Lets just repeat and highlight the important bit:
How does this manifest, as relatively poor power output before unacceptable distortion - but more importantly the variable feedback creates a nasty higher harmonic profile. In average use this is easy to ignore since the amp is operating at subwatt power levels - but as power rises the higher order harmonic rise and create a nasty edge to the sound.
Up to the fourth harmonic the RH designs perform almost identically to a pentode driven version.
Shoog
How does this manifest, as relatively poor power output before unacceptable distortion - but more importantly the variable feedback creates a nasty higher harmonic profile. In average use this is easy to ignore since the amp is operating at subwatt power levels - but as power rises the higher order harmonic rise and create a nasty edge to the sound.
Up to the fourth harmonic the RH designs perform almost identically to a pentode driven version.
Shoog
Looks like my objections to the name have been voiced before.
O.H. Schade's paper used series-applied voltage feedback, which I think is my preferred feedback scheme since it preserves the tube's high input impedance. I find it much easier to accomplish linear voltage amplification with tubes than linear V-to-I conversion with tubes, which is what you need with parallel-applied feedback schemes.
Schade used a transformer to apply the feedback and then he varied the amount of feedback by changing some resistances in his circuit. It was a novel circuit but you would never build an amp that way. I mean, Schade's goal was to write a white paper illustrating the effects of various feedback ratios on a 6L6, not to build an amp. It can be done much more simply and elegantly and in fact it has.
A very simple way to apply the same type of feedback in a practical amp is with an output transformer with a cathode feedback winding. It is the exact equivalent of the feedback in that paper. The only catch is that you have to know how much feedback you want before you buy or make your output transformer. Usually there are no simple circuit tweaks to play with the feedback ratio later.
It really isn't that difficult to add a LTP pentode front end to a PP circuit. A simple dropping resistor to the two driver pentodes screens makes an adequate g2 supply because the current draw when averaged over the two pentodes is almost constant. The added complexity is tiny.
However the draw back is that there are very few small signal pentodes capable of passing the required currents of 6-10mA.
The BabyHuey is another good solution to the problems and uses a triode driver (but it never claims to be a low distortion design).
Shoog
However the draw back is that there are very few small signal pentodes capable of passing the required currents of 6-10mA.
The BabyHuey is another good solution to the problems and uses a triode driver (but it never claims to be a low distortion design).
Shoog
Last edited:
I want to add 0.02.
My ongoing project (see the SHIII topic) is quite what you want, and I tried pentode mode in a balanced input stage. The first impression was very good, but it turned out be somewhat sensitive in regard of matched pair. It works not bad on unmatched tubes, but after a lot of trying I returned back to Mona's proposed CCS with triode mode, but with... added +6V bias on the 3rd grid, sounds very sweet for what ever silly it looks. I will post the final schematics later.
I also wanna note that the value of Schade FB resistors seems be unique for each schematics. Somehow I guessed it right on with 220K for mine, and a lot of trying/comparing within the range of 100-470K kicked me back to the initial 220K.
I came to the conclusion that value depends a lot on the plate load of both stages.
My ongoing project (see the SHIII topic) is quite what you want, and I tried pentode mode in a balanced input stage. The first impression was very good, but it turned out be somewhat sensitive in regard of matched pair. It works not bad on unmatched tubes, but after a lot of trying I returned back to Mona's proposed CCS with triode mode, but with... added +6V bias on the 3rd grid, sounds very sweet for what ever silly it looks. I will post the final schematics later.
I also wanna note that the value of Schade FB resistors seems be unique for each schematics. Somehow I guessed it right on with 220K for mine, and a lot of trying/comparing within the range of 100-470K kicked me back to the initial 220K.
I came to the conclusion that value depends a lot on the plate load of both stages.
I found this a very good read. I like the way he answered questions I was about to ask. He uses assumptions I use. He also respects proven designs. His conclusions not what I exspected and nearly where we are.
http://www.next-tube.com/articles/Veen2/Veen2EN.pdf
http://www.next-tube.com/articles/Veen2/Veen2EN.pdf
UL should be considered as UL option is easy to buy. I would love the cathode version included. Here is an in depth report with sensible warnings. This was about the time of UL being a selling feature. I suspect you might buy UL transformers at the time to upgrade a pentode amp. Looking at the Quad valve amp it might make a version of Super Pentode, the triode option. That is to use both cathode feedback and triode. That doubless would ask too much of the EF86's. Perhaps a LTP of ECC82 cascodes could do the same job in the same pin space. That's phase splitter amplifier with upto 10 mA. One could use double EF184. 12 very nice watts.
http://www.oestex.com/tubes/1955 Radiotronics.pdf
http://www.oestex.com/tubes/1955 Radiotronics.pdf
Please don't tell me what's wrong with this because I possibly already know. It's my own notes to myself to say this was an idea I tried. I show it because it was the point I departed with the Kitic design. I made the variant where the feedback was split into the cathode also. This version as a PYE Mozart inspired idea was better than I hoped. It is a bit of luck it could be this simple. I never tried the UL option. Hopefully my notes are correct.One can use 50 or 60 Hz and a cheap test meter to find ideal phase for feedback. There was a grid stopper into KT88 not shown.
