Princeton phase inverter mod

[casey73]

I rebuilt my old Princeton reverb SF. I looks stock, but is not anymore. I replaced the original power tranny with an Allen heavy duty one so I can use
6L6 or EL34 outputs, made the bias adjustable, replaced the original output tranny with a Mercury Magnetics Deluxe output and made several changes in the preamp and power amp coupling caps to cut out some of the flabbiness. I put it in a reproduction cabinet (Marsh) with a 12" baffle and installed a new Jensen Jet 100W alnico.

I've pretty happy with the tone, and it has amazing volume when using the larger output tubes. But I would like a little more clean headroom. I still have the original Princeton phase inverter set up, and am thinking about changing it to a long tail inverter, like the Deluxe. I never use the vibrato so wouldn't miss it.

Anyone ever done this mod? Would it give me the desired extra headroom? I'm open to suggestions.

[Chuck H]

I've never done it either. But from what I know and have read it will give you more clean headroom and more overdrive. That is, the amp will have a louder max clean volume but the PI will still have more voltage swing on tap to overdrive the power tube grids as the volume knob is increased.

A much easier mod is the Paul C mod. You could do it in a jiffy, and if it is enough improvement, your done. Not as big a leap as the LTP inverter mod, but it's good for a little more headroom and you get to keep the trem. The Paul C mod fixes the bias on the split load inverter to make it more efficient.

Here's a link with some easy mods to improve headroom in the PR, including the Paul C:

http://blueguitar.org/new/articles/b...ick_PRmods.pdf

Chuck

[loudthud]

IMHO the paul c mod reduces headroom. It will probably be worse with 6L6's.

[Chuck H]

Hmmm, everyone that does it reports a slight increase in headroom, smoother distortion onset and better PI balance (which should also help). I assumed this to be correct based on the nature of the mod. The stock phase inverter will have a tad less clean output because of bias shift. It's important to get the bias correct. In a PR you probably want the input grid about 1.4 to 1.6 volts lower than the cathode. I have read some reports where the mod was done incorrectly and there was a decrease in headroom though.

I'd be interested in your experience, since I haven't done it myself.

Chuck

[MWJB]

Converting to LTPI will increase gain. A good idea for a non-reverb Princeton which is short of gain to begin with.

Have you tried a 12AT7 or 5751 in V1, or a 12AX7 with a 2.7K cathode resistor at the first stage (10-15uf bypass cap)? Whether they get you the tone you want is another matter, but I'd try these quick, simple changes first before ripping out out your PI & vibrato.

[casey73]

Tube swapping

Converting to LTPI will increase gain. A good idea for a non-reverb Princeton which is short of gain to begin with.

Have you tried a 12AT7 or 5751 in V1, or a 12AX7 with a 2.7K cathode resistor at the first stage (10-15uf bypass cap)? Whether they get you the tone you want is another matter, but I'd try these quick, simple changes first before ripping out out your PI & vibrato.

Actually, I put in a switch in where the reverb footswitch jack was that allows me to go between a 25 uf/1.5k or .68uf/2.7k cathode bypass on the first preamp stage. I works well, and I find myself using the .68/2.7k setting most of the time. I've got 5751s and 12AT7s I'm going to try in the first slot this morning.

[Wakculloch]

Phase splitter gain.

Surely the concertina phase splitter will always have twice as much raw gain, when using the same type valve and anode load, as the long taile pair. The first stage is operating at full gain, then the second stage will be unity.
The long tail pair operates at half gain due to the feedback in the cathode circuit (as I understand it).

[casey73]

I plugged in several different tubes including a 12AY7, which was too much cut in gain. I settled on a 5751. It give me a little more space on the dial, so it's clean up to around 5 using humbuckers. At 6 it gets real touch sensative and spanky. From 7 up it's smooth overdrive. I've got and adjustable bias in it and set it up so I can use 6L6 or EL84 outputs. I'm getting a cleaner sound out of the Wing C EL34s than I was getting with the Wing C 6L6s.

I also A/B'd the Jensen Jet (100 W alnico) I installed against a Vintage 30 and Jensen
C12R. The C12R was the first speaker I installed, but after beefing up the power and output tranny and putting in 6L6's/EL34's it was too much for it to handle. The best description of the Jensen compared to the Vintage 30 is that it is much smoother. The vintage 30 is very agressive sounding in comparrison. I could only test the C12R at lower volumes so didn't do a high volume comparrison. The C12R was a little brighter sounding the the Jet, but just doesn't move as much air. The Jet puts out a respectable low end thump, but doesn't mush up too much, at least for my taste.

