Ok, finally, after going back and forth I have decided what type of amp I want to build.
I would like to build a amplifier with Push-Pull output using a pentode in triode mode. I don't want to confine it to a particular tube complement but I am thinking EL34 or KT88. With 6N1P/6922/6DJ8 in the driver/input stages. This will be used at home for listening to a wide variety of music - if that fact matters.
The one I have found so far, and I like, is the PP-1C from vacuumstate.com, and ofcourse there is the Williamson. I not sure whether the PP1-C schematic on the website is just a symbolic circuit of a commerical product with the trade secrets removed or it is indeed an actual complete and working circuit. By the way anybody know what is the PP-2CS?
Any other good ones that I could consider ?
Thanks!
I would like to build a amplifier with Push-Pull output using a pentode in triode mode. I don't want to confine it to a particular tube complement but I am thinking EL34 or KT88. With 6N1P/6922/6DJ8 in the driver/input stages. This will be used at home for listening to a wide variety of music - if that fact matters.
The one I have found so far, and I like, is the PP-1C from vacuumstate.com, and ofcourse there is the Williamson. I not sure whether the PP1-C schematic on the website is just a symbolic circuit of a commerical product with the trade secrets removed or it is indeed an actual complete and working circuit. By the way anybody know what is the PP-2CS?
Any other good ones that I could consider ?
Thanks!
There's an "improved" Williamson here: http://stiftsbogtrykkeriet.dk/~mcs/Williamson/
It shouldn't be too hard to change the design to use 6N1Ps
Best regards,
Mikkel C. Simonsen
It shouldn't be too hard to change the design to use 6N1Ps
Best regards,
Mikkel C. Simonsen
Thanks. Interesting website you have Mikkel.
I will look at the improved design in detail and let you know if I have any questions.
There were a couple other Willys I found on the net too.
http://www.dc-daylight.ltd.uk/Valve.../Williamson-Amp/Williamson-WW-April-1947.html
http://www.xs4all.nl/~ideas/amps/
Frank, what do you recommend ?
I will look at the improved design in detail and let you know if I have any questions.
There were a couple other Willys I found on the net too.
http://www.dc-daylight.ltd.uk/Valve.../Williamson-Amp/Williamson-WW-April-1947.html
http://www.xs4all.nl/~ideas/amps/
analog_sa said:
Frank, what do you recommend ?
The PP-2C schematic is proprietary but Allen Wright will probably share it with you if you email him. He's shared it with a number of people here and on AudioAsylum, on condition we don't redistribute it. It's basically a higher power version of the PP-1C, using the KT88, with some other modifications. I haven't built it yet, but I'm planning to. Brett here at DIYaudio has said some very nice things about it.
By the way, the AES Super Amp is another triode push-pull design you might be interested in. It's available in kit form, but is fairly expensive even then. One good thing about it is that because AES is a division of Cary, it has solid resale value.
By the way, the AES Super Amp is another triode push-pull design you might be interested in. It's available in kit form, but is fairly expensive even then. One good thing about it is that because AES is a division of Cary, it has solid resale value.
I sincerely doubt it was good even in its heyday. Especially if you are aware of what amps preceded it. It certainly was the amp which started the numbers competition and i personally blame it for several decades of bad sounding/ good measuring amps which followed.
What's bad about it? For one it doesn't sound very good although it's probably better than a lot of SS amps. I used to have a commercially built pair with all Partridge iron and i really spent time and effort trying to get it to sound good. Too much feedback, too many 6SN7 (mine had 6J5) and way too many stages.
What would i suggest? First of all establish if your speakers can play with a zero nfb amp. My current PP choice is a 2 stage amp: a driver stage, phase-splitting transformer and o/p stage.
PP1-C seems to be an excellent choice as well.
regards
peter
Just because Williamson used 20dBs of feedback you don't have to. The one I have (one of my many un-finished projects) sounds good to me. If sounds fine without any feedback also - and then it's sensitive enough for tape head signals (about 0.5mVrms) 😉
Can an amp have too many 6SN7s? 😀
Best regards,
Mikkel C. Simonsen
Can an amp have too many 6SN7s? 😀
Best regards,
Mikkel C. Simonsen
Depends on the listener. One is sometimes tolerable but repeating the same annoying tonal signature again and again can get very tiring. Which, to some extent, is true for any tube - even the good ones should be used not more than once in a system.
