Marshall Valvestate 8004 - Blowing TDA1514A's

[NickH]

Hi,

I have been working on this 8004 for a week or so now, taking my time & trying to solve it like a big boy on my own, but I have run out of ideas now & I am getting a little frustrated with it..

Both TDA1514A were blown (with huge cracks down the middle) I am guessing this is because someone had replaced the fuse that would have been blowing with a screw...

I removed the blown TDA1514A's and powered it up, the voltages seemed out and the BVD64 (TR2) was getting extremely hot, I replaced the BVD64 and then the (R3) resistor is was also very hot & finally the penny dropped and I replaced the (ZD2) Zener Diode that was actually causing the problem! (DOH!) It now gave me -27v.

I measured the voltages at the TDA1514A sockets (with the chips removed) and got the following:-

IC1

1 = 0
2 = -27
3 = -27
4 = -28
5 = +33
6 = +35
7 = +35
8 = 0
9 = +33

IC2

1 = 0
2 = -27
3 = -27
4 = -28
5 = +33
6 = +35
7 = +35
8 = 0
9 = +33

Obviously the + voltages are high, which I am pretty sure I fixed by replacing the + Zener Diode (ZD1) as well (I can't be 100% certain as I had soldered the new replacement TDA1514A chips in by this time [Don't ask... ] and my light bulb limiter is now 100% ON so there is reduced voltage across the board... (more info about this below) And I REALLY don't want to desolder the IC's if I can help it...)

The thing I find really worrying about the above voltages is the 5 & 9 pins both carrying lots of + voltage (what might be +27 now but was +33 at the time..) From what I can see from the schematic, these are the IC output, 5 being + & 9 being the -

As soon as I had soldered the new IC chips in & powered up my light bulb limiter came on 100% brightness, which made me think there is a direct short somewhere (the bulb wasn't lighting up before IC were put in).

I am guessing that it would be somewhere after the IC, but I am having real trouble locating it..

I have checked all the resistors on the board with my multi meter (most in circuit, some I had to lift one leg), and the electrolytic caps with my ESR meter, the output Jacks do not appear to be shorted ether.

Any ideas?

Cheers



N

[Enzo]

But I bet there is 33v of DC on the output jacks. I sure hope you are working without loads. Please disconnect any speakers or loads and just watch the outputs with a voltmeter.

The positive and negative supplies are the same curcuits. 27v zeners regulate the pass transistors - those BDV things. You need to get both rails cooperating. The zener is 27v, but if the transistor is shorted, the zener can;t do anything about it. Make sure TR1 BDV65 is not shorted. Once you get close to +/-27v on the two rails, we can move on.

Look at the schematic. On the TDAs, pin 6 is + rail, adn pin 7 is pulled up with only 168 ohms, so we expect +rail volts on both. Pin 5 is the output, and pin 9 is the Dc feedback. SO whatever is on the output will appear on pin 9. In this case it is +rail there. That means either the - rail is not getting to the ICs, or they are driven to +rail or they are shorted.

Note on the drawing it says DO NOT GROUND. The return leg of the speaker is not grounded, so don't connect a scope ground lead there either.

So, is - rail getting to pin 4 of each IC?

It is supicious both TDAs are blown, suggestinmg something they share, like powr supply. But who knows? Also pins 2,3 on each are refernced 6.8v + from -rail by a 6.8v zener adn a 22k resistor from + rail. Check all those parts. Make sure resistors R15,25 are not burnt open in the stability networks of both outputs.

[NickH]

Thanks for the help Enzo!

But I bet there is 33v of DC on the output jacks.

You are quite correct, +32 on both JS3 & JS4, the voltage slowly drops while you are measuring it, if that is significant..

I sure hope you are working without loads. Please disconnect any speakers or loads and just watch the outputs with a voltmeter.

Yep, no speaker or load...

The positive and negative supplies are the same curcuits. 27v zeners regulate the pass transistors - those BDV things. You need to get both rails cooperating. The zener is 27v, but if the transistor is shorted, the zener can;t do anything about it. Make sure TR1 BDV65 is not shorted. Once you get close to +/-27v on the two rails, we can move on.

TR1 BDV65 is shorted, well spotted I have just ordered a replacement..

Note on the drawing it says DO NOT GROUND. The return leg of the speaker is not grounded, so don't connect a scope ground lead there either.

Cool, sussed that & I haven't been...

So, is - rail getting to pin 4 of each IC?

