SVT 6550C =(C)= max. screen voltage

[novosibir]

Hello everybody!

The older data sheet of this tube says: screen voltage at idle 425V (absolute maximum ratings)
The newer data sheet of the same tube says: screen voltage at idle 400V (absolute maximum ratings)

Does anyone have real life experience, what would happen to these tubes, if I'd run a quad of them class AB PP with i.e. 470V on the plates and 466V on the screens, using independent screen resistors of 1K ?

No theoretically thoughts please, only real life experiences.

Thank you!
Larry

[Gingertube]

I fitted new winged C 6550 quads into a pair of Aranov LS9100 monoblocks Jan 2013. Ultralinear connection with B+ of +530V.
No problems.
If a DIYAudio member then I posted the schematic trace (with some voltage readings) at post #7 here:
Aronov LS9100 - any info ??? - diyAudio

Cheers,
Ian

[bea]

Theoretical thoughts are often necessary in order to deal with potential weaknesses of existing designs.

I would try to ensure that the max power on the grids remains in the permitted range.

Did You check if that's ensured with only 1k as grid resistors? Under overdrive conditions?

(In my EL34 amps i increased the screen grid resistors to 1.2-1.5 k in order to stay on the safe side.)

[Mike Sulzer]

Theoretical thoughts are often necessary in order to deal with potential weaknesses of existing designs.

I would try to ensure that the max power on the grids remains in the permitted range.

Did You check if that's ensured with only 1k as grid resistors? Under overdrive conditions?

(In my EL34 amps i increased the screen grid resistors to 1.2-1.5 k in order to stay on the safe side.)

That's a good idea. Speaking theoretically, I will reply to you, since I cannot reply to novosibir. Suppose we drive the amp with a full power square wave, and the plate current in the tube that is on is 300ma. Suppose that the screen grid current is about 20% of that as a purely theoretical rule of thumb. That is 60 ma, and we drop 60 volts across the 1K resistor. That leaves about 400 from screen to cathode. 400*.06 = 24, but this occurs half the time so the average power is 12W. The limit is 6W. So we have a theoretical melting condition, which might have nothing to do with actual experience. My theoretical course of action would be to start with 3K and make careful measurements before proceeding to lower values.

[novosibir]

Did You check if that's ensured with only 1k as grid resistors? Under overdrive conditions?

Couldn't check anything yet, because I'm still over building the first two amplifiers of this type, which do have dual bias for quickly swap from 6550/KT88 to EL34 just by switching to the other pre adjusted bias section.

Suppose we drive the amp with a full power square wave, and the plate current in the tube that is on is 300ma. Suppose that the screen grid current is about 20% of that as a purely theoretical rule of thumb. That is 60 ma, and we drop 60 volts across the 1K resistor. That leaves about 400 from screen to cathode. 400*.06 = 24, but this occurs half the time so the average power is 12W. The limit is 6W. So we have a theoretical melting condition, which might have nothing to do with actual experience. My theoretical course of action would be to start with 3K and make careful measurements before proceeding to lower values.

As you've pointed out theoretically, I agree and this would end up with dead output tubes. NOS Svetlanas are too expensive for such experiment, where you already can guess, how they'd end up.
Therefore I meanwhile have decided for a regulated 400VDC screen voltage supply, which will be activated in 6550/KT88 mode only and disselected in EL34 mode.
I've found a tricky and easy to build circuit for the screen regulation, what requires just a hand ful of components. This one:

Amplified music : Guitar valve amp service

Scroll down to the chapter "screen regulation"

Although I'll have only the screen on this regulation and not the entire preamp section, as shown in this example, to keep the preamp behaviour and 'modulation', caused by the sag of the power supply.
And yes, I'm really curious about the result, how it turns out and will report about it here.

Btw. thanks for your comments to this concern.

Larry

[Mike Sulzer]

Couldn't check anything yet, because I'm still over building the first two amplifiers of this type, which do have dual bias for quickly swap from 6550/KT88 to EL34 just by switching to the other pre adjusted bias section.


