by bordonbert Fri Feb 24, 2017 6:58 am
Sorry but that's not true at all 6 and the only way I can show this is to get technical. That may bore many people, perhaps yourself too, but I will put it here again for those others who pass and might take notice.
The 12A*7s can all be used interchangeably without damage in most cases and the H&K guys have confirmed this for the GM36 when I talked to them about it. But there is much more to the design of a stage than that. It isn't even guaranteed that when you put one in a 12AX7 slot the gain of the stage will drop just because the valve gain is lower! The gain is not the most important factor in using a valve, it isn't even a well specd value as it is derived from other values which are themselves well specd but which change in use. It doesn't even mean much in the real world, dropping in a valve with half the gain does not cut down your stage gain by half as other circuit features affect that.
The important factor is the fact that the Vg/Ia curves are totally different and the Ia will change sometimes drastically, always significantly. DC levels in the stage will alter because of that. If you look at the graphs and do your load lines you will find that putting the 12AY7 into the first stage of the GM36 with the way that has been set up for a 12AX7 with a 2.7k cathode resistor will increase the DC current from about .85mA to 1.75mA which is double. You haven't changed the load resistor so the DC voltage across that would double too.
Then there is one of the most overlooked parameters in the valve's Anode Resistance. This is the value of a parasitic resistance in the valve which appears in parallel with your Anode load resistor. You can't get round it. That drops the load of course and hence the gain of the stage. That might seem what you want but, along with the output DC blocking capacitor it also affects the upper bandwidth of the stage so that changes.
In the 12AX7 it is specd at 62.5k and up and 62.5k//100k = 38.5k. In the 12AY7, in common with all the other "low gain equivalents" it is lower, it is actually specd at 25k. With the same 100k anode load resistor that is 20k, a halving of your effective anode resistance, another big difference in real terms.
The 12AY7 is specd at 250V Anode voltage max (General Electric datasheet), the 12AX7 is 300V and that is often what it is designed for. In the GM36 the preamp 12AX7s run at 350V for V1A/B, 280V for T2A/B which is more crucial as one of them is a DC Cathode follower with no anode resistor to drop the voltage it sees, and 370V for T3A/B as phase splitter. I wouldn't put a 250V valve under the stress it could get in any of those positions and expect it to be happy.
I'm sure there are people who would go ahead with these substitutions on the grounds that "I put it in, it worked, it sounded just as I wanted". To them I would say, good luck with that I hope it stays that way. I would also add, there is a valve, the 5751. This literally is described in many official places as a low gain 12AX7 and it IS a good substitute with very close characteristics. [EDIT: I should have said that it is a low gain type too at 70. With two of these in circuit you would drop your gain by half. The 12AY7 is specd at 44.] It is the only one that I would ever put into a 12AX7 slot. I've looked at the spec sheet and done the maths on that one and can confirm that it is the best and only close substitute that I reckon you should consider. I use them in both my GM36, TM36 and my Marshall JVM205.
I'm with you in the desire to drop the gain in the stages, for my type of music that gives a much smoother sound, but I wouldn't just pop any old thing in there becuase it has always been done. The belief that these valves are equivalents with different gains is another music industry "common knowledge" myth built up over the years. That doesn't necessarily give good results in terms of reliability and as an engineer that matters to me.