12A_7 substitution - is it the tube or the bias point?

[thelonious]

Now that I've got my feet wet in amp building, I finally worked up the courage to pop open the SE boutique head I used to play through all the time and analyze it. Turns out that it is basically an AX84 Hi-Octane build with some modifications to gain pot values, tone stack, etc. All the voltages and the values for Rp and Rk are the same as for the Hi-Octane.

The original Hi-Octane schem calls for 12AX7s. This amp builder says to use a 12AX7 in V1 and a 12AU7 in V2. Other people online are raving about using a 12AT7 and a 12AU7, and in fact, that's what my amp came with when I bought it refurbished directly from the builder - so even he must have thought the 12AT7 in V1 sounded good enough to leave it there.

So I'm confused. V1a uses Rp=100k and Rk=2.7k. That's a pretty funky load line for a 12AT7, isn't it? Should I experiment with different values for Rp/Rk?  

I guess my question is... what is probably contributing the most character to the sound: the different tube type or the unusual bias point from using the "wrong" tube? Or is the answer, "yes, both, now go experiment"?  

[RicharD]

Need to know the B+ of V1 to plot a load line.  A perfectly balanced tube isn't necessarily what sounds best in a guitar amp.  You aren't gonna hurt anything swapping Ax, AT, AY, AU tubes around.

[thelonious]

The voltage at V1 is 234v. So the load line with a 12AT7 looks like:


But of course you're right, how it sounds is what matters, not how it looks on paper (er... monitor). I guess I'll start playing around with values for Rp and Rk and see what happens.

Still interested to know if, from everyone's experience, the tube type itself makes the big difference in sound when a tube is changed out, or if it's because when that new tube is dropped in without other component changes, it has an extremely different bias point, and that's what makes the big difference. So - one guy online said that putting a 12AT7 in "made the gain so much sweeter". Did the 12AT7 make that difference, or is it the cold bias it gets in that position in that amp? But that's probably a subjective, unanswerable question. Sorry! Ha!

Thanks!

[thermion]

i love this question and would like to see a few resident black belts comment. i also have wondered about this; i use a 12au7 in v1 on my fenders all the time without changing the plate load/ cathode resistor values and really dig the result.
inquiring minds want to know!

[RicharD]

I'm no black belt but I'll chime in a little bit suing TubeCad as an aid.  (See attached)
I am comparing the 12AX7 to the 12AT7 using values provided.  I assumed .01 for the coupling cap which affects the low frequency response.  I assumed 33k for the input Z which affects high frequency response.  I also assumed a Rl=500k which is 5:1 the plate resistor so it is negligible.  The AT pulls a little more current so in real world quiescent conditions, B+ should be a tad bit higher with the AX.  Gain is about 1dB different per stage.  Again not that big of a deal.  PSRR (power supply rejection ratio) is about double for the AT so it should be less susceptible to hum.  Zout is lower for the AT which in cascading stages is a non issue but it could be a factor if driving a tone stack although probably very slight.  Bias and threshold are higher with the AT meaning it can handle more input signal.  If you have a chain of stomp boxes in front of the amp, that extra 1/2 volt of headroom could be quite useful.  In cascading stages, again that extra 1/2 volt sorta comes into play but more so is the output swing.  The AX can make 90ish volts w/o becoming lopsided where as the AT can make 130ish volts w/o becoming lopsided.  Of course most of us drive our preamps hard and into the region where one side is clipped.  In both cases, the positive cycle clips about 50% before the negative clips. For the most part, both tubes are doing roughly the same thang in these conditions.  

Obviously the AT is loafing.  I didn't show it but the TubCad report says the AT has insufficient plate current, sometimes called plate starved which sometimes sounds kick donkey in a geetar amp.  I run a plate starved EF86 circuit.... but I digress.  Look at Imax.  The AT can run 10x the current of the AX.  RP for the AT is 15k where as the AX is 62.5k.  The AT could be happy with a much smaller plate resistor and therefore a smaller cathode resistor to get to a happy bias.  In most cases, an AT can sub for an AX with comparable results. This is not necessarily true the other way around.  If you have an AT circuit set to idle at 20mA and you put an AX in there, it's gonna fry.

Got a 12AY7 handy?  Stick that puppy in there.  Under these conditions Vout is almost perfectly symmetrical.  Threshold is 2.8V.  If you play clean, your brain might melt using this tube.

[thelonious]

Great explanation - thanks!    Gonna experiment more with plate starved conditions. My first build has a 5879 in it that might give some interesting results (which hopefully will not include being hopelessly microphonic...)

I do have a 12AY7 lying around and will definitely give it a shot in V1!

[PRR]

> Rp=100k and Rk=2.7k. That's a pretty funky load line for a 12AT7

Reference for geetar-amps is Fender's 12AX7 100K 1.5K.

Arguably THIS is a "funky" loadline, but we like it.

Knowing only that, we don't need no plots or computers.

The cathode bias should scale with Mu. Lo-Mu tubes will use larger cathode resistors.

100/60 * 1.5 = 2.5K. This will give Fender-like plate proportions with a 12AT7.

Nearest standard 10% value to 2.5K is 2.7K.

The plate side will overload a lot like a Fender, but the grid-plate gain will be lower. (Not even a lot lower.)

> Obviously the AT is loafing.

Yes. It is a 10mA tuner tube. Working it near 1mA means you paid for 10 times more tube than you needed.

> TubCad report says the AT has insufficient plate current
 
BS. You need to drive the load, not feed the tube. We know 1+mA in 100K drives Fender-like circuits well.

Specially on 12AT7. While designed to flow 10mA, it is also designed for AGC and doesn't do anything strange at lower currents.

[RicharD]

Tubecad reports are obsessed with current.  I agree it is bs and prr is correct.  I was simply sharing results.

[frus]

Reference for geetar-amps is Fender's 12AX7 100K 1.5K.

Arguably THIS is a "funky" loadline, but we like it.

why do you say it's funky? I always thought it's a "school example" of center-bias, reasonable, etc etc loadline

[John]

I think it depends on what you're biasing for. We treat tubes differently for guitar than we would for hi-fi. Hi-fi is all about clean and pure. Guitar is about coloring the sound.