6AF11 Push Pull Amp

[hayruss]

Hi All
I thought I would come to the place where the tube amp knowledge flows like wine. 
So to start thanks in advance for anyone that can assist with this build.

I have been toying with the idea of making a pushpull guitar amp using 2 of the 6AF11s after making the squirrel monkey which worked reasonably well except for the painful oscillation.
Having scoured the internet for examples of I P-P 6AF11 I found 1 build but no schematic was forth coming.
So I bit the bullet and knocked up a drawing of the amp using some of the methodology and configuration of the squirrel monkey and the super squirrel monkey amps.
Living in OZ it dont have access to cheap output trannies like a lot of you in the US so I am making do with using 100V PA line transformers.
These have worked well in the past for me having built a P-P 12AQ5 amp, P-P EL84 amp and a quad P-P 12V6 amp all sounding great (even if I do say so myself after much tweaking to remove the muddy farty sound). 
Using these transformers has allowed me to keep the price down, also I often use roasting pans or cake tins as amp chassis. I know I hear you snicker in the background but hey I use what I can.

I have just found plate characteristics of this valve so I am a little confused on how to achieve the correct bias for the triodes and pentode.
You may notice that I have used a single pot tone circuit which makes up a local feedback loop for the 1st triode. Might not be the best place for this?

I am open to any and all suggestions in an attempt to make this a nice balanced awesome sounding amp while keep the costs to a bare minimum. Call me cheap thats OK.

OK then to the diagram. Drum roll please ratta tatta ratta tatta.

Cheers Haydn

[jazbo8]

I did one awhile ago, it's not very powerful - couldn't get the max power out of it without oscillation... you can check it out here >> http://jazzbo8.wordpress.com/2012/08/15/6af11-push-pull-compactron-guitar-amp/. Also what's going on with the first stage in your schematic?

Jaz

[hayruss]

OK Cool thanks.

Do you have a schematic of your amp?
The tone control is from a garnet amp, having a look at the diagram and it is missing a 1nF cap above the 500k tone pot.
I am thinking it wont need the 1M resistor to ground from the input to the 2nd stage.
Also the layout of the triodes is tube 1 low mu triode to tube 1 high mu triode then to tube 2 high mu triode and then tube 2 low mu triode and the phase inverter.

Do you think it is necessary to by pass the cathodes of the other 2 gain stages??

[kagliostro]

Here an SE guitar amp

http://www.solorb.com/elect/musiccirc/squirrelmonkey1/

Interesting to see a PP version

2 minutes ago I was talking with my son about the SE version

K

[jazbo8]

OK Cool thanks.

Do you have a schematic of your amp?
The tone control is from a garnet amp, having a look at the diagram and it is missing a 1nF cap above the 500k tone pot.
I am thinking it wont need the 1M resistor to ground from the input to the 2nd stage.
Also the layout of the triodes is tube 1 low mu triode to tube 1 high mu triode then to tube 2 high mu triode and then tube 2 low mu triode and the phase inverter.

Do you think it is necessary to by pass the cathodes of the other 2 gain stages??

Here is the schematic, quite similar to yours, I tried bypassing all three stages but the oscillation set in too early - so the third stage was not bypassed. The plate to grid feedback caps on stage 1 and 2 were also needed to tame the oscillation but YMMV. It was a fun build - tried out a few different things, but if I were to do it again, I probably would just go for the Super Squirrel Monkey...

Jaz

[PRR]

IMHO, three (or more!!) stages of gain on one socket is just *begging* for oscillation.

Use more sockets. Spread out. Keep input away from output, not a half-inch apart.

_I_ think this tube, and 100V OTs, will be happier with ~~150V supply than 280V supply. The peak available current, and G2 dissipation, can be WAY too high with G2 at 280V. The 200++V plate swings are too high for a 50V-rated (per side) OT.

[jazbo8]

Yes, I found that out, as hard as I tried I could not get rid of all the oscillation (without making it sound like a tiny, muffled amp). These tubes were never designed for such application anyway...

