parallel SRPP in a Pultec...

[lolo-m]

Looking at the GPULTEC schematic I wondered what was the output impedance... It doesn't appear to me to be a 600 ohm standard output  ???.

What is exactly its output impedance ?

I need a maximum 1K output impedance so if it is about 2K, I thought that this design should solve the problem... but it's maybe completely stupid  ;D

Should this work ?

Am I right to think this is dividing by 2 the SRPP impedance ?

 

[NewYorkDave]

If you need low output impedance with good drive capability, you'd be better off with a White cathode follower.

 

[tk@halmi]

Try searching for Aikido amplifier on the web. It will get you many hits on ready made and tested examples of gain + drive.

 

[lolo-m]

Thank you both !!!  ;D ;D ;D
I will try other designs...

But who can tell me if the above design will work or not and if I'm right to think that parralleling 2 tubes will divide their impedance by 2 ?

Another question is : does anybody tryed one of the software there :
http://www.glass-ware.com/GlassWare_Programs.htm

 

[PRR]

> Am I right to think this is dividing by 2 the SRPP impedance?

Yes, this much is clear.

After that things are murky.

I should start by saying I think the SRPP is a silly folly. There is almost some other way to do it, and usually better.

EDIT: I just poked at this for an hour, and I think it was a waste. SRPP really sucks in low-Z loads. And I hit a point I was reminded of Broskie's analysis: there is a crossover impedance such that SRPP has different problems above or below that impedance.

Use ONE half ECC99 with a 15K plate resistor. Use 250V supply, bias so the plate sits near 100V (10mA). Probaby a 200R resistor. Bypass it. Zout will be very close to 3K, gain near 25. After your 2:1 iron this is gain of 12 and Zout of 750. Current consumption will be 20% higher, tube consumption will be half. Assuming your external load is really much higher than 1K, and peak levels more like 5V than 50V, linearity will be fine.

If you must throw bottles: half 12AX7 plus a whole ECC99 in WCF will give gain near 50 (25 after the iron) and Zout dominated by winding loss. If you get the WCF well balanced, linearity will be excellent down to VERY low loads, dominated by 12AX7 volt-amp which tends to be good-enough for a lot of things.

 

[NewYorkDave]

I should start by saying I think the SRPP is a silly folly. There is almost some other way to do it, and usually better.

Agreed. I don't know if this is still the case, but it was a really trendy circuit for a while. I always considered it a "cute" but not most-effective way to use a dual triode.

I suppose one reason for its popularity among audiophile DIY types is that in their application, it's almost always running into a fairly high Z load.

 

[jensenmann]

I always thought it´s reputation in HiFi application came from audiophool´s dislike of negative feedback.

 

[lolo-m]

SRPP really sucks in low-Z loads. And I hit a point I was reminded of Broskie's analysis: there is a crossover impedance such that SRPP has different problems above or below that impedance.

I'm just trying to understand things (and I've got a lot to understand  ;D ;D ;D!!!)... This crossover impedance you're talking about is the SRPP impedance or load impedance ?

 

[tk@halmi]

I'm just trying to understand things (and I've got a lot to understand  ;D ;D ;D!!!)... This crossover impedance you're talking about is the SRPP impedance or load impedance ?

The way Broski describes it: a particular SRPP design is optimized for a specific load impedance. Loading the SRPP with a higher impedance causes one set of problems, and loading with a lower impedance causes a different set of problems. If you browse his site long enough you will find detailed explanation and even the mathematical formulas for designing an SRPP stage.