Quad 6V6 Plexi

[Paul1453]

Since my lower voltage 6V6 Plexi sounds so good,
I thought I'd try to modify that design to work with the quad 6V6 parts I have from a Magnavox Belvedere. 

Here is the output section of the original Belvedere,
and my 1st draft of the new output section.

It's just my starting point and I know I left some things out, screen resistors etc.

I'm also considering individual bias adjustments for the 6V6s.
There is a Chinese board on Ebay for that for less than $15 delivered.

After trying to match tubes on my 1st build,
I think it might be best to have individual bias adjustments to make it easier to get the whole quad working at the same level.   

[Paul1453]

Original design is parallel PP output.

My change still has tubes working as 2 pairs.

To use this board to control each grid voltage independently would require more modification.

I'd have to split the input signals to bias each one on it's own.

In theory, it seems like a pretty good idea.

In the real world, this idea may be beyond my abilities to implement.

It might be easier to just ensure I have a well matched quad of 6V6s instead.

[Paul1453]

I think I have a good layout plan and most of a proven design that works with my parts, PT/OT chassis.

Hoping for a little more output circuit design guidance.   

[HotBluePlates]

Original design is parallel PP output.

My change still has tubes working as 2 pairs.

To use this board to control each grid voltage independently would require more modification.

I'd have to split the input signals to bias each one on it's own.

You'd have to have 2 coupling caps and 2 bias-feed resistors per phase inverter output, so that each 6V6 can have its own independent bias voltage. That's not hard, but it is more space.

The board you were looking at also assumes you have an a.c. voltage source, probably a winding of 50-65vac. It bridge-rectifies that a.c. voltage before filtering and sending to the individual dividers. You'll have to know what bias voltage you'll need and do the design a bias supply as though you were going to build it yourself, just to figure out what raw a.c. voltage you should add.

But I didn't like how the ad shows different-valued resistors in the various pictures of the board. There's at least 2 versions pictured. So you may still have to unsolder & replace some resistors to get the range of adjustment suitable for your amp.

I think I have a good layout plan and most of a proven design that works with my parts, PT/OT chassis.

Hoping for a little more output circuit design guidance.

Well, the design plan is simple, since you're going to do the 6V6 Plexi but double-up on tubes. Assuming your PT is roughly the same voltage as your previous build, you just need double the high voltage current, and an OT that's half the primary impedance.

How did the transformers end up where they are? I don't know what the layout of the original chassis looked like, but see where you already had 4 holes for the 6V6's.

I assume you won't be using tube rectifiers, so the PT could have sat over the area of the old 5Y3's & can caps, with the OT in between the PT & 6V6's. Then you'd have a bunch of real estate towards the front for the pots, board & preamp.

But maybe that big PT cutout was already made, etc... In that case what you have is workable. You might end up wishing you had a punch for the tube socket holes (or maybe you already have one).

[Paul1453]

The PT from another pulse generator fit right in the spot where the old one with 5VAC was.   
I'll save that for another tube rectified amp, maybe a HoSo???

When I looked at it like that, the holes in the back line up, and the front is undrilled so pots can go where I want them.
The wires between the tubes/board/pots all stay nice and short this way.
Bias pot or Bias board w/multiple pots and TPs will be on the back this time.  Likely get smaller TPs, and new pots and knobs.

Ran PSU for 2 of the chokes I have.  The one shown has higher resistance and causes the B voltage to be significantly lower.
It also gives me a warning, where the 100 ohm Triad does not, but on the Triad the B voltage spikes to 460 at startup.

Will want to do a real pro look, or as best as I can, on this one.
It should sound fantastic just like the 1st one does.             

[Paul1453]

Something like this maybe, HBP?

I was concerned that since I split the inputs,
that I would only be driving the output tubes with 1/2 the original input signal level.

It seems to me this would individually bias the output tubes. 

[HotBluePlates]

Something like this maybe, HBP?

Yep, that's how you would do it. there's a blocking cap & a bias-feed resistor per output tube. Now you just need to verify your PT has a dedicated winding just for the input a.c. volts to the bias board's bridge rectifier. (Or perhaps someone else can chime in to verify a bridge would need a separate winding and not just a tap on the high voltage winding; however, I also thought a tapped winding might be unlikely on your proposed PT).

Anyway, were pins 6 & 7 a 50vac winding on your HP power transformer?

I was concerned that since I split the inputs,
that I would only be driving the output tubes with 1/2 the original input signal level.

Not half the level (because you have a voltage output from the phase inverter and a.c. current will be limited to that through the bias-feed resistor plus bias pot wiper-to-ground resistance). But I guess since you have double the bias-feed resistors, you do have an a.c. load of half what is was before with 2x 6V6.

How are you set for 0.01µF caps? You could roughly double the bias feed resistors to 220kΩ and halve the coupling/blocking caps, and so maintain the same bass response and a.c. load.

[Paul1453]

Something like this maybe, HBP?

Anyway, were pins 6 & 7 a 50vac winding on your HP power transformer?

But I guess since you have double the bias-feed resistors, you do have an a.c. load of half what is was before with 2x 6V6.