Jaz
 

[darryl]

I also attempted to use the 6AF11, but in a single-ended amp. It appeared to be impossible to achieve worthwhile gain without instability. The most effective result was obtained by grounding all elements of the 'middle' triode section - effectively using it as a shield between the other triode and the pentode. This did tend to negate the whole purpose of using a three-section valve though . . . 

The entire tale of woe is here:  http://www.guitargear.net.au/discussion/index.php/topic,19889.0.html

[kagliostro]

In the SE version of the Squirrel

http://www.solorb.com/elect/musiccirc/squirrelmonkey1/

in the page of the author we can read he used a B+ of 180v

he say the primary impedance of the OT that better suited the amp was 3.5Kohm

and the output is only 0.8W that seems a bit low to me for a 5W plate dissipation tube

http://www.nj7p.info/Common/Tube/SQL/Tube_query.php?Type=6AF11

Do you think the choice to tame the output power is due to avoid oscillations and other problems ?

Or there is a way to push a bit more of power from the pentode, say 2W ??

The datasheet say Class A Amp Plate 250v - Grid 2 150v - Plate current 24mA

Sometime ago I've draw a turret layout for that amp, but the transformer I was thinking to use was broken then I do other stuff

as you can read in my previous post on this tread I was talking about this amp with my son just minutes before to read this tread, so a new interest comes to me for both versions SE and the "new" PP

I think that an SE version in the ZVex style, will be a nice little amp



So if someone has councils for an improvement of this amps SE & PP I'm very interested to read

@ Jazbo8

didn't you considered a 2 gain stage + LTPI + PP  with a TS 

or to use one triode as CF and drive a TS ?

this will tame the gain of the whole amp, but maybe it becomes less prone to oscillations

K

[jazbo8]

kagliostro,

I didn't keep all my notes, but when I tried the CF -> TS configuration (in fact that's what I started with), that made the amp wimpy and sterile sounding to me... so I opted for maximum possible gain through the preamp (that's why no TS), and dialed it back down a bit using the volume pot. This is a tricky tube to be sure, but YMMV.

Jaz

[kagliostro]

If you give a look to the Squirrel Monkey SE, the tone control is obtained via an NFB control of the amp

now I begin to understand a possible reason for that

K

[hayruss]

preamp biasing

HI guys I am still really confused.
I am going over the average plate characteristics charts for the triodes to attempt to calculate the bias points and notice the plate current scale goes up to 35mA but the max plate dissipation for T1 is 1.1W and T2 2W.
Is someone could check my calcs but say running at 250V plate attempting to get a bias of -1V you are drawing 18.5 ish mA.
Watts = amps x Volts?
0.0185 x 250 = 4.625W more than 4 times higher than max allowed plate dissipation.
Why even include such highly extrapolated charts if there is no chance of running the valve at that level?
Using the same charts to calculate a bias point for T1 I choose 330V as being max voltage 0 amps and using a 100K plate resistor I get 3.3mA and mark that on the chart at 0 V.
Connect the 2 dots with a line. Really low values.  As I am writing this I have been thinking and doing further calcs.
If I calculated the load line using 33K plate resistor 330V max V no Amps and 10mA no volts.  I can use 200V B+ fro the plate and draw 4mA.
200v x 0.004A = 0.8W which would give me around -2.7V bias.
Adding a small value cathode resistor 1K or less will drop the plate diss down a little further but I think I am getting there.

Am I on the right track???
Or have I missed the boat all together??

Cheers H

[hayruss]

IMHO, three (or more!!) stages of gain on one socket is just *begging* for oscillation.

Use more sockets. Spread out. Keep input away from output, not a half-inch apart.

_I_ think this tube, and 100V OTs, will be happier with ~~150V supply than 280V supply. The peak available current, and G2 dissipation, can be WAY too high with G2 at 280V. The 200++V plate swings are too high for a 50V-rated (per side) OT.

Thanks for your insight regarding the OTs.

Here is a link to a site mainly about building DYI budget Valve amps.
http://www.ozvalveamps.org/optrans.htm
They go into detail about the 100V PA line transformers.
I have built there lammington AVA101. 15W P-P EL84.
Also some more detailed info about the line transformers here.
http://home.alphalink.com.au/~cambie/6AN8amp/M1115.htm

I definitely see where you're coming from regarding the pentode section of the valve but max plate and screen voltage is 330V but I do notice there is a rapid decay in screen dissipation between screen V 200V and 300V.