Just to give a contrary view, mine is 100% opposite of analog's. I've built Williamsons, and with a very few modifications (mostly twiddling the LF time constants), it's a very linear, very good-sounding amp. Morgan Jones did a very nice deconstruction of this design in "Valve Amplifiers", which I would strongly recommend you read.
I recommend that you use a phase-splitting transformer or a combination input/phase-splitter like the Lundahl LL1676. It simplifies construction, makes mods easier, and sounds better to boot. I would personally rather pay the money than to ever have to solder up another phase inverting circuit.
John
John
Attachments
Hi,
Amen to that...
Too complicated for its own good it measures well but shows you everything you shouldn't do with an amplifier if you want a natural reproduction of an instrument.
Both timbral and spatial information are often severely altered with these types of amps.
With the advent of high output line level sources and the return of excellent high(ish) efficiency speakers it's hard chose a Williamson amp after you heard what a simple two stage PP amp can do.
One should at the very least have a listen to both topologies under comparable circumstances and see what one prefers.
Cheers, 😉
Amen to that...
Too complicated for its own good it measures well but shows you everything you shouldn't do with an amplifier if you want a natural reproduction of an instrument.
Both timbral and spatial information are often severely altered with these types of amps.
With the advent of high output line level sources and the return of excellent high(ish) efficiency speakers it's hard chose a Williamson amp after you heard what a simple two stage PP amp can do.
One should at the very least have a listen to both topologies under comparable circumstances and see what one prefers.
Cheers, 😉
Regarding the Williamson amplifier I think it is important to realise that there are many amplifiers described as "Williamson" or "Amplifier with Williamson driver" or "Ultralinear Williamson" that has very little in common with the original amplifier designed by Williamson. As an example "ultralinear Williamson" is something that doesn't exist, the orginal used triode coupled KT66 with a special output transformer, Williamson never design an amplifier with ultralinear coupling of the output transformer.
The Williamson amplifier has been subjected to criticism over the years but in most cases I think this is due to misunderstanding of the original circuit function or because many people have modified it order to �gimprove�h it.
The most common �gimprovements�h are using an output transformer other than the one specified, ultra-linear coupling of the output stage, change of tubes in the driver stage from original L63�fs to other tubes without modification of other circuit values or change of output tubes from original KT66�fs to 807�fs or other tubes. Most of these changes don�ft really offer any improvement of the original design; on the contrary most of these modifications will increase distortion, reduce bandwidth or introduce instability compared to the original circuit.
The name "Williamson amplifier" was obviously used as an indication of quality and performance, a good testimonial of the original amplifiers benefits.
Using the original circuit with KT66's, L63's and a correct output transformer I still think give a very good result. One interesting fact is that John Linsley Hood used an original Williamson amplifier as reference when he designed his famous 10W class A design transistor design and even he admitted that the Williamson sounds as good as his own transistor design.
One of the few things I think can be improved on the original circuit is the stability at low frequencies, this can easily be done using a phase compensating network and it is quite easy to increase stability margin from the original ~30 degrees to >70 degrees using a simple network, I am a bit puzzled why Williamson didn't do this himself.
Regards Hans
The Williamson amplifier has been subjected to criticism over the years but in most cases I think this is due to misunderstanding of the original circuit function or because many people have modified it order to �gimprove�h it.
The most common �gimprovements�h are using an output transformer other than the one specified, ultra-linear coupling of the output stage, change of tubes in the driver stage from original L63�fs to other tubes without modification of other circuit values or change of output tubes from original KT66�fs to 807�fs or other tubes. Most of these changes don�ft really offer any improvement of the original design; on the contrary most of these modifications will increase distortion, reduce bandwidth or introduce instability compared to the original circuit.
The name "Williamson amplifier" was obviously used as an indication of quality and performance, a good testimonial of the original amplifiers benefits.
Using the original circuit with KT66's, L63's and a correct output transformer I still think give a very good result. One interesting fact is that John Linsley Hood used an original Williamson amplifier as reference when he designed his famous 10W class A design transistor design and even he admitted that the Williamson sounds as good as his own transistor design.
One of the few things I think can be improved on the original circuit is the stability at low frequencies, this can easily be done using a phase compensating network and it is quite easy to increase stability margin from the original ~30 degrees to >70 degrees using a simple network, I am a bit puzzled why Williamson didn't do this himself.