How are you set for 0.01µF caps? You could roughly double the bias feed resistors to 220kΩ and halve the coupling/blocking caps, and so maintain the same bass response and a.c. load.

I thought you might have caught something I missed on my PT.
So, I powered up the Plexi and got 120VAC across Pins 6 & 7.
If I'm not mistaken, it seems like that is also a Primary winding?
If you feed it with 50VAC @ 1Khz you get the listed outputs???   

Maybe it could still work with the bias board?  I just ordered one of those.
It would not need as much resistance to drop it down to 60VAC as the secondary tap does.

I think I can come up with .01uF caps if that is what is called for.
I should be set for .68 and .022 caps when these come in.  LOL

I ran PSU for this again, and am leaning towards using my 4H 100 ohm Triad choke.
No warnings and my B voltage stays much closer to A with it.
Looks like they should be around 315 with that choke.
Initial voltage spike doesn't exceed my 450V rated caps, so I think that might work.

[HotBluePlates]

... I powered up the Plexi and got 120VAC across Pins 6 & 7.
If I'm not mistaken, it seems like that is also a Primary winding?
If you feed it with 50VAC @ 1Khz you get the listed outputs???    ...

I'd need to know what HP device it was pulled from, then go check the schematic to see how it's used.

I'm suspecting it's a secondary with a 1:1 ratio with the primary. Or it might be another primary winding which you could series or parallel for 240vac or 120vac operation, respectively. I also suspect the "50 1kc" really means it can operate with a power line frequency from 50Hz to 1kHz (I think 400Hz was common for shipboard power generation?).

Maybe it could still work with the bias board?  I just ordered one of those.
It would not need as much resistance to drop it down to 60VAC as the secondary tap does. ...

You should assume the marked "65v" is a hard upper limit.

Why? I think I see 100v-rated filter caps on that board. 65v * 1.414 = ~92v and you should assume the caps are rated for the peak of the incoming voltage. That's because absent current draw (which admittedly should occur as those caps try to charge at power-on), the caps will see that peak voltage.

If it were my amp, I would be looking to have a 60-65vac at the most from a winding. Though, you could probably guesstimate resistors to drop the voltage using PSU designer.

[Paul1453]

It is definitely a primary winding.

To me it appears to have been used with 230VAC somehow, and to run the fan.

Does this diagram help?

[HotBluePlates]

It is definitely a primary winding. To me it appears to have been used with 230VAC somehow, and to run the fan.

Got it. It's a dual-primary PT for 115v/230v.

If you want 115v operation, you run the 2x primaries in parallel, and connect them to the line. If you want 230v operation, you run the 2x primaries in series, and connect them to the line.

Theoretically, the PT would operate most efficiently if you energized both primaries. But I don't think you're anywhere near your PT's VA limit, so you could ignore that.

You could look at it as a 120vac secondary. I don't know how much current you could safely pull through it, but you don't need much current for a bias supply. I suppose you could use it with the bias board if you have a series resistor from pin 6 to one of the bias board inputs, and another from pin 7 to the other input. Come up with a notional value of resistance by adjusting the Ω of the PT in the PSU designer program; the resistance in each leg would be half of that total resistance which yielded <100v for the bias supply caps.

[Paul1453]

I only connected 12 and 13 with the switch and fuse, but still have my 120VAC on 6 and 7.
I was a little surprised by that.  So in effect, I have another 120VAC secondary to run the bias circuit. 

You are absolutely correct on running this PT well below her limits.
I played the amp for about an hour, and felt my major components to see if any were hot.
OT, choke, cap can, all room temperature, not hot at all.
Touched the PT and it actually felt cold, it has a rather large core.   

Ekdenton suggested I might have sacrificed some clean headroom running the 6V6s at this lower voltage.
I'm not so sure about that.  I know I have sacrificed some output power.
Biased the way she is, I estimate I'm only putting out 10-12 Watts.

To test the clean theory, I got out a single coil guitar and dimed the Master Volume.
If I keep Volume 1 and Bright Volume at 1 or less I have a very clean output.
Not super loud, but very clean.  Turn Volume 1 or the bright volume to 3 or more and I have both preamp and output tube distortion.
And she gets much louder.

Looking at the data sheet, and extrapolating higher voltages because it doesn't go to 400+V.
It seems I am at the sweet spot for THD.  As voltage rises, the 3rd harmonic climbs with it to the end of the graph.
Around my voltage level is where that 3rd drops down to it's lowest level.

If I put the MV around 2-3 and Vol 1 and or Bright volume at 3 or more I have nice 12AX7 preamp distortion.
I can get a pretty aggressive growl from her at very low volume.
Crank them all up and she will grunt and howl with the best of them.   

I'm hoping that 4 output tubes will bring that clean volume up to a gigable level.

I just replaced my close enough to give it a go used cathode bypass caps with new .68uF as specified ones.
I didn't test that yet, but expect it only gets better with new as designed parts.   

[Paul1453]

My Quad layout is starting to firm up.

I have repositioned V3 the PI to near the output tubes.

