Beefed up Princeton AA964

[acheld]

Continuing on from the discussion/decision making https://el34world.com/Forum/index.php?topic=28052.0  regarding a smallish amp with headroom and the ability to use either 6V6 or 5881 tubes in the output section, I've decided to proceed with something close to the Princeton AA964.

I was comparing the AA1164 and AA964 circuits, as well as Hoffman's Princeton Reverb (which is based on the AA1164)

https://el34world.com/charts/Schematics/files/Fender/Fender_princeton_aa964_schematic.pdf
https://el34world.com/charts/Schematics/files/Fender/Fender_princeton_reverb_aa1164_schematic.pdf
https://el34world.com/Hoffman/files/Hoffman_PrincetonReverb.pdf

This prompts some questions:

1.  Looks like the biggest difference between the AA964 and AA1164 -- aside from the presence of the reverb -- is the presence of an additional gain stage just before the cathodyne PI.   How much difference does this make in how the amp sounds?  Other than louder, of course.   Currently, my plan is to stay close to the AA964 circuit -- but plans are made to be changed, especially at this stage.
2.  Neither the AA1164 nor the AA964 circuits use grid stopper or screen resistors in the output section, though Hoffman's design does.   Merlin clearly likes both, and all of my other amps have had these.   As best I can tell, neither grid stoppers nor screen resistors should not materially affect the tone of the amp -- they are there purely to protect from oscillation and excessive screen current when overdriven, respectively.   Is this correct?   My inclination is to use these in my amp.

[mresistor]

I always use gs's and sr's in my PR builds. I've got a '78 B1270 PR that doesn't have them though. Leaving that one basically unmodified. I did work on a recap of a bf AA964 and the thing I noticed is it is not as powerful as an AA1164, it also didn't sound exactly like a PR either.

 

[SILVERGUN]

1.  Looks like the biggest difference between the AA964 and AA1164 -- aside from the presence of the reverb -- is the presence of an additional gain stage just before the cathodyne PI.   How much difference does this make in how the amp sounds?  Other than louder, of course.   Currently, my plan is to stay close to the AA964 circuit -- but plans are made to be changed, especially at this stage.

Are you planning to include reverb?
The extra triode is there to make up for the signal loss of the reverb circuit and drive the PI. Without the reverb circuit you are likely to create the overdrive you were trying to avoid. Without reverb, and since you want the amp to stay clean, you won't need the extra gain stage.
I still vote for 2 gain stages into a 12AU7 (or AT7 if needed) LTP PI into 6V6s.


Since you have the DR there, you could experiment a bit.
Pull the reverb and trem tubes (this will provide slightly higher B+ to preamp). Install a 12AU7 into the PI and see what you think.
This should put the normal channels 2 gain stages directly into the PI and see where you stand.
Just a thought, and I know how we all like to have fun.
I think I have an AU7 in my '68 DR right now because I liked it so much.

[jordan86]

Rob Robinette has some great reading material on the 1164 and AB763 models. I read these exhaustively before deciding to build my 1164.

https://robrobinette.com/AA1164_Princeton_Reverb.htm

https://robrobinette.com/AB763_Model_Differences.htm

As stated, the extra stage is a “make up” stage to compensate for the loss caused by reverb circuit. Keeping it without the reverb makes way too high gain. As a whole though, it nets to be a little more gain than the loss, so the reverb amps are capable of more drive.

I’d probably just build a mostly stock AA964 in your shoes with the following mods.

- Power tube Screen grids and grid stoppers
- stokes mod (PI on unused node)
- increased filter capacitance (JJ makes a 40/20/20)
- adjust dropping resistors on B+ (maybe 15K instead of 18k) to get preamp voltages a tad higher.
- oversized iron (at least 20w)
- maybe drop cathode bypass caps to 5-10uf to clean up lows

That should “beef it up”

[acheld]

