I have cut and pasted a section from one of my earlier posts because the subject was changing substantially and I thought we'd get a better overall discussion by seperating it.
Below is the thread I posted and the replies to date, I've cut it inot two as it was too long:
Has anyone got an opinion on the Walton Audio Amplifier?
http://indigo.ie/~walton/300bmk2.html
There is a more detailed section for his first version
http://indigo.ie/~walton/300b.html
It still appears to me that the 300B is the simplest to build at this stage.
__________________
Audio
I actually met Derek Walton at the Dublin hi-fi show where he had rented space to showcase his amplifiers more though in the spirit of offering his services for hand-made amps to interested parties but not really as a business venture.I had a chance to talk to him because I am also intersted in building a 300b and we spoke on the phone in order to arrange a further meeting.Very nice and enthusiastic person.However he has completely dropped out of sight in the last few months and all my efforts to contact him have had no result.
His website is great for anyone wishing to build a 300b and quite detailed and simple.
The only caveat seems to me that some experts recommend a good two stage driver for the 300b while he has adopted only a 6sn7 driver approach.
Another good site is the Plitron site where there are a couple of good articles reprinted on a stereo 300b set and a push-pull version with the seemingly excellent plitron trafos.
I would also appreciate the opinion of the "experts" here on what in their humble opinion is the best 300b circuit including the question on tube or silicon rectification , choke vs transistor regulation , grid bias regulation etc.
IP: 159.134.158.xxx
Yesterday, at 09:08 PM Quote:
Jose R []
diyAudio Member
Offline
Registered: Jan 2004
Location: NYC
Post #98
quote:
Originally posted by protos
I would also appreciate the opinion of the "experts" here on what in their humble opinion is the best 300b circuit including the question on tube or silicon rectification , choke vs transistor regulation , grid bias regulation etc.
I'm no expert, but as has often been stated on this forum, there is no *best* circuit or design; its more a matter of taste. Also, you have to address what your goals are, your budget, your abilities, and where are you willing to compromise.
Having said that though and with the necessary disclaimers, I will offer the following:
I would never recommend a 6SN7 as driver for a 300B.
I would recommend a 6V6 (triode strapped) or other pentode, or other triode as driver.
I would not recommend a two stage approach. "Everything should be made as simple as possible, but not simpler." Alot of people view the two stage amp as the simple and purer approach to designing. Afterall, less parts, less wire, shorter circuit path, what could be more perfect. For me, that viewpoint is flawed in the sense that you are now asking a smaller number of parts to do multiple duties. Which then creates complex interactions which then limits the number of available driver choices down to a few high gain tubes (like 6c45pi). And these tubes, if they aren't rare and expensive NOS, are current production russian which have their own idioyncrasies in terms of their application. So why invite unecessary headaches? And on top of that, a two stage amp in most cases requires you to have an active preamp. Which is basically the third stage which you dropped from the front of the amp because you wanted to *KISS*.
Therefore, for me a three stage approach is best; it allows you the freedom of choice in terms of tubes and topographies. And the use of a passive volume control like a TVC or autoformer which hands down beats all active preamps, period.
An unbypassed 1/2 6SN7 loaded with a 68K-100K resistor makes a darn good voltage amplifier as input tube. A triode-strapped pentode loaded with a choke in series with a resistor would make a decent driver. Cap coupling between stages. If you're game and are cathode self-biasing then direct coupling might be a choice to consider.
Keep the power supply for the output stage separate from the power supply for the input/driver stages. Basically, build two full supplies, complete with CLCLC filtering for each B+. The needs of the output tubes are different than the needs of the input/drivers. Separating the supplies allows each tube to be properly loaded. And while you're at it, and since at this point who cares about costs use fixed bias on both the drivers and output tubes. Thereby eliminating the big cathode resistors and electrolytic bypass caps; and since you have left the 6SN7 unbypased, you have just eliminated a number of parts which if left in would add nasty colorations the sonic signature of the amp. AND thereby also eliminating the need for matched pairs anywhere in the amp. Afterall, *matched pairs* of tubes is the biggest myth in tube audio. AND thereby, giving you complete control of the op of each tube.
Use hybrid rectification, combining tube and silicon rectifiers. Hexfreds or the recently introduced high voltage Schottkys. Better controlled bass response.
Use separate filament supplies for each DHT or shared supply for IDHTs. Use separate filament transformers. DC heated. RCLC filtering.
And finally, keep the power supply on a separate chassis than the circuit components. Connect with a properly shielded umbilical. Keep the last LC or RC section of each PS on the amp chassis. Place the last cap of the DC filament supply right at the filament tabs of the tube sockets. Star ground.
And one more thing, use high quality parts throughout. Especially iron.
Basically, I've just described to you my amp. And since its my first DIY amp, and also a scratch build, and when I started the amp, I knew nothing about electronics, much less amp building or design, I make no claim to being an expert. But, now after 2.5 years of working on this amp, I do have a little experience. So take my humble suggestions FW their W.
Here is a pic:
IP: 208.201.160.xxx
Today, at 04:00 AM Quote:
Positron []
diyAudio Member
Offline
Registered: May 2003
Location: Morton, Illinois
Well.... Post #99
Actually it is a matter of whether one wants to outboard the first stage or not; with an additional selector switch.