Input jack and preamp tubes V1 & V2 are in the far left corner still.
They are as far away from HV, OT, and the output tubes as this chassis will allow.
I have made a parts board layout for this pre-amp section.
All input wires, shielded wires, and connections to these tubes will be as short as possible.

I cut my parts board in two, to keep all the wiring connections as short as possible.
I'm actually considering having a third small board for the individual biasing circuitry.
I have made a parts layout for V3 and the NFB circuit.
It now seems best to me to cut that board down to that size,
and make another small parts board layout for my biasing components.
That will allow me to place that board in the best spot possible,
to work with my Chinese bias adjustment board while keeping connections to the output tubes logical and short.

Some of you may wonder why I am making all these changes to a proven design and layout. 
All of my tubes are vintage Mullard, Bell & Howell, RCA 6V6 tubes.
I am concerned about just burning these vintage tubes up with the higher voltages.
If I were buying new parts, I'd definitely go with the higher voltage PT and JJ 6V6s.
I'm hoping that by running these old tubes conservatively,
I can maintain that vintage tube sound and get another 50+ years of service out of them. 

[HotBluePlates]

... Some of you may wonder why I am making all these changes to a proven design and layout. 
All of my tubes are vintage Mullard, Bell & Howell, RCA 6V6 tubes.
I am concerned about just burning these vintage tubes up with the higher voltages. ...

How much higher is your expected "higher voltage" (I was, maybe mistakenly, assuming you were using another of the same HP PT's)?

I ran old production 6V6's in the '67 Princeton Reverb I used to have for years at 420vdc and more. I only ever swapped output tubes out of curiosity to see if other brands would sound different (they didn't really, likely because of the negative feedback).

I run old early-50's 6V6's in my 5E3 Deluxe copy, which probably has ~360vdc on the plates and something similar on the screens.

I also don't think you'll burn up any preamp tube, ever. I have a pair of McIntosh MC-30's which run a 12AX7 with a total of 490vdc from plate-to-cathode (plate is at 445v, cathode is at -45v). Again, I'm using mostly old production tubes, primarily because the ones I have I bought at prices lower than new production.

If you're not topping those cited supply voltages, I don't think you'll burn up a tube unless either:
a. You make a wiring mistake; we can all kill good parts that way, or
b. The tube already has one foot in the grave and the other on a banana peel.

[Paul1453]

Yes I am planning on using another PT from an HP 214A, same as 1st one.

I could still try to use the PT that the Belvedere had.
It powered 2 5U4s and I'm not sure what the 6V6 plates actually had on them.
I saw 320VDC off the rectifiers, but it then went through a field coil speaker and resistor.
The HP PT fit perfectly in the hole already there for the original PT.
I thought I would save the original PT to use with a GZ34 rectifier on a different build.
The Plexi uses SS rectification and saves me space in the chassis.

I guess I'll have to try higher voltages on a few of my lower performing 6V6s. 

My 1st Plexi also used an OT with around 8.5K at 8 ohms and that seemed to fit well with data sheet values.

I'm not sure exactly what this quad OT has for load values yet.
My 9 volt battery in my meter just went below operational voltage, and I'm getting goofy readings from it now.
I need to get some new 9V batteries so I can test this quad OT and a few others I'm making plans for.

I have some PTs with 5V windings that can give me a 450+VDC B+.
I have two 6L6 builds on the drawing board, and a pair of EL34s without a matching OT to consider.
I was trying to put my lower voltage, no 5VAC winding, HP PTs to good use.
I still have a 3rd HP PT, lower performing 6V6s, and a OT that would work for another of these Plexi's.
I think I'm getting a pretty good handle on this modified design/layout.

I realize I may not be getting everything I could out of this design and my parts.
But the 1st one sounds Really Good, even with my gumbo soup of used parts.

I'm making much more professional looking parts layout boards on this one.
The design visualization portion is almost complete at this point.

After staring at all my original amps layouts for untold hours,
and trying to figure out how to make them work with a guitar amp design,
and finally realizing you can't get there from here.

After you helped me remove my mental block,
I can now see where the parts need to go,
and how I can position/route things to give me a good chance at completing a hum/oscillation free build.
I really appreciate you removing the scales from my eyes! 

[HotBluePlates]

Yes I am planning on using another PT from an HP 214A, same as 1st one.
...
I guess I'll have to try higher voltages on a few of my lower performing 6V6s. 

My 1st Plexi also used an OT with around 8.5K at 8 ohms and that seemed to fit well with data sheet values.

I'm not sure exactly what this quad OT has for load values yet. ...

Since you're using the same PT, and the same circuit but with additional current because of doubled output tubes, you should have similar supply voltage (in part because that PT seemed to have a lot more current capacity than what the circuit demanded). If anything, output voltage might drop a bit because of the doubled output tube plate current.

Since the supply voltage will be about the same, OT primary for twice the output tubes should be half what it was in the 2x 6V6 amp. So you need ~4kΩ or so.


Were you thinking the output tube plate voltage would be higher because of the 4kΩ vs. 8kΩ OT's? If so, remember that this part is only "4k/8kΩ" for a.c. signals; it's d.c. resistance is much less (in the 100-300Ω range) so plate voltage drop due to current draw is small.

