I took on the low cost challenge, and then when the request for additional capability and recycleability came up, it was easy enough to accomodate. It was an interesting challenge to cut away the fat, yet leave the basic circuit alive.

The shunt reg's, done with gas tubes are on my list of 'expendables' for the time being. An active load ought to be able deal with the small amount of ripple left over from such a big choke and cap L-C filter easily. Since the load is not presenting exactly a constant load( the load is the regulated current + the current put into the amplifier and cables ), I am questioning the need for considering output impedance of the supply. Which one would have a lower impedance or more importantly, sonic effect is an open question for me. Neither sound bad, and I know the gas tube PS does sound really good.

The question remains, how much trouble is it worth? The VR tubes do look pretty good, no? The VR tubes also allow some more choices in rectifier. I'd rather see an indirectly heated like the GZ34, 6AX4's or 6CA4 if shunt reg's aren't used.

My linestage( Rev. 4 ) is very similar. I use 12B4A for the amplifier and gas tube shunt reg( two 0D3/VR150's ) for each channel. I used battery biased CCS instead of resistors so I could get away with minimum voltage headroom and heat generation. I also used battery biased active loads. I also used Mercury vapour rectifiers( two type 816's ), and implemented a 60 second delay to give the 816's time to warm up. Hg vapour type requires a warm up before application of voltage.

I buiilt Rev. 4 as a decompression project after turning in my MS Thesis last April. took about 2 days to get it singing.

I think that sooner or later I'm going to try an Iron coupled linestage. good optput TX are hard to come by. Fortunately a likely suspect has its drawings on file at Heyboer.
regards,
Douglas