Voltage drop/sag under load

[guitarmike2107]

You have it the wrong way round. The bias excursion from grid current pushes the bias colder.

sorry I know this is a bit of a playground answer, but, not according to Mr Kuehnel for fixed bias. I have tried to explain it as best I could without copying his text word for word, I can scan it and let you have a look at it.
unfortunately I dont have a fixed bias amp here to try and measure.

Cathode bias amps do move towards class B on both accounts so yes I agree there.

To avoid the above confusion with "increasing" a negative number, I always use the terms hotter and colder instead, referring to more or less idle current. Colder means an increase in bias voltage magnitude, ie from -27 to -28.

good idea!

[MWJB]

If HT voltage & bias voltage sag in the same proportion then bias goes colder, the bias voltage would need to drop (nearer to 0v) in a significantly larger proportion than the B+ to start going hotter under load. Lots of fixed bias amps take their bias voltage from the rectifier AC side of the B+ secondary, the B+ winding on other PTs is usually tapped from the main B+ winding, a large draw on any secondary causes the others to drop too...it strikes me that whichever way you look at it, the bias goes colder under load in a fixed bias amp.

[Steve Conner]

Yes, I guess that is true. If all the electrode voltages decrease in the same proportion, the tube current will still decrease according to the two-thirds power of the voltage, or the three-halves or whatever it is again.

I say poo-poo to Mr. Kuehnel. When grid current starts, you don't measure any change in the bias supply voltage. But the coupling caps to the power tube grids charge themselves up, and add a negative DC voltage to the bias, making it colder. The voltage decays away to zero as soon as you remove the overdrive, so it's hard to measure. It also happens in cathode-biased amps, but they also have their cathode bypass cap charging up, which is doing exactly the same thing, temporarily pushing the bias colder.


I have a signal generator that can do gated bursts, of say 10 cycles at 400Hz, repeating a few times per second, and have used it as a handy tool to study bias excursions in all kinds of amps. What I saw supports the conventional wisdom, that they are large, nasty, and responsible for a good bit of the "feel" and "dynamics" in the tone.

In tube RF power amps, as well as Class-AB2 amps like the old Fender 300/400PS series, the grid current actually flows back into the bias supply and tries to pump the voltage more negative. It has to be shunt regulated, or have a resistor across it to waste the power fed back into it.

[guitarmike2107]

Ok I agree, but what if you had a very overated PT, that did not sag, or had a seperate bias supply transformer that was not linked to saggy power rail, what happens then?

The book didnt take sag into account, which seems silly now as almost all amps have some sag and therefore overall would push the bias voltage closer to class B.

[guitarmike2107]

I say poo-poo to Mr. Kuehnel. When grid current starts, you don't measure any change in the bias supply voltage. But the coupling caps to the power tube grids charge themselves up, and add a negative DC voltage to the bias, making it colder. The voltage decays away to zero as soon as you remove the overdrive, so it's hard to measure.

I suspect I have done a really crap job explaining what was written so I will scan it and PM you a link to read if you care.

I understand what he was saying, and I understand what you are saying, so my only conclusion is that I am currently confused

[MWJB]

"Ok I agree, but what if you had a very overated PT, that did not sag, or had a seperate bias supply transformer that was not linked to saggy power rail, what happens then?" In the first instance you would delay the onset of sag...regarding the second, I guess you would have to set the idle current cooler to stop the tubes from redplating under load...much like you do already.

[kg]

sometimes with an amp that shows excessive power supply sag, a little bit more power can be squeezed out by increasing the primary impedance.

you can do this by putting a higher than "rated" load on a given tap. ie, 16r load on 8r tap.

this flattens the load line, causing more plate voltage swing rather than plate current swing.

in a multigrid tube the screens will tend to draw more current with the flatter load line as they will spend more time at a potential lower than that of the plates (when plate current is highest). keep an eye on their dissipation.

you're not going to turn a 100w amp into a 200w amp this way, but you might be able to get another 10-15% power output.

ken