MV, Powerscaling in 50W, 100W Amps

[soundguruman]

I agree with you. It just doesn't seem right to have these running at 1 or 2mA. And yes they go into crossover distortion very easily.

what you need is to control the bias with a more sophisticated system.
The crossover distortion should not be there at any wattage setting.
It needs to be clean down to 1 watt...
whenever that distortion is there the amp sounds gritty, and really defeats the purpose of scaling.
sure, you have reduced the wattage, but now it sounds like foo foo.

So how to do that?

[Gregg]

You can get a more sophisticated system by using a uCU. It would be interesting if someone measured all those voltages at different power settings in an actual AFD/YJM100.
The other way is determine it experimentally. I'll try playing with different bias settings at lower screen voltages and will let you know.

[markr14850]

I've used a multi pole rotary switch, to allow for customizing both the b+ and the bias to whatever I want.

[Gregg]

You can see the results from couple of experiments at low screen voltage settings. I was driving the PI with a 1kHz sine wave from zero to around 4 Volts. Because the pot sections didn't track each other very well this time the bias was set separately from the screen voltage/FB pot pair to match the values in the "% max Bias1" column in the table. This resulted in a 2 to 4 fold rise in idle current compared to the previously recorded data. Obviously idle current is very sensitive at those lower settings. After 150V tracking is better so I didn't bother:

AFD100EPA2.xls

I produced couple of scope/spectrum screenshot videos showing what's going on:

First video shows how the sine wave and corresponding spectrum behaves at full plate and screen voltage:

PSfullHV - Download, Play - 4shared

Screen at 100V:

PS100Vscopeandspectrum - Download, Play - 4shared

Screen at 50V:

PS50Vscopeandspectrum - Download, Play - 4shared

Here the screen is at 50V but I change the bias first to more negative and then back to less negative going down to zero. I'm not driving the input with maximum signal:

PS50VscopeandspectrumBiasVariable - Download, Play - 4shared

For better quality don't bother with their player but download.

[Gregg]

A Paul Ruby zener mod (12V at 50V screens so far) fixed those nasty waveforms and as expected fizz is gone too. Now it sounds more like it.

[soundguruman]

A Paul Ruby zener mod (12V at 50V screens so far) fixed those nasty waveforms and as expected fizz is gone too. Now it sounds more like it.

Yes, I tried that...I'm sure I did the zener mod before anyone else did...in the 1980s...

To get you up to speed, one of the first commercially produced guitar amps with power scaling was the Seymour Duncan Convertible amp....now many people should be familiar with that.

I built my first scaling circuit from the patent application. It was not from a kit or schematic...the first amp we installed it in was a Marshall 2000, in 1980s....

So now we know that the basic idea does work OK, and the circuit does track the bias voltage according to the screen voltage, in a primitive "rough" kind of way.

But now, we need to make the circuit work without the bugs.
There needs to be a single control, not 2.
And it needs to keep the bias consistent all the way up and down the scale.
Needs to keep the bias HOT, and distortion LOW, at ANY setting. The Zener mod DOES NOT accomplish this, but rather "fudges" the bias voltage at the low end of the scale.

Good idea, but how are you going to finish the design of the circuit and make it "practical?"

[ric]

I have a question about standby sw safety/best practice. I was told switching the HV CT was a possibility, and liked the sound of it and have used that method exclusively on my hobby stuff. With one side of the sw already at chassis grnd, it seems safer than it being in series with the HV. Any opinions on this?

[Gregg]

There needs to be a single control, not 2.

One way of doing it is using a uCU.

And it needs to keep the bias consistent all the way up and down the scale.
Needs to keep the bias HOT, and distortion LOW, at ANY setting. The Zener mod DOES NOT accomplish this, but rather "fudges" the bias voltage at the low end of the scale.

