Official M2 schematic

[TungstenAudio]

The M2, with the standard input of 'K170 + 'J74 is good for a supply range of 21 to 24 Volts. The higher voltage should be used with caution when the JFets have an Idss greater than 8 mA. Power transformers with 18V or 19V secondaries are a good match.
The final supply voltage depends of the type of rectifier used, in addition to the transformer secondary voltage. The somewhat typical GCPC3502 blocks will drop about 3.5V to 3.8V under load. Other options are available that will drop less.

I initially used FEP30 diodes in the bridge rectifier PCBs that come with the store PSU boards. With an Antek AS-4218 and a set of 22,000 uF caps, I got +/– 22.2V on the amp PCBs. That is an improvement over the 3502 rectifiers, which would typically yield about 21.2V. More recently I switched to a pair of LT4320 based 'ideal' rectifiers, which increased the supply voltage to +/– 23.4V and also improved the sound of the amp.

Overall, I find that the quality of the PSU has a significant impact on the sound of an amplifier. Yes, even Class A amps depend on high quality PSUs.

[Extreme_Boky]

Originally Posted by aljordan

I used a 2 x 18 volt 300 VA transformer. I thought the first watt power supply specified 18 volt. Does the M2 specify 20 volt?

If you go by specifications, then you'll get 21.2V, as you already mentioned. (I think Nelson used 300VA windings, on a 400VA core)

It appears that you want to go higher than that (21.2V, as I understand....) Power supply modifications will not get you there, you need to replace the transformer.

My previous post explains the most common factors in choosing the right transformer and the contributors that will affect the end result.

The main question you should be asking yourself is: "What am I trying to achieve?" Is it a hotter heatsink, more power, "better" sound...?

Some people want to measure 24V DC on rails, under the load, simply because the schematic reads: "24V"

You could increase the quiescent current, if the heatsinks are cool... that will give you a slight reduction in overall distortions, and hopefully a better sound than you'll be able to perceive as an improvement.

ACA builders reported an overall improvement in sound quality when they pushed ACA's from 24V to 30V... again, the idea sounds nice but this needs to be evaluated with (very) small ACA heatsinks in mind...

The best would be to get a 5A/30V dual rails laboratory power supply, signal generator, oscilloscope and the desired resistor value (similar to your speaker impedance...) at 50W or so dissipation. And then you can observe the onset of clipping vs. power supply rails relationship. Once happy with the result, just get the transformer that will give you a desired power supply rails' voltages