Regards Hans
You might find you don't need global negative feedback with triode-strapped pentodes. A bit of shunt feedback across the output stage might be useful but this will not introduce the risk of instability that global nfb can cause, so circuits with more than one coupling capacitor (e.g. Williamson) needn't be a concern.
My personal preference is for an all differential design because this helps to eliminate power supply noise and sounds good. My own amp (which I'm in the process of improving) uses 6SL7 LTP splitter, 6SN7 driver and EL34s triode-strapped PP output stage. The splitter uses a pentode (6AU6) CCS in the tail, to help the balance. I'm planning to try a transistor CCS in the tail of the driver as well, on advice from helpful members of this forum.
I'm currently trying an operating point for the EL34s that is halfway between Mullard's two recommended options of using 440 ohms in each cathode either bypassed (5k p-p load giving 19w) or unbypassed (10k p-p load giving 14w). I have 220 ohms in each cathode, below which is a 120 ohms shared resistor and a p-p load of 7.5k. Shunt feedback is provided by 470k resistors between the EL34 plates and the 6SN7 plates. It sounds good to me but I suspect it needs a shade more shunt feedback to improve the damping of the speaker.
Good luck with your project!
My personal preference is for an all differential design because this helps to eliminate power supply noise and sounds good. My own amp (which I'm in the process of improving) uses 6SL7 LTP splitter, 6SN7 driver and EL34s triode-strapped PP output stage. The splitter uses a pentode (6AU6) CCS in the tail, to help the balance. I'm planning to try a transistor CCS in the tail of the driver as well, on advice from helpful members of this forum.
I'm currently trying an operating point for the EL34s that is halfway between Mullard's two recommended options of using 440 ohms in each cathode either bypassed (5k p-p load giving 19w) or unbypassed (10k p-p load giving 14w). I have 220 ohms in each cathode, below which is a 120 ohms shared resistor and a p-p load of 7.5k. Shunt feedback is provided by 470k resistors between the EL34 plates and the 6SN7 plates. It sounds good to me but I suspect it needs a shade more shunt feedback to improve the damping of the speaker.
Good luck with your project!
Konnichiwa,
The first absolute key question you must ask is:
Class A or Class AB?
If class A then you are limited to maybe 12 Watt with the EL34. A class A Push-Pull Amp with trioded EL34 (or KT88) can be pretty good and will work well without any negative feedback. Driverstages are easy too and you can even add a 3rd harmonic cancellation, making for a VERY LINEAR output stage which if you add a suitably linear driver stage will do well.
Only problem is rather low power.
If you choose Class Ab you MUST use negative feedback. The best solution here is something like the German V69 TAB Studio Amplifier (even though that uses Pentodes of course).
The Actual Williamson is full of fundamental design flaws. Even if you correct them all it is at best a middeling.
Allan Wrights Amp's are likely very good, if you look closely they are very much like a V69 without feedback.... ;-)
It is an actual working circuit, however it is a pretty old variant of Allan's designs. His current stuff offers most likely much improved performance.
You can actually devise a "shades of the V69" Amplifier using Cascoded 6DJ8/6922 instead of pentodes with the V69's "fixed-selfbias" circuit retained and the Output Stage swichable between Pentode with NFB and Triode without or with less NFB.
I personally also can recommend Rickard Berglunds "Ultimate" Amplifier (dig out your sound practices backissues (winter 94/95 - SP8), which is a rather unusal "SE Soundalike" PP Amp. It does not really sound like an SE Amp, but it sounds pretty good, especially if you Bias the Output Valves for a modest current imbalance so that the Valve that catually receives the lower drive signal is biased for a current that by the stated imbalance limit of the output transformer higher than that of valve driven by the higher level drive signal.
For commercial implementations of Push-Pull I moved away from the zero NFB "Ultimate" to something rather different, but as simple DIY Amp where Damping factor and measured performance is not much of a concern it is an excellent, if off the wall and "crazy" choice.
Sayonara
percy said:
The first absolute key question you must ask is:
Class A or Class AB?
If class A then you are limited to maybe 12 Watt with the EL34. A class A Push-Pull Amp with trioded EL34 (or KT88) can be pretty good and will work well without any negative feedback. Driverstages are easy too and you can even add a 3rd harmonic cancellation, making for a VERY LINEAR output stage which if you add a suitably linear driver stage will do well.
