Switchable One Setting VVR/Powerscaling For Fixed Bias Amp

[Mason]

Is it possible to have a switchable vvr for just one setting of attenuation? I’m not really interested in being able to sweep through every different voltage setting, I just need one. I’m looking to drop power down to one tenth.

I saw something similar while searching through the forum but it was only for cathode bias amps. My amp is fixed bias.

Any ideas? Also would this be a simpler circuit than a normal vvr for fixed bias, or would it require just about as much space in the chassis?

[JB]

The basic VVR circuit is one supply in one variavble supply out, and would be easy to have a switched setting rather than continuously variable.

However your amp is fixed bias so you need the bias supply to track whatever the output valve anodes and screens are doing. 

There are circuits that will do that, but more complicated than the basic VVR.  An alternative might be a power transformer with more than one secondary so you can switch the whole amp to a lower supply voltage (assuming your bias is derived from the same, not a dedicated winding).

[sluckey]

Is it possible to have a switchable vvr for just one setting of attenuation?

Sure. You will also need a switch bias settings. May be as simple as switching a parallel resistor. A DPDT switch may be sufficient to switch B+ and bias.

[Mason]

Awesome. It’s for the Plexi 6v6 I just built. What value resistor would I be looking at to drop B+ by that much? My B+ right now is 450. And then are you saying to add another resistor to change bias?

[sluckey]

Awesome. It’s for the Plexi 6v6 I just built. What value resistor would I be looking at to drop B+ by that much? My B+ right now is 450. And then are you saying to add another resistor to change bias?

Build the simple vvr with pot for cathode bias circuits and connect it into your amp. Then adjust the pot for the reduced voltage level you want to operate with. Then without disturbing the pot setting, measure each end of the pot to the wiper. Replace the pot with two fixed resistors that closely match the measured pot values.

Then experiment with the bias. Put a resistor in series with the 47K resistor to increase the bias voltage and cause the tubes to run cooler. I would probably just use a 25K trim pot so I could easily adjust the low power bias point. Connect the pot on a switch so that the pot is in series with the pot or just goes straight to ground when in full power.

Be prepared to experiment a bit with this.

[2deaf]

. . . Then experiment with the bias. Put a resistor in series with the 47K resistor to increase the bias voltage and cause the tubes to run cooler.

If I understand this correctly, you want to increase the absolute value of the bias when B+ is reduced.  You would want to decrease the absolute value of the bias when B+ is reduced, so a resistor in parallel with the 47K resistor would be appropriate.

[sluckey]

If I understand this correctly, you want to increase the absolute value of the bias when B+ is reduced.  You would want to decrease the absolute value of the bias when B+ is reduced,

What are you saying? You can't have it both ways. 

[2deaf]

What are you saying? You can't have it both ways.

You want to increase the absolute value of the bias.  I want to decrease the absolute value of the bias, i.e. bring it closer to zero.

[sluckey]

What a difference a pronoun makes.

[2deaf]

Let's say you use a 470K/220K voltage divider on your 450V B+ and you set up a double-throw switch to control a MOSFET or IGBT or whatever and you wind up with a B+ of either 445V or 140V.  It is very plausible that 140V will result in one-tenth the power attained at 445V.  Let's also say that the bias is -40V at 445V B+.  My guess for an acceptable bias with a B+ of 140V is -7V.  Let's say that the bias is -40V when the 50K bias pot is at mid-rotation so that there is 25K in series with 47K going to ground.  If you shunted the 47K to ground so that there is only 25K from the bias point to ground, the bias will probably be about -16V.  You wouldn't be able to get the -7V needed by fooling with the 47K resistor alone.

[PRR]

Turn your pot(s) so you get the desired "low" sound.

Power down (bleed!) and measure the pots, both ways (top to wiper, wiper to bottom).

Use nearest standard fix resistor values.

[2deaf]

The problem with the Plexi 6V6 bias at the low sound setting is that the required bias voltage is not in the range of the bias pot.  The desired bias voltage can be attained if the 47K resistor is shorted to ground and the bias pot is adjusted.  But you would want the switch to give the correct low sound bias voltage with the pot still set where it was for the proper high sound bias voltage. 

Attached are two takes on the same idea.   

[Mason]

I'm still not sure I understand how to implement this.  Does someone have a link to the simple vvr you're mentioning?  Do I need specially rated resistors for this, and a 500v rated switch?

[92Volts]

There are a few ways to do VVR (Google will bring up schematics-- I don't know which ones are best) but it stands for variable voltage regulation-- it's not just a resistor.

If you don't need it to be variable, there are fewer obvious advantages over a simple resistor, but there is still an important one-- steady voltage with changing current draw.

The idea is you pass the B+ through a MOSFET or other giant transistor. The desired B+ (derived through a pot or resistors or zener diodes), is applied to the gate of the MOSFET. Like a tube in cathode follower mode, the source of the MOSFET will output very near the voltage applied to the gate and self-corrects if it's too high or low. But instead of getting a low-impedance signal from a cathode follower to drive a tone stack or whatever, this is our power supply.

[2deaf]

I'm still not sure I understand how to implement this.  Does someone have a link to the simple vvr you're mentioning?  Do I need specially rated resistors for this, and a 500v rated switch?

That big DPDT switch that Doug sells is fine.

[2deaf]

If you decide to go with the VVR, you have the opportunity to bring that 450V B+ down a little and give those poor 6V6's a break.

[jloewenheim]

The problem with the Plexi 6V6 bias at the low sound setting is that the required bias voltage is not in the range of the bias pot.  The desired bias voltage can be attained if the 47K resistor is shorted to ground and the bias pot is adjusted.  But you would want the switch to give the correct low sound bias voltage with the pot still set where it was for the proper high sound bias voltage. 

Attached are two takes on the same idea.

If you use this layout, will you not bleed off B+ voltage through the series 470K/220K resistors to ground?  Meaning in full power mode, you are still lowering voltage.

[sluckey]

If you use this layout, will you not bleed off B+ voltage through the series 470K/220K resistors to ground?  Meaning in full power mode, you are still lowering voltage.

Do you realize this topic is 6 years old?

The 470K/220K voltage divider is always in the circuit but the reduced voltage is not being used in full power mode. Full power mode uses the full B+ voltage as an input to the 100K so the output voltage will not be reduced.

[jloewenheim]

Hi Steve,

Thanks for the clarification and reply (Yes I do realize that it is 6 years old :) ). 

I am going to implement this in my Hot Rod Deville-to-cascaded plexi conversion. The plan is that I will just be adding VVR to the Screens.  For the Bias, the thought was to use the marshall style bias circuit with 2 resistor/pot combos.  Depending on which mode the amp is in (VVR or full), the ground connection to one of the pots will be disabled. I think this is how Suhr implements his dual bias circuits on the SL67/68 amps.

-Jake