Separate Power Supplies

[xcortes]

I'm finally thinking of rebuilding my Parabees. Since I have a pair of RS080 PTs I thought it would be cool to build separate power supplies for the drivers (I think I'll use the SR/driver stage  boards from the Paramounts that I removed to go to soft start). Are there any benefits or drawbacks from doing this?

Thanks

[Paul Birkeland]

The Parabee is very tightly packed into that chassis.  I would recommend buying a pair of Paramount plates to free up some room, then doing a nice, slow rebuild.  Getting the plate choke up top or well off to the side would be really helpful! 

Providing an isolated high voltage supply for the driver supply would most likely be negated by the effects of the shunt reg. driver setup. 

PJ could no doubt chime in about the values of R1 on each side of the CCS, as well as R2 on the higher current side.  I think regulating to 350V would still be the target. (I suppose you could always use the 12AT7 instead of the 6N1P and leave the board alone)

[xcortes]

Thanks Paul.

I do have a set of Paramounts chassis plates for this project. And I can use the old smaller ones for the driver PS if I go that route.

And yes, I was thinking of using 12AT7s to leave the boards as they are (I also have a bunch of nice 12AT7s).

One benefit of using a separate PT would be that I'll be using the 6v filament taps of the pgp 8.1 since I have the DC filament mods. And that way each tube filaments would have their own transformer.

I just can't believe how much you guys end up negating the benefits of ps improvements due to the SR. I guess SR works that well!

[Paul Joppa]

I have never met a power supply improvement that wasn't an audible improvement.

You are not "stuck" with the 12AT7, so choose the tube you want to hear.

[xcortes]

Ha! I was going to quote you since you've said that before! Separate power supplies then (and big chokes on the BP!).

What good be another alternative to drive the 300Bs cap coupled? How about 437As? The 080s have 20mA including whatever I dump through the regulator.

[xcortes]

What I'll do is build it first with a single ps and the 12at7s. Then I can easily add the scond power supply later. And maybe try a different driver then.

[Paul Joppa]

It's an interesting question, about the maximum practical driver plate current. Here's a calculation I've done but not documented before:

You have 20mA to play with. In drivers I've always done equal reg and driver current, so 10mA is available (you could do more, with less in the reg, if you use C4S driver plate load)

Now you want the driver plate at around 2/3 of the regulated voltage, for good and symmetric headroom - let's say 300v regulated and 200v driver plate. The "beam resistance" is then 200v/10mA or 20K. Typical operating point is Rbeam = 5*rp, so the plate resistance must be 4K or greater. The 6C45 is out.

Obviously you can vary the assumptions, this is a scoping analysis not an exact calculation.

[Atom Shop]

Paul, I'm well aware of the spec differences between the AU, AT and AX. I'm wondering what problems might be encountered, though, if I were to substitute a 12AT7 in place of a 12AX7 in an existing preamp circuit.

Atom

[Paul Joppa]

Paul, I'm well aware of the spec differences between the AU, AT and AX. I'm wondering what problems might be encountered, though, if I were to substitute a 12AT7 in place of a 12AX7 in an existing preamp circuit.

Atom

It really does depend on the particular circuit. I can offer some generalizations or examples though.

Suppose the 12AX7 is a conventional cathode biased, resistor loaded circuit. It will probably be running the 12AX7 at around 1mA - let's say 300v power supply, 100K plate resistor, 200v on the plate, and 1.6K cathode resistor. The AT7 would run at a similar point; I get about 150v on the plate and 1.5mA. The output current capability is improved 50% but the output voltage swing has become asymmetrical with one-sided clipping happening earlier. If it's a preamp, the reduced peak voltage is probably no problem and the performance is similar to the AX7.

But if you look at the plate curves, you will see that this is at the very bottom, in the region where the AT7 distortion is pretty bad. And the plate impedance is much lower, so if it's in an RIAA circuit the equalization will be thrown off.

If instead you ran the AT7 at the higher current it likes - perhaps 7mA at the same 200v on the plate - then it would have 600% more current capability and the same voltage capability as the AX7 did - a much more robust performance, in a way that may or may not be important in the particular application. It would of course need a more capable power supply.

Hope that gives some of the flavor of the various effects.