Help with power scaling bias?

[Speedypancake]

I'm about to build a jcm800 style amp and already have the appropriate power and output transformers.

I intend to investigate "power scaling" by driving the B1+ & B2+ from a variable (mosfet) HT power to see what effect this has on overall volume & tone.

I need help to understand how to bias the pair of EL34 for "safe" or optimum operation when driven from a lower plate voltage, eg. 350V instead of the stock 50W voltage of @ 470V.

Is this correct:
Idle dissipation is simply the product of the plate voltage and plate current at idle, i.e., when no signal is applied. This is easily measured and calculated.

Plate dissipation follows Ohm's law, in that the power is equal to the voltage multiplied by the current.

If I measure across a 1R / 1% resistor in series with the cathodes, we can read the mv which directly equates to the current flowing in the valve in mA.??

So if the Plate voltage is @ 350V and the bias voltage is variable between -25V to -50V, what range of idle current should I be aiming for to keep within a "safe" plate dissipation range for the valve.

I'm sorry if this is common knowledge here or a dumb question, but I've googled biasing push pull valve guitar amps until my head aches, so...

[martin manning]

Some more reading for you: https://tubeamparchive.com/viewtopic.php?f=6&t=28948 There are probably more threads on VVR here too.
Bias voltage is reduced along with plate and screen voltage. Idle dissipation goes down when the voltages are reduced, so no need to worry about that.

[Speedypancake]

Thanks Martin - that "power sliding" circuit looks interesting.. so a dual 1M pot which controls B+ & adjusts the bias to suit?

[martin manning]

Exactly.

[katopan]

I was going to post my opinions on the Powersliding circuit, but see that I've already done so in the thread that Martin linked.

[Speedypancake]

Thanks Katopan. Just signed up with Watkins and looking forward to trying the VVR fixed bias circuit in the pdf from the thread you linked to

[martin manning]

Thanks Katopan. Just signed up with Watkins and looking forward to trying the VVR fixed bias circuit in the pdf from the thread you linked to :)

Can anyone upload that .pdf? I was signed up at one time, my e-mail is taken, but it says it is not recognized. Catch 22.

[stephenl]

https://londonpower.com/electronics/pow ... ixed-bias/

This works best with a low impedance dc supply of about -80vdc...so a separate transformer winding of about 50 - 60 vac. He also has a raw bias supply kit that includes a small transformer.

I put it in every amp I build and I've retrofitted it into existing amps and they work really well. A nice added benefit is that it will likely extend your power tube life, especially compared to running the amp full out into a load box.

Another design consideration is to have two powerscale pots (with associated switching) - one for clean, one for OD. This allows you to set the poweramp headroom for each mode.

[martin manning]

I was going to post my opinions on the Powersliding circuit, but see that I've already done so in the thread that Martin linked.

Thanks Katopan. Just signed up with Watkins and looking forward to trying the VVR fixed bias circuit in the pdf from the thread you linked to :)

With the bias coming off a 1M pot wiper, it's possible that the grid leak impedance could be quite high. Any concerns about that?

[katopan]

With the bias coming off a 1M pot wiper, it's possible that the grid leak impedance could be quite high. Any concerns about that?

Absolutely! I assume you are referring to the Power_Sliding pdf? I made comments on the Wattkins thread about why I thought that was a bad idea, how the mosfet in the bias half isn't actually contributing anything and it's the same as just using the 1M pot. Do you have access to Wattkins Martin? I can cross post some of the info including the jpg circuit diagram, which I posted there in 2016 if that helps completeness of the discussion. Tillydog's post https://tubeamparchive.com/viewtopic.php?t=24565 from 2014 also discussed just copying the B+ side of cathode bias VVR but reversing polarities (P ch mosfet and reverse the zener).

[martin manning]

...I assume you are referring to the Power_Sliding pdf? I made comments on the Wattkins thread about why I thought that was a bad idea, how the mosfet in the bias half isn't actually contributing anything and it's the same as just using the 1M pot. Do you have access to Wattkins Martin? I can cross post some of the info including the jpg circuit diagram, which I posted there in 2016 if that helps completeness of the discussion. Tillydog's post https://tubeamparchive.com/viewtopic.php?t=24565 from 2014 also discussed just copying the B+ side of cathode bias VVR but reversing polarities (P ch mosfet and reverse the zener).