Yes -28v

Also pins 2,3 on each are refernced 6.8v + from -rail by a 6.8v zener adn a 22k resistor from + rail. Check all those parts.

C19 & C9 are ok with a ESR reading of 0.26 ohms each, On the Zener's I am getting

ZD4
Cathode Anode
-26.66 -27.86

ZD3
Cathode Anode
-26.98 -27.88

Does this mean that the Zener's are hosed? I couldn't get my head round how the zener's were working here, should the output be -6.8v? that seems like a lot of voltage to dump, and where is it dumping it to?

R17 & R5 are both ok, I will desolder C18 & C2 and check in the morning (1:30am here...)

Make sure resistors R15,25 are not burnt open in the stability networks of both outputs.

These are both fine & within tolerance...

Many, many thanks for your help again Enzo!



N

[Enzo]

Those two zeners are suspect. I'd just measure across them rather then each end to ground, but whatever works for you... Lift the zeners from the board and check as diodes. Are they shorted?

If course the zeners are across pins 3 and 4 of the TDAs, so if those are internally shorted, the zeners will appear so.

As I see it, there is a 22k resistor and a 6.8v zener in series strung between +rail and -rail. SO I'd expect 6.8v across the zener. The cathode end of the zener should be 6.8v more positive (or less negative if you like) than -rail. So if -rail is -27v, then pins 2,3 of the TDAs should sit at about -20. It provides a reference for the mute and protection features of the TDAs. AT last that is how it loos to me.

[NickH]

Thanks Enzo,

Sorry for the delay, it took a little while for the new parts to arrive..

Both Zener's were Fubar & I have replaced...

I am getting far more sensible voltages now:-

These reading were taken with the TDA1514A Chips removed

IC 1
1 = 0
2 = -21.27
3 = -21.27
4 = -28.03
5 = +26.18 (Dropping)
6 = +28.22
7 = +28.22
8 = 0
9 = +26.10 (Dropping)

IC 2
1 = 0
2 = -21.33
3 = -21.33
4 = -28.05
5 = +25.91 (Dropping)
6 = +28.25
7 = +28.25
8 = 0
9 = +25.89 (Dropping)

So the input voltages are happy, but the output is still wrong.

Looking at the schematic, with the chips removed the only component I could see that would cause this would be C24 (& C15)

I re-tested it and it seems ok, with the following readings:-

C24
211.2uF
0.20 ohm ESR
6.644 M ohm resistance

But just for kicks I powered it up without the C24 cap in & low & behold the +25v DC voltage is gone from the output of that channel!

Is it possible that the C24 is passing the tests but is failing under load?

Many Thanks



Nick

[NickH]

but the output is still wrong.

Looking at the schematic, with the chips removed the only component I could see that would cause this would be C24 (& C15)

I re-tested it and it seems ok, with the following readings:-

C24
211.2uF
0.20 ohm ESR
6.644 M ohm resistance

But just for kicks I powered it up without the C24 cap in & low & behold the +25v DC voltage is gone from the output of that channel!

Is it possible that the C24 is passing the tests but is failing under load?

Also, while I remember.. The schematic shows these caps should be 40V but the caps that are in there are 35V, is this significant?



N

[Enzo]

C24, C15 are responsible for that drifting voltage on the output, but I doubt they are bad. Normally there is a power amp chip keeping things straight. Plus if you stick a load on those outputs, that voltage would disappear. The super high impedance of your meter is what is causing them to slowly discharge. Well slowly charge, actually.

I'd say the thing is ready to try a new TDA or two. Try just one if you are timid.

And this is exactly where a variac and current metered mains connection are meant to be. Or lacking that, the light bulb limiter.

[NickH]

C24, C15 are responsible for that drifting voltage on the output, but I doubt they are bad. Normally there is a power amp chip keeping things straight. Plus if you stick a load on those outputs, that voltage would disappear. The super high impedance of your meter is what is causing them to slowly discharge. Well slowly charge, actually.

I'd say the thing is ready to try a new TDA or two. Try just one if you are timid.

Spot on Enzo! It's now working fine!

Many Thanks Again! Your advice was invaluable!

I think the important lesson I have learned from this one, is sort the supply voltages & don't move on until they are right!



N

[Enzo]

Amen, circuits don't work when their power isn't right.

[chriscool]

What is part# for the diodes?

[mullytron]

Amen, circuits don't work when their power isn't right.