Using this as a possible starting point all options seem to be available. It is mostly convention we call the transformer secondary that . The way they are wound makes it little different to a special third winding. 2.7 mA of current in the transformer I doubt is a concern. The only reason I show this is it was the only loop feedback vesrion that seemed to keep the harmonics in the order of a non feedback amp. Most look a bit ugly with loop feedback.
The idea that comes to mind is use my post No 56 combined with this and 4 x EL 84 in UL. We are looking at 20 to 30 watts and 0.7 % THD at the lower power without loop feedback ( purple ) . I suspect a UL transformer for EL 34 would be OK. In SE driven PP we don't need much drive so win plenty if SE driven. EL 84 is a very nice sounding device in UL ( Heath and Leak amps are supurb ).
If my hunch is right we can do things which should need custom components with off the shelf this way. This seems to be about 12 dB of feedback. I would not recomend this as a solution. It was just my way of seeing where it went. The ECC 81 looks a bit ugly at the 5 th harmonic I seem to remember.
BTW. I read somewhere a guy say the damping factor of UL was not like that of pentode or triode. It was less affected by the speaker load even without loop feedback. As I understood it part of UL was making the best possible matching between device and transformer and the happy linearity compromise between pentode and triode as a bonus. The idea that it could ignore the load better is not obvious.
Never make the mistake I did. The nasal sound of UL seems to be the loop feedback when added. Doubtless some ringing which if hunted down can be cured. Then you must look at the harmonics and ask is that as good as it was without feedback albeit it was higher. If you look at No 56 it is not too bad. It looks like a SE amp. The best SE amps using perhaps 211 are very special. I doubt No 56 would be far behind ( SE driven PP ). My No 88 in final form sounded better than a commersial 300B amp and that was with perhaps 8 types of 300B tried. Sorry to say 300B is not my cup of tea mostly. The Kron type was. He told me he could not sell it as it was not to the original design. Sounded great to me.
Also. I don't hold with that valves have 2nd harmonic and we like that. I for one don't. I prefer the Hiraga ideal of the harmonics decaying in the fashion of how our ears corrupt the sound. It is the shape of the inner ear alone that dictates this. It is not a mystery. I was told by Oxford University that the ear has up to 30 % distortion. I helped them with parts to do the testing. If the source is zero distortion or in the same style up to about 3 % we hear it as perfect. Usually this asks the amplifer to be less than 1%. I always get annoyed with people who sort of suggest making what the ear likes is wrong. It might just be in the true maths they are wrong. As Hiraga says it is relative difference of harmonics that counts. Some imply they have the hearing and training that makes them perfectionists. I bet you anything that's wrong if blind tested. My BBC friend Janine Elliot doesn't claim this ( her degree is music and sound engineer ). She has a Krell and a Leak EL 84 amp and we both think the Leak is better. It sounds less muffled. This is using BBC LS5/9 ( Graham Audio ) which might not be exspected to like valve gear. Her fun amp is Graham Slee. I really like it. Not unlike a Lab47. Sorry to say the Krell is not my cup of tea. Remember the Leak has tons of feedback so is a technical amp also.
Never make the mistake I did. The nasal sound of UL seems to be the loop feedback when added. Doubtless some ringing which if hunted down can be cured. Then you must look at the harmonics and ask is that as good as it was without feedback albeit it was higher. If you look at No 56 it is not too bad. It looks like a SE amp. The best SE amps using perhaps 211 are very special. I doubt No 56 would be far behind ( SE driven PP ). My No 88 in final form sounded better than a commersial 300B amp and that was with perhaps 8 types of 300B tried. Sorry to say 300B is not my cup of tea mostly. The Kron type was. He told me he could not sell it as it was not to the original design. Sounded great to me.
