gain stage fun

[phsyconoodler]

I've been experimenting with adding a gain stage before an existing amp's V1 with some good and not so good results.
  I added a 12AX7 before the V1 on my personal 'little bastard' amp.It's 2-6L6's,treble,mid,bass,master with basically a Fendery topology.
I use a gain control on the added tube.I tried cold biasing both tubes and it kind of works well on some settings and not on others.
  It seems to have too much gain too early so I throttled it back and then it's not enough.
 Are there any existing circuits or guys that have done this? I like it but the tweaking to get it right is excessive and the small changes seem to make too big a change.There has to be a good compromise when doing this.
  So you take a Fender Super preamp and add the gain stage in FRONT of the existing V1.It is marvelous at some settings and rather ratty at others.Get my drift?
  I tried all kinds of tubes,12AU7,5751,12AY7,12AX7 and many brands.I always come back to the 12AX7.

[kagliostro]

Have you tried to think of trying a gradual boost

Kagliostro

[phsyconoodler]

Interesting but I can't see how it couples to the previous and post stages.

[kagliostro]

I was thinking you can use something like that in your V1 and V0 tube (V0 or Vzero = the stage you add before V1)

Maybe that way you can progressively tune the stages each other

Kagliostro

[JayB]

It's kind of hard to get that smooth overdrive with out shaving some bass and the tinny highs. I suggest using a lower coupling cap. Work your way up from .001. Throw a plate bypass on the second triode.

You must be looking for that boost pedal in an amp but using a tube thing? Most of the overdrives like the 808 are mid rangy centered around 720hz.

[phsyconoodler]

Yeah it's sort of that feel I'm looking for.It kind of works like a peavey with a pre and post gain control.
   There are a lot of tones lurking in there to be sure,it's just hard to figure which stage to use the bypass caps on and which to not.
The master is kind of crappy so I'm going to implement a Lar-Mar master and see where that takes me.
  What do you recommend for a plate bypass value?
 

[topbrent]

Did you try the Mesa Lonestar arrangement?

Gainstage into AB763

[JayB]

Yeah it's sort of that feel I'm looking for.It kind of works like a peavey with a pre and post gain control.
   There are a lot of tones lurking in there to be sure,it's just hard to figure which stage to use the bypass caps on and which to not.
The master is kind of crappy so I'm going to implement a Lar-Mar master and see where that takes me.
  What do you recommend for a plate bypass value?
 

A 250pf to 1000pf cap. Just got to trust your ear what sounds good to you.

[Geezer]

Also, don't forget about trying some high value grid stoppers  to limit the signal current & cut down on blocking distortion (like the 220k in the signal path on the Lonestar, & like the Dumbles have).

G

[Tone Junkie]

Also, don't forget about trying some high value grid stoppers  to limit the signal current & cut down on blocking distortion (like the 220k in the signal path on the Lonestar, & like the Dumbles have).

G

And you will usually see them in combinations like your cap then 220k to ground then a 220k grid stopper which is sometimes bypassed with say a 220pf or 270pf cap or like the marshall the cap then a 470k to ground then 470k grid stopper often bypassed with the 470pf cap. I like to put my extra gain stages after the first gain stage . But thats just my personal preferance.
Thanks Bill

[phsyconoodler]

Thanks guys! The mesa approach is pretty cool.I'm going to try it and report back on my results.

Thanks again for pointing me in the right direction!

[Fresh_Start]

Thanks for pointing out that Mesa Lonestar circuit.  I can see how you might start with a Princeton Reverb, pull out one triode of the reverb driver for the "extra" gain stage, get frustrated with the cathodyne phase inverter and switch to a long-tailed pair and.... voila!  Mesa Boogie.

Chip

[topbrent]

Thanks for pointing out that Mesa Lonestar circuit.  I can see how you might start with a Princeton Reverb, pull out one triode of the reverb driver for the "extra" gain stage, get frustrated with the cathodyne phase inverter and switch to a long-tailed pair and.... voila!  Mesa Boogie.

Chip

Put it on a relay and it is a switchable gainstage, also.