I have some PTs with 5V windings that can give me a 450+VDC B+.
I have two 6L6 builds on the drawing board, and a pair of EL34s without a matching OT to consider.
I was trying to put my lower voltage, no 5VAC winding, HP PTs to good use.

I though the goal for this amp was "same as the 2x 6V6 Plexi, but louder".

You have a known-quantity in your last build. And for double-loud version of last build, double-tubes and half-OT-primary-impedance is a right answer with no other changes, again since the PT you're using had more than enough current available on the high voltage winding.

You could switch to 2x 6L6 or 2x EL34, but then you're having to re-think the bias supply and you already seemed to have committed to the bias board you got.

[Paul1453]

PSU suggests I will have maybe +318 instead of +325 for my A voltage.
I guess I'll find out for sure when it all gets hooked up.

I'll be using the OT from the quad Belvedere, so it should be the 4K you suggested.
I just like to test and verify exactly what values the parts I use actually are.
I read some info on the Belvedere which stated it could use it's internal speakers, external speakers, or both simultaneously.
There was some kind of switch involved which maintained the proper load on the output tubes.
So I want to ensure I correctly document the OT specs before I try to install it.

Yes, the intent is to have a louder version of the original.
It seemed to me the 1st one wouldn't quite be loud enough for a gigging musician.
But, I really don't have a clue about all that.

This one also had the purpose of making something usable out of my junk pile.
With the intention of bringing my workmanship up to a standard,
a professional musician might consider spending his hard earned money on.

I'm pretty confident this one will sound excellent, and last a very long time too.

[Paul1453]

Proposed layout v1.

I need to wait a couple of weeks for the bias board and other parts to arrive from Asia.

This is what I am considering as a reasonable layout now.

Exactly how that bias board can mount properly will also determine the placement of my last parts board.

As of now I am actually considering mounting that board vertically in that back corner.

However I'm flexible enough to consider other options if this location becomes problematic.

In the meantime, I'll clean and paint the chassis, and focus on making my parts board and wiring connections as professional looking as I can at this time.

I think this layout looks to have reasonably high chances of hum/oscillation free operation.

What do y'all think?   

[HotBluePlates]

On the outside view, what about moving those 2 preamp tube sockets from the lower-right corner of the chassis to the lower-left corner? Then move the choke to that old spot in the lower-right.

That would keep the position of the preamp tube sockets close to each other, and get the potentially-noisy choke away from them. It would also avoid a long signal wire run from those sockets way over to the phase inverter socket & component board.

The bias board doesn't need to be close to the output tubes themselves, or even the bias-feed resistors which might be on your circuit board. The bias board is just generating a relatively-clean d.c., and the wiring from it to the output tube-related components can be long.

[Paul1453]

Thanks for the feedback, HBP!

Putting the preamp tubes in a line over there was an option I had not considered.

Probably because I didn't visualize turning the parts board, and the resulting pot connections.

I'm not quite thinking outside the box enough on my layout designs yet.   

But I think I am getting much better.   

I have taken some actions that leave me pretty committed to my current layout.

My layout keeps the majority of my connections short and as far away from noisy trouble spots as I could visualize them.

The signal line from the MV to PI should be 3" or less, and come from the front of the chassis away from the noisy area.

I could use shielded wire on it if necessary.

I have made some changes to my chassis to improve its' appearance.

I think you will be quite surprised and hopefully impressed by the next pictures I post of my progress.   

[Paul1453]

Since the supply voltage will be about the same, OT primary for twice the output tubes should be half what it was in the 2x 6V6 amp. So you need ~4kΩ or so.

So if I understand this statement and I'm looking at any data sheet I can take the PP AB1 Plate to Plate (Effective Load) value and use 50% of that value for a Quad output section of those tubes?

For example I have a Bogen K10 with Dummyload's mods that uses a pair of 6AQ5 tubes.
It sounds very good, and I have a bunch of 6AQ5 tubes.
So let's say I wanted to try a Quad 6AQ5 output section with this circuit.
I look at the Data sheet and find a pair of 6AQ5s wants a 10K plate load to make 10W of output power.
To run a Quad of 6AQ5s I need a 20W OT with a 5K plate load with my desired ohm value of speakers.
So if I had a 20W OT that had a 10K load using 8 ohm speakers, I could use that OT with a Quad of 6AQ5s if I used 4 ohm speakers?

HBP, you are giving me the knowledge I need to turn my pile of junk parts into lots of great sounding guitar amps!   
That you very much! 

[HotBluePlates]

... So if I understand this statement and I'm looking at any data sheet I can take the PP AB1 Plate to Plate (Effective Load) value and use 50% of that value for a Quad output section of those tubes? ...

Yes, as long as you're also copying every other element of the operating condition shown (plate & screen voltage, bias, etc).

But some data sheets already show that if you don't need the maximum clean output power, or some specific balance of distortion components, then the load is non-critical. Check out the post I wrote on output tube loads. There's a 5:1 spread of loads shown on that sheet's graph, from 2kΩ to 10kΩ.

