Talk to me about screen resistors...

[jem]

It's coming.

-g

Awesome, can't wait to see it!

[Mooreamps]

Ummm. . .
How about a guitar amp?
M

Think of it this way, in a guitar amp, it's a square wave at a 100% duty cycle, assuming class A single ended. Hense, the need for a series resistor, after the filter cap.

-g

[Mooreamps]

It's coming.

-g

Awesome, can't wait to see it!

[Merlinb]

Anyone want to take a stab at why we might want a capacitor at the screen, not before the screen stopper? In my RCA manual, they put a cap at the screen, but they don't say why. Gary says it is counterproductive, but perhaps it helps stabilize current through the tube by maintaining a more constant DC voltage?
M

Because guitar amps want to maximise power they usually run at high screen voltages. If you put a cap there that voltage remains fixed. You need the series resistor to prevent current overload when the valve is overdriven. The RCA manual probably never intended overdriving!

Of course, if you make the screen voltage lower to begin with then you CAN fix the voltage with the cap, but you lose some compression, and waste more heat from dropping the quiescent voltage.

[zaphod_phil]

Because the grid curves of a pentode allow the signal voltage to swing almost to 0V on the anode, and because we would like to run the stage in Class A, there is a rule of thumb we can use....

I'm curious - are you ignoring the anode voltages below which the load line crosses the knee of the tetrode/pentode characteristics?

Hucky says, "the ultralinear circuit is acting like a local negative feedback on the tube. At the same time this feedback loop is overall with the primary OT and can minimize the "magnetic distortion"

Yes, that's correct. And triode mode is even more NFB.

.... and both the circuits they used as examples have caps AT the screen grid. We dont seem to do that...

We have a filter cap for the screen supply and then a current limiting resistor at the screen grid itself. This helps limit potentially harmful current surges into the screen grid(s) when the amp is cranking. If you have a cap right on the screen grid, then you lose that protection. The resistor is also recommended to help deter parasitic oscillations taking place, and is therefore preferably soldered right onto the appropriate socket pins.

In one place they say that it is OK to use a fixed resistor as a drop from the power supply during times when the tube is not being stressed by high current fluctuations), but in another place they talk about the need for a separate supply of constant voltage to minimize distortion during times when there are heavy currents in the plate and screen circuits. Hmm. . Is this the kind of distortion we want?

Amp designer Carl Berger did some tests with an SE 6550 power amp, running at around 400V with a regulated 250V on the screen grid. He discovered that this power stage sounded much better with a simple resistive dropper to the screen grid, which actually ran the screen close to anode voltage. There was a flattened compression of a sine wave signal which sounded really good.

Anyone want to take a stab at why we might want a capacitor at the screen, not before the screen stopper? In my RCA manual, they put a cap at the screen, but they don't say why. Gary says it is counterproductive, but perhaps it helps stabilize current through the tube by maintaining a more constant DC voltage?

This is often done with small signal pentodes, like EF86s, to help filter noise out. I don't think it would be a good idea with power tubes, for the reasons I just stated.

The RCA manual probably never intended overdriving!

That's another possibility - ie that the amplifier in question is only operating very gently, and preferably in true Class A. I still don't like it though. Even hifi amps can get overdriven when a 2-year old walks into the room and twists the volume knob, while either Mahler or Metallica is playing! IMO it's always good design practice to at least have a small safety resistor (like 100 ohms), on each screen grid, even in UL and triode mode.

[Mooreamps]

The RCA manual probably never intended overdriving!

Gee, you think ??

-g

[Manoteal]

OK, I will drop the cap at the screen for now. There really doesn't seem to be any benefit except that you need to have a lo Z run to ground for AC. But I guess you still get that through the stopper resistor then down through the filter caps.
What would happen if you ran the screen voltage real low? Like 100V, or 200? Or even less?
Manoteal

[zaphod_phil]

You would probably lose some power, change your amp's impedance match, stuff like that. I don't think it would be a real smart thing to do...

[Mooreamps]

What would happen if you ran the screen voltage real low? Like 100V, or 200? Or even less?
Manoteal

I believe it browns the tone. Also, you would need very high wattage low ohmic power resistors to provide a low enough impedance to support the current needs of the screen, and even more so for the plate. I'm sure because of this Kevin uses a power FET device rated for both the high voltage and current requirements.
-g

[Manoteal]

So how is that different that lowering the plate voltage? I assume that if you lower your plate voltage that you also lower your screen as well to not let it 'become the plate'. I was just thinking about leaving the plate circuit alone and lower the screen voltage with a voltage divider or by large series resistance. I know it messes up the Mu and current, but I was just curious about how low you can go without doing real harm. I don't mind the mismatch, I am trying to get less power in the first place. This will be an option that I can switch in to lower the level when I want so I don't have to pad down the output so much to bring it down to bedroom levels.
M

[Mooreamps]

If the impedance of the screen circuit is 22.5 K ohms, the source driving impedance should be at least 10 times less to drive it ; 2.5K ohms. 300 volts over 2.5K ohms is 120 millamps ; squared is 0.0144 times 2.5K equals 36 watts. So here is your voltage divider ; 2.5 k ohms at 36 watts, better make it a 40 watt resistor.