True, one would need an IC, but then you don't have to contend with signal interaction through the power supply. It is frequency related, bass and then highs the most vunurable. Really messes with the sonics. I prefer the separation and although it costs much more, the sound is better.
Steve
SAS Audio
__________________
Take care.
Steve
Last edited by Positron on 01-23-2004 at 06:37 AM
IP: 204.42.21.xxx
Today, at 06:30 AM Quote:
Kuei Yang Wang []
diyAudio Member
Offline
Registered: Nov 2002
Location: City of the Sanjian
Post #100
Below is the thread I posted and the replies to date, I've cut it inot two as it was too long:
Has anyone got an opinion on the Walton Audio Amplifier?
http://indigo.ie/~walton/300bmk2.html
There is a more detailed section for his first version
http://indigo.ie/~walton/300b.html
It still appears to me that the 300B is the simplest to build at this stage.
__________________
Audio
I actually met Derek Walton at the Dublin hi-fi show where he had rented space to showcase his amplifiers more though in the spirit of offering his services for hand-made amps to interested parties but not really as a business venture.I had a chance to talk to him because I am also intersted in building a 300b and we spoke on the phone in order to arrange a further meeting.Very nice and enthusiastic person.However he has completely dropped out of sight in the last few months and all my efforts to contact him have had no result.
His website is great for anyone wishing to build a 300b and quite detailed and simple.
The only caveat seems to me that some experts recommend a good two stage driver for the 300b while he has adopted only a 6sn7 driver approach.
Another good site is the Plitron site where there are a couple of good articles reprinted on a stereo 300b set and a push-pull version with the seemingly excellent plitron trafos.
I would also appreciate the opinion of the "experts" here on what in their humble opinion is the best 300b circuit including the question on tube or silicon rectification , choke vs transistor regulation , grid bias regulation etc.
IP: 159.134.158.xxx
Yesterday, at 09:08 PM Quote:
Jose R []
diyAudio Member
Offline
Registered: Jan 2004
Location: NYC
Post #98
quote:
Originally posted by protos
I would also appreciate the opinion of the "experts" here on what in their humble opinion is the best 300b circuit including the question on tube or silicon rectification , choke vs transistor regulation , grid bias regulation etc.
I'm no expert, but as has often been stated on this forum, there is no *best* circuit or design; its more a matter of taste. Also, you have to address what your goals are, your budget, your abilities, and where are you willing to compromise.
Having said that though and with the necessary disclaimers, I will offer the following:
I would never recommend a 6SN7 as driver for a 300B.
I would recommend a 6V6 (triode strapped) or other pentode, or other triode as driver.
I would not recommend a two stage approach. "Everything should be made as simple as possible, but not simpler." Alot of people view the two stage amp as the simple and purer approach to designing. Afterall, less parts, less wire, shorter circuit path, what could be more perfect. For me, that viewpoint is flawed in the sense that you are now asking a smaller number of parts to do multiple duties. Which then creates complex interactions which then limits the number of available driver choices down to a few high gain tubes (like 6c45pi). And these tubes, if they aren't rare and expensive NOS, are current production russian which have their own idioyncrasies in terms of their application. So why invite unecessary headaches? And on top of that, a two stage amp in most cases requires you to have an active preamp. Which is basically the third stage which you dropped from the front of the amp because you wanted to *KISS*.
Therefore, for me a three stage approach is best; it allows you the freedom of choice in terms of tubes and topographies. And the use of a passive volume control like a TVC or autoformer which hands down beats all active preamps, period.
An unbypassed 1/2 6SN7 loaded with a 68K-100K resistor makes a darn good voltage amplifier as input tube. A triode-strapped pentode loaded with a choke in series with a resistor would make a decent driver. Cap coupling between stages. If you're game and are cathode self-biasing then direct coupling might be a choice to consider.
Keep the power supply for the output stage separate from the power supply for the input/driver stages. Basically, build two full supplies, complete with CLCLC filtering for each B+. The needs of the output tubes are different than the needs of the input/drivers. Separating the supplies allows each tube to be properly loaded. And while you're at it, and since at this point who cares about costs use fixed bias on both the drivers and output tubes. Thereby eliminating the big cathode resistors and electrolytic bypass caps; and since you have left the 6SN7 unbypased, you have just eliminated a number of parts which if left in would add nasty colorations the sonic signature of the amp. AND thereby also eliminating the need for matched pairs anywhere in the amp. Afterall, *matched pairs* of tubes is the biggest myth in tube audio. AND thereby, giving you complete control of the op of each tube.
Use hybrid rectification, combining tube and silicon rectifiers. Hexfreds or the recently introduced high voltage Schottkys. Better controlled bass response.
Use separate filament supplies for each DHT or shared supply for IDHTs. Use separate filament transformers. DC heated. RCLC filtering.
And finally, keep the power supply on a separate chassis than the circuit components. Connect with a properly shielded umbilical. Keep the last LC or RC section of each PS on the amp chassis. Place the last cap of the DC filament supply right at the filament tabs of the tube sockets. Star ground.
And one more thing, use high quality parts throughout. Especially iron.
Basically, I've just described to you my amp. And since its my first DIY amp, and also a scratch build, and when I started the amp, I knew nothing about electronics, much less amp building or design, I make no claim to being an expert. But, now after 2.5 years of working on this amp, I do have a little experience. So take my humble suggestions FW their W.