... For example I have a Bogen K10 with Dummyload's mods that uses a pair of 6AQ5 tubes.
It sounds very good, and I have a bunch of 6AQ5 tubes.
So let's say I wanted to try a Quad 6AQ5 output section with this circuit.
I look at the Data sheet and find ...

The data sheet load is relevant to the specific operating condition shown, as well as the resulting power output and %THD. In this case, it's much simpler.

If you already have a working amp, and you like how it does what it does, skip the data sheet. You already know the plate & screen voltage for the output tubes, the bias used and the OT primary impedance.

Now if you want double-power with double output tubes, just halve whatever OT primary impedance you had for your smaller version of the amp. Keep supply and bias voltages the same. Ensure you have a PT which can deliver double the high-voltage winding current (at the same voltage) as the smaller amp.

If you look at schematics of Fender amps using 2x and 4x 6L6's, this is essentially what they did. The voltages on the output tubes are essentially unchanged, Fender added two more 6L6's, and the OT primary impedance dropped from 4kΩ to 2kΩ.

[Paul1453]

Look Ma, No Holes!   

Well almost.  The one in the midst of the output tubes, I guess I'll just call that a ventilation hole. 

One thin coat of primer and still verifying the parts fit properly.
My output tube socket holes could have been better, but will still work.

I'm considering using Bondo to completely level out the surface now.
I've got a can of glossy black paint, or might get some glossy red for final coat.
What colors do you guys like?

I really am trying to give this amp a professional looking finish.
I'm pretty confident it will sound professional enough.   

[Paul1453]

Picture of the guts progress.

I'm very happy with the layout and wiring so far. 
I don't think I could have made the wiring any shorter,
or ran it with a better chance of a successful outcome.
There are no wires from board 1 to board 2.
Board 2 gets Master Volume from the pot on the front.
Has about 1" wiring to the PI, and the only long wire runs are to the Presence Pot and NFB.

I only need to finish up the PS and output sections now.
The output section is dependent on the bias board, so I hope that gets here soon.
The PS is a piece of cake to finish, as it has plenty of space.
Short wire runs with excellent Grounding are my overall design objectives.

I think this layout might even be an improvement on my 1st one.
What do you guys think of my layout so far? 

[Paul1453]

HBP, so looking back at my output circuit modifications.   

You suggested, maybe I should change the signal caps from .022 to .01 and the bias injection resistor to 220K.
Is that still your opinion as to what I should do to best adjust for splitting my input signals in two?

The screen resistors are labeled as 3W, but the current flowing through the screens doesn't seem to justify that Wattage.

Would 1 or 2 W screen resistors be acceptable?

I'm looking to work on the layouts of my final parts board and PS.

Once my parts come in this should come together rather quickly.

[HotBluePlates]

You suggested, maybe I should change the signal caps from .022 to .01 and the bias injection resistor to 220K.
Is that still your opinion as to what I should do to best adjust for splitting my input signals in two? ...

I suggested 0.01µF in place of 0.022µF only because at the time you were out of 0.022µF caps and seemed unwilling to buy any.

Unless you've tested a circuit change and found you prefer the sound/performance, I'd suggest following the schematic values.

... The screen resistors are labeled as 3W, but the current flowing through the screens doesn't seem to justify that Wattage. Would 1 or 2 W screen resistors be acceptable? ...

You have to think about max signal conditions for screen current. Some of the typical curves (which appear to be endlessly copied) for 6V6's shows at a screen voltage of 250v, peak screen current might rise to ~20mA, or an average current of ~11-12mA. A nomograph in one of the RCA manuals suggests that an increase of screen voltage from 250v to 400v will double screen current compared to the 250v condition (take this with a grain of salt). So we might expect 40mA peak or 22-24mA average screen current, at max signal.

24mA through a 470Ω resistor implies 270mW, so a 1w resistor is fine. If you use 1kΩ screen resistors, the dissipated power will be a little more than double that, and a 1w resistor will start getting warm.

If the screen resistor were to burn open, tube current will drop to almost zero. If the screen resistor burns open, you would prefer you have a flameproof device. You fave suppliers may/may not have flameproof 1w resistors, so that may be a factor in your choice. Obviously, if you drastically raise screen resistor value (for less output power & more sag), then I²R should tell you you'll need to check & possibly raise the wattage of the resistor used.

[Paul1453]

I suggested 0.01µF in place of 0.022µF only because at the time you were out of 0.022µF caps and seemed unwilling to buy any.

Unless you've tested a circuit change and found you prefer the sound/performance, I'd suggest following the schematic values.

OK, I thought this had more to do with retaining the original signal frequencies and the resistor change to 220K for the levels.   

I'll stick with those values if the split doesn't really change the levels or frequency response.
That is what I'd be forced to do anyway, if I didn't have someone knowledgeable like you advising me! 

Yeah at the time I was out of salvaged .022s and didn't really want to dig through my junk pile to try to extract some more.   

Since then, I have received a huge lot of new .022uF 600V caps.
So no scrounging required now.

Metal resistors instead of carbon for the screens, I got it. 

