Princeton Reverb woes

[AtomicMassUnit]

I have a princeton reverb with a strange problem. Basically, the overall voltages in the amp are a good deal too low. The amp works, but sounds really nasty, it's not as loud as it used to be, and it distorts early, when it used to be very clean up until about 7 on the vol knob. It is not a stock one at all though. Here's the basic story.

It's an ebay amp that a friend of mine bought and it had a replaced power transformer, and ALL the wiring in the amp had been replaced in the most haphazard manner possible. I replaced the filter caps and other odds and ends that needed doing, and he gigged with it for a good while and enjoyed it.

When he came back after a break at a gig the amp was breaking up way early and just sounded crappy, so he brought it to me. The only thing that was off was the overall B+ voltage is 336 volts... and of course everythin after it is way off. I first checked tubes, then capacitors and resistors. I checked all the transformers, swapping in brand new ones, since I figured what else could it be but a bad power transformer? It didn't help. I sent it to the local tech, and he had the same result, couldn't find the problem.

I know this for sure... the filters, tubes, and transformers are all good. All the resistors in the power supply are spot on (new replacements). I rewired the entire amp with good wire. No changes. I get 336 volts out of the rectifier (GZ34) with the power tubes in. They're pulling only 7ma of current each. I'm getting -31v at pin 5 of the power tubes. When I pull the power tubes out, the voltage at the rectifier jumps up to 400. I've got 5.3vac at the rectifier tube heaters and 6.3vac on the other tubes. It seems like if I could just get the voltage high enough it would sound right. What would make that voltage lower?

What could be wrong with this amp?

[R.G.]

Many, many things.

I personally would disconnect the choke, then try it again with and without power tubes. Does it still drop? If so, the problem is in the output stage, if not, it's past the output stage.

What is the current through the power supply as measured by an ammeter? With and without tubes?

If 14ma is dropping your power supply by 64V, then the power supply impedance is 64v/0.014A = 4.5K, which is guaranteed to be wrong.

How did you measure tube current? Plate curent? Cathode current? Screen current?

Is the current going into the OT the same as the sum of the currents going OUT of the primary leads? The OT could be leaking to chassis.

Your ground wiring could be high impedance.

Your AC wiring could be high impedance.

Your FUSE could be high impedance. Does the AC line voltage on the primary drop when there are tubes in there?

There are zillions of things.

[AtomicMassUnit]

There isn't a choke in this amp, it's a plain old silverface princeton reverb. Could I possibly disconnect a resistor in the power supply to accomplish the same thing? It would be great to be able to at least figure out what part of the amp the problem is in, it just baffles me that I could replace every part (aside from pots and tube sockets) and still have the same problem.

How do I measure the current through the power supply and in and out of the OT? I don't think it's the OT since I swapped in a new one and it didn't solve the problem.

I measured the plate current using the shunt method on my fluke meter.

The line voltage on the primary side of the pt remains the same 118v with or without tubes.

Edit: I found something that looks wierd. On the schematic it shows the bias supply connecting to the rectifier on pin 6 and the high voltage connected to Pin 2. On this amp the bias is connected to pin 4 and the high v is connected to pin 8. I'm wondering what's going on here... Obviously the original owner had changed the rectifier socket since it looks all crazy, maybe that's what the problem is with this amp.

[Enzo]

The point of disconnecting the choke was to break the B+ path to the rest of the circuit. Lifting the resistor after the first B+ node does the same thing.

To measure current in anything, you either have to insert the ammeter in series with the load or you have to measure voltage across a series resistance and calculate current with Ohm's Law. This being the well known way we check tube current with a 1 ohm cathode resistor. To do a current reading, set the meter to DC amps, position the test leads accordingly - my FLuke uses diffferent jacks for current. Now disconnect the wire from the OT CT. Clip your meter to the free CT wire, and the other lead clip to the B+ where the CT was before. Now you are reading the current into the transformer. With tubes removed there should be no current flowing.

Likewise, you could lift the wire from the 5U4, insert the meter and read the current for the whole circuit. Lift the 1k resistor going from the first filter to the second filter, insert the meter between the free end of the resistor and where it was, read the current flowing through the rest of the circuit after the power tranny and output plates.

If the resistors are close to value, just reading the voltage across them lets you calculate the current through them with Ohms LAw. SO even if the voltages are low, if something is dragging them down, it would show.

For example, B+ is 420 volts on my print, then the second node is 400v with the 1k between. 20 volts across 1k means there are 20ma flowing through it. That means that everything in the amp except the power tube plates adds up to 20ma. Or should. Now even if the B+ is only 336v, 20ma through 1k is still 20v. SO if there remains that 20v drop, then the stuff to the left of it is drawing about what it should. On the other hand, if that resistor now has 40 volts across it, then 40ma is flowing - twice what it ought, and that would mean there was a problem over to the left.

That is the technique - or one of them anyway.

I have a concern about the power transformer. Without tubes to load it, your B+ is only 400. The print shows 420 loaded. SHould be much higher unloaded. So dropping well below 400 with tubes would seem likely.

YOu say you replaced the PT. Did you use an actual Princeton tranny? Schemo calls for 340-0-340 volts. You got that much? That ought to rectify to about 480 volts DC, then lose 60 volts in the 5U4 for the 420. What does yout HV secondary fall to under load?

[AtomicMassUnit]

The replacement I got was the generic 'princeton reverb, princeton, champ, tweed deluxe pt' from antique electronics. Now that I look at it, it says 325-0-325. Here's a link.

http://www.tubesandmore.com/new/scri...item=P-TF22772

Maybe that's the whole problem!

