Tweed Deluxe Build - Consolidated Discussion

[TheTinMan]

I tested it on the LTSpice schematic I have of my 'Double Deluxe'. Setting it well below clipping it gives 7.202V rms and 6.980V rms over the full sweep of the pot. That's a difference of 0.27dB.

Is that with 1 meg or 500K between the pot wiper and the "bottom" of the 1.5K cathode bias resistor?

With the pot wiper all the way to the left in my drawing, should the pot value be added to what is shown as 1 meg resistor to get the effective grid return resistor? If so, does changing the 1 meg to 500K change anyone's conclusion about the effectiveness of this idea?

I do not understand how introducing a voltage divider between the 2nd gain stage and the PI would not reduce the signal going to the power tubes, at least a little. I can understand how a 500K/1 meg divider might not reduce the signal enough to be worthwhile.

If this arrangement were between two normal gain stages, I'm reasonably sure it would work. However, I'm obviously missing something here. Please help me understand.

Dave - does your "Double Deluxe" use a cathodyne phase inverter?

Respectfully,
Chip

[TheTinMan]

OK I did a couple of drawings to compare and try to understand.

In the first attachment, is the value of the grid return resistor in Option A 1meg or 1.5 meg?

In Option B, the cathodyne' bias should be unaffected but I think that there is some attenuation of the signal going to the cathodyne grid. Is that correct?

The second attachment shows the pre-PI master volume that got me thinking about all of this. It's a normal voltage divider; however, the grid return resistor value also drops as the volume is turned down. Maybe that's why it works?

Thanks for any insights,
Chip

P.S. I have asked about this on the Hoffman Amps Forum as well. I understand that you all say it doesn't work well but don't understand WHY.

[Dave H]

Is that with 1 meg or 500K between the pot wiper and the "bottom" of the 1.5K cathode bias resistor?

It's neither It's from your schematic in post #6 which has a 1M between the pot wiper and the "bottom" of the 1k5. 7.202V in the sim is with the pot wiper set to the left on your scm and 6.980V is with the pot wiper set to the right. It doesn't attenuate as much as a normal potential divider because the bottom of the 1M isn't grounded. It's connected to the cathode circuit which is in phase with the grid.

Yes, my "Double Deluxe" does have a cathodyne PI.

[ThermionicScott]

I question the merit of a master volume before the PI on a 5E3, regardless of the circuit. Think about it -- most of the characteristic overdrive of a 5E3 happens in the phase inverter and output tubes. With a master volume after the second gain stage, you might eke out a little bit of clipping with the volume control all the way up and the MV barely cracked, but it's mostly going to act as another volume control. Ho-hum...

[Dave H]

The second attachment shows the pre-PI master volume that got me thinking about all of this. It's a normal voltage divider; however, the grid return resistor value also drops as the volume is turned down. Maybe that's why it works?

Sorry, the second attachment isn't a normal voltage divider either because the bottom of the pot is connected to the cathode not ground so the pot is bootstrapped. It will attenuate more than your circuit because the valve grid connects to the cathode when the pot is fully ccw but I don't think it will work well as an MV. The bootstrapping will make the taper be almost on/off. The volume will be full on as soon as it's turned up from the end stop. I must be missing something here. I can't understand why a standard MV circuit like the one below isn't used.

[Dave H]

I question the merit of a phase inverter before the PI on a 5E3, regardless of the circuit.

Change "phase inverter" to master volume and I'll agree with you

[ThermionicScott]

Change "phase inverter" to master volume and I'll agree with you

D'oh! No amount of proofreading before hitting the "Post" button ever seems to be enough. Thanks.

[TheTinMan]

Sorry, the second attachment isn't a normal voltage divider either because the bottom of the pot is connected to the cathode not ground so the pot is bootstrapped. It will attenuate more than your circuit because the valve grid connects to the cathode when the pot is fully ccw but I don't think it will work well as an MV. The bootstrapping will make the taper be almost on/off. The volume will be full on as soon as it's turned up from the end stop. I must be missing something here. I can't understand why a standard MV circuit like the one below isn't used.