Also. I don't hold with that valves have 2nd harmonic and we like that. I for one don't. I prefer the Hiraga ideal of the harmonics decaying in the fashion of how our ears corrupt the sound. It is the shape of the inner ear alone that dictates this. It is not a mystery. I was told by Oxford University that the ear has up to 30 % distortion. I helped them with parts to do the testing. If the source is zero distortion or in the same style up to about 3 % we hear it as perfect. Usually this asks the amplifer to be less than 1%. I always get annoyed with people who sort of suggest making what the ear likes is wrong. It might just be in the true maths they are wrong. As Hiraga says it is relative difference of harmonics that counts. Some imply they have the hearing and training that makes them perfectionists. I bet you anything that's wrong if blind tested. My BBC friend Janine Elliot doesn't claim this ( her degree is music and sound engineer ). She has a Krell and a Leak EL 84 amp and we both think the Leak is better. It sounds less muffled. This is using BBC LS5/9 ( Graham Audio ) which might not be exspected to like valve gear. Her fun amp is Graham Slee. I really like it. Not unlike a Lab47. Sorry to say the Krell is not my cup of tea. Remember the Leak has tons of feedback so is a technical amp also.
I found this a good read also.
http://www.asinus.dk/Asinus_Elektroakustik/Home_files/cb-amplifier_8wse.pdf
http://www.asinus.dk/Asinus_Elektroakustik/Home_files/cb-amplifier_8wse.pdf
This may be off the topic a bit? But if the two output tubes are 180 degrees out of phase couldn't you apply feedback between the two plates of the outputs? Or between the plates and the opposing cathodes? If the preamp tube is a triode it will have its own internal feedback. And if the feedback is very local, (between the outputs) wouldn't there be the minimum of time/phase smearing?
I think what this says is very local feedback is ideal. It is tempting to think we could use one speaker output to do the job. I noted on the SE it might even improve the design I have. That is I would slightly reduce the DC transformer bias if the output winding is in antiphase. My amp gets a bit ugly at 16 Hz although output is good. I suspect it would change that and not because of negative feedback. I have a 16 ohm winding that could be used. Thus I could present 16 Vrms at the cathode. I can have 8 or 4 V rms also. The maker ( Danbury ) says 16 ohms the better winding for any feedback use. Might include this type.
My big surprise is how EL84 is transformed by UL feedback. It should need only 16 Vrms to be driven in SE to long tail pair output ( whatever we call that ). What would be ideal is that an ECC83 could drive two sets of EL84. In paralell it might. If only needing 16 Vrms I dare say I could lower the anode voltage and up the current. I would aim for 500 mV input for 20 watts rms. ECC83 in ECC82 driver might be how to do it. Taking feedback to the SE driver is as local as I might get using parts I have.
The SE driver favours triode. What would be interesting is a power pentode driver with UL transformer load. This would offer gain for my post No 56. I think it would be safe to assume 43% UL . I would give it a gain of 6. 1:6 should be easier to wind than a typical output type. That would need 25 mA at a guess. Hard to say because it drives the next stage rather nicely.
My big surprise is how EL84 is transformed by UL feedback. It should need only 16 Vrms to be driven in SE to long tail pair output ( whatever we call that ). What would be ideal is that an ECC83 could drive two sets of EL84. In paralell it might. If only needing 16 Vrms I dare say I could lower the anode voltage and up the current. I would aim for 500 mV input for 20 watts rms. ECC83 in ECC82 driver might be how to do it. Taking feedback to the SE driver is as local as I might get using parts I have.
The SE driver favours triode. What would be interesting is a power pentode driver with UL transformer load. This would offer gain for my post No 56. I think it would be safe to assume 43% UL . I would give it a gain of 6. 1:6 should be easier to wind than a typical output type. That would need 25 mA at a guess. Hard to say because it drives the next stage rather nicely.
BTW. I added shunt feedback to an ECC81. I was rather happy with how is measured. Valve Wizard had advice which looks obvious once tried. I fancy these nasty valves only have high gain as the blame. OK ECC81 is not really an audio valve. It looks OK when asked to be an ECC82 standin. Someone got very silly when I said this. All the same it is a type of feedback most likely to mimic a better device without measuring equipement to hand ( own up, many don't ). His arguement was the spacing of the parts inside the ECC81 would never be ideal. Seems to me one wants op amps at that point. For example a TDA2050 type would drive 4 x EL84 rather well, LM1875 even is an OK op amp if asked to be one. It would have no crossover distortion as the load will use the class A output ( AB usually ). If not in the feedback loop I suspect it would be rather nice and need no heatsink.
This is a concept and might be poor on detail. It will need more gain. An ECC83 would be ideal. Or op amp to prove concept. The FET's unlike bipolars will run close to V max. 250 V types exist. Don't use IRF/IRFP types as they need specail biasing. The FET's shown I think have the gate zener and back EMF diode built in. The harmonic distortion without feedback of the FET's is 0.8% from memory. 0.05% if a complimentary feedback set up of NPN PNP + PNP NPN. If the amp was reduced to gain of 2 I suspect it would do fine. The a gain of 30 plus to stage one not shown.