Here is a pic:
IP: 208.201.160.xxx
Today, at 04:00 AM Quote:
Positron []
diyAudio Member
Offline
Registered: May 2003
Location: Morton, Illinois
Well.... Post #99
Actually it is a matter of whether one wants to outboard the first stage or not; with an additional selector switch.
True, one would need an IC, but then you don't have to contend with signal interaction through the power supply. It is frequency related, bass and then highs the most vunurable. Really messes with the sonics. I prefer the separation and although it costs much more, the sound is better.
Steve
SAS Audio
__________________
Take care.
Steve
Last edited by Positron on 01-23-2004 at 06:37 AM
IP: 204.42.21.xxx
Today, at 06:30 AM Quote:
Kuei Yang Wang []
diyAudio Member
Offline
Registered: Nov 2002
Location: City of the Sanjian
Post #100
Konnichiwa,
quote:
Originally posted by protos
I would also appreciate the opinion of the "experts" here on what in their humble opinion is the best 300b circuit including the question on tube or silicon rectification , choke vs transistor regulation , grid bias regulation etc.
Well, the subject is complex. First a few notes on gain. If you use the russian 6S45 or any of the following - EC8010, EC8020, WE 417A, WE 437A or triode wired Pentodes like E810F/7788 / E280F / E180F/6688 / E55L et al you can have a driverstahe that can drive the 300B at normal Operating Points (350V60...80mA) to full power with around 0.9 - 1.3V, so with a normal CD Player your systemn is in essence gainmatched with a unity gain Preamp (passive or active), meaning the Amplifier will never receive an input signal that is over around 6db above it's clipping point with the volume full up.
A TVC makes an excellent choice in this case. Daisychaning a pair of 6SN7 Halves or the like not only produces a lot of distortion prior to the outputstage with a lot of upper harmonics, in addition it will also drive the 300B to clipping with around 0.25 - 0.4V RMS, meaning between 8 - 12db attenuation will be always needed.
So you gain nothing by using a dual stage driver, but you loose transparencey and tonality, plus you are ging to be further away from the best sounding range of the TVC (contrary to common misconception the TVC is best turned all the way up!!!) should you use one.
So, unless you need to accomodate very low output sources having more than one driverstage gives no benefit.
The best solution to maximising the performance of such an amplifier IMHO is to use directcoupling with one of the possible variations of the DRD/Monkey/FreeLunch/Stacked Supplies approach, in all cases using a Anode load choke (Tangoe TC160-15 or suitable items from Stevens & Billington, Electraprint or Magnequest).
My gut feeling favours a "stacked supplies free lunch" with a seperate valve rectified supply for the driver and a suitable RL series combo to provide the bias for the Output Valve. In this case you need a suitable interlock that can be conveniently be applied from a Relais that will power up the HT for the output stage once the driver valve draws enough current to bias the output stage.
As for rectification, it is possible to make solid state rectifiers sound very close to Valve ones, but it takes a lot of effort and is poorely documented. The valve rectifier serves two sides for good sound. On one hand it's soft turnon and lack of rectifier noise are hard to match with solid state and secondly, the fairly high power AC heater of the rectifier provides a steady load for the Mains transformer, thus damping the tendency of the transformer to oscillate when current pulses from the rectification of the HT or Heater Voltage try excite the parasitic resonances.
Ideally you use a LCLC or CLCLC Filter with a small first C to bring the supply as close to a choke input supply as possible, with LCLC supplies you can eliminate for all practical purposes any +B supply related noise, so you get an inherently "quiet" rectification and a very clean supply even when using Film or Oil Capacitors (highly recommended).
So an option would be a 350V supply using a WE 274A/B or 5R4GY (or 5AR4/GZ34) in LCLC for the 300B. Add to this a supply with around 250V +B for the driverstage, rectified for example with a 6X5, WE420A or EZ80 and CLCLC filtering. Driver valve of choice would be E55L or E810F or WE 437A, for my taste....
If this is combined with a suitable LC filtered DC supply for the 300B Heaters you would be very close to "edge of art", assuming of course high performance Transformers, chokes etc.
If you have multiple taps on the +B for the Output Valve (maybe including more Current and Voltage for the "super 300B's" and maybe something like a Tango XE20S as Output Transformer plus a pot to adjust the Bias (in series with the anode load choke) and selectable Heater Voltages (2.5V AC, 4V/5V/6.3V/7.5V DC) you can use a wide range of output valves (if you make easily exchanged socket plates even european oddballs), de facto making the amplifier "universal" and making sure you can run whatever valves you get hold of....
Of course, such an Amp would be quite a monster in size and weight, fitted with 2 pcs Mains Transformer, 5 pcs substantial size input/filter chokes, 4 - 5 pcs large Film or Oil PSU Capacitors plus output transformer and anode load choke. And it will be a layout nightmare if you want to avoid magnetic coupling. Of course monoblock chassis with Bronze, Brass or Copper top plate and wooden frame.
Well, it would make for a stonking Amp and yes, it would cost an arm and a leg if you build this kind of design consequently enough, BUT I'm pretty confident it will require a Badge with the incription "Beware - Kicks Butt", despite giving only 1.5 Watt when fitted with a #45!