Cheers H

[jazbo8]

If you give a look to the Squirrel Monkey SE, the tone control is obtained via an NFB control of the amp

now I begin to understand a possible reason for that

K

Since I was not getting the expected output power, I did not bother with NFB which would have reduced the output even more but it certainly worked in the SM, so I may try the SE version when I get a chance.

Jaz

[jazbo8]

preamp biasing

HI guys I am still really confused.
I am going over the average plate characteristics charts for the triodes to attempt to calculate the bias points and notice the plate current scale goes up to 35mA but the max plate dissipation for T1 is 1.1W and T2 2W.
Is someone could check my calcs but say running at 250V plate attempting to get a bias of -1V you are drawing 18.5 ish mA.
Watts = amps x Volts?
0.0185 x 250 = 4.625W more than 4 times higher than max allowed plate dissipation.
Why even include such highly extrapolated charts if there is no chance of running the valve at that level?
Using the same charts to calculate a bias point for T1 I choose 330V as being max voltage 0 amps and using a 100K plate resistor I get 3.3mA and mark that on the chart at 0 V.
Connect the 2 dots with a line. Really low values.  As I am writing this I have been thinking and doing further calcs.
If I calculated the load line using 33K plate resistor 330V max V no Amps and 10mA no volts.  I can use 200V B+ fro the plate and draw 4mA.
200v x 0.004A = 0.8W which would give me around -2.7V bias.
Adding a small value cathode resistor 1K or less will drop the plate diss down a little further but I think I am getting there.

Am I on the right track???
Or have I missed the boat all together??

Cheers H

Some of the numbers look odd to me, I am not at home now, so can't look up my notes on what I did, but found the loadline for T2 on my dropbox, so take a look to see how it compare with your design.



Jaz

[PRR]

> drawing 18.5 ish mA.

Preamps are designed for high gain, not high power.

Preamp tubes use 50K-200K plate load resistors, set plate at *roughly* half of B+. Say B+ is 200V. Then preamp tube current is 2mA to 0.5mA, not 10+mA.

> Why even include such highly extrapolated charts if there is no chance of running the valve at that level?

It's a TV tube. Some TV circuits must *slam* from high-voltage low-current to low-voltage high-current, never lingering at medium-voltage medium-current where they would fry.

You are right: that 300V 35mA point would *never* be approached in any practical circuit. That graph didn't need to be drawn above 20mA (if that). Sometimes the back-room boys just torture tubes for fun?

Yes, a rational 1mA load-line is at the bottom of the graph where the lines are dubious.

You can rough-guess a G-K bias by picking a happy plate load resistor (somewhat less than the audio load, which is typically 200K) and a plate voltage (typically half of stage's B+). Divide P-K voltage by Mu. That's an approximate cut-off voltage. Cut that in half (or multiply by 0.6). That's a working guess of Vgk at that Vpk. Divide by current (other half of B+ divided by plate load resistor) and you have a good-guess for Cathode resistor.

200V of B+, 100K resistor, Mu of 41. Start with assuming 100V across tube, 100V across 100K. 1mA current. 100V across tube divided by 41 is 2.5V to cut-off. Half that is 1.25V. (0.6 times that is 1.5V.) So 1.25K, 1.5K, somewhere in there.

Mu is not really constant, declines at low current. So you may need a larger cathode resistor.

Guitar amps "often" bias plate above half of B+. (Fender tends to 0.7.) So you need less current, which means larger Rk.

So grab any resistor in the 1K to 2K range, stick it in (tack solder), and see how the plate biases. If 1/2 to 3/4 of B+, it will play.

[hayruss]

> drawing 18.5 ish mA.

Preamps are designed for high gain, not high power.

Preamp tubes use 50K-200K plate load resistors, set plate at *roughly* half of B+. Say B+ is 200V. Then preamp tube current is 2mA to 0.5mA, not 10+mA.

