 (10.4)
Q316 B = .49 (.3)
C315 = 12 (9.8)

Q303 C = 85 (41.7)
L301 = 85 (43.7)

Q311 B = 85 (120 mV)
Q312-14 B = 85 (160 mV)

My first guess is one or more of the upper output transistors has shorted.
Since one upper and one lower already tested bad, it may be best to replace all.

Which replacement is best? 2N3055
2N3055AG (ON Semi) = $3 2N3055G (On Semi)=$1.75 2N3055 (ON)=$2.10
2N3055 (STMicro)=$1.40

There's a tech note to use only matched RCA 2N3055's as other brands fail prematurely.
Vibroworld is the only place I've found them, but they're closed.
I also found some NOS RCA 2N3055HOM, which are supposedly better performers.

Q315 is failing the resistance tests .34 from the collector to emitter either polarity.

Is this correct, or should I be looking for something else?
(Also, would it be wise to remove all the output transistors until the DC voltages are correct during test?)

Update: Q302 is bad. I missed it by testing only BE and BC.

I'm worried about multiple transistor failures on this board.
Would it be wise to replace all the transistors?
Also, R322 and 323 have brown marks but test fine. Should I order replacements?

[PRNDL]

Fixed

Well, it's fixed.

I ended up replacing 20 components - all the output and a few driver transistors and a bunch on the preamp board, a few diodes and a couple of resistors. It turns out that when one bad transistor burns out it takes out its neighbors and works its way through everything.

I've learned a lot about solid state amps in the process.

This amp sounds incredible ... it sounds crystal clear with some kind of emphasis on the high end that's hard to describe. I'm using a crappy old Kustom PA speaker and turning down the treble a lot -- it must be real nice with a good 4-ohm bass cab.

Thanks for the help!!

[cbarrow7625]

PRNDL,

That's great to hear. Looked like that had to be the case pretty early on. Glad you had the chance to poke around and figure it out yourself. That's always the best kind of learning.

In discreet solid state amps where everything is DC coupled it's pretty hard not to take out everything downstream when one small transistor goes. It throws of all of the bias voltages in the entire power amp. There are complex protection circuits that can be designed in to handle such things but most folks think they never sound as good once all of the additional circuitry is hanging off the "working parts".

Good luck with it. I hope it stays together for ya.