So, anyway, that's my take for "the best 300B Amp I can come up with if money, space and time are infinite".
Sayonara
quote:
Originally posted by protos
I would also appreciate the opinion of the "experts" here on what in their humble opinion is the best 300b circuit including the question on tube or silicon rectification , choke vs transistor regulation , grid bias regulation etc.
Well, the subject is complex. First a few notes on gain. If you use the russian 6S45 or any of the following - EC8010, EC8020, WE 417A, WE 437A or triode wired Pentodes like E810F/7788 / E280F / E180F/6688 / E55L et al you can have a driverstahe that can drive the 300B at normal Operating Points (350V60...80mA) to full power with around 0.9 - 1.3V, so with a normal CD Player your systemn is in essence gainmatched with a unity gain Preamp (passive or active), meaning the Amplifier will never receive an input signal that is over around 6db above it's clipping point with the volume full up.
A TVC makes an excellent choice in this case. Daisychaning a pair of 6SN7 Halves or the like not only produces a lot of distortion prior to the outputstage with a lot of upper harmonics, in addition it will also drive the 300B to clipping with around 0.25 - 0.4V RMS, meaning between 8 - 12db attenuation will be always needed.
So you gain nothing by using a dual stage driver, but you loose transparencey and tonality, plus you are ging to be further away from the best sounding range of the TVC (contrary to common misconception the TVC is best turned all the way up!!!) should you use one.
So, unless you need to accomodate very low output sources having more than one driverstage gives no benefit.
The best solution to maximising the performance of such an amplifier IMHO is to use directcoupling with one of the possible variations of the DRD/Monkey/FreeLunch/Stacked Supplies approach, in all cases using a Anode load choke (Tangoe TC160-15 or suitable items from Stevens & Billington, Electraprint or Magnequest).
My gut feeling favours a "stacked supplies free lunch" with a seperate valve rectified supply for the driver and a suitable RL series combo to provide the bias for the Output Valve. In this case you need a suitable interlock that can be conveniently be applied from a Relais that will power up the HT for the output stage once the driver valve draws enough current to bias the output stage.
As for rectification, it is possible to make solid state rectifiers sound very close to Valve ones, but it takes a lot of effort and is poorely documented. The valve rectifier serves two sides for good sound. On one hand it's soft turnon and lack of rectifier noise are hard to match with solid state and secondly, the fairly high power AC heater of the rectifier provides a steady load for the Mains transformer, thus damping the tendency of the transformer to oscillate when current pulses from the rectification of the HT or Heater Voltage try excite the parasitic resonances.
Ideally you use a LCLC or CLCLC Filter with a small first C to bring the supply as close to a choke input supply as possible, with LCLC supplies you can eliminate for all practical purposes any +B supply related noise, so you get an inherently "quiet" rectification and a very clean supply even when using Film or Oil Capacitors (highly recommended).
So an option would be a 350V supply using a WE 274A/B or 5R4GY (or 5AR4/GZ34) in LCLC for the 300B. Add to this a supply with around 250V +B for the driverstage, rectified for example with a 6X5, WE420A or EZ80 and CLCLC filtering. Driver valve of choice would be E55L or E810F or WE 437A, for my taste....
If this is combined with a suitable LC filtered DC supply for the 300B Heaters you would be very close to "edge of art", assuming of course high performance Transformers, chokes etc.
If you have multiple taps on the +B for the Output Valve (maybe including more Current and Voltage for the "super 300B's" and maybe something like a Tango XE20S as Output Transformer plus a pot to adjust the Bias (in series with the anode load choke) and selectable Heater Voltages (2.5V AC, 4V/5V/6.3V/7.5V DC) you can use a wide range of output valves (if you make easily exchanged socket plates even european oddballs), de facto making the amplifier "universal" and making sure you can run whatever valves you get hold of....
Of course, such an Amp would be quite a monster in size and weight, fitted with 2 pcs Mains Transformer, 5 pcs substantial size input/filter chokes, 4 - 5 pcs large Film or Oil PSU Capacitors plus output transformer and anode load choke. And it will be a layout nightmare if you want to avoid magnetic coupling. Of course monoblock chassis with Bronze, Brass or Copper top plate and wooden frame.
Well, it would make for a stonking Amp and yes, it would cost an arm and a leg if you build this kind of design consequently enough, BUT I'm pretty confident it will require a Badge with the incription "Beware - Kicks Butt", despite giving only 1.5 Watt when fitted with a #45!
So, anyway, that's my take for "the best 300B Amp I can come up with if money, space and time are infinite".
Sayonara
Konnichiwa,
Some notes, the "Walton" 300B Amp is basically the JE Labs Amp which in turn borrows heavily from the Fi-Primer and Sun Audio. The only plus of this general design is that it is cheap to realise and extensively documented. It makes a fine first SE Amp build on a shoestring, using Hammond Iron and generic parts to be later relegated 9into a second system or to be passed to non-diying friends....
The basic circuit and parts quality is very similar to a middle of the road 300B Poweramp as manufacturerd by a range of companies, at least in the first Unit from Walton and with JE Labs. There are also strong similarities to the "standard" Audio Note UK Circuit, so if you want an idea what such an Amp will sound like audition an Audio Note Quest or the usual suspects from Cary & Cannary.