> Why even include such highly extrapolated charts if there is no chance of running the valve at that level?

It's a TV tube. Some TV circuits must *slam* from high-voltage low-current to low-voltage high-current, never lingering at medium-voltage medium-current where they would fry.

You are right: that 300V 35mA point would *never* be approached in any practical circuit. That graph didn't need to be drawn above 20mA (if that). Sometimes the back-room boys just torture tubes for fun?

Yes, a rational 1mA load-line is at the bottom of the graph where the lines are dubious.

You can rough-guess a G-K bias by picking a happy plate load resistor (somewhat less than the audio load, which is typically 200K) and a plate voltage (typically half of stage's B+). Divide P-K voltage by Mu. That's an approximate cut-off voltage. Cut that in half (or multiply by 0.6). That's a working guess of Vgk at that Vpk. Divide by current (other half of B+ divided by plate load resistor) and you have a good-guess for Cathode resistor.

200V of B+, 100K resistor, Mu of 41. Start with assuming 100V across tube, 100V across 100K. 1mA current. 100V across tube divided by 41 is 2.5V to cut-off. Half that is 1.25V. (0.6 times that is 1.5V.) So 1.25K, 1.5K, somewhere in there.

Mu is not really constant, declines at low current. So you may need a larger cathode resistor.

Guitar amps "often" bias plate above half of B+. (Fender tends to 0.7.) So you need less current, which means larger Rk.

So grab any resistor in the 1K to 2K range, stick it in (tack solder), and see how the plate biases. If 1/2 to 3/4 of B+, it will play.

Thanks PRR 

Fantastic information very useful.
I like that idea of torturing tubes for fun
I have been reading up a bit more on preamp load lines and things have become somewhat clearer and with the information that you have provided has help even more.

So this morning I have changed the diagram with my newly gained knowledge. 
Except for the power supply which I know I have to reduce the B+ and I still need to calculate the power amp bias resistor to an appropriate level.
I have moved the Garnet style tone stack (which is one of the designs that apparently does not chew all of your gain) to after the last gain stage.
Design for this came from the Adam amps web site. http://amps.zugster.net/articles/tone-stacks

I have calculated the pre-amp cathode bias using a 100K load line and the -3V grid line.
Rebiased the PI and added gris stoppers on the pentodes.

Can anyone see any problems here??
I have added any oscillation reduction components yet as I want to build it and then throw it on the oscilloscope to which stage is oscillating if any.
Yeah right not oscillate haha 

Thanks Again everyone keep the info flowing.

[hayruss]

Just trying to get the diagram bigger.

Also I adjusted the PI bias and used a 22K load line in an attempt to increase the current.
Not quite sure if this is right but it sort of makes sense to me.

I will be starting the build today. 

[kagliostro]

I'm waiting with interest your test 

about the TS there are a pair of it that as far as I know has a low gain loss

the James and the BoneRay TS





or you can try to split the tone control between gain stages as in this schematic

http://www.privateeramps.com/d6images/ak02sc2.jpg



K

[PRR]

> Can anyone see any problems here??

V2 T2 *needs* a blocking cap at the grid. (Grid needs to be up 50V-100V, the 1Ma gain-pot is at zero V.)

> which stage is oscillating

Any oscillation anywhere will show up in "all" stages (just like you want your guitar signal to happen in all stages).

Oh, remotely possible it oscillates in V2 P to V2 T1 loop only. But with V2 P throwing 200V swings, there's sure to be some coupling into T1. Only very detailed gain-analysis could tell cause from effect. And with gain of 30,000 from V1 T2 Grid to V2 T2 (and nearly a Million from V2 T2 grid to either power pentode!), the least coupling will give significant induced signals.

Most coupling will be capacitive. Lower impedances *may* defer trouble. These are pretty beefy triodes. And you have a ton of gain!! Maybe go to 50K plate loads, 100K even 50K pots, with 500pFd-1000pFd caps each grid to cathode.

Personally I think two stages of gain is a lot for such a compact build. Build it low-gain, put a pedal in front for dirt.

