VVR question

[Curbdog]

I am using a VVR2 to power scale my amp's push-pull 7591A power section and cathodyne phase inverter which is a 12AU-7 triode. Power to the rest of the amp's preamp circuit is unregulated. I implemented a master volume just before the cathodyne. I have found that even with a very large dropping resistor, like 150K, off the power tube screen supply feeding the cathodyne plate, the plate voltage on the cathodyne is about twice as high as it was before I implemented the VVR this way (300v vs 150v). In fact, all the premap voltages everywhere are now higher, and the amp is sounding worse (harsher treble tone) than when I started. What is going on here? This has proved to be tricky. I am contemplating going back to regulating the whole amp. Any other suggestions? Thank you.

[ampgeek]

Do you see the higher voltages even with the VVR set at it's highest setting?

I can see how this could occur at lower settings (eg lower voltages-->lower current draws = smaller voltage drops across the de-coupling resistors in the B+ string--->higher voltages at tube plates).

I have only implemented full amp VVRs so limited experience with what you are doing. However, my rudimentary understanding is that PA VVR applications work best with regulating only the power tube plates and screens and the inclusion of a PPIMV.

But...again....I have no practical experience to share on that one.

Cheers,
Dave O.

[Curbdog]

Yes, I do see that. I thought I would try regulating the PI and power tubes because it was convenient for this amp in regard to implementing the master volume and locating that POT on the panel, and because I wanted to try out the idea as described in Merlin Blencowe's Power Supplies book. I will try the traditional PPIMV next and see how that sounds. Thank you.

[Kagliostro]

In order to investigate about what is happening can you post a schematic of how you have done your VVR mod ?

Kagliostro

[Curbdog]

Here is my schematic - thank you

[Tillydog]

As standard, I guess node D would be fed from node C with the 5k6 resistor - is that correct?

A->(470R)->B->(100k)->C->(5k6)->D->(1k)->E->(1k)->F

You now have B+ nodes:

A ->(470R) ->B ->(100k) ->C fed from the VVR and:

D ->(1k)-> E ->(1k) ->F fed from the resevior cap via 5k6 to node D

In this case you need to increase the resistor between B & C to compensate for the reduced current through it because it no longer supplies nodes D-F, and increase the resistor feeding node D because it is now fed from the full B+, rather than the node C voltage.

You should be able to make measurements in the 'stock' configuration to determine actual current draw and then calculate the correct resistor values.

As a rough estimate:

D, E & F supply 8 preamp valves in total, so maybe 10ma current in total.

C supplies the PI, so maybe 1ma (probably quite a bit less!)

therefore you need to increase the resistor between B and C by ~10x the stock value (which I guess is nearer 22k than 100k) to get the same voltage drop.

You have said that the voltage at node C has gone up by 150V, so the resitor feeding node D from B+ needs to drop ~150v + 5k6 x 10ma = ~200V total.

If the preamp current estimate were correct, that would mean 200V / 10mA = 20k between B+ and D

Andy

[Curbdog]

Thanks Dog !

I made the changes to the power supply that you recommended and the anode voltages are now mostly in line (thanks for the Ohms law refresher)

I have noticed some other challenges with implementing power scaling of only the power tubes and PI with this amp.

The negative feedback is going to be affected, but more concerning is that I have noticed that the voltages of pins 7 and 8 of the PI are out of whack. I am expecting about 86 volts on pin 7 and about 105 volts on pin 8 of this 12AU-7 triode. I see 17 volts on pin 7 and 30 volts on pin 8. What could be causing this?

[Tillydog]

No worries! Glad you sorted it

....I have noticed that the voltages of pins 7 and 8 of the PI are out of whack. I am expecting about 86 volts on pin 7 and about 105 volts on pin 8 of this 12AU-7 triode. I see 17 volts on pin 7 and 30 volts on pin 8. What could be causing this?

Not sure if it is a typo on pin numbers:

Pin 7 is grid & pin 8 is cathode, so I would expect to see pin 8 a few volts higher than pin 7. I'm guessing you mean pins 6 & 8.

Looking over your schematic again, I think the PI has problems:

You have referenced the grid to ground with the master volume - (I've noticed that people do this a lot ! ). It should be referenced to the bottom of the 2k2 bias resistor in the tail.

You need to put a coupling cap between the PI grid and the wiper on the 'Master Volume' pot to block that path for DC and you need to put a grid leak resistor between the PI grid and the bottom of the 2k2 bias resistor to create a DC path to hold the grid at the correct DC bias voltage.

Exactly as the arrangement on the RIGHT of this picture from Merlin Blencowe's page on cathodyne phase inverters which is well worth a read if you haven't already - especially with regard to the grid stopper (likely you want to keep it.):

[img:313:321]http://www.freewebs.com/valvewizard/Cathodyne2.jpg[/img]

Make sense?

Andy

[Tillydog]

So while I'm thinking about it.....

I have noticed some other challenges with implementing power scaling of only the power tubes and PI with this amp.

The negative feedback is going to be affected....

You could move the MV to one stage earlier (before the first 12AU7) and add node 'E' to the VVR, so the whole of the NFB loop is fed by the VVR.

You could then DC couple the PI without any need for a new coupling cap or grid leak - just connect the 470k direct to the anode/plate of the previous stage.. If you do that, the 2k2 resistor in the tail serves no real purpose (AFAICT), and could be removed (if it isn't there for MOJO!).

If you are considering this then check on the max heater-> cathode voltage for the PI (read up on DC coupled cathode followers on Merlin's site; there are ways of dealing with it).

HTH

Andy

[Curbdog]

OK, I've moved the MV to one stage earlier (before the first 12AU7) and so the whole of the NFB loop is now fed by the VVR. Sounds better already! In this case, is the MV referenced to ground?

[Tillydog]

Yup, no need for a coupling cap because the grid of the first half of the 12AU7 is referenced to ground.

Andy