As you've pointed out theoretically, I agree and this would end up with dead output tubes. NOS Svetlanas are too expensive for such experiment, where you already can guess, how they'd end up.
Therefore I meanwhile have decided for a regulated 400VDC screen voltage supply, which will be activated in 6550/KT88 mode only and disselected in EL34 mode.
I've found a tricky and easy to build circuit for the screen regulation, what requires just a hand ful of components. This one:

Amplified music : Guitar valve amp service

Scroll down to the chapter "screen regulation"

Although I'll have only the screen on this regulation and not the entire preamp section, as shown in this example, to keep the preamp behaviour and 'modulation', caused by the sag of the power supply.
And yes, I'm really curious about the result, how it turns out and will report about it here.

Btw. thanks for your comments to this concern.

Larry

A few points:

1. Regulated 400 volts is still 400 volts. Why would the screen dissipation be any less with a full power square wave? (A stated maximum in a spec does not mean that that value is suitable for what you want to do.)

2. If this is a guitar amp, I suggest that it might not sound the way you expect with a (proper voltage) regulated screen supply. Remember, the screen grid should be considered a modulating element. Whenever you change the screen voltage, you get a new set of plate characteristics. When using a "sloppy" screen supply, that is, a series resistor, you are creating a compressed waveform, and so introducing a pattern of harmonics. This is not something I have blind ABed (my hearing would not survive), but I think you want to be careful about changing the nature of the basic design unless you really want to.

3. I think a reasonable way to design this circuit would be to find the screen resistor that, given the idle current screen voltage, lowers the screen voltage sufficiently at clipping so that you can get enough plate current to get the power you expect with the plate voltage not falling below something like 100volts. Note well: There is a very scary plot in the available on line Tung Sol fact sheet for this tube. It shows screen current (and thus dissipation) running away if you allow the plate voltage to fall too low. In effect, the screen becomes the plate if you keep its voltage too high with small control grid voltage, but it cannot handle the power. I would design this circuit by careful experimentation and measurement beginning with a resistor known to be too large, but someone else might have a better idea. But remember, there is a reason why regulated screen supplies are often at about 250 volts rather than 400.

[novosibir]

1. Regulated 400 volts is still 400 volts. Why would the screen dissipation be any less with a full power square wave?

As long as the plate voltage is higher than the screen voltage, the plate is sucking the electrons away of the screens (beam power plates). But if the plate voltage drops below the screen voltage, then is 'game over'.

2. If this is a guitar amp, I suggest that it might not sound the way you expect with a (proper voltage) regulated screen supply. Remember, the screen grid should be considered a modulating element. Whenever you change the screen voltage, you get a new set of plate characteristics. When using a "sloppy" screen supply, that is, a series resistor, you are creating a compressed waveform, and so introducing a pattern of harmonics.

Thank you for pointing this out and for opening my eyes again! I've discussed this on a Facebook group (Tube Guitar Amp Builders), where others lead me to the idea of a regulated screen supply for safety & reliability.
But yesssss, this amp shouldn't be a tame home HiFi mono bloc, where linearity is the highest target. No! It should be an angry beast and should fulfil his name, what is "Pure Metal Machine"

3. Note well: There is a very scary plot in the available on line Tung Sol fact sheet for this tube. It shows screen current (and thus dissipation) running away if you allow the plate voltage to fall too low. In effect, the screen becomes the plate if you keep its voltage too high with small control grid voltage, but it cannot handle the power.

I know about this desatrous effect and this is, what I fear most of all. But nonetheless the regulated circuit is cancelled and I'll go back to an idea I've had, before well meaning people have lead me to the regulator solution.
The idea is, to have right after the choke first a common screen resistor of 470ohms/10W to 1K/25W feeding the four independent screen resistors of 1K/7W (Welwyn) and will verify by slowly driving the amp from idle step by step to full blast, that in no situation the screen voltage will exceed the plate voltage AND that the screen current doesn't exceed healthy values. Will be measuring the screen current via voltage drop across the independent 1K screen resistors, because with an 1K it's an easy math

Larry