-g

[zaphod_phil]

Why would you want to drop a whole 300V across the screen supply resistor? That suggest you want to run the screen grids at 0V. Also, your approach would tend to hold the screen voltage pretty steady due to the low supply impedance. However, in a guitar amp, we usually want the screen voltage to swing, which is easier done with a higher impedance supply to the screen grids. The low supply impedance will also tend to make it more easy to fry the screen grids.

[Manoteal]

Why would you want to drop a whole 300V across the screen supply resistor? That suggest you want to run the screen grids at 0V. Also, your approach would tend to hold the screen voltage pretty steady due to the low supply impedance. However, in a guitar amp, we usually want the screen voltage to swing, which is easier done with a higher impedance supply to the screen grids. The low supply impedance will also tend to make it more easy to fry the screen grids.

That was Gary's math and suggestions. I would lose too much B+ doing what he suggest and a 40 watt resistor is not what I need.

My idea-
The stock bleeder resistor in the VJ is 220K and is less than 5 watts. So the amp is capable of that kind of load easily. In fact, it is 1.3 mA. So if I want to do, say 80K for my voltage divider from the screen supply, I change the 6 mA that is there stock (4mA for screen and 2mA for preamps) to 3.5 mA more than that to equal 9.5 ma. So, instead of dropping 28 volts across the 4K7 screen supply resistor with the 6 mA load, I am dropping 44.6 volts instead. That puts the screen voltage at 240 or so compared to about 280 now. So I could make a voltage divider of a total load of 80K which a part of it would be variable so that I can lower the voltage at will to the screen. I still have the 2K stopper resistor in there for peak current reduction. I would beef up the caps in the 'new' screen supply to help stabilise the voltage. Since the screen draws at MOST that I have measured as 40 mA, I don't see why I need 40 watt 2.5K load like Gary suggests. I think he is wrong in his calculations somewhere and I don't think he has tried this (unless this is exactly what he is doing in his power brake circuit). The idea is to reduce the screen voltage enough so that it controls the current through the tube, which lets the B+ go higher due to reduced load helping maintain proper loading on the tube. I am not trying to cut off the screen voltage entirely, and I certainly am not doing it as a series resistor to the screen. I am making a low current voltage divider that allows for some control of the output tube parameters. The current draw is not so large as to cause problems, but at most I might have to decrease the value of my R12 and increase the wattage. But even without changing it I get less than 1/2 watt dissipation on this resistor (9.5mA x 44.6V= .418watt).
Now.
Any questions or comments?
Manoteal

[Mooreamps]

I was not sure what you wanted to do. It's clearer to me now. If you just wanted to drop an extra 40 volts for the screen, which is a very good idea, then you can use a series resistor, or my preference to use a choke with a higher winding resistance. In fact, one choke I tend to favor the most is a Hammond rated at 20 Henrys with a series resistance of 1.6 K ohms, rated for 20 mills, which does exactly that.

As far as how I do my power brake, I know I've been doing a tap dance around that issue for a few months now. I shouldn't say how far I go to hide the circuit by concealing an extra tube inside the chassis to make that thing work. But if one were to read the sum collection of all my posts in the last few months, I practically have already given out the circuit already.

-g

[Manoteal]

Hi Gary,
I am glad to see what I am trying to do. From what I gather, you can reduce power quite a bit by changing the screen voltage, and I just wanted an adjustable way to do that. I bought a couple of 5 watt pots (one at 30K and the other at 50K) to try and see what I would need as far as adjustment range. I would then connect the pot to the screen supply feed to the tube when I am in pentode and use a high wattage (3 or 5 watt) resistor at 50-100K to ground to complete the circuit. Like Phil says, I do want it to fluctuate a bit with the signal, because I will only use this in high gain settings when I want to reduce the output. I still have triode and ultra linear for my other 'full power' choices that just have the 2K stopper going to the screen and will not upset the voltage balance normally used.
I was just seeing if you guys were awake!

Another power tube in your circuit for power brake. Hmmm instead of a mosfet, I would guess you are using the tube as a variable voltage source for the plate. Since you call it electronic power brake, I guess you have to have an active device in your circuit. But I think my way will work over a range that will prove useful and I will call it an electric power brake, since it doesn't have an active device in the screen supply.
Since neither you nor Phil (or anyone else) has told me that I will blow up the amp by doing it this way, I guess I am cleared to try it.
Thanks for all the input!
Still looking forward to your thermionic power brake, Gary. That is a great idea!
Manoteal

[Mooreamps]

I did try it that way once. It lived for a couple days, but then after playing the amp hard, it burned a hole inside the pot. Three of them, and two of those were reostats. Screen current is just too high when your running over-drive. It sounds like it would have been a simple idea, but in practice you will not get very many miles out of it.

-g

[Manoteal]

Ok, I see that it may not be the best idea. What about having a couple of switches or a rotary switch to select different screen resistors or taps in the voltage divider? That would avoid the burned pot thing.
M

[Mooreamps]

Ok, I see that it may not be the best idea. What about having a couple of switches or a rotary switch to select different screen resistors or taps in the voltage divider? That would avoid the burned pot thing.
M

The rotary switch idea seems to work as that seems to be the set-up on the WestAmps "Ultimate Upgrade" VJ head. But it's unclear if it's switching the plate voltage or not.

-g