Concert Lead Power Amp Problem

[george]

http://i284.photobucket.com/albums/ll33/bonosurf/conLeadPA.jpg

The issue is Q3 is getting very hot at turn on with no signal input or load. R303/304 get really hot (+200F) in about 30 sec.

I don't measure 1/2 VCC at emitter - only around 4.6V with the base at 5V. Supply voltages correct and all new transistors in the output but only Q3 gets hot at idle (yes I swapped another device and it still gets hot). Bias trim pot is measuring correct resistance but has on change on the issue. Something is wrong with the bias but I can not see it - changed or inspected all components on this board 

This poor amp was a pig's breakfast when I received it - someone pulled off a stack or dropped it and then some tweeker tried to "fix" it.

[loudthud]

There could be a short in the driver transformer. Easiest way to check is disconnect the primary wires and ohm the primary to ground. I've only seen shorts from primary to one of the secondarys but there could be a short between two secondarys.

Remove Q2, Q3, Q4 and Q5. Power the amp up with no speaker and see if the output goes close to zero volts. Check the voltages at the output transistor sockets and replace any bad resistors.

Note that the output transistors should be matched for beta within 5% or as close as you can get.

[george]

The interstage transformer appears to check out okay. The primary reads 18.1 DCR and there is no short to any secondary windings that I can measure or ground. The secondaries all appear to be isolated from each other (2.5 DCR).

I'll map out the voltages at the output transistor sockets. Don't get why VE isn't half of VCC?? Almost as if the 22 ohm resistors aren't connected??

I've always used the same date code when replacing output transistors on these amps without an issue. I can test these on a curve tracer at work to match HFE but I think I'm missing something very simple

[george]

Okay I measured the voltages on the output transistors and I'm seeing an issue with the bias on all devices.

        Q2    Q3    Q4     Q5
VCE 4.6V   46V  46V    3.9V
VBE 3.4     4.8V  3.9V  4.8V
VE            4.2

Okay the voltages above are referenced from either the "-" terminal of C2 or the "-" terminal of C1.

The values for VCE should read around 23V for Q2 and Q5 but they're not. The base bias is really wrong - should be 1/2 of VCC. I didn't measure IE (didn't have time). I'm almost 100% certain that the transistors are fine - they're new. I did notice that the TO-3 sockets have loose pins (low insertion force) for the emitter and base but the devices are making contact.

Check the grounding scheme and found no issues - though it looks like this design lands chassis ground at the output connector. All the bias resistors check out good and I replaced a few for good measure.

Frustrating since this is a simple common emitter circuit - where's the problem? It has to be simple but I don't see it 

[loudthud]

Frustrating since this is a simple common emitter circuit - where's the problem? It has to be simple but I don't see it   

In cases like this I try to go back to the basics. Do the rails balance with the transistors removed (no speaker connected)? Any one bad connection to B or E could cause problems. Repair or replace those sockets or just solder the parts in. Is there any possibility one of the driver transformer leads is going to the wrong terminal on the PCB? Once things start working without the transistors, try it again with the transistors installed but still no speaker. Try to get the DC on the output as low as possible.

Okay the voltages above are referenced from either the "-" terminal of C2 or the "-" terminal of C1.

This confuses me. You realize the minus side of C1 isn't grounded right? Ground is the E of Q3 and C of Q2.

       Q2    Q3    Q4     Q5
VCE 4.6V   46V  46V    3.9V
VBE 3.4     4.8V  3.9V  4.8V
VE            4.2

Again you are confusing me. The voltage between B and E should be around 0.6V

[george]

Yup I'm going to replace the transistor sockets - just found a good local source here in town.

Yikes I thought node 21 was ground but you're right.

What should VC read on Q2 and Q5?
I'm assuming that VB should be 1/2VCC on Q2-Q5 correct?
In other words should I reference all my voltage/rair measurements to node 33?

Thanks

[loudthud]

This is a bridge drive circuit with a floating power supply. Quite advanced for it's day.

The rails are generally around +/- 25V or thereabouts. Any problem with the bias resistors or transistor matching will upset the balance. That's why I suggest making sure everything is good with the transistors removed. Once that is ok you can install the transistors and swap them around to try to minimize the DC on the output.

[george]

Loudthud thanks for the insight - here's where I'm at now:

I have removed all the output devices and R304 and R303 are still heating up. I did find a short that the previous "tech" had snuck in - connecting node 2 to 4. The VC (collector voltage) on all the output sockets is now correct at 1/2VCC = 48/2= 24V. I replaced all the bias resistors on Q2-Q5 for good measure (some had drifted) but that's not the fix.

What else should I look for to see if the bias is balanced? Since the secondaries on the interstage transformer are floating - I'm assuming that, unless there is a short to ground on S1,T1 or T1,S2, they don't figure into this. Yes it did run a continuity check on the above secondaries and they don't appear to be shorted to ground. What else he asks??

[loudthud]

Node 2 to 4? that should really light things up, but not in a good way. It's time to check the 0.27 ohm resistors and install the transistors. Turn the amp on with a variac or light bulb limiter and see how much DC is on the output without the speaker connected. Swap the transistors around to see if you can get the voltage as low as possible. I've never seen a spec from Sunn but try for under a volt of two. The amp should be ready to go.

[george]

1) DC Voltage across
R303  = 24V
R304  = 24V
R301  = 24V
R307  = 24V
R302  =.58V
R305  =.58V
R322  =.58V
R306  =.58V
 
TP 16/17 =24.5V
TP 12/21/22 = -24.6v
With amp on, no signal, no speaker, no transistors

R303/304/307 are all heating up to >100F in about 45 sec. Something is still wrong and installing the output transistors doesn't make sense until I understand what's going on. Could C307 be the problem (though it would have to be a short)?

[loudthud]

Those 910 ohm resistors could get a little hot. With 24V across them they are dissipating .63 watts. Isn't R301 getting hot? They should heat up fast but stay at that temperature and not burst into flames or melt the solder. They may get slightly hotter when you play the amp. Nothing to worry about. I always mount resistors like this up off the circuit board for good air circulation and so they don't turn the board black.

There is very little voltage across C307. If it is shorted, R323 will burn up when a signal is coming through the amp.

The voltages look good. Did you replace the sockets yet? The amp should work now.

[george]

Replaced sockets and I have a set of new On Semi 2N3055's that I matched on a curve tracer at 1A/20V (use pulsed mode for collector supply and work FAST) - I'll install and give you the results. I've repaired about half a dozen of this series of Concert amps but never noticed that the bias resistors were this hot. Thanks