It is such amplifiers that are often heard at shows and commented upon, as only a few small companies (best known among them are Wavelength Audio and Border Patrol, also Wavac & Audio Note Japan) make SE Amplifiers not severely compromised by having to be commercial, sellable products.
The best DIY'ers can make when they simply wave any commercial considereations is miles if not lightyears ahead of any semi-sensibly priced commercial SET Amp. Hence my take is: Why spend a lot of money to copy a Ford Ka when you can build something that is better than a Ferrari?
Sayonara
Some notes, the "Walton" 300B Amp is basically the JE Labs Amp which in turn borrows heavily from the Fi-Primer and Sun Audio. The only plus of this general design is that it is cheap to realise and extensively documented. It makes a fine first SE Amp build on a shoestring, using Hammond Iron and generic parts to be later relegated 9into a second system or to be passed to non-diying friends....
The basic circuit and parts quality is very similar to a middle of the road 300B Poweramp as manufacturerd by a range of companies, at least in the first Unit from Walton and with JE Labs. There are also strong similarities to the "standard" Audio Note UK Circuit, so if you want an idea what such an Amp will sound like audition an Audio Note Quest or the usual suspects from Cary & Cannary.
It is such amplifiers that are often heard at shows and commented upon, as only a few small companies (best known among them are Wavelength Audio and Border Patrol, also Wavac & Audio Note Japan) make SE Amplifiers not severely compromised by having to be commercial, sellable products.
The best DIY'ers can make when they simply wave any commercial considereations is miles if not lightyears ahead of any semi-sensibly priced commercial SET Amp. Hence my take is: Why spend a lot of money to copy a Ford Ka when you can build something that is better than a Ferrari?
Sayonara
Konnichiwa,
I have published the odd "small compromise" design and there are some semi-published ones, including this:
http://www.jogis-roehrenbude.de/Leserbriefe/TH-Loesch-Amp/Th-Loesch-Amp.htm
(Sorry for the german text - I'll eventually publish in english when I find the time)
and there are others to be found in older issues of Sound Practices and Glass Audio as well as Valve. I would reccomend to buy the (quite inexpensive for the content) Valve & Sound Practices CD Archives and to read the Glass Audio Back issues index.
There is much practical experience, theory and background in these, plus some absolutely stonking designs such as Ciro Marzio & Christiano Jelasi's Direct Coupled SE Amp (very prosaic Amplifier circuit with ECC83, ECC88 or others driving 2A3 or 300B at low voltage operation, but a seriously overvbuild PSU witout electrolytic capacitors and Chimera Labs 211 Amp.
So basically, my recommendation, read some more before building if you want to build "one" final Amp and learn from others or simply start building the "Standard" (Fi-Primer, JE Labs, Walton, Sun et al) 300B Amp on a shoestring as learning excercise and then build a few more amps and learn from your own mistakes....
Sayonara
DVDHack said:
I have published the odd "small compromise" design and there are some semi-published ones, including this:
http://www.jogis-roehrenbude.de/Leserbriefe/TH-Loesch-Amp/Th-Loesch-Amp.htm
(Sorry for the german text - I'll eventually publish in english when I find the time)
and there are others to be found in older issues of Sound Practices and Glass Audio as well as Valve. I would reccomend to buy the (quite inexpensive for the content) Valve & Sound Practices CD Archives and to read the Glass Audio Back issues index.
There is much practical experience, theory and background in these, plus some absolutely stonking designs such as Ciro Marzio & Christiano Jelasi's Direct Coupled SE Amp (very prosaic Amplifier circuit with ECC83, ECC88 or others driving 2A3 or 300B at low voltage operation, but a seriously overvbuild PSU witout electrolytic capacitors and Chimera Labs 211 Amp.
So basically, my recommendation, read some more before building if you want to build "one" final Amp and learn from others or simply start building the "Standard" (Fi-Primer, JE Labs, Walton, Sun et al) 300B Amp on a shoestring as learning excercise and then build a few more amps and learn from your own mistakes....
Sayonara
No need to translate Thorsten...
the www will do that reasonably well
http://world.altavista.com/tr
Take the rest to http://www.freetranslation.com
Cheers,
Bas
the www will do that reasonably well
http://world.altavista.com/tr
Take the rest to http://www.freetranslation.com
Cheers,
Bas
With the limited knowledge that I have of valve amps the translation is pretty meaningless.
What is it about that particular amp that makes it no compromise. It appears to me from the many circuits available that there is little difference by way of designs - without building them all how on earth do I know which is better?
The Walton design has a revised version which changes the supply by including additional inductors CLCLC type config - the parts have been significantly improved using Piltron TFR and better passive components plus some changes in input stages. The cost appears quite high - think he spent near $4000.
Anymore thoughts on a top design or should I just focus on the best components?
Kuei Yang Wang,
Where are the designs that you have published?
Thanks
Ralf
What is it about that particular amp that makes it no compromise. It appears to me from the many circuits available that there is little difference by way of designs - without building them all how on earth do I know which is better?
The Walton design has a revised version which changes the supply by including additional inductors CLCLC type config - the parts have been significantly improved using Piltron TFR and better passive components plus some changes in input stages. The cost appears quite high - think he spent near $4000.
Anymore thoughts on a top design or should I just focus on the best components?
Kuei Yang Wang,
Where are the designs that you have published?