[casey73]

Since I didn't do the Paul C mod, I can't speak to that, but I do know I get more clean headroom out of the EL34's than the 6L6s, so I'm sticking with them.

[loudthud]

Most people with no test equipment gauge headroom by how far they have to turn up the volume control to get distortion. If you have an amp that distorts at 5 with a 20dB log volume pot, they will all tell you that the amp has less headroom if you install a linear volume pot because it will distort at 1 on the volume knob. The amp has exactly the same amount of headroom!!!

The paul c mod biases the concertina's cathode at slightly less than 1/3 the supply voltage. That's very near zero bias for a 12AX7. Lower gain tubes will have a little more headroom, but even a tube that can saturate to zero volts should be biased at 1/4 the supply voltage. The paul c mod sounds warm because it clips the positive half of the wave going to the bottom power tube. This is perceived as a gain reduction.

For maximum headroom, the cathode of a 12AX7 used as a concertina phase inverter needs to be about 1/6 the supply voltage when the sum of the split load resistors is about 100K. Use a 4.7meg instead of the 2.2meg.

[[ula58]

I thought I'd resurrect the thread.

Why bias the cathode at 1/6 the supply voltage? Seems to me that 1/4 gives the most headroom. Please explain!

P

[loudthud]

1/4 B+ would work if the tube could "saturate" near zero volts. Draw a 112K load line on the 12AX7 plate curves and you'll see where the load line crosses the zero volts bias line that the tube has over 100V plate to cathode (depending on B+). Subtract that 100V from B+, divide by 4 and that should be the cathode voltage for max headroom.

[[ula58]

Ah...I get it..the anode and cathode can never get closer to each other than 100V or so! So, you divide that by 4. So, with 350V power supply voltage (for example), and 56K cathode and anode R's you'd have a 62.5V grid bias...or roughly 1/5.6 of the supply - which is darn close to 1/6!


Thanks for explaining it to me!

I tried the PaulC mod with 2Meg and 1meg R devider and it seems to reduce the volume (I think because the input impedance of the PI becomes lower due to lack of feedback between grid and cathode that exists in the origonal PI. The larger resistor divider ratio that you suggest would have slightly higher inout impedance than the standard "Paulc" mod using 2.2Meg and 1Meg. That would be good.

Right!

P

[Bruce / Mission Amps]

Most people with no test equipment gauge headroom by how far they have to turn up the volume control to get distortion. If you have an amp that distorts at 5 with a 20dB log volume pot, they will all tell you that the amp has less headroom if you install a linear volume pot because it will distort at 1 on the volume knob. The amp has exactly the same amount of headroom!!!

The paul c mod biases the concertina's cathode at slightly less than 1/3 the supply voltage. That's very near zero bias for a 12AX7. Lower gain tubes will have a little more headroom, but even a tube that can saturate to zero volts should be biased at 1/4 the supply voltage. The paul c mod sounds warm because it clips the positive half of the wave going to the bottom power tube. This is perceived as a gain reduction.

For maximum headroom, the cathode of a 12AX7 used as a concertina phase inverter needs to be about 1/6 the supply voltage when the sum of the split load resistors is about 100K. Use a 4.7meg instead of the 2.2meg.

Have you ever taken your dual trace scope and put one probe (channel) on each coupling cap to the power tube grids and set the scope for summing the two channels? ... They're out of phase of course.
Then gas the amp up to full output power and beyond .... into a dummy load and watch the pips.
Try that test stock with no PI mod, then go 2M2, 3M3 and +4M2 etc... it is very interesting.
After you do this you'll know which method or what resistor to use if you like the fixed bias PI mod.

[[ula58]

Hi Bruce, .

The sum of the signals should be constant ideally, so, I wonder what you're driving at? Well now you have gotten me curious! I shall have to try that experiment.


I have wondered if it makes sense to put the same resistor value as the cathode and anode resistor in series with the anode and coupling cap to the output stage load. That way the output impedance of the circuit is balanced. That is, the anode output resistance is the anode resistor and the cathode output resistance is the same. Otherwise the cathode output impedance is roughly 1/gm, not even close to the cathode anode resistor value of 56k.

Ever try that?

Paul