Carvin AG100D power supply issue

[Mikestorm]

Hi to all members
I've got a Carvin AG100D, an acoustic guitar amp, with a power supply issue.
The amp powers on, status led is flashing (I hear a ticking noise on the board at the same freq of the led flashing), output voltages from the regulators are lower than expected and unstable.

I've checked if the 78xx and the 79xx on the circuits were shorted to ground, but nothing. Their in/out are not shorted.
Attached is the schematics of the power supply part.

Anyone got a similar issue?
Any idea or tips where to start troubleshoot and nocking down where the problem is?

Thanks in advance
Mike

[The Dude]

Do you have a scope? If so, see if the supply is pulsing- starting/stopping. My bet is that the supply is trying to protect/regulate itself and that possibly there is a short somewhere- maybe an output device.

[dmeek]

Look at the electrolytic capacitors in the power supply. If any are even slightly bulged on top, replace them. Use low ESR type capacitors

[Mikestorm]

Do you have a scope? If so, see if the supply is pulsing- starting/stopping. My bet is that the supply is trying to protect/regulate itself and that possibly there is a short somewhere- maybe an output device.

Thanks for the reply
I've got a cheap dso scope, should do the trick. Where in the circuit you suggest to look for?
by the way I've disconnected all secondary boards(preamp and dsp) and still got the issue

[Mikestorm]

Look at the electrolytic capacitors in the power supply. If any are even slightly bulged on top, replace them. Use low ESR type capacitors

Thanks for the reply

All the electrolytic seems to be fine, not bulged and testing them with continuity they seems to be not shorted

[The Dude]

Thanks for the reply
I've got a cheap dso scope, should do the trick. Where in the circuit you suggest to look for?
by the way I've disconnected all secondary boards(preamp and dsp) and still got the issue

That is another story, then. Without the complete schematic, I can't be sure, but it appears that there are some feedback lines from another circuit that attach to the PS primary circuit. It may be that the supply won't run without the rest of the unit connected. Please post the entire schematic.

[Mikestorm]

That is another story, then. Without the complete schematic, I can't be sure, but it appears that there are some feedback lines from another circuit that attach to the PS primary circuit. It may be that the supply won't run without the rest of the unit connected. Please post the entire schematic.

You were right, the problem is in the power amp section. Just for test I desoldered the 2 power amp transistors, and the issue is gone.
Offcourse only the preamp and dsp were working, but the power supply was working fine.
The two transistors are bdw84c and bdw83c. Q1 and Q12. Found the bdw83 bad, replaced, but after less than 30 seconds the issue came back(I think I burnt the new bdw83...).

Need to check the power amp area.

[J M Fahey]

You must find what is burning it.
And by the way, I do not trust its partner so replace both.

Turn amp ON, measure +V , -V and speaker out rails.
Also what you measure at each of 3 transistor pins to ground, so E 1-2-3-4-5-6

Since SMPS negate the use of Variacs and lamp bulb limiters for testing since they MURDER to get output voltages "right", I suggest you locate traces leading from supply +V and -V sources to amp +V and -V points and cut them at some convenient point with an X Acto type knife.

Scratch solder mask both sides of each cut to expose copper and bridge cuts with 2 100 ohm 1W resistors, which now become your current limiters.

Have a good supply handy because these will now become your "Fuses"/current limiters and will often burn.

Canīt believe I am using this ooooolllllddddd trick from my earliest days but power supplies that are more intelligent and stubborn than me drive me crazy.

For now, measure the voltages suggested above without them (unless you have a couple handy) but when you order fresh power transistors add them to the order.
Say 20 of them for future use, they are the "new normal" as far as supploycurrent limiting goes.
Cost is peanuts.

HOPE you bought transistors from a respected supplier (Mouser, Digikey, Farnell) and not EBay or AliBaba .

Also fill in your location data, suggestions depend on that.

[The Dude]

^^^^^^ yes, all of that, especially, "And by the way, I do not trust its partner so replace both". Also check driver transistors Q6,Q8, and Q11. If the amp was momentarily putting out DC before it shut down (likely), it could have also damaged Q9, so check that. Also check emitter resistors R113 and R114. The thing about direct coupled amps is that anything left uncorrected can upset the apple cart and cause the smoke to come out again. It's best to take your time and be thorough the first time. I will usually take a few minutes and check all solid state components in the circuit for shorts. It doesn't take that long. Then, if you find shorted components, look for open resistors that are in the current path of those shorted devices.

[Mikestorm]

^^^^^^ yes, all of that, especially, "And by the way, I do not trust its partner so replace both". Also check driver transistors Q6,Q8, and Q11. If the amp was momentarily putting out DC before it shut down (likely), it could have also damaged Q9, so check that. Also check emitter resistors R113 and R114. The thing about direct coupled amps is that anything left uncorrected can upset the apple cart and cause the smoke to come out again. It's best to take your time and be thorough the first time. I will usually take a few minutes and check all solid state components in the circuit for shorts. It doesn't take that long. Then, if you find shorted components, look for open resistors that are in the current path of those shorted devices.