[ATTACH=CONFIG]37686[/ATTACH]

Dave - sincere thanks for your time and explanation.

The "pre-PI master volume" idea came from Rob Robinette. Exchanging a 1 meg pot for a 1 meg resistor has appealing simplicity. However, Rob doesn't mention anything about the pre-PI MV having more of an on/off effect. I partially understand why it would now though.

I've read multiple places that cathodyne PI distortion tends to be nasty - and not nasty in a good way. Valve Wizard explains why and offers a grid stopper as a good solution. Lots of Princeton Reverb discussions address this perceived problem with the Paul C mod (fixed bias for cathodyne) and the Stokes mod (higher B+ for PI anode resistor). But the Paul C mod was not intended for cathode biased power amps and I don't want to add another power rail node with higher voltage for the PI for fear of messing up the feel of the 5E3 circuit. So I focused on the grid stopper and came up with my relatively unworkable idea.

There isn't a lot of extra room for mods/experiments inside a 5E3 chassis, so I'm trying to narrow down possible mods to incorporate in my layout.

Thanks again guys!

Chip

[ThermionicScott]

...for fear of messing up the feel of the 5E3 circuit...

I think the above is an excellent argument for building up the circuit as stock as possible first, giving it an honest appraisal, and then fixing any "deficiencies" you find. One unfortunate side-effect of all of the wonderful information and analysis of classic amps available online is that it has inspired a sort of hypochondria among guitar amp builders.

Don't forget that Fender sold (and still sells) thousands of amps with cathodyne phase inverters -- evidently, some people think the overdrive sounds fine as is.

[TheTinMan]

Thanks for all of your input so far.

Someone on the Hoffman forum reports that the pre-PI "master volume" doesn't do much of anything until the bottom of rotation. IOW only when you get to the bottom of the pot rotation and effectively start connecting the grid to the cathode - exactly as you all said. So I've abandoned that idea.

I also abandoned the idea of switching to the higher voltage PT secondary taps or having a tube/solid state rectifier switch. I will use a terminal block for the PT secondaries so that I can test the amp with higher voltages 355-0-355 but have no expectations of improving tone or feel that way.

OK - I have drawn up my schematic. Many thanks to Steve for his Visio file with both schematic & layout drawings!

Please feel free to fly-speck the schematic. Any corrections or comments will be greatly appreciated.

The cathode switching arrangement is based on Geezer's idea discussed in this thread on the Hoffman forum:
Geezer's Cathode Power Switching

The concept is to use a 3-pole, triple throw switch. "H" for high power should be stock 5E3 with a shared RC network. "L" for low power is separate and un-bypassed cathode resistors. "M" for medium power is separated RC networks for the power tubes with a trim pot adjustment between the bypass caps.

This extra switch will be in the "Power" hole and the power switch will be in the "Ground" hole.

I added an elevated DC voltage reference for the heater's center tap. Since the cathodyne PI cathode is at about 40 volts, I'm shooting for around 60 volts on the reference point. The "humdinger" may be a complete waste but what the heck.

The diodes prior to the rectifier tube are based on R.G. Keen's "Immortal Amplifier" articles.

The 330K grid stopper on the cathodyne is supposed to reduce PI distortion when the power tube grids start conducting current. Highly recommended on the Valve Wizard (Merlin) site, and it will be easy enough to test whether it helps or hurts.

I separated the cathodes of V1 to allow for experimentation with the bypass cap value for the Bright channel. On the Normal channel, I'm going to build it stock first. Then I'll try a 220K plate resistor with 2.2K and 2.7K cathode resistors just for the Normal channel. I really liked that sound in another amp so will have to give it a listen. It's also present in the Brown Deluxe 6G3 circuit and, yes, I recognize that's a 12AX7 not a 12AY7 in V1 of that amp.

I also plan to experiment with the first coupling caps, cathode bypass caps, etc. But only after I play the amp for a while and really get a feel for it.

Chip