When I started making tube gear (about a couple of months ago) I looked around for a good SS B+ regulator and I soon discovered the scarcity of "good" and "simple" regulators out there. After gathering a bunch of different designs, including the ones published in your journal, I tried these designs out using my Electronic Workbench 5 software and found them to be extremely sensitive to load changes and a bit noisy. Then I tried to tweak these designs to make them more "stiff" and/or less noisy.
    At the end the design I am sending you won because of its simplicity and extremely low noise (it is fairly stiff too.) To make sure that the simulation actually was correct I made three supplies. The first one was the un-tweaked version of the attached design. Second was the tweaked design (see the attached gif) and the third was your circuit of the month (GlassWare Feb '99). I tried them with my newly built 5687 based SRPP pre-amp. I also measured them using my Goldstar 20MHz Oscilloscope. The measurements, though not the same, were pretty close to the EWB predictions. Most importantly, this design passed the listening. Everyone preferred the sound of this regulator.
   I think anybody looking for a simple yet effective B+ regulator should give this circuit a try.

Shahzad I. M.
Pakistan

     I have to admit that I am puzzled by the results of your tests and by your regulator. The first thing I note is the use of an 8 Henry choke. This choke must have a large DCR, which must displace some voltage across its leads. Given that the raw B+ voltage is only 350 volts and that the output volt is 330 volts, just how much headroom does your regulator or any other regulator have in this arrangement? 

    A line voltage drop of only 3% will probably force the regulator out of regulation. And given that you are using such a large valued choke, why bother with the solid-state regulation? How does the line stage sound without the regulator? And given the miniscule voltage differential that the active portion of the regulator sees, is it possible that other topologies would perform better given more voltage in which to operate?

< PREVIOUS

NEXT >