- Power tube Screen grids and grid stoppers -- Yes for sure
- stokes mod (PI on unused node) --  I'm not sure this is necessary. I know it's a popular mod . . .  It should be easy enough to play with when I get to that stage.
- increased filter capacitance (JJ makes a 40/20/20)  -- My plan was to go with 40/20/20/20 with some film caps left over from another project.
- adjust dropping resistors on B+ (maybe 15K instead of 18k) to get preamp voltages a tad higher.
- oversized iron (at least 20w)   -- I haven't decided which PT I'm going to use, but it will either be a Hammond 290CX, most likely, or a BX.  For certain I will have to adjust the dropping resistors to get to the correct voltages.   I will be using 5881 power tubes and bias voltage range will be addressed.  IIRC, the 290BX does not have a dedicated bias winding, where the CX does. 
- maybe drop cathode bypass caps to 5-10uf to clean up lows. -- Maybe 10, but I'm going to start at stock, or close to, as I want it to sound like a Princeton, but "with some beef" that is going to come from the power tubes.

And yeah, I need a better name for this puppy.

Time to fire up JSCHEM.  Joy.

[SILVERGUN]

Both the BX and CX have a bias tap (which is not considered a separate winding)

I'm following along, interested in how this turns out for you.

Reverb?

[acheld]

First draft schematic for Princeton 5881.

No reverb, but yes! tremolo. 

The input stage and PI are straight off of the AA964 schematic.  The output stage is borrowed from an AB763.

I'm not sure about the tremolo and bias circuit.   All the amps I've built prior to this had two tubes for the Trem circuit -- this has one.  I'm wondering if this is the time to introduce the Trem-O-Nator into the mix.  I assume this injects its magic just before the PI?  or just after the volume control? 

I have on hand both Hammond 290CX and a 1760H (6.6K:4,8,16).  I think these will work, but am willing to change if needed.   With the 1760H, I plan to use 5881 tubes, but want the ability to move down to 6V6.  I did not draw the 4/8/16 ohm taps on the schematic yet, as all I have is JSCHEM, which works fine, but is a pain to make new parts. 

[sluckey]

That single triode tremolo may not have enough muscle to wiggle the 5881s.

[66Strat]

You might try adding in a source follower. The attached trem circuit mod borrows from the 6G16 Vibroverb.

[ac427v]

Hmmm. This thread caused me to look at all versions of the AA964 schematic in Hoffman's library. One version shows one fewer power supply nodes. That is like a "Stokes Mod" for the PI and the preamp which gets 230 volts on the plates. Which schematic will you follow?

[acheld]

I am no expert in the Stokes mod, but as I understand it:

There is an unused "node" in the power supply of the AA964 -- in my schematic the "missing" C+ node.  No one really understands why that design was used.    The Stokes mod uses that "unused node" to power the PI at a somewhat higher voltage than in the stock AA964 configuration.   This gives the PI a modestly higher gain.  It might be beneficial to my ears, it might not.    I have no aural experience with the Stokes mod; the only Princetons I have heard have not had this mod.   That said, people seem to like what it does.

The Stokes mod appears to be bone simple to implement.  I will try this when I get this amp off the ground and on the grid.  LOL, may take awhile.

My main consideration over the next days is: 
a) is a tremolo worth the effort?,
b) since I want to use up my 5881 stash, I will need an additional 12A_7 to wiggle the higher bias voltage of the 5881, and how to fit that in (I think the tremolo circuit  just becomes like an AB763 trem then, and that's fine), and
c) should I make it a little more complex with a Trem-O-Nator circuit with the additional tube, or
d) try the circuit proposed by 66Strat, which I found interesting.   

[tubeswell]

a) is a tremolo worth the effort?

I vote ‘Yes’. Gotta love that swampy blues sound


If you’re concerned about not getting good enough wiggle on the 5881grids, try a vibrochamp trem (bias vary trem on a preamp tube).

[sluckey]

c) should I make it a little more complex with a Trem-O-Nator circuit with the additional tube, or

The TON is actually much simpler than your bias wiggle trem. Sounds better and has none of the bias interaction problems.

[acheld]

The TON is actually much simpler than your bias wiggle trem. Sounds better and has none of the bias interaction problems.