Also one way of doing it is using uCU but no matter how hot you keep the bias the amp will still go into crossover at least that's what my experiments show (for example 0V at 50V screens) unless I miss something and I'm not an expert, just a guy with a soldering iron in the hand doing experiments.
Getting clean undistorted output is not a problem but here we're not into HiFi. We want the amp to distort "nicely" when overdriven at low volumes. If a zener helps I'm OK with that. How to get it working at different bias settings is another question. Maybe a variable zener circuit will help, I don't know maybe I'll take a shot at it.
IMHO scaling the amp to 50, 30 or 10 Watts doesn't help very much because it's still VERY LOUD for and apartment (and if you plug into a 4x12" cab) so I'm wondering whether a 2-3 position scaling would not be more practical and of course being much easier to do.

[loudthud]

To get you up to speed, one of the first commercially produced guitar amps with power scaling was the Seymour Duncan Convertible amp....now many people should be familiar with that.

The one of those amps I worked on back in the 80's did not have Power Scaling as we know it today, it used a 12AU7 to feed current into the plates of the phase inverter. The remarkable thing about it was that it made what I call the Marshall wave. A non symmetrical square wave with an inverted shark fin on the bottom, and it maintained the wave shape as you turned the power down. There may have been more than one version of the amp, schematics were hard to come by.

[Alan0354]

Powerscaling by Marshall found in their AFD100 amp:

[ATTACH]22239[/ATTACH]

It's digitally controlled self bias circuit for each tube. Only screen voltage is variable (which was suggested by Steve in another thread) in 21 steps.

I cannot open the link.

[Alan0354]

I did some PS experiments with my power amps and I couldn't hear much difference between when plate and screen voltage are decreasing simultaneously and when only screen voltage is varied. In both cases the measured power level corresponding to the PS pot position is very close if not the same.

So you can't hear difference of the sound quality between the amp that you vary plate and screen voltage simultaneously vs the amp that you vary the screen voltage only? Then just varying the screen is much easier!!! Power dissipation is a lot lower.

I did the screen voltage power scaling last year, a friend of mine said it does not sound as good. But I don't have one that vary both plate and screen simultaneously, so I can't tell. I did linear scale the fixed bias ( well approximately!!!).

[Gregg]

I tried to upload the schematic back but for some reason it wouldn't show when you click on it. THis link should work:

AFD100MainPCB - Download - 4shared

So you can't hear difference of the sound quality between the amp that you vary plate and screen voltage simultaneously vs the amp that you vary the screen voltage only?

I couldn't distinguish any particular advantages of the full scaling over the screens scaling only. Maybe if I spend hours comparing them I will be able to find the difference but I'm not that picky. It's a compromise anyway.

[Alan0354]

I tried to upload the schematic back but for some reason it wouldn't show when you click on it. THis link should work:

AFD100MainPCB - Download - 4shared



I couldn't distinguish any particular advantages of the full scaling over the screens scaling only. Maybe if I spend hours comparing them I will be able to find the difference but I'm not that picky. It's a compromise anyway.

Thanks to confirm this. I did a power scaling using variac ( keeping filament and preamp separate) long time ago. And I did the screen scaling, I think the screen sounded just as good. My friend said he did not like the screen control, I just need a third person to confirm.

I used Zener to do the step dropping for the screen voltage. I know the screen grid draws current when the tube clips, that's the reason I did not use resistor divider as the screen voltage drop as soon as you clip. Do you think that makes a difference? I scale the grid voltage approximately the same ratio.

About the download, it is an .exe file!!! Do I install it or what?

Thanks

[Gregg]

About the download, it is an .exe file!!! Do I install it or what?

It's not an EXE file - maybe you clicked on the wrong button. Usually you shouldn't install any EXE files from download sites.
Click on the Download button - the one which is next to Share, not the blue one! Then choose Regular download, not Premium.

I used Zener to do the step dropping for the screen voltage.

That's one way to do it.

[Alan0354]

It's not an EXE file - maybe you clicked on the wrong button. Usually you shouldn't install any EXE files from download sites.
Click on the Download button - the one which is next to Share, not the blue one! Then choose Regular download, not Premium.



That's one way to do it.

I tried, there is no button to download the file, they all points to an .exe file!!!

But how do they do it?