[HotBluePlates]

I suggested 0.01µF in place of 0.022µF only because at the time you were out of 0.022µF caps and seemed unwilling to buy any.

Unless you've tested a circuit change and found you prefer the sound/performance, I'd suggest following the schematic values.

OK, I thought this had more to do with retaining the original signal frequencies and the resistor change to 220K for the levels.   

I'll stick with those values if the split doesn't really change the levels or frequency response.
That is what I'd be forced to do anyway, if I didn't have someone knowledgeable like you advising me!

Yeah at the time I was out of salvaged .022s and didn't really want to dig through my junk pile to try to extract some more. ...

Oops! You're right. I did say switch to 220kΩ resistors, because you will have twice as many paths to ground, and it would maintain the same load on the phase inverter. And so you'll lower the 0.022µF caps to 0.01µF caps, because the bass roll-off of the coupling caps is set by the R-C interaction of those 2 components.

I lost track, as it's been almost 2 weeks, and I have my own build gearing up for assembly.

[Paul1453]

Built as advised.

Positioned as visualized.

Wire color coding transcribed!   


They may be odd colored wires, but salvaged from HP gear.

It is the finest wire I have had the pleasure of working with.


Things still seem on track to me.
No obvious roadblocks in my way.
Placing bias board and PS components, are the only design issues left.   

[Paul1453]

Bias board arrived today.

I'll need to remove the pots and put them on the back of the board.
The caps are too tall, and I want the pots easily tweaked without sticking a screwdriver in a dark hole. 
Figuring out how to mount this board, what resistors to use on the AC supply, and then building the PS nodes are all that's left.
I hope to have this completely finished by the end of the month. 
If it sounds as good as my 1st 6V6 Plexi, it should be one hell of a nice amp. 

[Paul1453]

Swapped the pots to the back of the board.

Gave the board 6.6VAC from my bench PS and adjusted each supply to -7.3VDC as the max.

Had to make some changes to my planned visualization for this part, but I think I got it laid out well now.

Final chassis modifications/improvements under way now.

Almost ready for painting and final assembly. 

[Paul1453]

I've got my output circuit built as advised by HBP, and I'm looking at a few other designs that have a quad of output tubes.

I don't think that the way I was advised to build this is wrong, I'd just like to understand the different circuits configurations better.

How about comparing mine to the Fender 6G8?

There are a number of differences I'd like to know more about.

Anyone willing to discuss quad output circuits and the different proven designs with me?

There seems to be a number of different circuit configurations that work well.

I'd like to better understand the differences of these.

[sluckey]

The only differences are in the grid circuit. Your differences are because you wanted individual bias control for every tube. The 6G8 circuit is much simpler because Fender provided no bias control/adjustment for any of the tubes.

[Paul1453]

I get that I added complexity with individual bias.

Another thing I notice is that the Fender only has grid resistors on 2 of the 4 tubes.

This bias supply resistor for 2 tubes is 220K.
The Plexi circuit was modded from 100K to 220K resistors, but it has 4 instead of 2 resistors.

I've seen another quad circuit where the 2 tubes with their plates tied together only had the signal input on the grid of the 1st tube.
the 2nd tube of the pair had a grid resistor that the bias voltage came in but no signal from the PI on the 2nd tube.
I think I understand how that might work, where the 1st tube ends up transferring the input signal by having it's plate tied to the 2nd tubes plate.
This would use less components, and simpler wiring, if it accomplishes the same thing.

I'm trying to understand how the coupling cap, grid resistor, and bias supply resistor all work together in the quad output circuit.   

[sluckey]

I've seen another quad circuit where the 2 tubes with their plates tied together only had the signal input on the grid of the 1st tube.
the 2nd tube of the pair had a grid resistor that the bias voltage came in but no signal from the PI on the 2nd tube.

I think you're mistaken. Could you post that schematic or a link to it?

[Paul1453]

Mistaken?  That wouldn't surprise me, it happens quite often.   

This is gathered from other DIY web sites that may contain errors.
I also could have just misinterpreted this schematic.   

It seemed the original circuit my parts came from used something similar to the Fender design.
Where the 2nd tube doesn't have it's own grid resistor.

[sluckey]

All 4 of those output tubes get signal.

[Paul1453]

I'm still in the experimentation/learning phase of this tube amp design thing.

The individual bias circuit appealed to me to be able to balance dissimilar performing tubes.

Another application that I just discovered and intrigued me was:
Matching a 6V6 with a 6L6, or 6L6 with EL34, or 6V6 with EL34 for one pair and 6V6 and 6L6 for the other.
As long as the individual bias supply could accommodate the different voltages required by the different tubes,
they might be able to be mixed and matched to get different sounds from the same amp.

Maybe that mix and match, results in poor sound or other issues?
Maybe it allows each tube to impart it's own flavor to the sound that sounds very good, and different than all other amps?
At a minimum, there is much for me to learn experimenting and understanding both successes and failures.   

[Paul1453]

I am also studying books.

The Radiotron 3rd edition, as recommended by PRR, doesn't seem to be as sleep inducing as the 4th edition.   

