OT Impedance Match with VVR

[martin manning]

Seems to me that when a VVR is used to reduce power the optimum impedance for the power stage will go up. For example, if the anode and screen voltages are reduced to 25%, Child's law would suggest that the current will be reduced to 12.5%, implying Zpri should be doubled. I traced some 6L6GC and EL34 curves at 100V Vg2, and the theory seems to hold.

Is this discussed in the power scaling literature?

[brewdude]

Ive never seen it discussed before, but I have experimented with some speaker impedance mismatching and have had good results plugging a 16 ohm speaker into an 8 ohm OT tap. I get confused as to whether this doubled or halfed the reflected load.

[pdf64]

That's correct, but I suppose that as the purpose of VVR / power scaling is to reduce the power output of the amp below that which it is capable of, then such an optimum is no longer applicable?

[JazzGuitarGimp]

Doubling the reflected impedance further decreases power output by 3dB(or is it 6dB?). It sounds like a win-win to me.

[martin manning]

Ive never seen it discussed before, but I have experimented with some speaker impedance mismatching and have had good results plugging a 16 ohm speaker into an 8 ohm OT tap. I get confused as to whether this doubled or halfed the reflected load.

A 16 ohm speaker on the 8 ohm tap will double the primary impedance, which is in the right direction. I was hoping that someone had or could try this out. Does it sound better that way?

The optimum bias point at low voltage is another question, but in theory it should be proportional to the anode and screen voltage reduction, and that is what VVR's with bias tracking attempt to do. The anode curves I traced at low voltage seem to bear that out. I guess I'd be surprised if someone like KOC hasn't investigated this, but I don't have much of his material.

Pete, this is about getting the best sound possible at these far off-design operating conditions, and maintaining the relative positioning of the load line seems like the right thing to do.

Doubling the reflected impedance further decreases power output by 3dB(or is it 6dB?). It sounds like a win-win to me.

Actually it would increase the power output since maintaining the original Zpri would be a miss-matched condition, eh?

[norburybrook]

I'm going on holiday next week but when I return I think I'm going to add a VVR to my rocket, so I can report back my findings on this in about 2 weeks.


My friend has a VVR on his express and it sounds and works great, hence my thought's about adding the same to my rocket.



Marcus

[David Root]

I have been thinking about doing a VVR Express with 7591s.

According to the datasheet this tube has a very flat impedance curve vs plate voltage, so presumably making the power output more or less linearly related to plate voltage, while maintaining the optimum impedance if you use a 6K6 OT.

Did I miss anything? Would the VVR itself affect the relationship?

6550s also have a fairly flat impedance curve, so maybe they would work well with a VVR too?

[martin manning]

This effect is just the physics of electron flow from cathode to anode, so it's the same for any tube type. Switching down one click on the impedance selector (doubling Zpri) rematches it when the voltage has been reduced to 25% of normal. In between 100% and 25%, I guess it's whatever sounds better. I'm curious about how moving the bias point at low voltage settings affects the sound too. At 25% voltage, doubling the impedance and running at about 40% of the full-power idle current would result in Class A p-p operation.

[David Root]

Martin, I see that, but I should have been a bit clearer in my post. I would be maintaining the OT secondary at the same setting, eg 8 ohms, not switching it around.
The bias voltage changes though, at least in the 7591 datasheet page. See top of page 2. I thought the VVR does that as well as change the plate voltage. I'm not quite clear on all the inter-relationships I guess.

[martin manning]

Yes, the fixed-bias type VVR brings the bias voltage up (makes it less negative) as the B+ comes down using a dual gang pot. I gather most people are not changing the impedance, but I'm wondering if doing so can improve the tone at very low power settings.

[David Root]

OK, I get it now. So "very low power settings" is probably off the left hand end of that range shown in the datasheet?

[martin manning]

So "very low power settings" is probably off the left hand end of that range shown in the datasheet?

Yes definitely, like going from full-power plate voltage of say 400V down to ~100V.

[andresound]

While changing plate resistors to dale on an express to reduce hiss, I decided to try the impedance mismatch using the VVR. To my surprise, using an 8ohm EVM 12L, I changed to the 4ohm tapping, and the tone definitely improved!! And the feel!!

Thank you again Martin

[andresound]

While changing plate resistors to dale on an express to reduce hiss, I decided to try the impedance mismatch using the VVR. To my surprise, using an 8ohm EVM 12L, I changed to the 4ohm tapping, and the tone definitely improved!! And the feel!!

Thank you again Martin

[martin manning]

So the theory works in practice! Thanks for the report, Andre.