I'm looking at your Bandmaster build documents, and it does look straightforward.   Just connect the circuit right just after the volume, right before the PI.   You used node "B" at higher voltage than the PI, which uses node C.  Is this needed?   

I may also use that bias circuit, which looks like a Marshall style (maybe?). 

[sluckey]

Fender used the screen node. I just followed suite and didn't ask questions.   

[acheld]

The 5881 T-O-N is taking shape.

I've changed the bias circuit and changed the trem circuit to sluckey's fabled Trem-O-Nator.   I am not 100% sure of the connection to the signal path -- the insertion point is correct, but do I need a coupling cap between the trem and the PI ?   My estimate is no, since there should be no high voltage on the Vactrol.

Edit:  schematic errors fixed, see below.

[sluckey]

You messed up the tremolo circuit. Better do it like this...

[acheld]

Thanks!

Bonehead error has been fixed.

Edit:  Faulty schematics were deleted.   See below.

[sluckey]

Better start working on revision 3.   

[acheld]

LOL, my days integrating large medical systems has taught me that good enough is . . . good enough.  Go for perfection and you are toast!

Now is the work I don't particularly enjoy, but important nonetheless, which is to organize a layout on a board and catalog parts.  I may have to learn just enough DIYLC to create a drill guide for Doug.  We'll see. 

[sluckey]

I'd work on getting the schematic right before tackling a layout.

[acheld]

OK, looks like the NFB is not inserted correctly. (How did I do that??) 

I will look at it in the light of day tomorrow or Monday to get it right.  Thanks sluckey.   

[sluckey]

The tremolo ain't right!

[acheld]

The tremolo ain't right!

That's for damn sure it was not.  Serves me right for drawing JSCHEM while watching Michigan football. 

Rev 3 is better.  Whether it is good enough remains to be seen.

I have a couple of notes/questions:

1.  In Reply #17 your schematic shows two possible connections to the signal path -- either just after the volume control, or just before the PI.  I chose to connect just before the PI, mimicking the Princeton design.   Is there a disadvantage to this?  I convinced myself that if we placed the trem signal before the gain recovery stage (following the tone stack), you would be amplifying the trem signal more than desirable -- not sure if that reasoning is correct.  I wanted to stay close to the Princeton design in any event.
2.  I copied the bias circuit from your Bandmaster AB763.  I assume that should be ok, though I know I will have to check for voltage range -- since I want to experiment with 5881s and 6V6s, I know I'll need a wide range of available bias voltage probably from -30 to -55.
3.  My panic comment from last night about Feedback insertion was not correct -- I drew it right the first time.   

[PRR]

Is that LFO phase shift resistor really returned to an unbypassed cathode? Is this from a known-good plan?

Sucking the tremolo in the *middle* of a NFB power amp, the NFB tries to correct the drop of gain.

This LDR is working with 40V signals, and will distort very audibly. I had a compressor working on 3V signals, so only hurting on loud points, and it was a near-mistake.

Moving the LDR+pot to the (lower signal) grid of V1-b will suck-out a ton of gain.

I don't have answers, just playing grumpy-man.

[sluckey]

Rev 3 is better.  Whether it is good enough remains to be seen.

Looks good.

1.  In Reply #17 your schematic shows two possible connections to the signal path -- either just after the volume control, or just before the PI.  I chose to connect just before the PI, mimicking the Princeton design.   Is there a disadvantage to this?  I convinced myself that if we placed the trem signal before the gain recovery stage (following the tone stack), you would be amplifying the trem signal more than desirable -- not sure if that reasoning is correct.  I wanted to stay close to the Princeton design in any event.

There is no tremolo signal at the top of the INT pot. The only thing there is a resistance that changes at the tremolo speed, so there's nothing to amplify. I just offered two different places to connect into the signal chain to see if you prefer one over the other. Connecting to the volume control is the easiest since all you have to do is use a short jumper wire from the INT pot over to the VOL pot. And you don't need to use an extra cap.

Read my Trem-O-Nator page again, with special focus on the second paragraph. You may have to experiment a bit.