I've pulled out the two power transistor(Q1 from my stash to replace the old bad one, and Q12, the original one).
Voltage at the power transistor pins are:
Q1) E:101/103mV, C: 46.7V and B: unstable from 28V to 30V.
Q12) E:100/103mV, C:-45.04V and B: unstable from -28V to -30V

As you suggested I've checked R113 and R114, both fine, and seems that Carvin at last opted for 0.47ohm ones, instead of 0.33ohm
Q6, Q8, Q9 and Q11 all fourt are not shorted to ground and seems that pins are not shorted together.
But I can see that 100mV on the output, without the power transistor, should say to me that something is giving me 100mV dc.
On similar types of transistor based fender amp, usually the max dc that is considered "accepted" is 15/20mV.

So I measured voltages accross R117, and thosa are -45V on the V- rail, and as a dropping voltage, the other side I see -42.8V. So at Q9 I see 99mV on emitter and base, and -42.8V on collector.

As another member and you too, suggested to replace both, I was searching on Mouser and Digikey, where usually get my parts, but seems that the BDW84/83C are obsolete, so, need to find a substitute, maybe beefyer. Maybe a Tip 142/147 will work fine, any tips?

I was thinking on pull out Q9, to see if the 100mV dc desappers, bytheway searching what is Q9(the majority on board are smd), it should be a multipurpose pnp, like the mmbt5401, so maybe momentary I can use a similar non smd replacement.

What you think guys?

[Mikestorm]

You must find what is burning it.
And by the way, I do not trust its partner so replace both.

Turn amp ON, measure +V , -V and speaker out rails.
Also what you measure at each of 3 transistor pins to ground, so E 1-2-3-4-5-6

Since SMPS negate the use of Variacs and lamp bulb limiters for testing since they MURDER to get output voltages "right", I suggest you locate traces leading from supply +V and -V sources to amp +V and -V points and cut them at some convenient point with an X Acto type knife.

Scratch solder mask both sides of each cut to expose copper and bridge cuts with 2 100 ohm 1W resistors, which now become your current limiters.

Have a good supply handy because these will now become your "Fuses"/current limiters and will often burn.

Canīt believe I am using this ooooolllllddddd trick from my earliest days but power supplies that are more intelligent and stubborn than me drive me crazy.

For now, measure the voltages suggested above without them (unless you have a couple handy) but when you order fresh power transistors add them to the order.
Say 20 of them for future use, they are the "new normal" as far as supploycurrent limiting goes.
Cost is peanuts.

HOPE you bought transistors from a respected supplier (Mouser, Digikey, Farnell) and not EBay or AliBaba .

Also fill in your location data, suggestions depend on that.

I took measurements on th E's point after pulling out the power transistors, and are the followings:
E1: 46.7V; E2: unstable between 28 and 30V, E3: 103mV; E4: 103mV; E5: unstable between -28.9V and -30V and E6: -45.04V.

As I mentioned to The Dude, I usually buy parts from Mouser or Digikey. The BDW83C I replaced(didnt get fried during testing), is a brand new that I found in my parts, and if I recall right, it was bought years ago for another project from Digikey.
Unfortunatelly those seems to be obsolete, and I'm checking both datasheets if can be replaced with Tip142/147 pair...

Any tips and idea are wellcome

[The Dude]

The MJH6284G and MJH6287G would be good crosses. That said, I wasn't able to find anyone with the latter in stock.

Also, Digikey lists the TIP142 and TIP147 as crosses, although the data sheet says its max collector current is 10 amps instead of your original 15 amp devices.

[Mikestorm]

The MJH6284G and MJH6287G would be good crosses. That said, I wasn't able to find anyone with the latter in stock.

Also, Digikey lists the TIP142 and TIP147 as crosses, although the data sheet says its max collector current is 10 amps instead of your original 15 amp devices.

I've lifted Q9 from the board, and still have 100mV dc on the output, the only one thing on that path(power transistors arent on board, waiting for the tip142/147 from mouser) is the base of Q4, the transistor of the long tail pair that manage the feedback.

At this point, if I lift Q4, and there is still present the dc on the output, really dont know from where those 100mV come from.

Any ideas?

[Enzo]

I don't generally concern myself with 100ma on the load at idle. Does it change if you present it a load?

[Mikestorm]

I don't generally concern myself with 100ma on the load at idle. Does it change if you present it a load?

Hi Enzo
thanks for reply and the help

If I connect the load, across the speacker I measure 0mv

I took some measuraments of voltages, with speaker connected, without power transistors and without Q9 on board.

Checked all resistors in the power amp section, and they all seems fine, accordin to the schematics, and checked all transistors too, and none of them are shorted together.