Carver FTM-25

[lowell]

Thanks Jazz. So I can leave the outputs in and ADD a 1k 1watt in parallel with b/e? Or should I remove one of them?

I have the output ballast resistors lifted. Could this be why those resistors are burning up? There is no current through the output bases right now. This could be my misunderstanding with solid state vs tubes.

[Jazz P Bass]

The idea is to remove all of the outputs & simply bridge the B/E junction of each rail to complete the circuit.

Mind you, this is only for testing.

[lowell]

So to be super clear. Disconnect Base, Emitter, and Collector of each output transistor. Then solder in a 1k/1w resistor from Base to Emitter. ? (ignoring any paralleled devices)

I think that's what you are saying.

Finally, am I right in that this is the error in my thinking? Current needs to flow from Base to Emitter in order for the previous stage to work correctly? So if the output bases are conducting, the bias resistors for the Vbe multiplier for the outputs would not be attempting (and failing) to carry the current that would otherwise be going thru the output's bases?

[Jazz P Bass]

Not too sure what is going on with the amp.

If you are afraid of blowing the outputs, taking them out of the circuit & inserting a 1K/ 1 W resistor B to E will take awy that fear.

The resistor simply completes the circuit.

For sure I would remove the limiting transistors.
They shouldn't even come in to play until you are at full power.
It does seem that they may be the reason for the overtemp in the two resistors that you mentioned.

[lowell]

Ok I pulled the limiters completely.
Disconnected all 3 outputs on each side. Added 1k/1w B/E resistors. All voltages listed are in reference to ground with no speaker connected.
With bulb limiter and variac I was then able to bring up the voltage to where I could take some measurements.
I ended up finding that R209 (Re) for Q201 was open. Replaced it.
Then found that it was smoking again, this led me to find that Q237 had a short. Replaced it.
OK, NOW WE'RE GETTING SOMEWHERE
I now bring the variac up to about 18v and have bipolar supplies throughout the PA.

Q201
C +5v
Q203
C -5v
Q207
E +4.4v
Q209
E -4.7v
Q211
E +4v
Q215
E -4v
Q217
B (will not go above 200mv)
Q219
B (will rise to -3 volts and more at which point I turn the dial back down)

So the Base of Q219 is going too high.
R243 and R245 are still smoking.

TO SUM IT UP - I have bipolar voltages throughout, albeit HIGH voltages. I turned VR201 bias pot as cold as I could get it and the voltages posted are as cold as can be.

At this point I've checked everything I could think would cause this. My guess is that Q205 may not be conducting enough...or at all.

Maybe at this point I should reinstall the outputs? I'd think the base voltage could not exceed ~700mv if they were installed. Or should the base voltages be in spec wirh the 1k resistors installed?

[lowell]

Can I put a 100ohm 2watt resistor acrosa the Vbe and see if that fixes it? I have a feeling the vbe isnt conducting regardless of my meter saying it's ok. I figure there's no more than about 100ma thru it when working and I need about +/-4v at the current mirrors' collectors.

I'm guessing a little at the math. But if the current mirrors are around 100ma, and the rail is +90v that means the collectors should be at .6% of 90 = 4v.

4v/100ma = 40ohms. But it is from rail to rail so 80ohms or 100ohms and 8v is more like it.

8v X .1 = .8watt so a 2w will suffice.

Am I crazy? Or onto something?

[lowell]

Ok no one is into this thread haha. Understandable cause I feel like I lost everyone with my post updates. ANYWHO -

I replaced the Vbe Q205 (already replaced a shorted Q237) and the AMP IS WORKING now... with the 1k/1w resistors in.

I just put the outputs back in and things are skewed a bit. The base of Q219 PNP output is at +100mv. Q217 is on point at ~500mv. I traced this all back to the collectors of the current mirrors. They are about 1 diode drop difference.

My QUESTION: Do these mirrors have to be complementary pairs? I checked their Vbe and the rail voltages. Both identical. I'm thinking Q201 is pulling more current. I replaced both of them with suitable subs, but need to know if they should be complementary pairs. THANKS!

[MarkusBass]

Ok no one is into this thread haha. Understandable cause I feel like I lost everyone with my post updates. ANYWHO -

My suggestion is that you post your questions in a different way. When you list 20 voltages (and the voltages change all the time because you measured them incorrectly, or shorted the amp), it is to difficult (at least for me) to follow what is the actual problem. I have to open the schematic, search for every component you listed and try to verify whether the voltage is correct or not. Instead of such approach, I suggest that you make a copy of the schematic and list the voltages directly on the schematic. And attach this schematic as image. This can be done with almost any graphic program, even as simple as MS Paint.

