In the EL84 SE design reccomendations thread, amplifers using 'partial feedback' - feedback from the plate to grid have been discussed. Seeing as other valves than the EL84 are starting to be discussed, I've started this thread to continue the discussion.
Some information about this type of feedback can be found in this article from the TubeCad Journal.
The RH84 and RH807 amplifiers can be found on Alex Kitic's site. The RH34 design is in this post. Of course, many more designs are in the works...
Jason
Some information about this type of feedback can be found in this article from the TubeCad Journal.
The RH84 and RH807 amplifiers can be found on Alex Kitic's site. The RH34 design is in this post. Of course, many more designs are in the works...
Jason
Have you guys seen this 300B design by John Broskie and Steve Bench. Partial Feedback used in this design.
http://www.enjoythemusic.com/magazine/equipment/0105/trex.htm
http://www.enjoythemusic.com/magazine/equipment/0105/trex.htm
Dave,
In AC terms, the plate of the driver valve is basically the same point as the grid of the output valve since they are coupled with a capacitor (which is an effective short to AC), and a gridstopper (which is a small resistance against the effectively infinite grid input resistance).
That was what I was thinking, but it was actually sloppy terminology on my part.
It is more correct to say plate to plate.
Jason
In AC terms, the plate of the driver valve is basically the same point as the grid of the output valve since they are coupled with a capacitor (which is an effective short to AC), and a gridstopper (which is a small resistance against the effectively infinite grid input resistance).
That was what I was thinking, but it was actually sloppy terminology on my part.
It is more correct to say plate to plate.
Jason
PLATE TO PLATE!
Hi to all... I see "all the usual suspects" are already here!
Well, first of all, the thread is named incorrectly -- it should be plate-plate feedback amps. But, since "partial feedback" is a better known name, I still think it is "politically correct". More people will get interested
Let me now mention the T-Rex amp. My attention was drawn to this one a few days ago -- therefore I did a quick read and download of everything related. First of all, that is a partial feedback amp -- unlike RH amps which are plate to plate. Secondly, between RH amps and the T-Rex there is a huge difference in the type of feedback (if I am not mistaking, beside the fact that in the T-Rex feedback is only in AC, I think it is also "voltage feedback" -- maybe I am wrong, I did not consider that one too long).
Beside that, there is another personal thing separating T-Rex and my designs -- simplicity. I tend to make them as simple as possible, so that eventual "necessary" wisdom is "secluded" at first glance (so what, is anyone going to kill me for saying it). The T-Rex, interesting as it is, is too complicated for my taste (not that I do not understand it and therefore dismiss it!!!). I do not think regulation is necessary (especially shunt) and I dislike the large first cap in the PS. Further, the "driver stage" is too elaborate for its own good (no particular drive needs there) and the feedback loop (partial feedback loop, with the cap) looks to me as something that is out of place... which was put there just to make it more interesting. Maybe I'm wrong?
Further news... I think the RH6550 (but could be named RH88 as well...) is now "finished" in the sense that I have tested everything possible in the PS department (schematics has remained unchanged) and I definitely think it lives up to expectations: the promise of approx. 10W/1% distortion, "RH sound" (yes, the bassy one) etc.
Regards to all...
Aleksandar
Hi to all... I see "all the usual suspects" are already here!
Well, first of all, the thread is named incorrectly -- it should be plate-plate feedback amps. But, since "partial feedback" is a better known name, I still think it is "politically correct". More people will get interested
Let me now mention the T-Rex amp. My attention was drawn to this one a few days ago -- therefore I did a quick read and download of everything related. First of all, that is a partial feedback amp -- unlike RH amps which are plate to plate. Secondly, between RH amps and the T-Rex there is a huge difference in the type of feedback (if I am not mistaking, beside the fact that in the T-Rex feedback is only in AC, I think it is also "voltage feedback" -- maybe I am wrong, I did not consider that one too long).
Beside that, there is another personal thing separating T-Rex and my designs -- simplicity. I tend to make them as simple as possible, so that eventual "necessary" wisdom is "secluded" at first glance (so what, is anyone going to kill me for saying it). The T-Rex, interesting as it is, is too complicated for my taste (not that I do not understand it and therefore dismiss it!!!). I do not think regulation is necessary (especially shunt) and I dislike the large first cap in the PS. Further, the "driver stage" is too elaborate for its own good (no particular drive needs there) and the feedback loop (partial feedback loop, with the cap) looks to me as something that is out of place... which was put there just to make it more interesting. Maybe I'm wrong?
Further news... I think the RH6550 (but could be named RH88 as well...) is now "finished" in the sense that I have tested everything possible in the PS department (schematics has remained unchanged) and I definitely think it lives up to expectations: the promise of approx. 10W/1% distortion, "RH sound" (yes, the bassy one) etc.
Regards to all...