Thanks
Ralf
Konnichiwa,
Which brings me back to suggesting you read a little more, so you can make informed desicions based on your understanding, instead of following others. The other way to gain knowledge is of course empirical (eg build something) but I find it helps to understand the theory of something reasonably well before attempting to practice.
2-Stages, use of the best available Valves foir each respective Job, attention paid in Powersupply design, LCL Coupling between stages, or in sort, attention to detail....
I would not say that. The basic 6SN7-6SN7-300B and my design could not be more different in circuit and execution unless you change the output valve.
And yet he kept that dreadful 6SN7 Cascade. I could probably think of a worse sounding solution, but that's hard work.
All over. There is a list in my Thunderstoneaudio Yahoo group where also a few "un-published" rather extreme designs are available.
Sayonara
DVDHack said:
Which brings me back to suggesting you read a little more, so you can make informed desicions based on your understanding, instead of following others. The other way to gain knowledge is of course empirical (eg build something) but I find it helps to understand the theory of something reasonably well before attempting to practice.
DVDHack said:
2-Stages, use of the best available Valves foir each respective Job, attention paid in Powersupply design, LCL Coupling between stages, or in sort, attention to detail....
DVDHack said:
I would not say that. The basic 6SN7-6SN7-300B and my design could not be more different in circuit and execution unless you change the output valve.
DVDHack said:
And yet he kept that dreadful 6SN7 Cascade. I could probably think of a worse sounding solution, but that's hard work.
DVDHack said:
All over. There is a list in my Thunderstoneaudio Yahoo group where also a few "un-published" rather extreme designs are available.
Sayonara
I have been in touch with Derek for a few years. I have a pair of Tango XE20s3.5 and a pair of Tamura 10 H chokes. I am going to build this simple amplifier. I just moved again and I now have a dedicated spare bedroom for my vintage and new audio projects. For the power transformers I will prolly use the suggested universal power ones from Angela at $99 each. I have a few GZ37 rectifiers already in my tube stash. The mesh plate 300b tubes are getting rave reviews so for under 300 a pair they will be my 300b tube of choice. What speakers will you be using with these? To me the most important part of any stereo system are the speakers. My speakers are 93dB efficient and Derek told me his speakers are around 87dB efficient and his amp works with those speakers. The only other things I need are the chassis, and tube sockets, which will prolly be the wood ones from Wellborne with brass top plates. These are simple enough to build that they are worth a try even for me. I have a flex drill that will work to cut the chassis on my small wooden bench. If I screw up I will just pay a local metal shop to cut the top plates for me. I think punches seem to leave extra marks on the chassis. There is another project that Gordon Rankin gave for free to DIY the audio community called the Rankin 2A3, A.K.A. "The Baby Onganku." If your speakers only need 3 to 4 wpc these look really interesting. There is a complete suggested layout for that too with the plans in PDF format. I simply blew up the images to make them easeir to read for me. I would not use the silver transformers, just the regular Magnequest iron. I am also looking into "James" iron that is supposed to be very good sounding and quite reasonably priced. I am thinking of these for my second project for my Dad's Tad TD-4001 drivers. All they need is a few watts and the Rankin 2a3 should work well with them. I could not think of two easier first time projects that either Derek's version of the 300b, or the Rankin 2A3 mono blocks. Just keep a sense of humor and remember audio is supposed to be fun. If you are thinking of building these and want to save some dollars I have some info on a few other output transformers, and a new Tamura made c-core power transformer available at under $100. Email me and I will look for the links for you. It looks interesting but I want to use my iron. It took me long enough to save and scrounge for it.. Maybe the inexpensive iron for the 2a3 project just for fun is worth a try. Grab your soldering iron and have some fun. The only thing left for me to decide is whether or not to change the power supply to try to reduce the 5mv of hum. Again with my 93dB efficient horn speakers that would not be a problem for me anyway.
After spending some time researching the various designs it would appear that the SJS Amp below meets many of the improvements recommended here.
http://www.users.globalnet.co.uk/~valveamp/images/SE108IT2.GIF
The amp employs no feedback, uses an interstage driver TFR and a single voltage amplifier stage. The only compromise appears to be a DC heater on the 300B but I may consider an AC supply initially and update it if its too noisy (can I use the supply on the GZ37 or does it need to be a separate winding?)
What are the thoughts around this design?
Lot of Iron in this design - think I'll need a hand lifting this if I build it.
Regards
Ralf
http://www.users.globalnet.co.uk/~valveamp/images/SE108IT2.GIF
The amp employs no feedback, uses an interstage driver TFR and a single voltage amplifier stage. The only compromise appears to be a DC heater on the 300B but I may consider an AC supply initially and update it if its too noisy (can I use the supply on the GZ37 or does it need to be a separate winding?)
What are the thoughts around this design?
Lot of Iron in this design - think I'll need a hand lifting this if I build it.
Regards
Ralf
Konnichiwa,
Yes, this is a really good one, the limiting factor is the IT, I have yet to find a SE IT I find good enough in terms of resolution and bandwidth though.
This is not a compromise but a necessity. The trick is to get DC to sound good. I recommend changing the DC Heatersupply to CLC with an added common mode filter, as shown in the schematic I have referenced further up.
The Rectifier heater winding is connected to +B so using it to power the 300B heater would be quite disasterous. PLEASE read up on Valve theory before making ANY changes on established, tested designs, however innocent these changes may appear to be.