> about the TS there are a pair of it that as far as I know has a low gain loss

The James is 10:1 (20dB) loss (there is a 2:1 7dB variation).

That Garnet gives-up the added gain of cathode bypass cap, so is typically 2:1 (5dB-7dB) loss.

Any low-loss scheme has very little boost. If you only cut, that's fine. Many players like some boost for some work. The James throws-away 10:1 of what your tubes give, but can put it back at 50Hz or 5KHz if you turn the knobs all the way. The Garnet throws-away half the tube's potential gain but gives it back in mid-treble when turned all the way.

However I think this plan is already way too much gain, and throwing-away some gain in tone-stack isn't a bad thing. (OTOH, the long wires to tone-pot may pick-up or transmit more stray signal than tight on-socket wiring.)

> trying to get the diagram bigger.

It was big enuff. Click on the attachment thumbnail, it gets bigger. 2000+ pixels wide is plenty big enuff. (And for line-drawings, GIF or PNG is better than JPG. JPG smears fine detail for more colors and smoother skin-tone: better for people and landscapes, can be blurry for line drawings.)

[hayruss]

Hi All

Thanks Jazz and K for your input the load line you provided now makes sense after doing a fair bit of reading.

I have added a few pictures of the build so far, I haven't completed any of the circuit yet as it is still very fluid.
You can see from the pictures that it is built on a square cake tin and the 100V PA line output transformer in attached inside.
I am using a toroidal power transformer small in size but very adequate for the job.
After this build I am seeing if I can build a parallel push pull 6AR11 guitar amp but I am getting ahead of myself, I haven't even finished this build yet. (barely started really)

I will tame down the 2 high mu triodes a little (as suggested by PRR) and drop the plate resistors to 56K which will also mean I will have to drop the cathode resistors to 1.2K.
See how it goes I may even increase the cathode resistors.

PRR

V2 T2 *needs* a blocking cap at the grid. (Grid needs to be up 50V-100V, the 1Ma gain-pot is at zero V.)

Not sure what you mean here as it has a 22nF 630V coupling cap between the 3rd gain stage and the master vol.
If I was to remove the tone stack would that change things and make it OK?

Hang on hang on, I think I stuffed up the phase inverter. 
Yep I did attached is an updated schematic.

Now has that sorted it all out?

Cheers H

[hayruss]

Another picture

[hayruss]

One more picture.

[sluckey]

Not sure what you mean here as it has a 22nF 630V coupling cap between the 3rd gain stage and the master vol.

There must be a blocking cap between the wiper of the MV pot and phase inverter grid (V2 T2 pin 3) to prevent loading/killing the bias voltage on the grid. The PI grid bias is bootstrapped to the cathode.

He's not talking about a blocking cap between the previous stage and the input to the MV pot.

I like the looks of your amp.

[kagliostro]

My preferred kind of cake 
K

[hayruss]

Hi All

I have installed the filter caps and UF4007 diodes in the amp.

I am attempting to calculate the voltage dropping resistors between the filter caps for each part of the circuit.
Can someone run their eye over it please just in case I have gone about it all wrong?

Checking the schematic
The 1st resistor will be placed between the rectifier and the B+ A supply
So B+ A=250V, B=200V, C=180V and D=170V.
Each B+ section will use approx 80% of max wattage.
I = P/V
B+ (W) A = 8W @ 250V =32mA
B+ (W) B = 2W @ 200V = 10mA
B+ (W) C = 1.6W @ 180V = 9mA
B+ (W) D = 3.36W @ 170V = 20mA

200V AC from transformer full wave rectification x 1.414
200 x 1.414 = 283V DC

R=V/I
B+ A 33V drop to 250V - 33V/0.071A = 465 Ohms
B+ B 50V drop to 200V - 50V/0.039A = 1282 Ohms
B+ C 20V drop to 180V - 20V/0.029A = 690 Ohms
B+ D 10V drop to 170V - 10V/0.020A = 500 Ohms

P=I^2R (I squared x R)
B+ A  - 0.071A^2 x 465 Ohms = 2.35W (3W)
B+ B  - 0.039A^2 x 1282 Ohms = 1.95W (3W)
B+ C  - 0.029A^2 x 690 Ohms = 0.58W  (1W)
B+ D  - 0.020A^2 x 500 Ohms = 0.2W    (1W)

Many of the resistor values are non standard so select a standard value that is close.