Only problem is rather low power.
If you choose Class Ab you MUST use negative feedback. The best solution here is something like the German V69 TAB Studio Amplifier (even though that uses Pentodes of course).
percy said:
The Actual Williamson is full of fundamental design flaws. Even if you correct them all it is at best a middeling.
Allan Wrights Amp's are likely very good, if you look closely they are very much like a V69 without feedback.... ;-)
percy said:
It is an actual working circuit, however it is a pretty old variant of Allan's designs. His current stuff offers most likely much improved performance.
percy said:
You can actually devise a "shades of the V69" Amplifier using Cascoded 6DJ8/6922 instead of pentodes with the V69's "fixed-selfbias" circuit retained and the Output Stage swichable between Pentode with NFB and Triode without or with less NFB.
I personally also can recommend Rickard Berglunds "Ultimate" Amplifier (dig out your sound practices backissues (winter 94/95 - SP8), which is a rather unusal "SE Soundalike" PP Amp. It does not really sound like an SE Amp, but it sounds pretty good, especially if you Bias the Output Valves for a modest current imbalance so that the Valve that catually receives the lower drive signal is biased for a current that by the stated imbalance limit of the output transformer higher than that of valve driven by the higher level drive signal.
For commercial implementations of Push-Pull I moved away from the zero NFB "Ultimate" to something rather different, but as simple DIY Amp where Damping factor and measured performance is not much of a concern it is an excellent, if off the wall and "crazy" choice.
Sayonara
I'd love to build a Class A1. Class AB is fine too as long as there is low(6db?) amount of nfb. What matters to me right now is that I want a design that I could use with no or very little modifications.
Something that has been done before, proven.
I was really expecting to see more designs but looks like from what has been suggested so far, other than the PP-1C, there really arn't any that stand out convincingly.
That is re-assuring. I wish the PP-2CS was available, then I wouldn't ask anything else from life ever 🙂
Now just a thought
- and this might prove to be a stupid question so please bear with me 
There are quite a few "driver boards" out there. Pete Millet's, Curcio Audio, Welbourne, etc. Can I Mix'n Match ? Is it possible to select one of these driver boards, then select or design an output stage seperately - say for example a Push-Pull Class A1 KT88 in triode mode and "tie" them together ? Can that become a happy marraige or a disaster ?
What factors should be considered in determining whether a third-party stand-alone driver board will work with your output stage or not ?
The proposed match should keep in mind the following:
1. What is the swing needed to drive the output stage? Can the driver provide that?
2. What is the input capacitance of the output stage and the source impedance of the driver stage?
3. What is the necessary gain from input to output of the amp? Will the two stages provide this, taking into account any desire to run feedback loops?
1. What is the swing needed to drive the output stage? Can the driver provide that?
2. What is the input capacitance of the output stage and the source impedance of the driver stage?
3. What is the necessary gain from input to output of the amp? Will the two stages provide this, taking into account any desire to run feedback loops?
Check out the Plitron website, Menno VanDerVeen has an article on pentode and triode strapped pentode amps with cathode feedback windings in the output transformers. Look under "news" and toward the bottom there are several articles in pdf form. The article is titled "New Push-Pull Tube Amplifiers".
If I am looking at the right article isn't that an ultra-linear ?
My first preference is to find a triode mode.
BTW, do you know of any groups in Twin Cities area that are active in diy audio (speaker building, tube/ss amps etc...) ?
My first preference is to find a triode mode.
BTW, do you know of any groups in Twin Cities area that are active in diy audio (speaker building, tube/ss amps etc...) ?
If you read the article you'll see that just about every type of PP amp is represented, but that the author prefers circuit #7 (out of eight) that is a triode strapped pentode (or beam power tetrode).
There's the Minnesota Audiophile Society, they have some members that are into DIY. I think once a year they have a meeting where people bring in their own stuff and demonstrate it. They meet at the Pavek Museum of Broadcasting, one of the coolest places as far as tubes go...
There's the Minnesota Audiophile Society, they have some members that are into DIY. I think once a year they have a meeting where people bring in their own stuff and demonstrate it. They meet at the Pavek Museum of Broadcasting, one of the coolest places as far as tubes go...
- Status
- Not open for further replies.