More Grunt from Aikido Needed to drive PL509's

[Philwait]

Hello All,

I've been quietly reading in the background and this is my first post here. A few years ago I built a Murray amplifier using EL509's and giving about 35W/channel at very low distortion.

The Murray amplifier is an unusual Australian design from the 1960's using distortion cancellation in the output stage. (see Tubecad on Murray Amplifier from around 2004 and I've attached a pdf).

Although the amp sounds very clean to me, I'm making some enhancements, as you do.

The 2xPL509 output stage requires almost 140Vp-p drive for full output, and the existing 12AX7 SRPP input stage is stretched to the limit.

So I am experimenting with an Aikido circuit, using a horizontal tube configuration allowing split heater supplies and a power supply voltage of 500V. The tubes are 12DW7's /ECC832's which provide a 12AX7 a equivalent stage (1st Aikido stage) and a 12AU7 equivalent stage (2nd Aikado stage).

The Aikido is working well and output voltage swing is huge, much more than I need and distortion measures at about 0.2% at 150V p-p output, (may be less in practice due to my very poor 'rats-nest' construction). Frequency response is -1dB at 33kHz, -3dB@53kHz into a 47k load.

These results are very good indeed and it seems to be a much better solution than the original SRPP circuit.

Only problem is that it needs about 1.3VRMS input to drive the PL509 amp into clipping, which is just a little more input level than I think is desirable for a power amp.

So, I'm thinking of using a low gain 12B4 grounded cathode stage after the Aikido, with a constant current source in it's anode. That would allow me to reduce the gain and the output voltage requirement of the Aikido. Maybe using all 12AU7's instead of the two 12DW7.

This would also remove the requirement for such a high power supply voltage on the Aikido , and therefore the need for the horizontal tube configuration and split heater supplies.

I know the PSSR of a common cathode amplifier is poor, but as it's operating at such a high signal level, and low gain, I suspect this will be easily overcome with some extra supply filtering.

My question is, am I about to stuff-up all the good qualities of the Aikido by adding a GC 12B4 post amp runing at such high output voltage levels? At say 20ma and 250V a-k it would be running pretty hard.

Or is it better to add another 2-stage Aikido amp, divide the gain between the two, and wear the heater supply difficulties?

Or any other better solution???

Or maybe I'm right off-track and I should stick to the knitting..

bye
Phil
Sydney, Australia

[SY]

Only problem is that it needs about 1.3VRMS input to drive the PL509 amp into clipping, which is just a little more input level than I think is desirable for a power amp.

That's intriguing, since most power amp sensitivities run from 1.5-2.5V. Could you explain a bit what you mean here?

How much more gain are you looking for?

I know the PSSR of a common cathode amplifier is poor

Doesn't have to be if the source impedance at the cathode is low (e.g., LED bias) and the plate load is a high quality current source (easy). Even better with a tube having a low plate resistance.

For example, if you have a 4R cathode source impedance and a tube with a mu of 100, rp of 60k, the effective rp is still about 60k. With a CCS source impedance of 20G, the ripple is divided down by about .060/20000 = 1/333,333, equivalent to 110dB.

[Philwait]

[QUOTE]Originally posted by SY
[B]

That's intriguing, since most power amp sensitivities run from 1.5-2.5V. Could you explain a bit what you mean here?

How much more gain are you looking for?

I'm thinking that 0.5 - 1V RMS is normal sensitivity for a power amp. I know that my DVD/CD player seems to put out that level.


Thanks for the other information. I have not considered an LED string in the cathode.

Phil

[SY]

That's much more sensitive than normal. But if that's what you want, I'd do it at the input; my first choice would be a 1:2 stepup input transformer.

[Philwait]

Thanks,

I'll try it at the higher sensitivity first and see how it is first.

I agree with you that an input transformer would be a good way to go, can you tell me a good supplier?

Phil

[SY]

Cinemag and Jensen are the American gold standards. I've used input transformers from both with excellent results. In England, Sowter rules.

I don't know if Magnequest currently has input transformers, but the ones I've tested from them were superb; the pricing can be a bit breathtaking. No experience with Lundahl, but people who have used them seem quite happy.