Hopefully this is close? 

Cheers H

[jazbo8]

Good job! To double check, you can also punch in the numbers using PSDII and see if any discrepancy pops up.

Jaz

[PRR]

>> blocking cap at the grid.
> Not sure what you mean here as it has a...

[hayruss]

HI All

Made it through the weekend got a bit done on the amp.

Thanks for the clarification on the blocking cap. 

I have updated the schematic with a couple of small changes will post once I know what it sounds like. 

I do have a couple of pics of the progress so far.

Starting to get messy as it is a bit cramped for space.

I only have the star grounding and pentode cathode to install.

Ohh well here are the pics.

Not long now to see if it squeals like a stuck pig my guess is it won't because I am awesome  just kidding I think it will go off to oscillation land.

any bets?? anyone nah didn't think so its almost a sure bet.

Cheers H

[hayruss]

another picture

[kagliostro]

Hayruss

only one question

I didn't see any kind of spacers between the board and the chassis

to me seems that the board is too close to the chassis and there is contact between circuit and chassis

is this a temporary situation ?

K

[hayruss]

The board is mounted on 6 screws with a 2 - 3 mm gap between the chassis.

I generally hand wire directly on the sockets but wanted to try something different this time, a semi turret board.

Cheers H

[kagliostro]

The semi turret board is fine to me

I wasn't able to see the 2-3mm of space between the board and the chassis

my council will be to use a thin plastic foil between the chassis and the board as to avoid any problem

ultimately 2-3mm isn't a so large safety space having B+ around

Ciao

K

[hayruss]

Hi All

I finally got it all wired up made the last few connections about 40 mins ago.
I left the valves out of it to start with double and tripple checked the wiring checked for little solder dags and wisps of wire that might make connections to something that it shouldn't.
None could be found.  Re-positioned the components to ensure there was as much as space between them as possible.
Another final check over just in case.
OK time for some power. Connected the power cord and flicked the power switch on.
Proceeded to check voltages 310V B+(A) 305V B+(B) Hmmm very high considering the voltages I was aiming for then it dawned on me no load.
Checked other voltages like the heaters 3.2V per pin just right.
Left it running for about 5 mins to settle in and no apparent problems.
So far so good (fingers crossed).
OK time to throw the valves in, disconnected the power cord and discharged the filter caps installed the valves hooked up the DMM, 8 ohm dummy load, scope and signal generator.
Powered on valves light up quite bright really should check voltages and pin orientation just in case I inadvertently connected something incorrectly.
I did check several time though.
Slowly opened the gain pots to about 10% along with the master vol and waa-la (viola) a signal appears on the scope.
Wound the gain stages to about 40% and slowly increased the master vol.
Signal looked good a little angular squared on the bottom of the signal but nice and rounded on top of the wave.
Would it open further and started to do the same thing on the top half of the wave.
To describe the wave shape: the rising waveform peaks with at a point (squares off but at 45%) continues down at a 45% angle and then sharply turns downwards to repeat the inverse shape at the bottom of the waveform.

AnyWho took some voltage readings at the best somewhat but not too much clipping shape.

MA 8V into an 8 ohm load = 8W Hmmm a bit high me thinks.
At the pentode
Plates 205V each
screen 172V each
Cathode 7.4V
Cathode resistor 150 ohm

If my calculations are correct and they might not be then
7.4V/150 = 49.3mA
205V x 0.0493A = 10.11W plate dissipation OUCH 

I thought hurry up and plug a guitar into it.

2 words is all I have to say.

FIG JAM (I shouldnt have to translate that)

As I turned up the gains and master vol I was expecting to hear ear piercing squeals but all I heard was clean awesome sounding guitar amp.

I am still pretty chuffed over the whole thing.

Cranking it up to about 30 - 40% brought out that nice rock overdrive not bassy or muffled at all. 