2.  I copied the bias circuit from your Bandmaster AB763.  I assume that should be ok, though I know I will have to check for voltage range -- since I want to experiment with 5881s and 6V6s, I know I'll need a wide range of available bias voltage probably from -30 to -55.

That's actually a Marshall design bias circuit and will probably give you plenty of range to cover both type tubes.

[acheld]

I think I get it.  The Vactrol is a variable resistor dumping guitar signal to ground, in conjunction with the voltage divider formed by the INT pot. 

So, I think I will simplify and move the Trem insertion back to just after the VOL pot.

[sluckey]

Don't forget... You may have to experiment with how the VTL5C1 actually connects to the signal chain. The circuit you have works very well with AB763 amps but may need some tweaking to work well with your amp without losing too much guitar signal. Here's a modified circuit a guy at TAG used in his ODS style amp...

     https://ampgarage.com/forum/viewtopic.php?p=407279#p407279

Ignore his oscillator circuit. The important part is on the right side of the VTL5C1. If you look at the message directly above the schematic you can hear his sound clip.

[acheld]

Got it, and understood that it will need tweaking.

I was going to re-locate the T-O-N insertion site back to the volume pot, but that may make layout even more of an issue, so I guess I will stay with my rev3 drawing.

[sluckey]

Any progress?

[acheld]

Busy with mandated home projects. 

Happy with the schematic.  There are a few very minor errors in labelling and I will post an update when it makes sense. 

Working on layout now, and I have never been particularly good with that.   So, it will take some time.   Also working on procuring some parts and a faceplate.   

LOL, I am slow as a turtle.      Reviewed your Bandmaster files, which helped me figure out a reasonable layout for my Trem section. 

[shooter]

never been particularly good with that

same.
once I dialed in the schematic, I'd go steal layouts with similar stages
get them blown up, cut n paste, literally.  used a fine point colored felt tip to tweak/notate/value changes ettal.
then i'd boot up the computer n use a pcb creator to make it whole

[acheld]

I'd go steal layouts with similar stages . . .

Precisely my (vaunted) technique.

It takes me forever, and since I rarely use DIYLC, well, you get the idea.   

A self filling beer glass --  what a great idea!

[acheld]

Spent quite a bit of time with DIYLC working on the turret board and layout.   That is tedious stuff, though necessary to produce the board.

The Vactrol tremolo section is designed to be removed from the main circuit easily.   I anticipate experimenting with various resistor values and/or changing the insertion point of the tremolo (which is really just a pulsing dump of the signal path to ground) to get the intensity correct.   

As always, comments are appreciated.  I'm sure there are errors present in this first effort . . .

I'm attaching most up to date schematic, layout pdf and layout diy. 

[sluckey]

A while back a forum member asked how to modify Hoffman's Princeton Reverb to add the Trem-O-Nator. Don't remember who or when, but I stripped out the standard tremolo circuit and added the TON without changing any turret locations or adding more turrets. I don't recall if the member followed through and actually added the TON.

Anyway, I just took the modified Hoffman layout and stripped out the reverb section, and also shortened the board. The end results is a AA964 Princeton with TON. The circuit checks out with the original AA964 schematic except the TON circuit. Of course I left Hoffman's grid stoppers, screen resistors, and bias resistors. This project will easily fit in Hoffman's 17x6.5x2.5 blank chassis. These documents might help you with your DIYLC layout or you could easily make your own board using the included drill guide.

Link to docs... http://sluckeyamps.com/misc/Princeton_AA964_with_Trem-O-Nator.pdf

[ac427v]

I've had great success with the Trem-o-nator and the AB763 circuit where the 50K intensity pot replaces the original 50k roach intensity pot. Other circuits needed some tweaking as Sluckey mentioned. The 50k pot connecting a 1 meg volume control to the trem-o-nator circuit (and circuit ground) is a huge load on the signal. This occurs even when the Trem-o-nator is switched off. It  makes for a very quiet amp. One solution is to copy the AmpGarage variation that Sluckey linked. I found a 250k pot connected as a variable resistor to work very well. See attached schematic.

EDIT... Converted jpeg image attachment to png format so it will display on the forum... sluckey