But book learning only gets me so far in my understanding.

There is something about constructing a circuit and making some noise with it,
that allows that book information to take root and stay with me.
Failures become etched in my mind, from the other sensory stimulation like smoke and melting wires/components.
Successes make more sense, when I vary a components value and get sensory feedback as to why certain values are more or less successful.

IMO, hands on experience is far more valuable than just pure book learning.   

[HotBluePlates]

... Another thing I notice is that the Fender only has grid resistors on 2 of the 4 tubes. ...

Grid stoppers.

At the time, Fender usually didn't put grid stopper resistors on any output tubes. When you look at the layout for the circuit you sampled, you'll see the tubes with a direct wire from the phase inverter coupling caps got no grid stopper, while the "remote tube" for that side did get a grid stopper.

What have you read about grid stoppers? When reading, what was the intended function for those? What do they do for the tube relative to stray wiring capacitance & inductance? Are long grid wires a good or bad thing?

This bias supply resistor for 2 tubes is 220K.
The Plexi circuit was modded from 100K to 220K resistors, but it has 4 instead of 2 resistors. ...

You call it "bias supply resistor." But if that output tube was cathode biased, the resistor would run to ground instead of a bias supply; under those circumstances, I usually call it a "grid reference resistor" (because the grid needs to be referenced to ground to function).

The phase inverter has a plate load resistor. But there is also a coupling cap connected to the plate, and for high enough signals that looks like a piece of wire. The coupling cap connects to the grid reference/bias feed resistor, which for either form of biasing runs to an a.c. ground (either "actual ground" or a cap-to-ground). So the grid reference/bias feed resistor is in parallel with the plate load resistor, making the effective load lower, which reduces phase inverter gain and probably output voltage swing.

For the phase inverter, the ideal case would be an infinite grid reference/bias feed resistor, as gain would be unchanged from that determined by the plate load resistor. But an infinite grid reference/bias feed resistor is bad for the tube (normal grid current in the microamp or lower range will counteract the intended bias voltage) and for the tube's ability to recover from grid signal overload (big signal transient causes grid current which charges the coupling cap and shifts the tube bias, resulting in blocking distortion, while the big RC value due to grid reference/bias feed resistor and coupling cap results in slow discharge of the cap).

Most everything within an amp involves a design trade-off where best-operation of one part of the circuit requires component values which hinder the best-operation of another part of the circuit. Understanding why a particular amp used certain values boils down to understanding what the designer prioritized in the tradeoffs they made.

[Paul1453]

Grid stoppers help reduce or eliminate parasitic oscillation and should be connected right at the tube socket.

So grid stoppers on all the output tubes should be a good thing.

And just double checking my mods,
The 220K bias feed resistors keep the PI's gain and voltage swing up.

If I pull 120VAC from that other tap as planned for the bias board,
just stick with the 220K resistor and see what it gives me?
If I don't have enough - bias, then slowly reduce that resistor?

I don't have another 25K pot yet, closest I have measures 17.7K.
So if I use that for the Mid pot, should I add an 8K resistor between it and Gnd to retain as designed TS values?

I've just about finished patching holes and sanding and painting.
About to begin wiring it up soon.   

[Paul1453]

Someone has a tag line "Plan to be wrong about something".

Mine should be "Plan to be wrong about most everything".

So, I'm getting close to final assembly.

I need to verify everything, and I finally test out the quad OT.

I attempted to do this before, but hit a snag.
I was testing my stock of transformers, and killed 2 of my cheap Chinese meters.
You can't give those cheapo meters more AC than they are labeled for without causing their chip to go goofy.
If you go from end to end on a 380-CT-380 or greater PT, you will destroy a cheap meter. 

I got a new meter and test my quad OT.  Z is way out of line for 8 ohm speakers.
It used a field coil speaker originally, so I have no idea what the speaker's voice coil's were.
An 8 ohm speaker will give me 12K plate load, and I need around 4K.  No good.

It seems many things are counter intuitive in tube amp design.
You have a PP output circuit that wants an 8K load.
You give it twice the tubes to push the load, so you need twice the load right?
No, that's wrong.  You need half the load.   

Luckily, I have a number of other OTs, and a PP 6L6 OT I have will give me like 4.7K at 8 ohms.
So, a last minute change has me putting this OT over the hole for the quad OT.
It fits, and looks to be good in relation to the PT and choke orientation.

What is the consensus on putting a light coat of paint over an uncovered OT?
I would expect this not to be a real problem, but then again I seem to be wrong about almost everything.
Is it OK for me to put a light coat of spray paint on my OT and continue with the build?
I can put it in unpainted, but was really trying to make this one look as good as I hope it will sound.

[Paul1453]

... Another thing I notice is that the Fender only has grid resistors on 2 of the 4 tubes. ...

Grid stoppers.

At the time, Fender usually didn't put grid stopper resistors on any output tubes. When you look at the layout for the circuit you sampled, you'll see the tubes with a direct wire from the phase inverter coupling caps got no grid stopper, while the "remote tube" for that side did get a grid stopper.

