Fender Princeton '65 Bias (?) Problem

[John R Smith]

Since I'm in the same time zone I'll chip in, although Tom is guiding you expertly.

Reconnect P18, keep the tremolo off and take three sets of voltage readings at TP19, C22 and TP13 with the 'intensity' control full Clockwise, in the middle and fully anti-clockwise. I'm thinking the control/ribbon/pcb might be breaking down once voltage is applied to it, unlikely though that seems. The readings will tell us which way the current is flowing.

Here are the results, Nick -

P18 connected, V4, V5 and V6 removed.

Intensity control full clockwise - TP19 -31.6, C22 -12.3, TP13 -12.4

Intensity control mid position - TP19 -31.6, C22 -12.4, TP13 -12.6

Intensity control full anti-clock -TP19 -31.7, C22 -12.5, TP13 -31.7

I hope this means something to you By the way, I now have a pack of five C22 type caps so I can replace it if necessary.

[52 Bill]

Replace the bias capacitor?

I don't know if I'd discount this idea yet, with no load even a bad cap will get the voltage up to that level. Isn't the Zener rated at 56 volts or something like that.

I'd try a cap in parallel with it just to see what happens.

[nickb]

Here are the results, Nick -

P18 connected, V4, V5 and V6 removed.

Intensity control full clockwise - TP19 -31.6, C22 -12.3, TP13 -12.4

Intensity control mid position - TP19 -31.6, C22 -12.4, TP13 -12.6

Intensity control full anti-clock -TP19 -31.7, C22 -12.5, TP13 -31.7

I hope this means something to you By the way, I now have a pack of five C22 type caps so I can replace it if necessary.

Yes it does mean something. It confirms what we had already surmised - that current is flowing all the way through the 250k pot and through P18 pin 1. But we pretty much ruled out C22 so the next most likely thing is the current is flowing through your meter. Tell me what is the make, model of your meter and what range did you have it on when you made the C22 measurement?

[John R Smith]

Right. My meter is a UNI-T UT2002 and it was set to the 200V DC range when readings were taken. I also have a UNI-T UT30B and I get the same results.

[Tom Phillips]

Given the current status I'd be trying the following:
1) Clip additional capacitance across C25, the bias supply filter, to rule out a problem with that part. Any cap of 25uF or more with a voltage rating of 50V or more will due. No need to remove C25 for this test.
You can also make a capacitance reading of C25 while it is in the circuit provided that everything is powered down and discharged. Because of the parallel resistance in the circuit the meter will report a higher reading than if the capacitor was measured out of circuit. You may read 150uF or more. This is just a confidence test. If you read significantly less than 100uF then we will suspect that C25 is bad.
2) Lift the side of C22 connected to P18A-1. This will isolate the tremolo circuit even when P18 if plugged in. This does mean, however, that you will be committed to working on the PC Board but I think you are already there.
3) Another open item is a good physical inspection under the PD Boards. A pain to do but maybe there is something strange going on there. CAUTION: If you do fold back the main board you need to minimize the movement & flexing of the flying wire leads that connect to the tube sockets. Those wires are not strain relived at the PCB and can easily break off if flexed too much.

Report on bias voltage readings as these tests are done.
If we find something, have a strong alcoholic Christmas drink to celebrate progress.
If we don't uncover any new findings, have a strong alcoholic Christmas drink to console yourself.
Edit: Drinking starts only after amp work is completed for the day. Brilliant ideas conceived during drinking may be written down for reference the next day.

Cheers,
Tom

[Tom Phillips]

Right. My meter is a UNI-T UT2002 and it was set to the 200V DC range when readings were taken. I also have a UNI-T UT30B and I get the same results.

Those meters have a 10 M Ohm input impedance spec for all DC ranges. Therefore, the meter should not adversely load the circuit nor affect the readings.

[nickb]

Right. My meter is a UNI-T UT2002 and it was set to the 200V DC range when readings were taken. I also have a UNI-T UT30B and I get the same results.

The UT2002 has a 10Mohm input impedance so that cannot be where the current it going so that was another dead end.
Edit:I see Tom just beat me to it!

[Jazz P Bass]

My best guess is that when the PCB is flipped over, you will find that the board was soldered by a 2 year old.

I have seen it often enough.

Reflow a main PCB: 1 yard of .031 solder.

[Tom Phillips]

My best guess is that when the PCB is flipped over, you will find that the board was soldered by a 2 year old...

This brings to mind a procedure that you can employ when checking installed components. After you measure the component by clipping directly to that component's mounting leads move the test leads to another circuit node which should be electrically equivalent to the point where the component lead is installed. For the type of measurements that we are making the reading taken directly on the component leads and the reading taken at the other points should be the same. Some examples for our troublesome Princeton circuit:
The negative side of C25 = the P18A-3 pin.
The positive side of C25 = the R2 lead that is grounded. (And many other ground points)
You can devise many additional examples once you get the idea.

