Solving “Flabby” Bass of a 1967 Fender Princeton Reverb

All of the tubes were tested on a TV-7 D/U and are all within specification.  There are some very good tubes in the amp; notably a Telefunken 12AX7 in V3 (this tube is valued as high as $100), and a Mullard Great Britain GZ34 rectifier (also about $100).  The original rectifier in this amp would (probably) have been a 5U4GB.

The JJ 6V6S output tubes were running at about 70% dissipation, and they were fairly well matched – the inboard one at 22 mA and the outboard one at 18.5 mA.  The amp is running about 17 Watts.

The amp has been modified in the past.  The power transformer is not original to the amp. Because it is running a GZ34 instead of the original 5U4GB, the original power supply transformer may have failed due to the higher voltage/current requirements of the GZ34.  The electrolytic capacitors had all been replaced, including the “can” filter capacitors (incidentally, the can cap was installed incorrectly – it doesn’t change anything, because the values are all 20 µF, but it made troubleshooting harder).  The coupling caps have been changed to Sprague orange drops, the values are as original.  The resistor values all seem to be original values, except for the bias section, where the original 27K has been changed to a 22K.  This was to set the bias of the output tubes.  The bypass cap of the bias board had been changed from 70 µF to 100 µF.  This is a common modification to smooth out the bias ripple.

The 25 µF cathode bypass capacitors of V1 have been paralleled with 1µF Tantalum capacitors.  This is more commonly seen in Hi-Fi amps, and it is employed to improve high frequency response.  It can also act to tighten the bass frequencies.  This may have been done to try to correct the bass “flabbiness” issue at some point.

The speaker most likely to reproduce the tone of the Princeton Reverb is the Weber 10F150T.  This is a 10” speaker, but is rated at 50W, and will do a better job of reproducing bass.  Going to a 12” speaker is an option, but you will hear more changes to the tone, and of course, you will have to replace the existing baffle board.

I think that there are some things to do before we try a new speaker.

Oscilloscope

I put the amp on my 100MHz oscilloscope and checked some things, running into a dummy load.  The amp behaves for the most part as a Princeton reverb should.  With a 600 Hz signal at 104 mV (like a fairly strong mid-range guitar signal), the output starts to clip (distort) with the volume at 4.5.  With the volume on 6, Treble 4, and Bass 4, the distortion is evident on both halves of the sine wave, and it looks like a square wave from all the harmonics in the signal.  The positive half of the sine wave clips first, and this is typical of a cathode-follower like the cathodyne phase inverter of the Princeton, and is said by some to be the penultimate amp distortion – creamy and smooth and still touch-sensitive.  On your amp, this “best” distortion would be between 4 and 5 on the volume setting.

However, at 600 Hz, the Treble control greatly influences distortion.  With the Treble at 7.5, there is crossover and clipping distortion with the Volume set at 4 and the Bass at 4.  You of course can hear this when you mess with the knobs while playing the amp.  It’s typical of a Fender amp, and the Princeton Reverb.

I injected an 84 Hz signal at 104.5 mV, which is about the low “E” of a guitar.  Settings were Volume 4, Treble 7.5, Bass 3.

The positive part of the sine wave is clipping, and the slanted line at the top peaks means there is phase shift going on, probably due to the capacitance time constant of the bass circuit.

Bass is more distorted and there is a peak on the leading edge.  This peak may be harmonics, but it may be a capacitor somewhere trying to boost the bass signal.  This might also be the output transformer.  It might also be the cathodyne phase inverter distorting.  Also might be blocking distortion.  (sigh).

Same waveform, same amplitude.  I suspect that this is the most Bass that the Princeton can reproduce.

This may be the dreaded blocking distortion.  Blocking distortion occurs when the output tube grids are driven to the point of conduction.  This makes a nasty, “farting out” or “cutting in and out” sound.  I could confirm that blocking distortion is occurring by watching the waveform at the output tube grid, but it is difficult to do because of the very high impedance at that point. (the scope pulls down the signal).  I will try this later.

The solutions to blocking distortion are:

1. Lower the value of the coupling capacitors. In the PR they are 0.1 µF.  (In a Marshall, they are typically 0.022 µF).
2. Increase the grid resistor values. (The PR doesn’t have any grid resistors).  It does not change the tone of the amp to add 1.5K grid resistors to the 6V6 tubes.

This may be the cathodyne phase inverter distorting, which is also sort of like blocking distortion.  The solution here is to add a grid resistor to the cathode follower side of V4.  Phase inverter distortion on a Princeton Reverb sounds just like blocking distortion – “blatty”.  There is a modification called the “Stokes Mod” that just moves the voltage of the phase inverter one stage higher, by moving one wire.  This gives the phase inverter higher voltage and so more gain, and is said to rectify the distortion of the stage.  It is said that the reason Fender didn’t do this on Princeton Reverbs is that they didn’t want the Princeton to compete with the Deluxe Reverb.  Could be true.

Conclusion and Summary

As with any surgery, we want to be minimally invasive, so here are some options:

1. The Stokes Mod. It’s just one wire.  Easy to un-do if you don’t like it.
2. Reduce the 0.1µF coupling caps of the phase inverter to 0.047µF. Easy to undo.
3. Add more filtering to the first stage – I didn’t mention this above, but it is easy to do and gives more bandwidth “strength” to the amp. Easily undone.
4. Add grid resistors, 1.5K, to the output tubes. Easily undone.
5. Add a grid resistor (big) to the cathodyne.
6. Change the first gain stage cathode bypass cap. It is currently 25µF with a Tantalum 1µF cap in parallel, making the capacitance 26µF.  The calculated frequency cutoff is 4.08 Hz.  Well below anything the guitar can produce.  If we change the 25µF cap to a 3.3µF, and leave the Tantalum cap there, the resultant 4.3µF capacitance gives a cutoff frequency of 24.7 Hz – still less than a guitar can produce (Low “E” is 82.41 Hz), but at least we ignore a lot of useless bass, and the amp won’t spend energy trying to amplify stuff that we can’t hear.

HardWay Solution

Number 6, above, solved the flabby bass at higher gain conditions.  The big change to the first bypass capacitor prevents the amplifier from trying to reproduce low bass frequencies that the speaker and the circuit cannot replicate.  Also, we can’t hear those notes, anyway.