What have you read about grid stoppers? When reading, what was the intended function for those? What do they do for the tube relative to stray wiring capacitance & inductance? Are long grid wires a good or bad thing?

Did I pass my Grid stoppers quiz?
I got my answer directly from Radiotron 3rd edition, Chapter 2, page 16, item (8) Parasitic Oscillation.
I hope I got it right, and was close to the answer you were looking for. 

[Paul1453]

I think I have found the right value for the bias board resistor.

I started with the 220K and only got like -2.7 to -.6 Vdc out of the board.

I had 3 of the pots set for max - and 1 set for min - to get my range.

Reduced resistance slowly, range moves slowly more negative.

Finally got to what I think should be a good bias range for a number of output tubes.

With a 10K 2W resistor, I get an approximate range of -39 to -11 Vdc from the board.

The 6V6s with 325 B+ on the 1st build biased at around -19V.

With this range, I think I might be able to try pairing a 6V6 with a 6L6 or EL34 and get it to work properly.   

Almost time for a chassis picture.  I've got an artistic daughter, and have her adding some unique artwork to really set this amp apart.

I think people will be very impressed by the finished product.   

[Paul1453]

Well I learned something.

You can't rush an artist.

Especially a female teenage artist.   

I can probably finish wiring this thing up in a couple of hours.
But I can't do that with wet oil paint on it.

You know the old saying?
"Good things come to those who wait".

I know this will be worth the wait.
I'm just an impatient grumpy old Dad,
who wants to hear his new creation.   

On the bright side,
It has given me some time to test out other ideas for the best use of my parts pile.   

[HotBluePlates]

... What have you read about grid stoppers? When reading, what was the intended function for those? What do they do for the tube relative to stray wiring capacitance & inductance? Are long grid wires a good or bad thing?

Did I pass my Grid stoppers quiz?
I got my answer directly from Radiotron 3rd edition, Chapter 2, page 16, item (8) Parasitic Oscillation.
I hope I got it right, and was close to the answer you were looking for. 

Yes, but let's think like a Radioman (Leo Fender started out as with a radio repair shop):
Wiring has stray reactances; stray capacitance to ground (and other wires), stray inductance along the path of the wire. Both kinds of reactance are small, maybe in the pF and µH range, but pF and µH may have a point of resonance in the RF range. So how do you typically squelch resonance? You reduce Q by adding R.

Fender may have noticed occasional problems with the long grid wires to the output tubes, and slapped a resistor mounted right at the grid pin for the output tubes positioned at the end of the grid wire furthest from the coupling cap. Later Fender amps got a grid stopper on each output tube socket.

[Paul1453]

Well I had to send my eldest son to go get this amp out of my daughters bedroom closet.   

Her painting is not finished yet, but she can finish that later.

I want to hear this thing make some noise, and see how well I did modifying the Plexi circuit for the Quad 6V6s.

It has been so long I had to spend some time just re-familiarizing myself with where I left off.
Connected up the individual bias wires to the Chinese board last night.
I will mount the last little board, and complete the wiring of the output section next.

Then I will mount all the pots and jack on the front and finish wiring up the pre-amp section.

Lastly I will return to the PS section.
I'm doing point to point wiring of those nodes.
I'll finish up with the rest of the PT wire connections and Power switch.

Maybe 4-6 hours of dedicated work left before I can verify my bias is set on pin 5 of all my output tube sockets.
Then I'll plug in all the tubes and check and document my voltages throughout the circuit.
Then I'll be anxious to hear how the "Black Widow" sounds.   

The plan is to get this up and running this weekend.
Once that is finished, my artist can complete her (not to be rushed by Dad) artwork.

I'm hoping (fingers crossed) and somewhat expecting this to be one hell of a nice sounding 6V6 Plexi amp.      

[Paul1453]

She's alive, and she's a real killer.   

I had a few issues after completing the wiring.
That is one of the challenges with re-using salvaged parts.
Rectifier diodes were the major problem.
Exploded an e-cap to find that out.
1 signal cap was leaking, so I needed to replace 4 parts to start.

Then she squealed something awful.
Swapped the OT leads, no more squealing.   

I really don't understand how that OT lead swap works.   
But it definitely does. 

Some fine tuning and the completion of her paint job are all that's left.

The "Black Widow" is a killer amp!

Thank you for your guidance HBP!   

[pompeiisneaks]

My understanding of the squeal is that its only when you have a NFB setup.  if the OT's are out of phase with the side that the NFB comes into it causes that horrid squealing (You create a PFB instead of NFB).  I didn't ever run into this on my first few builds as they had no NFB, but my recent Bassman build hit that problem :D

~Phil

[Paul1453]

My understanding of the squeal is that its only when you have a NFB setup.  if the OT's are out of phase with the side that the NFB comes into it causes that horrid squealing (You create a PFB instead of NFB).  I didn't ever run into this on my first few builds as they had no NFB, but my recent Bassman build hit that problem :D

~Phil

Thank you!

That does make some sense now.  A Positive feedback loop is a real squeal generator.   

You still have a 50/50 chance of getting it right the 1st time.
But whenever I use the leads that best reach the tubes, I usually get it wrong.