Home Tube DIY Asylum

Do It Yourself (DIY) paradise for tube and SET project builders.

[ Follow Ups ] Thread: [ Display  All  Email ] [ Tube DIY Asylum ]
[ Alert Moderator ]

Re: Too Much HIGH GAIN - Just Build 807 Williamson A515 - Use of NFB?

As others have already pointed out, w/o the global FB the circuit you posted has about 20dB more gain than designed for and about that amt more noise and distortion too.

The original premise behind the Williamson design was low distortion thru use of triodes and global feedback and both were factored into the original circuit. This was somewhat revolutionary in 1947 when the design was published (1947 is also the date of your circuit which is for all intents and purposes the original Williamson design with US tube types and some PS improvements). Leaving out the feedback really results in something other than a true Williamson amp.

However...as you seem to realize, corrective feedback is not without issues. One circuit element of the Williamson I don't think is really appreciated today is the totally excellent opt that was specified. Stable operation under real-world load conditions was totally dependent upon the quality of the opt. The reason I mention all of this is that while I think you should consider closing that feedback loop as the schematic shows, unless you have used a really great opt you may experience instability esp if your load is reactive with a capacitive element to it.

The authors of the article/schematic you posted did indeed experience instability in the form of a transient 60KHz oscillation under certain conditions of load and signal. Their cure was apparently to add capacitance betw screen grid and ground on the output tubes. This is not conventional but may be effective. I have no experience with this fix. Williamson also became aware of stability issues with his original design and when he updated it in 1949 he attempted to adddress it. His fix was to place a phase compensation network across the plate load resistor of the input triode (50K resistor feeding the plate of V1 in your schematic). Williamson's network was a 4.7K resistor in series with a 200pF cap. If you look up any of the contemporaneous clones of the Williamson design, odds are that you'll find some variation of this network as well as a small value cap across the feeback resistor. The Heath W5 comes to mind here but there are others. Also, there were/are countless articles on improving the basic Williamson design and most of them address the stability issue too. IOW the original design, while brilliant for the time had real world flaws.

The reason I mention all of this is that the circuit you posted will most likely be unstable under some real world condition of operation if the feedback loop is closed because it doesn't incorporate stability improvement measures even Williamson himself later came to unerstand as necessary. If you decide to close the feeback loop, you need to verify that the amp is stable or you might run into issues ranging from unexplainably poor sound to speaker damage.

One way to verify stability is to short the input, load the amp with a resistor of appropriate wattage (4, 8-16 ohm etc./20W), put a scope across the output and then parallel the load resistor with a 1uF cap. If the amp is stable, nothing appears on the scope either before or after the cap is added. If the amp is unstable, it will usually break into oscillation at somewhere betw 40KHz and 100KHz when the cap is added. If it does oscillate you probably need to take corrective action of some sort. If it doesn't oscillate, consider yourself fortunate.

Good luck!

This post is made possible by the generous support of people like you and our sponsors:
  Kimber Kable  

Follow Ups Full Thread
Follow Ups
You can not post to an archived thread.

[ Top Top of Page ] [ Contact Us ] [ Support/Wish List ] [ Copyright © 2025, Audio Asylum, all rights reserved ]
[ General ] [ Speakers ] [ Tubes ] [ Vinyl ] [ Digital ] [ Hi-Rez ] [ Video Asylum ] [ Cables ] [ Tweaks/DIY ] [ Music ] [ Films ]