[phsyconoodler]

The biggest thing I'm having trouble with is taking a whole tube,making it into a pre-gain stage and inputting the guitar into that stage and then coupling it with the existing stage.
  Assuming it works like a pedal would,then I would leave the existing amp circuit untouched and simply add a 1meg grid stopper to the input grid of the existing tube then add the extra stage before it.
  I did that in my test amp but I used my decade box to get the best tone from raising and lowering the grid stopper.Unfortunately I didn't think about trying it at several settings on the gain pot and it's kind of ratty and extremely noisy from the excessive gain.
  So I'm kind of at a loss how to tune this thing.Marshall's 2204 is rather easy to implement but it uses a cathode follower tone stack and I have a plate driven one and I don't really want to change it.
  Maybe the 470k/470k bypassed by the 500pf cap would work in this setup as well?
If it doesn't work I'm just going to do it 2204 style.
   

[phsyconoodler]

Ok,last try at this: gain stage ala weber's drawing,coupled with .02uf cap to pin 2 of existing stage,1meg grid stopper added. 1 meg gain pot.
  That should work.

[JayB]

Honestly, that's still going to sound ratty as hell.

A triode can put a nice swing compared to what's coming out of a pedal. Try the marshal 470k/470k 470pf arrangement (which happens to be a high shelf centered around 720hz) with a one meg pot. That should do it nicely.

[phsyconoodler]

So use the Marshall shelf arrangement on the grid of the existing V1,right?And leave the diagram just like it is?

[phsyconoodler]

Ok,tried the Marshall style divider and it works.However,I don't like the tone as it's too Marshally.So I switched it back to the way it was,added a type 1 PPIMV and throttled back the gain on v1 to 2.6k cathode resistors and the amp sounds fantastic.I guess I'm not a high-gain kind of guy after all.The purity of a nice clean tone and natural power tube distortion gets me going any day.
  The gain was great but for me I don't use that much pre-amp gain so it didn't float my boat.
The topology of the amp is 2-6L6's,430v,fixed bias,no negative feedback,a Fender AB763 tonestack with a 25k mid control,and a type 1 master volume.
  Extremely responsive amp and the tone is to die for.
 I did the experiment with the gain just to see if it could be done that way and it certainly can.But without a bypass input to bypass one gain stage like a 2204 it had too much preamp gain for my tastes.
  I had a Mesa Black shadow speaker laying around and I thought I'd try it out.It sounds excellent!I A/B'd it with a Mesa Lonestar with the same speaker and my amp sounds so much better it's unreal.The Mesa had the gain,mine has the sweetest tone I've heard in a long time.
  With a pedal like a Bluesbreaker II and a tele it's fabulous.With my Epi Riviera it's heaven.I'm leaving this one alone and moving on.

[topbrent]

Pictures?

[phsyconoodler]

You've seen it before.Here you go.

[JayB]

Nice looking amp. I can't blame you. I love high gain and nice clean tones just on the edge of break up but sometimes you just can't have both.

[phsyconoodler]

Ok,I don't give up easy when I'm on a roll!
  What about adding one triode before V1 and simply adding that gain into the circuit rather than two triodes?
 I'm liking the results of two triodes but it's too much gain for me and one triode would be better I think.
    The reason I'm pursuing this approach is because it's easier than inserting a gain stage into an existing circuit and even with no room for another socket you could add a project box to the side of the amp cabinet like they do with the top boost mod on early AC30's.
  I tried lower gain tubes but the tone wasn't up to par and I kept coming back to a 12AX7.
I guess it's just do a breadboard tweak and find the right combination. Surely someone has done this?