By employing this technique you are verifying the integrity of the circuit path in addition to the value of the component. The technique is especially useful when you don't have access to the back of the PC board.

Tom

[nickb]

My best guess is that when the PCB is flipped over, you will find that the board was soldered by a 2 year old.

I have seen it often enough.

Reflow a main PCB: 1 yard of .031 solder.

A bad joint would explain a lot.

[John R Smith]

Well, I trust that you all enjoyed a very good Christmas. Here is the latest update on the Fender Princeton Reverb ‘65 reissue problem, and I think I have made some progress but I am still nowhere near finding the cause of the fault. Just to recap -

The amplifier has an intermittent fault where the bias voltage to the power tubes goes high (from –45V to –31V) and the tremolo stops working. When this happens the power tubes red-plate and everything starts to smell very hot. Testing so far has therefore mostly been with the power tubes V5 and V6 removed.

On Christmas Eve I took stock. The last suggestion from Tom was that I should look again at C25 in the bias supply circuit, but C25 is mounted vertically and so tight to the board that there is no way to test it from the top. At this point we still don’t really know for sure whether the problem is in the bias supply itself or in the trem circuit, so I decided to do something more drastic -

* I first checked that we had the fault condition, with –31V at TP19 with P18 plugged in. At TP13 we also have –31V in this condition.

* With P18 unplugged we have –45V at TP19, but of course no voltage at TP13. So we cannot run the amp in this state.

* To enable running with all the valves installed and P18 unplugged I then powered down and soldered a flying lead from the hot side of the bias pot R22 directly to one end of R18 (which is TP13). With P18 unplugged I powered up, checked the voltage at –45V at TP13, powered down, installed the power tubes V5 and V6, and bought the amp up on test with speaker and reverb hooked up.

* Everything was fine, I ran it like this all evening (3 hours or so), played lots of guitar through it, everything worked except of course the tremolo.

* From this I assume that there is nothing much wrong with the bias supply circuit and C25. I had a glass of vintage port.

The next step was to confirm that I was still testing the amp in the fault condition.

* I powered down, removed V5 and V6, plugged P18 back in, powered up and tested TP19 at –31V. So the fault was there still and it looks pretty certain that it is in the tremolo circuit.

Then of course Sod’s Law swung into action. While I was testing some more stuff around the tremolo circuit (TP25 and TP29 which were OK) something made me check TP19 again and it was –45V. I unsoldered my flying lead but it was still –45V. So of course the fault had decided to take Christmas off.

On Christmas Day I re-installed V5 and V6 with P18 connected, checked the bias and everything was fine. I had it up and running for a couple of hours both in the morning and the afternoon, played loads of guitar through it, everything worked great including the tremolo. I have had it on soak test all day today with the power valves out (so I can walk away and leave it) and so far I cannot get the fault to return so I can’t do any more testing. I have tapped and banged everything on the board and even lifted each end of the chassis 3 inches and dropped it several times to try and dislodge a flaky connection or track but it’s running just fine. In despair I have had another glass of port.

The good thing is that I have been able to take some voltage readings with the circuit running correctly. The big difference is that, other than the correct –45V at TP19 and TP13, the reading at P18-1 is quite different. In the fault condition, the reading here is –12.5V steady. When working correctly, the reading at P18-1 cycles around and around from –55V to –33V which of course is my meter sampling the effect of the oscillator on the voltage. So there should be a clue here – when the fault kicks in it’s as if the oscillator stops working and voltage on this side of intensity pot (P18-1) goes up to –12V when normally it is never more than –32V. And when the fault happens the tremolo does indeed stop working.

Well, sorry this has been such a long-winded post. At the moment I am just waiting for the fault to kick in again, and of course as things are I daren’t put it all back together.

Regards, John

[Jazz P Bass]

I believe that you indicated that you now have spare 0.1uf capacitors.
If so, have you replaced C22 yet? (it really sounds like that cap is duff.)

When you get this animal licked, it will be a call for more vintage port.

[Tom Phillips]

...C25 is mounted vertically and so tight to the board that there is no way to test it from the top...

The technique I describe in post #69 will allow you to do this test. I think that it is unlikely that C25 is the problem but we are just keeping an open mind to all the possibilities.

[g1]

have you replaced C22 yet? (it really sounds like that cap is duff.)

I had missed the part about the trem quitting when the fault occurs. This makes me even more suspicious of C22. If C22 was ok and the bias fault was elsewhere, I don't think it would stop the oscillator action at P18-1.

[John R Smith]

I had missed the part about the trem quitting when the fault occurs. This makes me even more suspicious of C22. If C22 was ok and the bias fault was elsewhere, I don't think it would stop the oscillator action at P18-1.

Can I test C22 by strapping another .1 cap across it as Tom suggests for testing C25?