Going passed the 45% on any of the gains or master vol found the signal reduced dramatically and the radio type buzz/fuzz appeared.

But the amp is not a squealer.

Any suggestions as to what is causing the sudden drop off in signal??
I think that one of the valves is receiving too much/strong a signal and has gone into cut off or RF, most probably the pentode because it happens with the master volume as well??

I am having a guess here.

Any and all suggestions welcome.

Thanks again

Cheers H

[kagliostro]

Read this post (especially replay #11 from HotBluePlates)

http://www.el34world.com/Forum/index.php?topic=15517.0

and make your considerations

I think that the approach can be to tame (using trimmers) the signal between the preceding gain stage and the gain pot

HotBluPlates explained very well this technique in detail

K

EDIT: Sorry, previously I forgot to add the link  
Without the link the sense of this post was completely different  

[PRR]

> passed the 45% on any of the gains or master vol found the signal reduced dramatically and the radio type buzz/fuzz appeared.
> But the amp is not a squealer

It's squealing. Just above the audio range.

Too much gain, particularly high frequency gain, in too little space.

[kagliostro]

If I can repeat my previous suggestion

you can tame the gain via split resistors and/or with the use of Tone Control circuit

(in this case may be needed both)

The Tone Control suck a part of the signal as well as give a control on frequencies

personally I would try a "splitted" Tone Control as in this amp



this will tame two gain stages

There are also techniques to prevent high frequency oscillations and safe here in the forum someone could suggest something  to try to do this

here some hint

http://www.tone-lizard.com/Mods_and_Odds.htm

K

p.s.: I edited my previous message, please read it now (Sorry)

[DummyLoad]

Too much gain, particularly high frequency gain, in too little space.

+1

my thoughts are that the only way you're going to have any hope of making this amp work well is to bond the two triodes together.

go in the other direction - use single triode (6AV6/6AT6/6DR4/6C4/6AB4 etc.) tv tubes or use dual triode compactrons (6AG11/6AY11). ground the diodes. you could use a 9 x 13 or even 12 x 18 cake pan...

respectfully,

--pete

[kagliostro]

@ Pete

Do you mean to join the first two triodes a la Matchless ?

What about the use of two triodes as two gain stages and the other two triodes as LTPI ?

K

[DummyLoad]

Do you mean to join the first two triodes a la Matchless ?

yes.

if i were on this project - i'd experiment as follows:

v1 hi-mu gca >> v1 lo-mu cf >> ts&vol >> v2 hi-mu gca >> v2 lo-mu concertina PI >> v1 & v2 pentodes in P-P. 

--or--

v1 hi-mu gca >> v1 lo-mu cf >> ts&vol >> v2 hi-mu gca >> v2 cf >> v1 & v2 paralleled pentodes for SE.  <--- probably the path i'd take.

--pete

[hayruss]

HI All

Sorry for the lack of replies but I have been flat out like a lizard drinking with work.

OK 1st things 1st.
This amp does not self oscillate and what I mean by that is that it doesnt go off to radio land or higher with the gain and volume about 50% while playing.
So it is controllable.
As PRR has said it has way tooooooooo much gain definitely correct.

So I paralleled the 68K plate resistors (Halving the resistance) on the 2 high mu stages (stages 2 and 3) but it sounded really thin so I removed them.
I have paralleled the cathode resistors on the same 2 stages which allowed me to crank the 2nd gain up a little before radio land set in.
I have also added a 220K signal dropper resistor?? between the wiper and ground in an attempt to bleed some of the signal off which allowed me to crank it up a little more without affecting the output volume.
I have also paralleled the 2 x 470K resistors on the input to the pentodes to drop some of the signal.
I increased the Pentode cathode resistor to 300 ohm to achieve 6.83W plate dissipation.

The amp does sound good though nice overdrive.
I measured the output V on the scope just playing cords on a telecaster so I could hear the max volume level before oscillation.

Between 6V AC MA normal playing jumping up to 13+V AC MA (mean average) agressive playing into an 8 ohm speaker.
So at least 5W output.
Plenty loud enough for home or bedroom use with a nice bite and overdrive definitely not a gigging amp.

