Regulated screen grids in guitar amp

[catalin gramada]

Hello. There is a while I started to experiment with regulated screens into a project seems never ended. On short. my curiosity was to use regulated voltage screen into a p-p project. Started with power zener shunt regulators and ended with independent low impedance power supply sources the results was always disappointed in respect with 'classic' series drop resistor derived from plate voltage supply.- from a sound point (hard to describe by audio attributes, - like lack of dynamics, dull attack, smoothly sound etc... The power supply is very low impedance, rugged hefty one with a excellent 'self regulation' 3% around , so the plates and control grids voltage are almost stone stable. Even so the screen regulation seems it not work for me for guitar applications.
It seems I failed
My question is : Did anyone experienced the same results with regulated screens grids or should I reconsider the whole experiences I did and move in other direction please , simply cause don't sound right to me ? Thank You.

[g1]

Some of the Leslie amps used a gas regulator (OC3) for screen regulation. Have you ever tried?

[J M Fahey]

Main "feature" of Tube amps in Guitar use (and to some extent on Hi fi too) is that overdriven Power Pentodes (and their cousins Beam Tetrodes such as 6L6) *compress*, big time, when driven hard, by a "hidden in plain sight" compressor.

Two mechanisms for that.

#1 is grid rectification which creates extra negative bias, which along other side effects lowers gain.
Tubes, and specially pentodes, are so sensitive to that that bias variation is used for Tremolo, go figure.

#2 is screen voltage: lowering screen voltage not only lowers plate current (screen grid is a hidden in plain sight second control grid) but also lowers transconductance, so *directly* lowers gain too.

And screen voltage drops significantly when amp is driven hard, so gain drops, so compression happens.

Now regulating that voltage negates that compression, so amp becomes "stiffer", more "solid state"..

For a Guitar player, not a good thing.

[Mick Bailey]

I have an amp head I built about 20 years ago where I thought I'd try out regulating the screen supply and I've never been happy with it. In every other respect there's nothing at all unusual. It's a bizarre coincidence that I dusted it off this morning prior to reading this thread with a view to doing 'something' with it. Parts aren't cheap any more and it has some decent components and covered with genuine Rexine bought a few years before the factory closed. (it has the same smell as early 'smooth' Vox amps).

[Tom Phillips]

Many years ago there was an article published in Glass Audio magazine describing how an Audiophile "Improved" the circuits of a Fender amp. I believe it was a Princeton. The improvements included regulated power supply rails including the screen supply. The "improved performance" was demonstrated by taking test bench measurements. At the time, I remember thinking that the resulting amp must have sounded very sterile. The author, of course, totally misunderstood the concept that a musical instrument amp is part of the creation of the sound.

[catalin gramada]

...any advantage if introduce a bit of separation here, like this please...?

[Mick Bailey]

What tubes are you using?

[catalin gramada]

What tubes are you using?

There are 6550 right now but it works well with whatever I puted in. 480v plates with 1.9kRaa. The series 1k resistors are in respect with max dissipation allowed to the screens (used 470 ohm for shared one)

[catalin gramada]

...but to be honest this configuration with undecoupled series resistors sound best to my ears. I did also a quick A/B test with a strap wire, meant a common 500ohm (2×1k parallel) in all screens- aka hiwatt dr103 style , but I prefer this one by far...

[catalin gramada]

...it is clear it is not just a matter of voltage drop effect only, but is also a form of feedback induced in circuit which shapes the response....not to clear for me yet...because the voltage drop the same in all instances I tried it...I even used the same components...but sound very different in respect of screen resistor is decoupled or not...

[Mick Bailey]

I've been giving this some thought but can't arrive at a reason why this arrangement would make such a noticeable difference. I would have thought the difference would be maybe marginal, but I haven't heard the amp. Scoping the output under signal conditions would give something to compare, but that's reliant on capturing an exact A-B snapshot which isn't easy.

[tedmich]

of course you can regulate a lot of places...probably all bad for guitar

[Tom Phillips]

I've been giving this some thought but can't arrive at a reason why this arrangement would make such a noticeable difference. I would have thought the difference would be maybe marginal, but I haven't heard the amp. Scoping the output under signal conditions would give something to compare, but that's reliant on capturing an exact A-B snapshot which isn't easy.

My thoughts are that the stiff regulated supply will affect the dynamics (Sag, recovery, feel etc.) All things that would be difficult to show via signals captured by a scope unless you could do side-by-side capture of pulse response or something similar. It comes down to a guitar amp creating an imperfect response that results in the creation of desirable sound and dynamic feel for the guitar sound. Such a response would be considered a fault for a sound system reproduction amp.

[Helmholtz]

Why is this thread in Hang Out/Lobby and not in Theory&Design?

[loudthud]

Here's a way to measure a power amp's change in gain.

You use an Audio Precision THD Analyzer. (any model with do.) You set the units for the measurement side to dBg. That's dB with reference to the Generator's Output. Set the Generator frequency to something 200Hz to 1KHz. Next, setup a graph to measure the amp's output (with the appropriate load impedance) on the Vertical and Generator Output (units don't matter, I use Volts) on the Horizontal. You set the sweep to Log steps from perhaps 100 microVolts to 1 Volt. Number of steps to something like 50 or 100.

You want to set the amp's controls for heavy distortion with a 1 Volt input. All controls at Max is OK.

Here's the tricky part. Set the Generator to Auto-On. (It's possible to do a sweep with the Generator set to Off, Doh!) For the first sweep, set the Generator to Off. Then do a sweep to plot a graph. The analyzer will turn the Generator on, plot the graph as it sweeps and turn the Generator back off. Next, set the Generator to On. Set the Generator's amplitude high enough to distort the output. Now do another sweep without erasing the previous graph. As the sweep begins, the Generator's output will be set to 100 microVolts and you will see the difference in Gain caused by power supply sag. The power supply will slowly recover over the first few measurements and the plot lines will converge.

An interesting experiment would be to change the amp's Mains Voltage by 5% to 10% to see what influence that has on Gain.

FYI an Audio Precision analyzer is a precision instrument with an Audio Generator and Analyzer built into one unit, all controllable from a program running on a PC. The PC program allows you to plot and print many types of graphs. Nelson Pass uses one on many of his projects over at diyAudio forum.

Addendum: the AP analyzer does a true RMS measurement on it's input (distorted or not) and uses that to compute the gain.