12ax7 gain

[lowell]

Been referencing Merlin's book and online forums. It seems the basic approximation of the gain of a 12ax7 is Ra/Rk, assuming Rk is bypassed.

If taking ra (internal plate resistance into account) it seems that changing the plate resistor doesn't really affect the gain that much.

If we have the internal resistance, the plate resistor, and the following stage grid leak is 1meg: (Rk as 1.5k)
62k/100k/1m all in parallel
about 32k

32k/1.5k
A=23

So if we increase the plate resistor:
62k/470k/1m
about 52k

52k/1.5k
A=35

So really not that big of a difference. Is this correct? Thoughts? I've seen online references of A=60 with a Ra 470k. I think that those posts are ignoring real world values, like ignoring the internal plate resistance and following stage grid leak.

[Leo_Gnardo]

Not to undermine Merlin's book which is very good -

About 17 years ago I made up a quick n dirty rig to test voltage gain on 12AX7 and similar dual triode tubes, borrowing the circuit from a typical Fender first stage preamp. What I've found from measuring, not calculating, (I know, I'm such an empiricist) max Av around 75, average about 60, and many in the high 40's to mid 50's sound just fine in spite of having a low gain reading. It's easy enough to make your own, with a +300V supply, an oscillator to send in say 100 mV at audio frequencies, and a reasonably accurate AC voltmeter.

[lowell]

So what did you find when swapping plate resistor values? I left my tube proto board when I moved years ago and need to build another. But appreciate any insight meantime.

[Leo_Gnardo]

So what did you find when swapping plate resistor values? I left my tube proto board when I moved years ago and need to build another. But appreciate any insight meantime.

I haven't changed plate R's but sure, you can squeeze out more gain by increasing their value & vv. OTOH comes a point you're starving the tube for current. I often find resistor biased 12AX7 triodes 68k to 200K. Grid-leak bias 270K - 470K and they tend to sound "compressed" with a dynamic signal, which can be an attraction depending on what you're looking for.

[nsubulysses]

Plate resistor values change the sound a lot IMO. Maybe the numbers don't seem significant but surely you would say switching between 82K/100K/120K/220K/etc is a reasonable sound difference, right?

Get a resistance substitution box and see. Or a few resistors and some alligator clips. Maybe in clean circuits the change is less but I find changing plate resistor values changes the sound pretty considerably. Especially when you're straddling the line of good pick attack, sustain and note clarity, but if you go too far with the plate resistor all these unpleasant distortion sounds start to sneak into the mix, and it gets mushy or harsh

[Malcolm Irving]

Been referencing Merlin's book and online forums. It seems the basic approximation of the gain of a 12ax7 is Ra/Rk, assuming Rk is bypassed.

If taking ra (internal plate resistance into account) it seems that changing the plate resistor doesn't really affect the gain that much.

If we have the internal resistance, the plate resistor, and the following stage grid leak is 1meg: (Rk as 1.5k)
62k/100k/1m all in parallel
about 32k

32k/1.5k
A=23

So if we increase the plate resistor:
62k/470k/1m
about 52k

52k/1.5k
A=35

So really not that big of a difference. Is this correct? Thoughts? I've seen online references of A=60 with a Ra 470k. I think that those posts are ignoring real world values, like ignoring the internal plate resistance and following stage grid leak.

Sorry - I don't get it. Where is that formula coming from? I don't see it in Merlin's chapter. Merlin's formulae for 'bypassed' don't include Rk.

[Dave H]

The data sheet is a good place to look for this sort of info

[lowell]

That formula is not in Merlin's book. His formulae are more intricate. I was using this much simplified formula found online in more than one place.

But that's beside the point. The 62k ra internal resistance holds things fairly rigid in terms of calculating gain.

The point of my post was to solicit OTHERS' way of calculating gain. A "what's your process and what have you experienced" sorta thing.

[lowell]

I am building another amp right now and will report my real world findings using a sub box.

[Old Tele man]

ACTUAL (in-circuit) gain is ultimately dependent upon what the tube "feeds" into, ie: output circuit impedance.

[lowell]

Which datasheet is that from? JJ does not include such info...this must be from a specific manufacturer.

[Dave H]

The point of my post was to solicit OTHERS' way of calculating gain. A "what's your process and what have you experienced" sorta thing.

In your OP you said "It seems the basic approximation of the gain of a 12ax7 is Ra/Rk, assuming Rk is bypassed". Did you mean assuming Rk is not bypassed? If Rk is bypassed it's a short circuit to AC and its value doesn't affect the gain.

[lowell]

Ok right. But the tube is still biased by Rk and I would think bias point effects gain. But maybe it does not. I don't know. Although looking at the grid curves it does change symmetry and headroom of the output.

[Dave H]

Merlin says the formula for the gain of a triode with Rk not bypassed is -

A = uRa/(Ra+ra+Rk(u+1))

It looks like your formula of A = Ra/Rk is an approximation that is more accurate for large Rk (cathodyne say) when Ra+ra << Rk(u+1)

[pdf64]

...The 62k ra internal resistance holds things fairly rigid in terms of calculating gain...

I have a notion (due to laziness it's completely unfounded on either theory or experiment) that the anode resistance may be affected to some degree by the tube's operating point.

For the 12AX7 performance table provided by Dave H, see p2 of http://www.mif.pg.gda.pl/homepages/f...3/1/12AX7A.pdf
When searching electron Tube Data sheets - Search GE tube info is generally my first choice.