Sayonara
DVDHack said:
Yes, this is a really good one, the limiting factor is the IT, I have yet to find a SE IT I find good enough in terms of resolution and bandwidth though.
DVDHack said:
This is not a compromise but a necessity. The trick is to get DC to sound good. I recommend changing the DC Heatersupply to CLC with an added common mode filter, as shown in the schematic I have referenced further up.
DVDHack said:
The Rectifier heater winding is connected to +B so using it to power the 300B heater would be quite disasterous. PLEASE read up on Valve theory before making ANY changes on established, tested designs, however innocent these changes may appear to be.
Sayonara
DVDHack said:
Yes, it's a very good design; but rather than just tell you this, we should show you why and help you work it out for yourself.
Separate PSUs for driver stage and output stage are good for dynamics and clarity. I use this in my 'Valentine' 300B amp, right down to using separate mains TXs (supplied by Steve Shilton funnily enough).
I guess that with a shared PSU, interactions between the stages under load will give a crosstalk effect, possibly lowering dynamics due to sag of the PSU. Also, I guess the crosstalk may be a touch out of phase, giving an interference effect. OK, the effects are slight; but we're all looking for 'no compromise' designs (even if we never get there!)
The design uses 'fixed' bias, i.e. a negative voltage applied to the grid of the 300B. A very good method, often reckoned to be 'best'.
The main alternative is cathode bias (as you'll have seen in most designs) ... this generally needs a cathode bypass cap and because it has to be large, it is almost always an electrolytic cap (and these are usually of marginal quality).
Fixed bias eliminates the need for this cap.
Note a funny thing; the bias is adjustable by a pot, but it's still called 'fixed' as it's not cathode bias .... no, I don't understand why either.
Note the elegant way the interstage transformer does two things.
It applies the bias voltage to the 300B grid; and it couples the signal without needing a coupling cap.
To appreciate the design further, it's a very good idea to look up the curves for the driver valve; see if you can see why that valve was chosen. Don't be afraid to consider others or ask for suggestions; but I'm confident Steve S has made a good choice.
Think about the overall design. An issue is that we want (IMHO) to be able to drive the 300B grid by about + and - 150V (!!). Yes, I know that's 2 x the bias point ... but people say it gives a good result and my highly limited experience confirms this.
Now at first sight, 190V B+ on the driver stage looks like it won't do it? That's my only query with the design.
However, I guess that dynamic effects due to the inductance of the IT means that it's OK; I'd love an expert to comment on this!
PSUs look good but one could consider smaller caps in some places (allowing use of film type) and one of the PSUs might be better with two stages rather than a single choke; but then again it might not!
And these are small points anyway.
For ITs, I'd suggest you look into Lundahl?
I like this design a lot and am considering the same schematic for my next power amp ....
SJS looks good.Hey TL and others how about posting existing links to designs that you think exemplify the best out there.
How do you critique this design ?http://www.plitron.com/PDF/GLASSkoby.pdf
What I gather general agreement up to now is
1.Separate PS for driver
2.Fixed bias
3.AC heating
4.Valve rectification
5.Best OTX you can afford(Plitron,Tamura etc)
6.Useful to get power tx with heater windings and fixed bias windings.
I think the jury is still out on one or two driver stages or IT tx.
How do you critique this design ?http://www.plitron.com/PDF/GLASSkoby.pdf
What I gather general agreement up to now is
1.Separate PS for driver
2.Fixed bias
3.AC heating
4.Valve rectification
5.Best OTX you can afford(Plitron,Tamura etc)
6.Useful to get power tx with heater windings and fixed bias windings.
I think the jury is still out on one or two driver stages or IT tx.
Good points
"Separate PSUs for driver stage and output stage are good for dynamics and clarity. I use this in my 'Valentine' 300B amp, right down to using separate mains TXs (supplied by Steve Shilton funnily enough).
I guess that with a shared PSU, interactions between the stages under load will give a crosstalk effect, possibly lowering dynamics due to sag of the PSU. Also, I guess the crosstalk may be a touch out of phase, giving an interference effect. OK, the effects are slight; but we're all looking for 'no compromise' designs (even if we never get there!)"
DVD, I wrote an article several years ago addressing the problem of feedback through the power supply, as Mike explains above.
I agree with Wang about IT use about FR etc, but also from a distortion standpoint.
At high signal levels, transformers produce distortions, which when combined with the previous tube and output tube creates higher orders of distortions (I haven't even included the OPT).
For example, combining the 3rd harmonic from the first tube with the 2nd from the IT, and the 3rd from the output tube already produces the 18th harmonic, although fairly small. 2nd, 2nd, 2nd, produces 8th. 3rd, 2nd, 2nd produces 12th.
The amp actually is a combo of fixed and cathode bias, with a hum balancing pot.
Cheers,
"Separate PSUs for driver stage and output stage are good for dynamics and clarity. I use this in my 'Valentine' 300B amp, right down to using separate mains TXs (supplied by Steve Shilton funnily enough).
I guess that with a shared PSU, interactions between the stages under load will give a crosstalk effect, possibly lowering dynamics due to sag of the PSU. Also, I guess the crosstalk may be a touch out of phase, giving an interference effect. OK, the effects are slight; but we're all looking for 'no compromise' designs (even if we never get there!)"
