Parallel feed (choke load) in output stage

[yagoolar]

Hi,

Just want to listen to your impressions or remarks on choke load parallel feed in output stage of an SET amp. Any experience ?

[rcavictim]

That setup can work extremely well! The choke also benefits you by getting rid of more power supply ripple at the same time while getting the DC issue out of the output xfmer. You should also experiment with placing the signal return capacitor from the output transformer directly to the output tube cathode rather than chassis ground.

[yagoolar]

Hi rcavictim,

Originally posted by rcavictim
You should also experiment with placing the signal return capacitor from the output transformer directly to the output tube cathode rather than chassis ground.

Which is ? Secondary winding ? I think I don't get that setup. Can you point me a schematic ?

[pedroskova]

For parallel feed, you can replace the choke with a ccs like Gary Pimm's which can greatly simplify PSU design.

Re: cap to cathode, Mr.Pimm also discusses this somewhere on his site.

[audiousername]

Originally posted by yagoolar
Which is ? Secondary winding ? I think I don't get that setup. Can you point me a schematic ?

I'm not rcavictim but I think he means something like this

[rcavictim]

Originally posted by audiousername



I'm not rcavictim but I think he means something like this

Audiousername,

Yes that is exactly the circuit that I was referring to!

[yagoolar]

Hi,

I once found a pdf with such a schematic. It was published on bottlehead web page. The amp is called Lizzy (by Anthony Dockrill). I am curious what you think about the output stage and xformer connection. Has anyone tried it ?

[rcavictim]

Originally posted by yagoolar
Hi,

I once found a pdf with such a schematic. It was published on bottlehead web page. The amp is called Lizzy (by Anthony Dockrill). I am curious what you think about the output stage and xformer connection. Has anyone tried it ?

The circuit you showed has the 6L6 output tube wired as a cathode follower in parafeed mode. Since the output tube now has less than unity gain an extra (noise generating) gain stage must be added to what could be a simpler total amplifier circuit. This is very uncommon, at least from my sheltered experience. This would allow the use of a rather low primary Z output xfmer, but he has placed a rather high Z transformer of 3K here. I am a bit baffled by this. I see he places the speaker winding in series with the primary to turn the OPT into a autoxfmer. This further increases the Z of the output transformer albeit by only a small fraction. This is fine since there is no DC and must increase fidelity a bit. IMO the 1 uF output coupling cap is way undersized.
Where do you guys find this stuff?

[audiousername]

Originally posted by rcavictim

The circuit you showed has the 6L6 output tube wired as a cathode follower in parafeed mode. Since the output tube now has less than unity gain an extra (noise generating) gain stage must be added to what could be a simpler total amplifier circuit. This is very uncommon, at least from my sheltered experience. This would allow the use of a rather low primary Z output xfmer, but he has placed a rather high Z transformer of 3K here. I am a bit baffled by this.

Looking at the schematic, it appears that the designer had 10Y valves around, wanted to use them but also required low output impedance, which the 10Y couldn't provide because of its high-ish rp. So, the 6L6 is used in CF to prodvide low output Z.

So the question is... why wasn't a lower rp valve used to start off with? I guess it's because of the lure of old, rare and EXPENSIVE valves... argh... an ordinary $6 12E1 in triode may very well have performed better.... (and given more power) But I could be wrong, wouldn't be the first time...

Rant mode over...

Yagoolar, that is not the typical shunt-feed (parafeed) output stage. Check this out: http://members.aol.com/sbench/outstru.html It's an article by Steve Bench outlining the most common (and some of the less common) output structures for valve amplifiers.

[ReinoutdV]

Good morning all,

i do fully agree with rcavictim's remark about the 1 uF value; way to low !

The proposed schematic is a nice combination of ultrapath (Jack Elliano, Electraprint, etc) and parallel feed. This combination is sometimes referred to as "ultra parafeed". It is certainly not new, Western Electric used it already in 1938 in their 92B-amplifier.

Ultra parafeed does have some nice features:
- the coil will take care of DC-component. So you can optimize the OPT with more exotic cores. Or use a smaller winding to decrease the resistance. Either way......the OPT can be better (and there is no air gap as well).
- the extra cap does seem to smooth out a lot of power supply irregularities.

Of course there are trade-offs no surprise:
- the coil is not a standard choke !
- there will be an extra cap between the OPT and the cathode. This must be a quality item ! I do favour a nice oil cap here......but that is a matter of taste.

Lynn Olson wrote in VTV (issue 16) a very informative article about these topologies under the name "Ultrapath, parallel feed and Western Electric". This very article made me base my 833 amplifier on the proposed topology "ultra parafeed". And it does work superbly !

Included a picture of the inside of the 833 amplifier. In the left upper corner you'll see the cap. A 4 uF / 4 kV affair....i did not take any changes with the anode voltage of 1400 V....just to be on the safe side.

Reinout