Sag or "compression" from a cathode biased power amp?

[LooseChange]

Using a cathode biased power tube (or two), can you get sag?
I notice that the tubes will not draw anymore current when being pushed as compared to idle.
So, how do you get the sag?

[FYL]

You need to go active. For instance you can sample the small current variations in the output stage, then use them as control voltages for a tracking supply. KOC details a number of circuits in TUT and also offers full kits. Check for instance http://www.londonpower.com/catalog/product_info.php?cPath=7&products_id=81

[Fresh_Start]

What about the screen grid resistor?  Won't the voltage on the screen grid drop as it pulls more current?

Chip

[jjasilli]

This is a recurrent topic.  Last time I remember was a discussion with Panhead, who seems to have vanished.  I think [/b]cathode bias[/b] is not the issue; class A, is the issue.  Class A is almost always cathode biased and ususally SE, but that's not the point, though sometimes the termsm are loosely (no pun intended) used interchangeably. 

Anyway as LC points out, in Class A the plate is always drawing full current, even at idle.  But as FreshStart points out, the screen will draw more current at hi signal, thereby causing a voltage drop and maybe sag.  FWIW this is the position of Dan Torres.  But if I remember, Hotblue thinks this to be true, but so small as to be of no material affect. 

Meanwhile, despite my earnest pleas, no one has done any scientific tests.  So, in a fit of OCD I have just acquired my 4th Heathkit IM-18 VTVM, and my 3rd Fluke DMMl; and have a 'scope (which you guys have taught me to use).  I've been re-planning my workbench for a year now, and hope to have it all up and running presently.  Then perhaps I can engage in such tests - w/o electrocuting myself 

[FYL]

What about the screen grid resistor?  Won't the voltage on the screen grid drop as it pulls more current?

Yes, but the effect is quite small with beam tetrodes - a few mA at best with 6L6GCs, slightly more significant with high slope pentodes - maybe 10 mA for an EL34 PP.

[38Super]

Anyway as LC points out, in Class A the plate is always drawing full current, even at idle. 

This isn't quite accurate.  Class A amps are setup with a quiescent bias current (no signal).  The signal applied causes the amp to instantaneously draw more or less current as the AC waveform varies.

If your class A SE 6L6 is drawing 35mA, the cathode current might vary during play from 20-50mA, varying around the bias point of 35mA.

I always thought power amp sag comes from the voltage drop associated with tube rectifiers.  There's a compression effect at the onset of note attack that causes B+ to sag and decrease signal voltage due to increased current.  Then as note dies out, B+ starts to rise again, increasing power and sustain.

cheers,

rob

[bluesbear]

Does anyone know what this "Sustain & Compression Enhancer" from London Power is and what it's supposed to do? The web site is pretty vague... as if he wants you buy his book so you can understand what he's selling!
Thanks,
Dave

[FYL]

This isn't quite accurate.  Class A amps are setup with a quiescent bias current (no signal).  The signal applied causes the amp to instantaneously draw more or less current as the AC waveform varies.

Class A means that the amplifying device is on for the whole cycle. Current draw can't go below idle and usually doesn't raise by more than 15 or 20% on peaks. If not the amp isn't class A...
 

If your class A SE 6L6 is drawing 35mA, the cathode current might vary during play from 20-50mA, varying around the bias point of 35mA.

More like 35 mA min and maybe 42 mA max.

I always thought power amp sag comes from the voltage drop associated with tube rectifiers.

Not only : it comes from all resistive elements in the supply. Mo' current demand => higher losses => lower plate & screen voltage => sag.

[FYL]

Does anyone know what this "Sustain & Compression Enhancer" from London Power is and what it's supposed to do? The web site is pretty vague... as if he wants you buy his book so you can understand what he's selling!

Minute current variations are probed, then used as command voltages for a crude tracking power supply. Mo' current drawn => less voltage out.

[Willabe]

Hey LC,

TUT 2, p. 6-42 -- 6-45, fig. 6-52 (Altec 436/8 compressor) and 6-54, simple (two tube, 12a_7, 6AL5) transformerless compressor. Do a search for the Altec 436, 438, look at the feed back loop in it, KOC's using the same type of thing on some of his,(feed back loop using a few diodes, C's and R's, a pot or 2 for adjustment of attack and/or threshold) output plate signal back to the input G1 of same tube.

You can tap signal from PP or SE from plate, or tap PI, PP plate or preamp stage plate (Champ driver).  

TUT 4, Sag chapter 4, fig. 4-35, (adding comp. side-chain to SE amp). TUT 6, Sustain chapter 7, fig. 7-15, (grid modulation for sustain control in cathode-biased output stage).

SUS-2 kit is for cathode bias amps.

        
             Brad                  

[Willabe]

Does anyone know what this "Sustain & Compression Enhancer"

bluesbear, I did'nt see your post till after I finished the above. KOC's got a few ways of implamenting a "sag" control circuit in a tube amp. Look at my post above and look up the Altec schemo. that's the bais for at least one type (side-chain) he uses.

IIRC, LooseChange has done a few amps with them in the B+ supply which is another way of adding it.

and if I'm right , he should be able to tell you more about them.

KOC's Power Scaling (VVR) adds a "soft" current limit circuit right after the mosfet reg. consisting of a small transistor and 2 R's, which is then made adjustable with a pot and 2 more R's. IIRC you can't use this type with high bias current amps. You need to use the side-chain type for SE or high bias and class A bias amps.    


            Brad            

 

[bluesbear]

Okay, now I'm really confused! Willabe says KOC's power scaling is the same as VVR. London Power's web site has paragraphs saying they aren't the same. So, I have two questions about all this:

1) What is the difference (assuming there is one) between power scaling and VVR and how does each affect tone (browning, etc)?

