Tweakers' Asylum: Superregulator downgrade? (long) by carlmart

It was interesting to see a thread here over Walt Jung's superregulators, as I think some issues should be questioned or mentioned at least. DIY qestions.

First of all I wonder how many Asylum readers have read the original articles: Sulzer's supply, Walt Jung's supply and the latter's latest improvements.

W.Jung's supply is basically a clever upgrade of the original Sulzer's supply, published in 2/80 in "The Audio Amateur". Sulzer designed a high quality power supply that used huge amounts of feedback to lower the impedance of the power output. Lower impedance was directly related to better audio quality from the powered circuit.

For about 15 years it was considered the best you could rely on if you wanted the highest quality in audio. It was probably implemented in some commercial designs, but I am not sure on which.

Enter new chips, comes a new version. BUT with a few problems as well. Reports on oscillations, expensive/hard to find new chips, expensive double-sided pcb, etc. Things you never heard from those that built Sulzer's supply.

In my opinion Sulzer's still is a remarkable supply. If you look at the plots W. Jung showed in TAA 2/95 the measurements you could get with the Sulzer supply, using a cheap/easy to find NE5534, were just a little bit worse than the new supply using the problematic AD797. The Sulzer/5534 supply was better than the new supply using the AD848, which BTW was more stable than the 797.

One question I raised to W.Jung is why not use an NE5534 instead of the AD797 on that same circuit, but he made a valid point on why questioning the work or findings of someone that spent several years working on something. This he certainly did and I certainly admire his work.

But the points I raised above are still valid too. Should we upgrade to this new version? Spending lots of money in new chips and double-sided pcbs isn't for DIY. What are the trade-offs?

Sulzer's supply used single sided pcbs, (now cheap) NE5534s and seemed to be quite stable. So what happens if we try to implement some of the upgrades of the new supply. Like using a current source on the output transistor for lower dropout voltage; or re-routing the pcb to sense the connection right at the audio IC to include it on the feedback.

The word now goes to any designer that might be reading these pages. Are these valid issues?

Carlos