[tubeswell]

How about swapping V1 for a pentode? With an EF86you could use the morph control in Merlin's 1st book to get a 'gain' control. (The little SE amp I did a few months ago has a nice amount of gain with the EF86 morph control at full pentode mode. http://www.nzguitars.com/forum/download/file.php?id=8347 )

[phsyconoodler]

Well I've mucked with this idea of mine without having any theory about it in my head and have been relying on my ears alone.This is time consuming.What I've found out so far is that too large a coupling cap between stages transfers too much bass to the next stage and too large a grid stopper resistor adds too much gain.
  I still think the dual triode can be used but it needs to be cold biased and only the input grid needs a bypass cap.The existing V1 should have a voltage divider like a marshall 470k/470k with a 500pf cap and for precise control fiddle with the 470k to ground to get the sweet spot.I use a decade box to come up with the sweetest spot and then sub in a resistor.The 470k is not always the best to use.
  Now of course if I totally understood impedances between gain stages then this wouldn't be such a time consuming task!
Learning is always a good thing,even when your brain is already full from a lifetime of odd knowledge being stuffed into it!
  Time for a memory dump so I can cram more in there I guess!

[echuta13]

Have you thought about using a split-load setup on the plates to control gain? 

Maybe something like this?

[JayB]

Have you thought about using a split-load setup on the plates to control gain? 

Maybe something like this?

That was my next idea. Maybe from two 47ks.

[tubeswell]

Now of course if I totally understood impedances between gain stages then this wouldn't be such a time consuming task!

Well lets see if we can help.

Output Impedance
- For a triode with a fully bypassed cathode resistor:

Zout (Rk bypassed)  = Plate resistor||plate resistance. This expands to Ra x ra / Ra + ra.


- For a triode with an unbypassed cathode resistor:

Zout (Rk unbypassed) = Plate resisitor|| (plate resistance + Cathode resistor x (the tube amplification factor-1). This expands to Ra.[ra + Rk.(u + 1)] / Ra + ra + Rk.(u +1)

Input Impedance
Input impedance of a triode = the grid load resistor (you can raise the input impedance by bootstrapping the grid load resistor to the stage's cathode)

Impedance bridging
is important between gain stages - Rule of thumb is to keep the input impedance of the following stage at least 5 x higher than the previous stage's output impedance, in order to preserve a good bandwidth.

Freq rolloff (or half boost point)
 - for cathode bypassing [where the sum of the Plate Resistor and the tube's plate resistance, is greater than the product of the Cathode Resistor multiplied by (the tube's amplification factor +1):

f(half boost) ~= 1/(2Pii x Cathode resistor x Cathode bypass cap)


- for a grid stopper is:

R = 1 / (2Pii x desired freq cutoff x Input Capacitance of the stage) where total input capacitance is the sum of the Grid to cathode cacitance and the grid to plate capacitance)


 - for a coupling cap is:

C = 1 / (2Pii x desired freq cutoff  x the sum of the following stage's grid load resistor and the previous stage's output impedance)


There you go

[Tone Junkie]

Tubeswell the math looks really good, to bad it went right over my head can you break that down in laymans terms. Im sorry Im still wet behind the ears .
Thanks Bill
I know if I could understand it you would have just taught me a bunch

[phsyconoodler]

Yes,I'll have to really buckle down to understand that math!It hurts my head to think that hard.

[tubeswell]

Ra = 'plate resistor value' (in Ohms)

ra = the tube's plate resistance (in Ohms). This can be estimated when you know the tube's gm (transconductance) and u (amplification factor) according to the equation u = ra x gm (or various resolvings thereof). (Transconductance is the tube's ability to translate 'grid voltage change' into 'plate current change', and is measured in mA/V (or milli-Siemens, or 'umhos')

u = the tube's amplification factor (as I said just now) e.g.: a 12AX7 is '100', a 12AY7 is '45', a 12AT7 is '60', a 5751 is '70', a 12AU7 is '19' etc etc...

The tube's input capacitance you get from the tube's datasheet (where you also normally get the gm and the u)

Rk = the cathode resistor value (in Ohms)

C = the cap value (in farads - so 0.000001 = 1uF, 0.000000001 = 1nF, 0.000000000001 = 1pF )

f(half boost) is the desired roll-off mid point of the freq boost (in Hz). (There is a chart somewhere in Merlin's site that explains what guitar and bass strings/frets correspond to which freqs)

|| means 'in parallel with'. For example "Ra||ra" literally (in 'math' terms) means "Ra x ra / Ra + ra" (like you do when you expand any parallel equation)

'Pii' is what it's always been (i.e.: 22/7 or 3.142857... ), and '2Pii' is twice that much.