2) Does this "Sustain & Compression Enhancer" actually work as advertsed? Will it work with the VVR. THey have a model they say will work with power scaling but.....

Thanks,
Dave

[Willabe]

Hi bluesbear,

I'm sorry for the confusion, my fault. The early PS was the same type as the VVR, but KOC has new types that are not the same. The older style was called SB1 (for fixed bias) and  SB2 (for cathode bias). The newer ones have more options and do not need a large, mil spec. 24m.m. 2w. pot because the high DC voltage is not on the PS pots element. You can have multiple PS pots like one pot each for PA, PI, PRE, or one PS pot for each channel of a two channel amp, two pots for a single channel amp so you have two volume levels of PS, each with their own B+ trim pot, instant headroom foot switch,(which you can have/use with the SB1/SB2) ect. He's got 97 pages on PS in TUT4, and 78 pages in TUT6 on what he calls "DC Power Scaling", which is the newer style kit's. There's a lot of things you can do with these, if you need different "tools" to work with on stage. The older style works fine, just depends what you want/need to do/get from the PS/VVR.

Belive me when I say, I have a hard time with just the basics of this stuff    

One differance between the old style PS and the VVR is KOC uses the larger pot, and a larger mosfet case/package (TO-3?) to help keep the mosfet cooler. The mil spec pot is rated at 500vDC on the resistive element and 900vDC between element to pot case, for safety. I think a more standard pot (alpha) is only rated for 350vDC on the element and 500vDC element to pot case.

Now Dana might have changed these two parts, I don't know.

OTOH, it might not be needed to up the values on these two parts, depends on how high the B+ is and how much power/volume you want/need to drop in the amp your installing this in.  

I belive LooseChange has a few builds in which he has used/installed the sag kit.  

              Brad          

I forgot one more thing. KOC (SB1, fixed bias)and Dana use a different bias tracking setup. There is debate over this that VVR tracking does not track through the control pot's rotation as well as the PS's version. Which causes the VVR not to be able to keep the tubes transfer curve the same at different settings and this changes the tone. This is the main reason it's (PS/VVR) supposed to work, it does not change the shape of the transfer curve. It's not supposed to be just another form of attenuation, which does change the sound of the amp as it's over all volume is lowered.

And no I can't read or trace out a transfer curve        but there seems to be a lot of guys who are really happy with these, which ever one/style/maker they choose.  

The last point about the tracking regulator won't matter to you if your using cathode bias.
  

[HotBluePlates]

Anyway as LC points out, in Class A the plate is always drawing full current, even at idle. 

This isn't quite accurate.  Class A amps are setup with a quiescent bias current (no signal).  The signal applied causes the amp to instantaneously draw more or less current as the AC waveform varies.

You're right, but what people are really talking about here (whether or not they're aware of it) is the "average current draw" stays the same.

That assume a distortionless class A amp, so that the current variation above and below the idle point are equal. And in a push-pull class A amp, you could argue that the instantaneous increase of current in the push side is offset by the instantaneous decrease of current in the pull side. Either way, the current draw as measured on a meter stays fairly constant unless there is distortion (and there's always a little, and a bit of rectification effect, so the current does rise somewhat).

But as FreshStart points out, the screen will draw more current at hi signal, thereby causing a voltage drop and maybe sag.  FWIW this is the position of Dan Torres.  But if I remember, Hotblue thinks this to be true, but so small as to be of no material affect. 

I haven't measured, but I'm fairly certain I've heard the effect, as Tubenit has built a bunch of amps with largish screen resistors. I've heard so many of them, I can't remember which are fixed-bias and which are cathode-bias, but he gets just the right amount of sag at less than ear-splitting volume.

And I just posted in another thread describing (yet agin) the mechanism by which this sag occurs. And FYL is right that the effect could be more noticable in true pentodes (or other high slope tubes) which have a bigger screen current variation. The downside of my past suggestion to use big screen resistors is that the resistance value is fixed, and so you have to choose how much output power you need; that's because the sag effect is noticable because it is limiting the tube from making the full output power it would otherwise be capable of.

Anyway, the theory is there, so the real goal of scientific experiment should probably be to find what values of resistance give an audible effect. But in any case, the effect will probably only be noticeable, if my thinking is right, when G1 is driven positive. That's when the biggest increase in screen current happens, although any look at data sheets will show average screen current rises from zero output to full output.

And... if memory serves, O'Connor's standard Sag control inserts a MOSFET in series with the B+. As you turn a control, it adds series resistance to the power supply, which allows B+ voltage to drop with increased current draw. The intent is you build the amp with solid-state rectifiers, and can have minimal sag to rediculous sag at the turn of a knob. The Sag Control kit (at least when I got one way back) was not much more than a few resistors, a pot and a MOSFET. And maybe a small transistor...

And I haven't disappeared... I'm just swamped with work at my new duty station. 

[FYL]

What is the difference (assuming there is one) between power scaling and VVR and how does each affect tone (browning, etc)?

VVR is based on first generation Power Scaling. The only difference is that it works on clean filtered DC instead of pulsating DC. VVRs come between the first filter section and the R/LC sections downstream, with stages fed thru caps.

Today's Power Scaling works on pure DC, modules are connected between plates/screens/... and caps - stages are fed by the PS modules.

Does this "Sustain & Compression Enhancer" actually work as advertsed? Will it work with the VVR. THey have a model they say will work with power scaling but.....

It works very nicely in most amps and is quite subtle in others.