I have a couple of questions if someone could pose an answer that would be great.

What effect would it have if I increased the cathode resistors to say 15K on the 3 gain stages??
Is the placement of the signal dropping resistors correct?

Thanks All

[John]

What effect would it have if I increased the cathode resistors to say 15K on the 3 gain stages??

Short answer, give you even more gain.

If I understand DL correctly, he meant to parallel the first tube- it will give you much less gain than if it's cascaded. Just connect your grids together, your plates together, and your cathodes together. Then, use 1 resistor for both plates, 1 for both cathodes. Also experiment with the plate resistor. You can lessen gain quite a bit by lowering the value of that. To make it easy to find your sweet spot, carefully connect jumper wires and clip varying resistors in and out, saves you a lot of soldering.

Is the placement of the signal dropping resistors correct?

If I understand the way you did it, not really. Try connecting a fairly large (330K to 470K) resistor directly to pin 6 of V2 (assuming you are going to parallel V1) In my limited experience, it doesn't affect volume much at all, but does attenuate high frequencies (that you can't even hear, many times) and may help with the oscillation.

I'm basing this looking at your schematic, so if things have changed I apologize for not thoroughly reading the whole thread. Also, keep in mind I'm still learning too, so don't take any of this as gospel!

[sluckey]

What effect would it have if I increased the cathode resistors to say 15K on the 3 gain stages??

Increasing the value of the cathode resistor will DECREASE the gain.

Is the placement of the signal dropping resistors correct?

Not really. It will only decrease the gain a little bit. A much better way to lower the gain in this amp would be to put a 470K resistor between the .022µF coupling cap and the top of the gain/volume pot. (This is what Marshall does on the JCM-800 2204 amp). Do this for the first and second gain stages for sure and maybe even do it for the third gain stage.

Radio noise... Your amp would benefit from using a grid stopper resistor between the input jack and grid of the first gain stage. Using a stopper between the gain pot and grid of the second and third gain stages would also be a good idea. I'd use something like a 33K, 47K, 56K, or 68K. And physically connect the resistor directly to the tube grid in all cases.

Did you ever put the blocking cap on the grid of the PI as PRR suggested?

[hayruss]

Excellent info Thanks guys.

I checked out the link that K put in for the tone lizard and it also said the same thing regarding increasing the cathode resistors.
Thanks K and Sluckey

I have removed all of the dropper resistors that I just connected increased each gain stage cathode resistors to approx 12K (just added a 10K in series with what was there.
Tested it and now I can turn the gain pots up to about 90% without HF being evident not as ballsy as it previously was but still good.
I also added a 2.2uF bypass cap on the 2 gain stage which increase volume by about 10% maybe noticeable increase though.

Cheers Sluckey I actually did have a 33K grid stopper between the input jack and grid of 1st triode but I hadn't updated the diagram yet.
Yeah mate I used a 22nF blocking cap on the grid of the PI but I am still unsure as to exactly why I need it as I have never used one in this position before.
I have checked quite a few amp diagrams and have not really noticed it before. There mest be a very good reason though otherwise PRR would not have suggested it.

I am also going to change the 2 resistors back to 470K that connect the pentode grids to ground as this should boost the signal a little as I have removed a bunch of gain just from increasing the triode cathode resistors.
I have included an updated schematic as how it is now. Note that the tone circuit is not connected yet.

Thanks again everyone for your comments very helpful.

Cheers H

[sluckey]

Yeah mate I used a 22nF blocking cap on the grid of the PI but I am still unsure as to exactly why I need it as I have never used one in this position before.
I have checked quite a few amp diagrams and have not really noticed it before. There mest be a very good reason though otherwise PRR would not have suggested it.

I told you the very good reason in reply #24. If you need to prove it to yourself just measure the voltage on pin 3 of that tube. You'll see a healthy positive dc voltage. That cap prevents the gain/volume pot from upsetting or outright killing that voltage. Just for kicks, put a jumper across that cap. That dc voltage will change considerably and the amp may not sound very good anymore.

[John]

Increasing the value of the cathode resistor will DECREASE the gain.

D'oh!