Aleksandar
Re: PLATE TO PLATE!
That's exactly what I was thinking, Plate-to-plate feedback sounds a little boring compared to "partial feedback"
On second thoughts, I should have called the thread "Inverse Feedback Pair Amplifiers" Much more interesting.
It is a little complicated for my liking as well. I like two-stage amps for that reason. I do not pretend to understand circuits as well as John Broskie or Steve Bench, but I too think the driver stage (a cascode, cascaded into SRPP) is a little overcomplicated.
In any case, one design choice I don't completely understand is that the I/V output stage is driven by the low-ish output impedance of the SRPP (about 300 ohms). As Alex has noted in his designs, to achieve near 100% local feedback around the output stage, you need to drive it with a high output impedance (hence the driver valves as triodes unbypassed cathode resistors, or pentodes). I think the design was never intended to achieve a large amount of local feedback around the output stage, which makes the partial feedback loop and the claim of a 'virtual cathode follower' output stage seem a little out of place.
Name it the "RH88".... More people will be interested in it that way
Jason
Alex Kitic said:Well, first of all, the thread is named incorrectly -- it should be plate-plate feedback amps. But, since "partial feedback" is a better known name, I still think it is "politically correct". More people will get interested
That's exactly what I was thinking, Plate-to-plate feedback sounds a little boring compared to "partial feedback"
On second thoughts, I should have called the thread "Inverse Feedback Pair Amplifiers"
Much more interesting.Alex Kitic said:
It is a little complicated for my liking as well. I like two-stage amps for that reason. I do not pretend to understand circuits as well as John Broskie or Steve Bench, but I too think the driver stage (a cascode, cascaded into SRPP) is a little overcomplicated.
In any case, one design choice I don't completely understand is that the I/V output stage is driven by the low-ish output impedance of the SRPP (about 300 ohms). As Alex has noted in his designs, to achieve near 100% local feedback around the output stage, you need to drive it with a high output impedance (hence the driver valves as triodes unbypassed cathode resistors, or pentodes). I think the design was never intended to achieve a large amount of local feedback around the output stage, which makes the partial feedback loop and the claim of a 'virtual cathode follower' output stage seem a little out of place.
Alex Kitic said:
Name it the "RH88".... More people will be interested in it that way
Jason
Why SRPP driver...
The reason is that the author of the circuit (which I suspect to be John Broskie) -- or at least of that part -- likes SRPP a lot when used properly, i.e. as a driver into a known impedance. But, in this case the "modus operandi of the amplifier" is different, as I already wrote, in the fact that the driver does not connect to the anode of the amp except in AC: this means that it does not function as a current source, therefore the highish impedance is not at all desirable!
When we are at that, JB likes cascodes as well... and since the other participants in the project might have asked of him, he has inserted as well the partial feedback loop Well, besides this joke, I think that the feedback loop in this case might have something to do with having a fixed and permanent input impedance of the output tube, so the SRPP can work at its best. Most probably... I think that I have already mentioned JB too many times, more than necessary... creating a "liable case"
Did you get my mail?
Regards to all,
Aleksandar
The reason is that the author of the circuit (which I suspect to be John Broskie) -- or at least of that part -- likes SRPP a lot when used properly, i.e. as a driver into a known impedance. But, in this case the "modus operandi of the amplifier" is different, as I already wrote, in the fact that the driver does not connect to the anode of the amp except in AC: this means that it does not function as a current source, therefore the highish impedance is not at all desirable!
When we are at that, JB likes cascodes as well... and since the other participants in the project might have asked of him, he has inserted as well the partial feedback loop
Well, besides this joke, I think that the feedback loop in this case might have something to do with having a fixed and permanent input impedance of the output tube, so the SRPP can work at its best. Most probably... I think that I have already mentioned JB too many times, more than necessary... creating a "liable case" Did you get my mail?
Regards to all,
Aleksandar
mebbe just one thought from me-
I don't like SRPP ,no matter what sort of is in question .
my first preamp was SRPP based (looong time ago when I was in short pants ) (Anzai preamp from Elektor) .
with that preamp,fortunately ,I made every possible step of improvement and it suffer also from every possible trouble one preamp can suffer.
just to mention whistle ,hehe........
so ,in case that some stage need those flawors SRPP can bring, I use another tube in more "common" arrangement.
Alex's I/V conversion is sort of novelty these days ,even if is only "boiled water" ; in that case looong time ago forgoten boiled water.
same as WOT preamps in this days-nobody remember ole plain topology till recently ,but everyone who listen to it is just blown away with bloody imediacy which only energy preservation of signal can bring.
in that case (or in that light ) for me almost every SRPP is passe ;
hard to preserve energy ,even in first stage of phono preamp.
