Compression and Touch Sensitivity

[LOUDthud]

I was reading an old thread about the highish 6K6 impedance of the Express and the reason for it.

Link to thread: https://tubeamparchive.com/viewtopic.ph ... nt&start=0

j-po had an interesting observation.

The highish 6.6k impedance causes more screen current to flow when the tube hits saturation thus making the distorted sound rounder. The low B voltage helps the screens survive the larger current.

One of the things I've been studying is how B+ regulation (sag) in the preamp actually changes the gain of the preamp. Between idle and when the power amp begins to clip the gain of each preamp stage falls by about 1dB per stage. My theory is that in the Express, the big 1K screen dropper resistor paired with the large screen current contributes substantially to the compression and tough sensitivity.

Measuring the gain of a guitar amplifier is somewhat tedious because you have to measure the input, then the output and do the math. I recently obtained an Audio Precision System One that makes it easy to do the measurement. You simply setup a sweep of output vs generator level but set the output to display in Volts dBg (dB in reference to the generator) and it does the math for you. I set the gain/volume/tone/presence controls of the amp so that the output clips when the input is 100mV. This is just an arbitrary reference so I know at what generator level clipping of the output begins and how much compression has occurred. Generator frequency and tone control settings don't mean much because it's not the absolute gain I'm interested in (although I'd like to know), it's the amount of gain compression that I want to measure. A resistive load is used.

I know that touch sensitivity involves attack and recovery time constants. At this time, those time constants will have to be determined by experimentation unless someone can figure out how to measure them.

Below, the attachment is a compression measurement I made of a 5E3. Can someone do the measurement on an Express?

[geetarpicker]

I've even noticed on my original Express amps sometimes (it seems somewhat tube dependent though) if you crank the amp far enough when you chomp down on a chord and then quickly stop playing you can hear that the hiss from the gain has almost disappeared, then the hiss ramps back up but at a rate slow enough that you can actually hear it. It almost sounds like a studio compressor with a fast attack and slow release time (maybe 1/2-1 second?) recovering from a strong transient.
I've never heard this on any other type of amp, though an old Fender tweed comes close but in sloppier way.

[iknowjohnny]

This is what i always figured was the reason that when i used seperates in the early 90's for about a year or 2 i always noticed that of the 3 different preamps i used with 3 different power amps, all of them felt stiff and undynamic. I always attributed that to the fact the power section's demands weren't affecting the preamp since they weren't sharing a single PSU. It just makes sense.

[Reeltarded]

I use that for an effect with my lil '56 Retsch amp. Soft power supply sags like a wet noodle when it's cranked. The res can't keep up with current demand and that is real sag on that lil amp. Saggytarius. lol

[Blackburn]

I want to see that amp. Send it to me so I can borrow it for a while. I won't make it sag any more than it already does, I promise.

[katopan]

The thing I'm more interested in, and I've posted this before, is studying how the Express transitions from clean to dirty and the theory that the drop in gain and headroom you talk about after the screen supply means one side of the PI has it's headroom threshold drop below the clipping threshold of the power grids. If this turns out to be right then the output distorted waveform could turn out to actually be predominantly power clipping on one side and predominately PI clipping on the other as well as being assymetrical.

I'd be happy to make the same measurement on an Express, but I'm only just starting to build mine and measuring it on my 10W version with 6P1Ps in the output stage wouldn't be valid or the same. But I think it will just show you that as you go into clipping, the gain drops. I suppose the shape of that transistion might tell you something....