The Tube Swapping Thread

Please, while I have got into the same way of thinking as yourselves on the value of this, just a cautionary note here. The idea of swapping 12A*7/ECC8* types around has been around for a long while. So much so that people have got into the way of thinking that they are completely equivalent. Sometimes it's ok to substitute but sometimes it most assuredly is not! There is more difference in these valves than just the gain. Circuits which are designed properly for 12AX7 do not automatically suit 12AT7 and will never suit them perfectly. Permitted voltage levels are different, anode load resistances are different, current levels are different, and frequency responses will change along with gain in the same circuit.

Look up online for the 12AT7, 12AU7, 12AX7, 12AY7 swaps and actually read some of what the technically knowledgeable posters say rather than looking for "someone" with an "expert" posting status who tells you its ok so you go ahead. And for Pete's sake take it seriously. It's dangerous to choose to believe the guy who says "well I just did it and it sounds great for me". Firstly his ear may be severely flawed and completely different to yours. He thinks it sounds great while the rest of the world and all of its cats and dogs are holding their ears and howling. Secondly, one month down the line he may be looking at a smoking pile of components, as could you!

Here is just one example. Try to sort out the advice from those who "do it" and those who "understand it".
http://www.diyaudio.com/forums/tubes-valves/211582-12ax7-12at7-recommendations.html

For the record, I'm running a NOS Mullard 12AT7WA in my V3 PI slot to drop the gain a tad, and I took pains to talk to someone who knows a lot more about specific valves than I do. This is the only one he would recommend. But I'm watching the situation closely for any signs of stress on the amp and I'll go back to my JJ/Electro Harmonix if I spot any.

I really don't mean to put this whole issue into a red alert state or to suggest anyone's opinion is stupid but as I said, there's more to it than just swapping equivalents.  I'm not a valve types expert like some who have spent many years working with them, but I do know my electronics so the principles behind this at least are clear to me.  There is nothing hard in understanding the technical side if you really want to, but there are so many threads out there just the same as your own where this sort of question is posed and the advice is always the same, "go ahead and do it, this one will sound great, this one will sound rough" and so on.  The voices who say "if you do that you may lose bottom end because of a shift in the frequencies of the various time constants" or "that valve was never designed to do that sort of job, it doesn't have the plate voltage to take it reliably" or "that valve will run with too much current if put into the biasing circuit of the original" are never listened to.  Everyone wants to listen to the expertise of the non-techy guy who has just plain gone ahead and done it.  No one wants to listen to the tech savvy guy who knows what can or even will happen.

Bear in mind that the TSC only takes care of the bias of the power valves, the 3 preamp valves are in positions where there is no auto biasing function so their bias will depend on their own characteristics and these could change drastically for different valve types, even ones usually considered close like 12AX7/12AT7.

The answer is simple:

1)  Start off by sticking to the advice of people who will tell you WHICH 12AX7 you should use, not which OTHER valve type you should replace it with.  This is the safest option and therefore a best starting point.  If you can achieve the sound you want this way then it's the best possible solution.  I know from my own experience that you can make all of the changes you need to remove that harsh top end "fizziness" you describe, (which others have reported as well as us), by swapping out the correct valve types for other makes and getting your speaker choice right.  (Read the threads on speaker choice in both the GM36 and TM36 forums, that is the most important thing to cover).

2)  Make sure you only take advice based on YOUR OWN amp.  What is the point of someone telling you what difference a valve swap made in a Marshall if your question was about a H&K?  I've even read posts where people swap out "the valve" in a solid state modelling amp and advise on the basis of that change about valves in a full valve amp.  There is no guarantee you will get any of the alteration they report.  It's like saying that a Ford car responded well to changing the suspension geometry setup so your Peugeot would too.

3)  Make sure you do not just change the valves out and then automatically tell yourself that it sounds better!  I've been in the audio electronics game for many years now in both guitar and hifi design and this is the most common error we all make.  It's in our nature to be unable to respond to a change honestly and critically even to ourselves once we have paid out money for it and spent time making it.  We WANT to find it makes a big improvement therefore we HEAR one.  And we advise others on the strength of that and send them down the same path as ourselves.  If this is the case it doesn't last!  How many people have you come across who make the most astounding improvement to their sound with a change of some sort, then they are back in a month or two asking about more changes of the same sort?  Give yourself a significant time using the amp before you make a final decision as to whether the change really made an improvement.

The real answer to that one is blind testing.  That is where you have the change you are testing, (swapping a valve for example), set up so it can be swapped easily and invisibly so it is the only change that could occur, (not easy for the valves).  An operator sets it up for one or the other at random and so the test subject cannot see which is in use at any time.  Then he plays it for a while and says which he thinks it is.  He is given no info about whether he is right or wrong.  The results of a largish number of test guesses are then analysed to see whether there was any degree of correctness in the guessing.  If there is no difference things should work out close to 50:50 of course and the more tests you take the better these results become.  47:53 would not really be significant, but 38:62 may well be.