What does it mean when the power is right (+22, -21.8), and the TDA is still intermittent? I'm working on a 40V/8040, no bad solder joints, all the operating voltages look good, the unit worked intermittently (played fine, then cuts out to a whisper) when it came in, so I played it a bunch and poked around until I got it to cut out. Couldn't find any bad diodes, etc etc, so I swapped in a new TDA and it played fine for 10 minutes, now it's back to whispering at me. All controls work, all the gain structure is there up to the output, but there's no volume. Did I roach it? They're only 8 or 10 bucks, but that was the last one my local parts house had.

[DrGonz78]

Wow this is an old thread...

What does it mean when the power is right (+22, -21.8), and the TDA is still intermittent?

Probably means that might not be what is intermittent. How can we confirm that the TDA1514A's are blown? Have you measured if there is any DC on the outputs? If there any DC on the output then a light bulb limiter is needed. Have you verified the voltages with the TDA removed, measuring all the pins? Although it seems you have confirmed that the voltages are correct, so why assume the chip is the problem? When you say that you poked around and got the amp to fail... Was that only a coincidence or did you hit on some component that made it fail? I guess I just have more questions than any answers.

Edit: One more question... Did you clean the output jacks to be sure they were not dirty?

[Enzo]

This is not precision circuitry, these are guitar amps. By bot right, I mean we have to have good powr supplies. The difference between 22 and 21.8 is 2/10 of a volt. What is that, a 1% difference? I expect the V+ and V- to be more or less the same, and fairly clear of ripple. I don't really care if they are 21v or 24v so much as both are within a volt or so of each other. And clean.

But if you are troubleshooting a system, and you have +22 and -12, THAT is a problem, or your -15 is there but +15 is missing. SO anything you try to do to an amp under those conditions will be meaningless. One has to get the power supply right before anything else can happen.


You have an 8040, not the 8004 of this thread? Very different amps. How did you decide the TDA was at fault. That seems less likely to me.

When this happens, turn any reverb down, and ball up your fist and whack the top of the amp. Does that wake it up or does it react at all? If so, some connection is loose. And the VERY common loop jack trouble. PLug a spare cord from FX send to FX return. ANy help? If so, clean the cutout contact on the FX return jack.

Not that simple? OK, isolate the problem. Turn the reverb up half way, and rock the amp to crash the reverb springs. Is that loud and clear or is that also weak and low? The reverb return is just before the power amp IC, so it is a good test of the power amp itself.

Plug the guitar into the FX return jack, is that clear and strong or is it weak and low?

Plug the guitar into the regular input, and connect the FX send to some other amp. Is that FX send sound clear and strong over there or is it low and weak?

Are both channel equally affected?

The speaker does not return to ground, it returns through a low value resistor. LOoks like R65 maybe, and maybe 0.1 ohm 2W? Hard to read from here. Check the solder on that part.

And feed a steady signal into the amp. I use a feed from my shop stereo, but anything that frees your hands. Now use an insulated probe like a wooden chopstick, and go along the board, pushing against each part, and also pushing down on the board here and there to introduce some flex. Looking for anything that reacts.

And of course reflow solder on all the controls and jacks.

You have a line out that is derived from the power amp output. SO while it is working, listen to the line out, then when the amp fails, again listen to the line out. If the line out remains OK, then the trouble must be in the speaker and its wiring and connections. If it fails along with the speaker, then the problem is farther back.

And what Gonz said. To change the power chip, you have to go onto the board, moving all manner of stuff around, and when you put it back together, it might work, but I suspect if you just reinstalled the old chip, the same thing would happen. In other words, often it is the ACT of changing the chip rather than the new chip that brings it back to life for a short time.

[mullytron]

Thanks fellas.

Amp came in, customer said no output. I plug in and play, amp works fine. After 10min amp cuts out and whispers. I poke around with my chopstick, amp still no good. Reflow all jacks and pot joints, check tube, check all operating voltages, check diodes, check bridge, etc. All good. I found a TDA locally, install, amp works fine! After 10min, amp cuts out and whispers.

So I am thinking the original TDA was ok and there's an intermittent somewhere, but I just haven't found it. Either that or I am the one somehow breaking the TDAs... There is .2VDC on the output, I'm thinking that's OK, just offset, not a problem?? Reverb return is just as quiet as everything else, even cranked, causing me to suspect something in the TDA area. Will try the send trick.

[Enzo]

And the line out test?