Yes, the sliding circuit mentioned above. I did have access to the Watkins forum at one time, but I've long since forgotten my PW. Now it says my e-mail is taken, but it won't send me a PW change request link. If you can post the preferred circuit that would be great.

I suppose the PMOSFET does have some benefit since it allows the use of an available dual pot to control the bias voltage, which is then fed in with a normal grid leak value.

[katopan]

Like Tillydog's post here on TAG, it's just the standard VVR circuit, reversed polarity and smaller mosfet. People get hung up on mosfet selection and if you know how to spec it, there are many to choose from. I picked the model shown in the diagram below because it was one of the cheapest options from RS Components that fit the bill at the time I bought some.

I've left the 12V zener for mosfet gate protection, but with the short circuit current limiting removed the threshold voltage becomes not critical. It just needs to be above the normal forward voltage gate to source, but below the maximum rated gate to source voltage. I know some people have just copied the cathode bias VVR set up including the current limiting, but it's not needed as almost always the bias supply upstream of this has at least one supply chain dropping resistor which will limit current anyway.

Keep the existing bias circuit including bias adjustment all upstream of the VVR, then the output of the VVR feeds the power grid leak resistors.

Ok, don't use this circuit. The B+ side is fine and just standard VVR. But the bias side is a bit funny. It'll work but it's as good as using the pot just by itself (which might actually work just fine too!).

Funny what you see with a second look. I don't know if Dana's VVR3 is the same circuit as this (with the different resistor values) or not. Photos of his PCB only show the top. Unfortunately the 'bias in' being next to the onboard trimpot and the 'bias out' being the other side of the dual 1M pot suggest it could be. So VVR is a standard source follower regulator circuit, except where in a regulated power supply the gate is tied to a fixed reference voltage (like a biased zener or smaller voltage regulator chip), here it's just a pot giving some proportion of the input voltage. Simple and beautifully elegant. The fixed bias add on circuit in the above power sliding information has the gate connected to the incoming voltage and a zener for gate protection (which does nothing during normal operation). Holding the gate at the incoming voltage means the mosfet source sits at a few volts lower (in this case less negative - p channel mosfet) than the drain & gate. Source followers effectively 'follow' the input voltage on the gate, but also have the effect of providing a lower source impedance. But that relies on the drain being connected to a node of lower source impedance, usually further up stream. With the gate tied to the drain all the current to any load on the output has to come from the input. Therefore in this arrangement it doesn't lower the source impedance or buffer upstream from the downstream load. The mosfet is acting as a 3V or thereabouts zener. In other words it's contributing nothing to the effectiveness of how the bias regulator is working and you might as well just use the pot by itself.

Now if we want to have the buffering benefits of a proper source follower, it needs to be changed to be the same as the circuit used for the B+ (without the current limiting sensing resistor). I wired it like that last night but in a moment of tiredness put it in the middle of the Express bias circuitry between the two caps instead of at the end. The funny stuff I was seeing with voltages is due to the bias supply resistive divider operation all being stuffed up with variable loading of the 1M pot. If I move it as is to the end of the bias supply, so it's between the bias supply cap node and the power grid leak resistors (like both the power sliding info and the VVR info say) all should be well. By the way the B+ side of things worked fine as expected. Got something else on tonight but I will hopefully get to swap these wires over on the weekend. I really want to have it working and proven for next Tuesday's jam session.

I've attached what I think would be a better arrangement than what is shown on the above power sliding circuit for the fixed bias add on circuit.

[katopan]

I've never fully understood why Dana included that diode. It protects the mosfet from reverse bias across drain to source during power off, when the filter caps are all discharging at different rates. But the mosfets that are used all have an internal protection diode (albeit parallel instead of series) anyway. I've never included the extra series diode and never lost a mosfet in amps from 2W to 35W. If you use a mosfet that doesn't have the parallel reverse bias protection diode then you would need it. Even as an external diode I think parallel is better than series because it still allows the discharge currents to flow like normal instead of isolating the slower discharge of the preamp filter caps from the faster power end filter caps. Obviously on the bias side it would be turned around backwards compared to the B+ side.