The perfect issue for this approach is guitar tone capacitors.  You take a magic mojo capacitor and measure its value very accurately.  You then find a plain old cheap everyday type which measures up at exactly the same value.  You solder them both into the tone circuit of a guitar with a small switch to swap between them and make sure the guitarist subject cannot see it in any way.  This way the only thing in the whole experiment which varies at any time is the capacitor type.  You toss a coin where the subject can't see it and if it comes down heads the switch goes one way, tails the other.  You always make a play of changing the switch even if you don't!  You let the guitarist play until he has decided which one he thinks is in circuit.  When he gives his result you give him no information as to whether he is correct or not.  After a large number of results you look to see how many times he was right and apply a couple of formulae to see if it is significantly off the beam one way or the other.  You know what?  When this is done as I have described no one can ever....  Well you know what I'm going to say don't you.  And it would even be possible to prove this way that the cheaper option is preferred even by those who initially thought the special option was better!

That said, I have swapped for JJs in the EL34 (edit: errm, EL84 of course) positions and that maybe smoothed out some of the roughness but it wasn't really blinding.  I have tried Electro Harmonix and JJ 12AX7s in the V1 position.  Of the two I prefer the JJ and will stick with that.  I have an Electro Harmonix in V2 at the moment and would like to see what a JJ would do there but I don't have another one to hand and it isn't a crucial issue so I can wait.  I have also gone for a NOS Mullard 12AT7WA in the V3 position.  (This is sometimes listed as CV4024 and sells for under £15.00.).  I was advised by a very experienced valve techy guy that this was a feasible swap and the job was suited to the capabilities of the Mullard.  A Mullard WA type was the only one he would recommend, he said that others sounded thin due to resistances changing and frequency response being altered.  I can't say it was night and day but the amp became a little easier to dial in "classic" sounds rather than full on "high gain modern".  They were still available, I just had to wind things up slightly more to get to them which gave me a little more control lower down where I wanted it.

To round off (at last), talking here to others about what they found when they swapped out valves is the start of your answer.  We are all GM36 users and our own findings are therefore all directly relevant to your own situation.  We can certainly report on what changes we heard when we did a change ourselves, but this is highly subjective.  One man's total lack of bass is another's balanced Heaven.  We DON'T know all of the results of swapping out valve types, no one yet has any schematic for the GM36 to study for any implications.  Bear in mind H&K say that they only guarantee the product if: "Assembly, extension, re-adjustment, modifications or repairs are carried out by the manufacturer or by persons authorised to do so".  You are not one of those persons and valve type swapping is not a sanctioned mod!

Keep us up to date with your progress, we all learn from it.  And don't stress over it, it isn't something you HAVE to get right straight away.  Hang fire until you are pretty sure of where you want to go.  The GM36 is a superb amp even with the original valves in there.

Don't know if these have been mentioned but I found this:

5751 is a GE "Five Star" version of the 12AX7. It is electrically very similar to 12AX7, but with slightly less gain, more rugged internal construction, etc. It's not much like 12AT7. It's really more like a variant of 12AX7. One can almost always directly substitute a 5751 in a 12AX7 circuit, but again, the gain (mu) is a bit lower (5751 mu = 70, 12AX7 mu = 100).

12AT7 is a quite different tube, with lower rp, higher gm, less gain (mu = 50), etc. You can put a 12AT7 in place of a 12AX7, but now you're into modification territory, and the circuit will now act very differently from the original.

I've had a quick glance at the datasheets for the 12AX7, 12AT7 and 5751.  Here are a few of the highlights:

The 12AX7 is designed as a "hi-mu audio voltage amplifier and phase inverter".  The 12AT7 is a "hi-mu grounded grid radio frequency amplifier or as a combined oscillator and mixer at frequencies below approximately 300MHz".  The 5751 is "ruggedised hi-mu adaptable to low level input and high voltage gain" such as "voltage amplifiers, phase inverters and multivibrators".  So which two are closest to each other there?

Valve:	12AX7	12AT7	5751
mu:	100 / 100	60 / 60	70 / 70
Plate R:	80k / 62.5k	15k / 10.9	58k / 58k
Plate I:	0.5mA / 1.2mA	3.7mA / 10mA	0.9mA / 1mA

.
The dual values are quoted at 100V and 250V.  You can easily see how much better a match for the 12AX7 the 5751 is than the 12AT7. Look at the difference in Plate Resistances, and that is an important parameter with respect to the bandwidth of the stage.

Here is a superb link for anyone interested in the technical side of this and wanting to understand the bigger picture.  Why do tubes sound different?  It clearly describes why Plate resistance is the important parameter for overall performance and not mu!  And I have also posted on the H&K Facebook site asking them for guidance as to whether tube rolling is a good idea for the GM36 from a reliability point of view.  They have passed the query on to the designers and promised to get back to me.  I'll post when they do.

That's a good decision Luke. When you get used to the JJs you can always borrow a couple of alternatives from someone to try out without lashing out more money that may just sit in a draw. I rate those JJs too.