I don't like SRPP ,no matter what sort of is in question .
my first preamp was SRPP based (looong time ago when I was in short pants
) (Anzai preamp from Elektor) .with that preamp,fortunately ,I made every possible step of improvement and it suffer also from every possible trouble one preamp can suffer.
just to mention whistle ,hehe........
so ,in case that some stage need those flawors SRPP can bring, I use another tube in more "common" arrangement.
Alex's I/V conversion is sort of novelty these days ,even if is only "boiled water" ; in that case looong time ago forgoten boiled water.
same as WOT preamps in this days-nobody remember ole plain topology till recently ,but everyone who listen to it is just blown away with bloody imediacy which only energy preservation of signal can bring.
in that case (or in that light ) for me almost every SRPP is passe ;
hard to preserve energy ,even in first stage of phono preamp.
Hi everyone,
Just found this thread about local feedback (my prefered name for this kind of feedback). Thought I'd throw in another design in for comment that uses local feedback and is DC coupled.
I've been working on this local feedback design for a couple of years now and have so far built 2 amplifiers using DC coupled local feedback. The basic design uses pentodes for both the input and output stages. The input stage uses 6AU6's. In the output stage I have used 47's in a low power version and 1624's in a higher power version. Both the 47's and the 1624's are directly heated. The 47's are DHP's and the 1624's are DHT's (Directly Heated Tetrodes)
The 47 version puts out about 4 watts, the 1624 version is almost 10 watts. The local feedback does a good job of controlling the output impedance of the amplifier. The output impedance of the 1624 amp is 1 ohm on the 4 ohm tap, 1.75 ohms on the 8 ohm tap and 3.5 ohms on the 16 ohm tap.
Sound is super fast and transparent, not at all like a "normal" williamson type pentode amp. Several folks on the forum got to listen to the 1624 amp at the ETF this year.
Here is the schematic of the latest 1624 version.
To see the development story of this design (so far but not including the 1624 amp) check out the 47 P-P link on my web page.
Gary P's DIY Page
Gary
Just found this thread about local feedback (my prefered name for this kind of feedback). Thought I'd throw in another design in for comment that uses local feedback and is DC coupled.
I've been working on this local feedback design for a couple of years now and have so far built 2 amplifiers using DC coupled local feedback. The basic design uses pentodes for both the input and output stages. The input stage uses 6AU6's. In the output stage I have used 47's in a low power version and 1624's in a higher power version. Both the 47's and the 1624's are directly heated. The 47's are DHP's and the 1624's are DHT's (Directly Heated Tetrodes
)The 47 version puts out about 4 watts, the 1624 version is almost 10 watts. The local feedback does a good job of controlling the output impedance of the amplifier. The output impedance of the 1624 amp is 1 ohm on the 4 ohm tap, 1.75 ohms on the 8 ohm tap and 3.5 ohms on the 16 ohm tap.
Sound is super fast and transparent, not at all like a "normal" williamson type pentode amp. Several folks on the forum got to listen to the 1624 amp at the ETF this year.
Here is the schematic of the latest 1624 version.
To see the development story of this design (so far but not including the 1624 amp) check out the 47 P-P link on my web page.
Gary P's DIY Page
Gary
Gary, I like your elegant feedback implementation. In my own little world, I many years ago consigned feedback to the bin of fidelity-detracting ideas to perhaps be researched at some later time with perhaps historical interest only. Ahem. Putting a small question mark to the assumption I'd adopted, I recently read certain of Johathan Carr's posts where he discusses using as much as 50dB feedback without, in his opinion, sonic detriment. From this inauspacious start, I have thus, and only recently, began investigating feedback possibilities, beginning by looking at certain fundamentals why feedback might detract sonically (it's advantages, well known, need no mention). My tentative observation is that feedback taken over, across or through one or more phase shifting passive components---worst offenders being capacitors and inductors/transformers---must by some definition meaningful to me be worse than feedback taken across non-phase shifting components, errr, componenT, being a resistor. A slightly larger version illustrating this view is that if the 5KHz part of a given transient is fed to a grid with 14 degrees phase lag and the 8KHz part with 22 degrees, and if part of the amplified version of that "transient" emerging in amplified form from the tube in question is fed back to the input of the previous tube, there'll be trouble in Dodge, as they say (what is the sound of a 5KHz etc transient added to the same transient somewhat out of phase?). Low, direct-(resistive-)coupled-phase-shift feedback, if you will, doesn't seem so troublesome as the only phase difficulty involved is the small shift---linear across a wide band of frequencies---created by the amplifying device or devices in question (ie, the phase shift, relative to feedback input node, created by the time it takes for a tube to amplify a signal).
I just installed the partial feedback resistor in my DC 2a3 amp after hearing about it from BTW. Needed to re-adjust the operating points on the driver and output tubes else amp goes into severe clipping - not to nice. once up and running it sounds much cleaner with far less bass overhang. i agree with Gary P's observation in that the amp sound much faster as well.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.