I built something like this recently to power a turntable albeit transformer driven and no loop feedback. It is rather good. FET's need about 1.4 V DC gate to gate to give the standing current. Not the 5 V people say, that's industrial FET's. .
Nige
Check this hybrid amp pdf article ,basically is one OK hybrid amp design but not ideal , for example I don`t like that OP-amp inserted in GNFB loop , and authors claim that the same OP-amp work as DC servo is wrong .
Absolutely. It's asking for trouble.
Here is a concept of using my parts. I think UL is better than Kitic idea. Notice OPA453. This should sound good. Notice 2 x EL 84 are grid to ground. This requires double the voltage swing to drive it. However the current is small. The g3 would be to cathode if EL84 and various grids would be stopped. The wire to the output is just ground. The op amp would be very happy at gain of 40. It must have >5. I have 8 matched 84's. Seeing as everything is op amp these days perhaps one more is no big deal. Said he who uses a Quad 33/303.
Distortion would be < 1% at 20 watts rms. If it followed an ideal law 0.1% at 5 watts. That's not impossible. If it could do that my personal hi fi standard is met.
Here is a concept of using my parts. I think UL is better than Kitic idea. Notice OPA453. This should sound good. Notice 2 x EL 84 are grid to ground. This requires double the voltage swing to drive it. However the current is small. The g3 would be to cathode if EL84 and various grids would be stopped. The wire to the output is just ground. The op amp would be very happy at gain of 40. It must have >5. I have 8 matched 84's. Seeing as everything is op amp these days perhaps one more is no big deal. Said he who uses a Quad 33/303.
Distortion would be < 1% at 20 watts rms. If it followed an ideal law 0.1% at 5 watts. That's not impossible. If it could do that my personal hi fi standard is met.
Absolutely. It's asking for trouble.
Here is a concept of using my parts. I think UL is better than Kitic idea. Notice OPA453. This should sound good. Notice 2 x EL 84 are grid to ground. This requires double the voltage swing to drive it. However the current is small. The g3 would be to cathode if EL84 and various grids would be stopped. The wire to the output is just ground. The op amp would be very happy at gain of 40. It must have >5. I have 8 matched 84's. Seeing as everything is op amp these days perhaps one more is no big deal. Said he who uses a Quad 33/303.
Distortion would be < 1% at 20 watts rms. If it followed an ideal law 0.1% at 5 watts. That's not impossible. If it could do that my personal hi fi standard is met.
Nige
Everything is one matter of compromise , mostly between desired max. out power and desired performance ,
Your schematic suggest that four EL84 works in differential mode , so four 84 cathodes have to be connected via common CCS to ground line , or even better via CCS to some negative supply line ,
one alternative solution is all four G1 connected directly to ground line and cathodes to be PP driven via some medium power NJFET`s , that will be one very linear pure grounded grid mode of operation , in that case EL84 -G2`s can be connected , 1) to anodes , 2) to UL taps , 3 )or to B+ line ,
btw , EL84 is one nice sounding tube , I have used them even for DIY FM transmitter 1W final stage , EI-NIS works fine there , but Russian counterpart not ,
btw , A, Kitic solution is nothing new , I have seen similar solution in many ancient tube amp schematics , here is the one ,
Attachments
IMHO that hybrid amp is OK designed but only from tube side , where pentode part is used as VAS and triode part as CF driver for BUZ complementary pair ,
but that OPS complementary pair can be connected in different way to avoid that OP-amp in GNFB loop .
Hey Guys,
I am planning an amp with an input transformer driving a pair of 6N6P triodes into a custom inter-stage transformer into a pair of pentodes, probably 6550. Nigel recommends UL as it is a form of local feedback. I was planning to use some Eico ST70 transformers for this amp but now I'm not so sure? If I were to order new outputs with 43% ul taps for the screens what about a set of dedicated taps for feedback to the 6550 cathodes? And if that were possible where would they be located on the primary winding? I.E. what percentage?
I am planning an amp with an input transformer driving a pair of 6N6P triodes into a custom inter-stage transformer into a pair of pentodes, probably 6550. Nigel recommends UL as it is a form of local feedback. I was planning to use some Eico ST70 transformers for this amp but now I'm not so sure? If I were to order new outputs with 43% ul taps for the screens what about a set of dedicated taps for feedback to the 6550 cathodes? And if that were possible where would they be located on the primary winding? I.E. what percentage?
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.