Is that any further help?

[phsyconoodler]

Thanks tubeswell! That is very close to where most people can understand.
  This will make designing stages a whole lot easier!

[terminalgs]

Here is a preamp pedal that I've made.  Its been on loan to a studio and gets lots of use, even if the boost knob is set to near unity.   You could add it sans the CF in front of the first stage of your amp.


the first pedal  I gave him  was a 3 stage overdrive pedal.  I had a trim pot on it,  but I trimmed it to amplify a 1V P2P to about 20V P2P and said "never turn the drive and volume knobs both to ten,  its about +13dB"   he understood, but a guitar player recording with it didn't get the message.  I heard the recording, and it sounded great until the EL34 quartet gave out.

I built the same circuit into a "overdrive or boost pedal" with a 12au7 for the boost and I didn't need a trim knob.

For this one I trimmed it to be about +6dB at 1V P2P,  and I never heard about melted tubes again.  

[phsyconoodler]

thanks a bunch! I think I'd like to build that pedal!
  So if I were to use this circuit in front of my existing preamp tube,I would use a 12AU7 with no trim pot?
What about the Weber gain stage?Why is it so finicky?
   So in my mind I want to view this like a pedal built into the amp right?I would leave a 1meg grid resistor hanging off the input grid of the existing V1 or?????
  I don't know why I'm having so much trouble envisioning this??
IHere's the way I THINK I should do it: build this circuit into a box and plug the input via a shielded cable into the box.Leave the existing V1 completely alone and couple this circuit's output into the input grid.Done.Right or wrong?

[VMS]

Question to terminalgs:

Why are the bridge rectifier and two caps not connected to ground?

[terminalgs]

yeah,  you can that circuit (or one like it) into a separate enclosure,   put a signal by-pass switch on it so you can hear the difference.    If you have an o-scope, you can trim the output to be +6dB or +12dB or whatever you want.  without an o-scope, you'll have to do it by ear.   a 12au7 in that circuit  won't really need a trim pot,  build it with a trim pot,  and you can trim it for a 12au7, 12ay7, or a 12ax7.

I like to do it pedals or separate enclosures so it can used on multiple amps, and If it doesn't work out, I didn't invade a nice amp.

here are some enclosures I've made: http://www.dougcircuits.com/

[terminalgs]

Question to terminalgs:

Why are the bridge rectifier and two caps not connected to ground?

ugh,  thanks for catching that.  I updated the link.  It was incorrect because I didn't proof read it....  thanks again.

[phsyconoodler]

With a 12AU7 does it have enough gain?

[terminalgs]

with a 12au7,  its good for about  +10dB  @ 1V P2P.  If you want more than that,  12ax7 and a trim pot.

I should also add (in my experience) that not all amps take boost as well as others (good reason for the separate enclosure!).   It seems like Voxes don't take boost as well as Ampegs, Fenders, and Marshalls.   I built the original Vox AC4 EF86/EL84 circuit into a little SS Pathfinder cabinet.   instead of using a 12ax7 for a trem circuit, I put it in as a "gain stage->level-knob->CF" as a 2nd input that if used, inserts itself  in front of the EF86.  I never got good results from it.  It always sounds better guitar straight to EF86. 

I don't think it was a pentode thing either.  I built a 12ax7/6SJ7 pedal.  at first it was 6SJ7 overdriving a 12AX7 stage,  but it sounded better as a 12ax7 stage overdriving the 6SJ7.  I've got a 90s Korg era AC15 that doesn't seem to sound better with my any of my boost pedals or even a fulltone OCD pedal (gold standard!).  I loaned the OCD to a guy with a HW AC30 and he loves it.  Voxes seem to be picker about boost pedals...(that's all I can figure).

also,  if you build it as a separate enclosure,  be sure to use a shorting output jack and have the signal shorted to ground when not plugged in.   Otherwise the coupling cap right before the jack can build up a charge and "surprise" you (or your amp!).    I use three prong plugs and  put a ground lift switch on the center ground plug.  you can get ground loop noise issues between the shielding on your cable to the amp and the wall circuit common ground,  a ground lift cancels it.