I replaced the Vbe Q205 (already replaced a shorted Q237) and the AMP IS WORKING now... with the 1k/1w resistors in.

The question is how it is possible that you haven't look at Q205 until now .

I just put the outputs back in and things are skewed a bit. The base of Q219 PNP output is at +100mv. Q217 is on point at ~500mv. I traced this all back to the collectors of the current mirrors. They are about 1 diode drop difference.
My QUESTION: Do these mirrors have to be complementary pairs? I checked their Vbe and the rail voltages. Both identical. I'm thinking Q201 is pulling more current. I replaced both of them with suitable subs, but need to know if they should be complementary pairs.

You haven't mentioned what current mirrors you are talking about . Do you mean Q201 and Q203? They are not current mirrors - they are current sources (which is not the same as current mirrors). If you have 1 diode drop difference, I would immediately look at D201, D203 (D202, D204) - is one of them shorted?

Mark

[lowell]

Mark,

Thanks for the reply.

My main question is unanswered. Should the current sources be complementary pairs?

I believe I had stated already that Q205 measured correctly all along. Only after replacing it did the bipolar voltages come back.

D201/203 are not shorted. D202/204 are in the other "working" channel. But D205/207 are for the negative rail current source and are also not shorted.

Thanks!
I agree about posting the schematic with voltages.

[lowell]

Also, can you suggest some references for advanced troubleshooting of direct coupled power amplifiers?

[MarkusBass]

My main question is unanswered. Should the current sources be complementary pairs?

I haven't answer to this question because you asked about current mirrors and you didn't specify which transistors you were talking about. If you want to find out the answer, go to wikipedia: https://en.wikipedia.org/wiki/Current_source and read it. However, I think that the question is irrelevant because you have differences like 100mV and 500mV. I don't think that such a difference can be caused by unmatched transistors. The current provided by current source depends on voltage drop on the diodes, and on emitter resistor. Are you sure that the diodes are OK and the resistor is OK? Basic test would be measuring the voltage drop on each diode and each emitter resistors. Are they the same?
I cannot recommend and book on advance troubleshooting of such amps but I think that required knowledge is available for free on the internet (eg. on wikipedia). You already know that you have to identify functional blocks of the amp, understand how each of them works and verify whether actual behavior is the same as expected. And if not, find out why.

Mark

[lowell]

In my previous post I stated that the current "sources" had identical and good Vbe and rail voltages. It wasn't so clearly stated though I admit And to be clear I'm saying that yes those diodes are all reading good and have good diode drops.

Do these mirrors have to be complementary pairs? I checked their Vbe and the rail voltages. Both identical

Yes the Re resistors for them both have identical voltage drops across them too and the resistors are good.

I just discovered that Q215 is reading suspect. I lifted it's biasing resistors R243/245 and still read very low. Pulled it from the board and still very low. I had already replaced this so it seems during troubleshooting that I did AFTER that repair it has fried again. Replaced it and now the output bases look good!

I brought it up to full voltage without the limiter and voltages are good.

I am playing music through it right now and it sounds good too.

Thanks for your help with this Mark. As is always the case on challenging repairs I learned a lot on this one. Mostly that transistors are most likely suspects when all surrounding components SEEM to be reading correctly on the meter. Also Jazz's suggestion with the 1k/1w resistors from B/E was a big help!

I have a few other power amps that have stumped me thus far and I feel more competent so will be returning to them and will most likely have some more threads on similar repairs. I will try to be more clear with my posts. I think it'll be easier to be clear with my posts when I have more confidence in what I'm actually measuring and what my results SHOULD be.

[lowell]

The service manual specifies the biasing voltage with "no input signal." It does not specify a load. Should a load be connected when biasing? I ask because the bias changes when a load is connected.

Also what are your processes for testing a power amp after replacing transistors and checking bias? Monitoring transistor temps with infrared gun? What temp is acceptable at what wattage?

I'm aware of the deg C/W ratings. The 2SA1302 has Tj 150deg max rating. I have it running at 36 watts on just the left channel and my gun reads about 56degC at the hottest spot I can find.

56deg - 25deg = 31deg

So it seems to be raising in temp less than 1deg per watt. Does this approach work? And does this look normal?

[The Dude]

If it doesn't specify, I would bias with a load. It's normally specified.

[lowell]

Hmmm... the +Vb of the outputs doesn't go DOWN if I turn the bias pot WITH a load. The -Vb DOES go lower. In any case I'm convinced it's fine. The -Vb is a hair lower on both channels at about -450mv. I've run the thing at about 36watts at 4 ohms and the temp seemed to go up < 1deg per watt. From what I could find and have read online this seems ok to me. Ofcourse if I were to run the thing at 200watts that would not be good so now I'm wondering.