Using 12AT7 instead of 12AX7

[Chevy]

I see there are still a fair number of 12AT7 tubes out there at a fair price, even NOS / JAN types. . I haven't heard anyone recommend subbing them for 12AX7's... wondering why ? I would think the biasing and gain and all that would be "close enough" for a direct plug in replacement for the AX.
Has anyone tried a bunch of different types and can recommend for or against this substitution ? Any particular 12AT7's sound great for use in standard preamp stages ? The JAN types are very tempting to get a hold of, while still available at a decent price, but not sure if they are sweet sounding tubes or not. Very interested in used old stock or nos tubes, as they seem to be much better made than new production tubes.
I would've thought that some boutique amp maker(s) would've already used them in preamp stages if they were good for that application. They're common enough in driver stages, but not preamp stages.

[Tooboob]

"Biasing and gain and all that" don't usually get you into trouble with reliability, though Carvin's single-triode reverb driver tends to melt, so you can try them in most preamp gain stages.

They don't bias up the same though. Sane people bias amplifier tubes so that they idle at the middle of their operating range. Lead preamp designers usually make adjustments to this in their gain stages to yield clipping that's more assymetric. When you plug in higher-current members of the 12A_7 family, the operating point shifts, changing the distortion. This isn't inherently bad, and it's an easy way for anybody to play preamp designer. You also get a lower plate resistance, which lets a few more notes of high end get through the blocking cap.

Actual gain in-circuit is higher than expected for the lower mu tubes because of their lower plate resistance. See here for more info and a handy table:

The explanation you must read

The other members of the 12AX7 family aren't low-cost rejects. Their lower plate resistance and higher power rating yield more potential power output.

[tedmich]

subbing in a 12AT7 is a very common method for knocking down the gain of a preamp, ie "making my 5150 sound more like a Bassman" but you can also buy lower gain sets of std 12AX7s for this purpose. It is sometimes possible to find excellent old NOS 12A*7 tubes for cheaper than the std 12AX7 but this is not news to most and the prices tend to compensate quickly...good luck!

[teemuk]

Replacing a higher gain tube like 12AX7 with lower gain tube will not only decrease overall gain but will have other side effects too. Since gain decreases the overall difference between gains of cathode-bypassed and cathode un-bypassed circuits decreases as well. If the stage is designed to provide a hi-pass filter by bypassing the cathode resistor with a smallish cap then such treble boosting effects are almost lost when a tube like 12AT7 or 12AU7 is used as a replacement for 12AX7.

So, tube swaps not only decrease the overall gain, they can have a huge effect on amp's frequency response and hence tone.

[Audiotexan]

Sane people bias amplifier tubes so that they idle at the middle of their operating range. Lead preamp designers usually make adjustments to this in their gain stages to yield clipping that's more assymetric. When you plug in higher-current members of the 12A_7 family, the operating point shifts, changing the distortion. This isn't inherently bad, and it's an easy way for anybody to play preamp designer. You also get a lower plate resistance, which lets a few more notes of high end get through the blocking cap.

Actual gain in-circuit is higher than expected for the lower mu tubes because of their lower plate resistance. See here for more info and a handy table:

The explanation you must read

Awesome post! Those charts will soon be above the bench as a reminder...
Things you learned, but details are far too easily forgetten when seldom called upon.

Man I love this place.

[Richard]

Just from my own limited experience, the 12AT7 sounded a bit "dull" in V1 of my Super Reverb but a 5751 lowered the gain appreciably and sounds good.

[trem]

Whether it be Fender, Marshall, or Vox.....My ears always seem to prefer an AT7 in the PI of just about any amp I hear.
I would have to do a lot more blind testing, but I think if you are wanting to tame V1...I pretty much prefer a 5751 over the AT7 most every time.
best

[PeanutNore]

One caution - for reasons I don't quite understand, the 12AT7 doesn't work very well in a cathode follower circuit designed for a 12AX7 or 12AU7. I believe this is related to the transconductance? In plate driven circuits it seems to work fine, though.

[nickb]

"Biasing and gain and all that" don't usually get you into trouble with reliability, though Carvin's single-triode reverb driver tends to melt, so you can try them in most preamp gain stages.

They don't bias up the same though. Sane people bias amplifier tubes so that they idle at the middle of their operating range. Lead preamp designers usually make adjustments to this in their gain stages to yield clipping that's more assymetric. When you plug in higher-current members of the 12A_7 family, the operating point shifts, changing the distortion. This isn't inherently bad, and it's an easy way for anybody to play preamp designer. You also get a lower plate resistance, which lets a few more notes of high end get through the blocking cap.

Actual gain in-circuit is higher than expected for the lower mu tubes because of their lower plate resistance. See here for more info and a handy table:

The explanation you must read

The other members of the 12AX7 family aren't low-cost rejects. Their lower plate resistance and higher power rating yield more potential power output.

Seems like this has run it's course so I wanted to comment on the 'explanation'.

I have to say I'm struggling with this 'explanation'. It's not that it's wrong but I think it's terribly misleading. How can you say rp is more important when, as he correctly states mu = gm . rp ? Seems like gm has equal weight.

Also consider also that the effect of rp is diminished by the plate resistor & load then gm seems to have even more importance.

Now look at the absurdly simplified circuit model. It has an ideal input, a plate resistor and load that is a resistor in parallel with a capacitor. Any useful stage is going to have gain and therefore Miller capacitance and a driving stage. Therefore, the lower mu' stage will have proportionately wider high end bandwidth due to the lower Miller capacitance. Not to mention all the frequency shaping and non-linear things that go on. Sure, rp is a factor but it's just one. To single it out as the holy grail of tube differences is going to cause confusion.

[Tooboob]

I can find charts at various places on the web that list the mu values for all the members of the 12A_7 family, and I read about people changing part numbers to decrease gain all the time. You can certainly hear a difference in a lead preamp, and the change is in the distortion, which lots of guitarists call "gain".

There really is an assumption that the ratio of the mu values of different family members is the same as the ratio of their result output swings for a given input, in all circuits.

Old Tele man is trying to address this misunderstanding. I think he does a good job, as long as the reader understands the usual AC current and voltage models for triodes, which probably eliminates much of the intended readership.

nickb points out that not all real world factors are considered. He's right. Miller capacitance at the input will change. I'm more concerned with bias voltage. I expect that most of the family will draw higher bias current into a given cathode resistor compared to a 12AX7, and this can have a big effect on the operating point at the plate, and the resulting bypassed supply voltage.

So yes, the tubes sound different, but it's not just a matter of gm. Since substitutions are generally safe in tube amps, you can try them, and sometimes you'll prefer the results. You can also have fun randomly changing resistors and capacitors, but most folks can substitute preamp tubes. Far fewer can remove a chassis, find the right parts, and solder them in, and perform enough calculations to make sure that flames won't result.

There's a reason why the Vgc = -1V curve for a 12AX7 is only drawn up to 3.2mA, while the 12AT7 line is drawn up to 23mA. The 12AT7's lower Rp really is a big deal, making it much better for higher current circuits.

Strangely, a full analysis in a given application requires a full analysis. Changing preamp tubes is sort of a random way to tune an amp to begin with (the criteria is the subjective "sounds better?"), so we usually don't need to do it. You can stare at speaker response and impedance curves, and get some idea of how a speaker might sound, but you plug it in to make your choice.