[Bruce / Mission Amps]

What is the actual AC voltage coming out of the PT secondary at lugs 4 and 6 of the rectifier, while the amp is on and the power tubes inserted?
I think 70ma and 325vac in this amp would still get you up in the 375vdc to 390vdc range with power tubes nominally biased.

[Enzo]

Yeah, we're only talking about 20VDC here then, not 60.

[AtomicMassUnit]

I get 350 vac on pin 4 of the rectifier with power tubes in.

I wish I had a way to record and post a clip of the amp. It has a pretty loud low frequency hum and when you play through it, low single notes sound really out of tune, like almost a flat 5th interval even on single notes.

[Enzo]

May I request you not go back and add things in edit. I just now noticed the addition at the end of post #3. Many of us don't go back and look for changes. Just add a new post.

Edit: I found something that looks wierd. On the schematic it shows the bias supply connecting to the rectifier on pin 6 and the high voltage connected to Pin 2. On this amp the bias is connected to pin 4 and the high v is connected to pin 8. I'm wondering what's going on here... Obviously the original owner had changed the rectifier socket since it looks all crazy, maybe that's what the problem is with this amp.

The amp not quite wired like the schematic? The bias supply taps off one side of the HV winding. The fact that it also goes to the recto tube is a separate issue. Doesn't matter which side, they are equivalent. Same with the cathode/heater of the recto. The B+ really ought to come off pin 8, which is what you have.

Look at the wiring layout - I am looking at AA1164 - B+ comes off pin 8, the schematic is drawn with the numbers reversed. It also shows the bias tapping off pin 4.

I think what happened is they drew the circuit, and someone filled in the pin numbers without looking at the layout. AA964 is the same way, the two sides of the page don't agree. Your description sounds stock.

I have two versions of AA1164 - the one with the GZ34, and one almost identical but with 5U4. The 5U4 version also adds the pair of 100 ohm resistors from the heater lines to ground at the pilot lamp socket. Oddly enough the grounded heater center tap is also still there.

How about this: how clean is the bias? Scope it or set your meter to AC volts and measure the ripple on the bias. There should be almost zero ripple there. If there is substantial AC voltage, then the bias filter cap is not working.

ANd one other thought. Scope the output, or at least the signal on the plates of the output tubes. Is it clean or is the amp cranking out 40 watts of RF that you can't hear? My experience with amps that went RF is that they still pass signal, but it is weak and very hummy.

No scope? Your amp is dragging down the B+ - possibly - even without a signal, so we can mess with the signal path during test. You can probably kill any oscillation by tacking a big old .1 cap or something grid to grid across the output tubes. Or a couple of them to ground from either grid. If the voltage goes up where we want it, then the amp was unstable, and THAT is what we have to fix. A long shot, but...

[AtomicMassUnit]

I don't have a scope, but I measured AC ripple on the bias and it measures .110v, and I tried a .1 from grid to grid, and from grid to ground and got no difference in the voltage at the plates, it's still dead on 336v.

I just tried pulling all the preamp tubes out, and with them out the B+ comes up to 346v. You would think that replacing all the caps and resistors in the power supply would fix this thing, but it didn't. What could I try next?

[AtomicMassUnit]

I just tried the suggestion from RG, and disconnected the resistor after the first B+ node, measured the plate voltage and it's 236vdc, with or without power tubes. Shouldn't it be way higher than that?

[AtomicMassUnit]

Ok, I went and disconnected the next resistor down the line (the 18k) and measured the B+ and got 346v... same with the one after that, the 2nd 18k.

[Enzo]

236??? That doesn't fit with all the other readings.

How many rectifiers have you tried? What is the ripple on the first filter?

Try this. The brute force troubleshoot. Take a spare filter cap, the ever popular 22uf/500v would be perfect. Disconnect the wire from the recto tube pin 8. So the wire to the whole circuit is disconnected. Now tack that spare 22/500 from pin 8 to chassis. That way the recto is sitting there totaly unloaded with a 22uf filter. What voltage appears across that cap?

If you have 350VAC on pin 4 AND pin 6, that filter cap ought to have a lot more than 336v across it. I'd expect well over 400. If the volts are still low, I suspect the recto or the tranny. Stick a different recto in there and see.

Voltage aside, power down and measure resustance through the HV winding. The CT is hopefully grounded, make sure. Then measure pin 4 and pin 6 to ground. Each ought to be half what pin 4 to pin 6 measure.

Likewise verify the 5VAC is indeed that and verufy it is not open or something. I assume both sides of the GZ glow brightly heaterwise?

If you can get my 400 volts or better with just the one test cap, then run the wire to that original first filter, but disconnect the 1K/1W and the OT CT. SO the first filter is connected to the recto all by itself just like our test cap was. At this point we have removed the test cap. If that voltage stays high, good. If not, then the only difference was the cap, and yours seems bad then.

If it does hold up, then connect only the center tap of the OT. This gets us to a condition you already did - essentially lifting the 1K. But we have now verified the stuff before it independently.

In like fashion bring one element at a time back on line to see what thing brings us down.

[Amp Kat]

Hey if you don't fix it I'm only 45 minutes away so feel free to drop it by and we can scope it out. I'm there on weekends except for this one (LSU) but I'd be happy to help you out dude.

[AtomicMassUnit]

Enzo, that found the problem. When I put the test cap alone from the rectifier to ground I got 470 volts across it! Then when I did the same with the first filter cap like you said, I only got 200 volts! So, it must be the can cap. What's really strange is, that can cap is only about a year old, its from CE distribution. I did try subbing in a good filter cap in each position once, but the problem must be with all the of the caps inside the can, because that didn't solve the problem, leading me to think it couldn't be the filter caps. So, I'm ordering a new one and I'm going to hope that fixes it. I'll post back shortly.