DVD, I wrote an article several years ago addressing the problem of feedback through the power supply, as Mike explains above.
I agree with Wang about IT use about FR etc, but also from a distortion standpoint.
At high signal levels, transformers produce distortions, which when combined with the previous tube and output tube creates higher orders of distortions (I haven't even included the OPT).
For example, combining the 3rd harmonic from the first tube with the 2nd from the IT, and the 3rd from the output tube already produces the 18th harmonic, although fairly small. 2nd, 2nd, 2nd, produces 8th. 3rd, 2nd, 2nd produces 12th.
The amp actually is a combo of fixed and cathode bias, with a hum balancing pot.
Cheers,
Konnichiwa,
Currently I am not really aware of a truely "no-compromise" design. Closest would be a combination of the SJS & Hoekstra Axiom Design, with two LCLC Supply filter sections.
I am just "specing out" such a design for myself. Intended is the following basic outline:
Medium MU/High GM Triode like WE417A/437A, russian 6S45/6S4, german EC8010/8020 or triode wired High Transconductance Valves with an Anode load choke (50% Mu Metal laminations), ultrapath cathode decoupling with variable cathode resistor and easily changed Valve mounting plates to allow a wide range of valves to be acopmmodated, operating point probably around 10-12mA.
The +B will be switchable around 350V, 280V & 250V for 250/180/150V anode voltage with up to 100V Bias for the Output Valve. PSU will be valve rectified and LCLC using 2-section chokes trimmed for minimum parasitic capacitance within each section and maximum leakage inductance between section to serve as combined differential & common mode chokes, PSU entierley using Film Capacitors with silver mica bypass..
Output Stage direct coupled with the anode current of the driver Valve providing the output stage Bias due to the voltage drop in the drivers anode and using a 5K nominal Output transformer (but with a secondary similar to the Tango X20S, but with an improved secondary winding arrangement set to minimise the unconnected sections of the secondary).
The +B Supply will be adjustable to 450, 400, 350, 300 & 250V to allow a range of Output Valves apart from the 300B to work as well as any of the "super 300B" Valves that work at higfher Voltage & current, again dual section Chokes as above and only Film Capacitors in the Supply and the usual .
The output valve heaters will be selectable 2.5V or 2.5V/5V/7.5V DC with the DC schottky rectified and CLCLC (only electrolytic capacitors in the Amp) smoothed. The two seperate mains transformers will be made with "balanced" primaries and multiple electrostatic screens to maximise the common mode rejection and with a "low flux" design and a small airgap to elimiate susceptibility to DC and potential saturation or high distortion of the core. In both transformers a conventional rectifier heater windiong is used to damp any potential rininging in the transformer/rectifier circuit. The driverstage +B path will include a suitable relais as interlock to switch on the output Mains transformer once the driverstage is operational and the bias voltage is present.
Looks like all Iron for this project will have to be scratch developed ith each item specifically tailored to the job (like first choke as swining input choke, second choke as super wideband filter choke etc.), including a silver secondary and possibly primary on the output transformer and possibly silver winding on the driver's anode load choke.
Completely traditonal and not very interesting. Neither Mu-Follower nor White Follower (excellent measurements nonwithstanding) sound all that good and having used active regulation I know how hard it is to get good sound from it.
IF I where to recommend a DIY Design right now it would be a modified Jack Elliano DRD Amplifier from VTV, especially with a more relaxed 300B operating point, a PSU free from electrolytic capacitors and valve rectification. Or build the Marzio/Jelasi direct coupled Amplifier from Sound practices.
Not neccesarily, but surely helps. I found a RCRC filter from the main supply often sufficient occasionally even a CR filter, assuming the timconstant is high enough.
Often more trouble than it's worth. Fixed bias rarely sounds as good as well implemented self bias scheme due to the clipping behaviour. The advantage of fixed bias is the elimination of the cathode bypass capcitor. If fixed bias is used it is VERY ESSENTIAL to minimise the grid circuits (DC) resistance which increases the demands on the driver or that failing increases the overload recovery of the circuit.
Usually completely unacceptable for 5V and higher voltage valves due to excessive Hum and the IMD produced "veil" across the music. DC on the other hand can sound clinical and unpleasant. Solutions are various and will not be detailed here, look at several of muy designs to see what I found to work well.
Absolutely. At least unil High Voltage schottky diodes that glow in the dark become cheaper than valve rectifiers.
Yes. However, what constitutes the "best" output transformer and what ingrediences are needed is a matter of taste or so it seems.
Probably. Mains Transformer design is another often overlooked science. Most commonly found mains transformers are not well suited to real world operation neccesitating extensive (and expensive) external power conditioning to allow good sound. The suceptibility of a given piece of gear to mains cable changes and mains conditioners as well as the "Midnight Hour" phenomenae are strong indicators for ****poor powertransformer design. Of course, a well designed power transformer is larger, more complex to make and subsequently more expensive than an excellent SE Output transformer.
Two driver stages invariably lead to a much increased level of higher harmonics and related IMD in the signal input to the Output Valve. It is no problem with modern linelevel sources to use only two stages. With IT's I find that those which require to cope with DC bias tend to be too copmpromised in many areas to perform as intended. At low frequencies they run out of inductance and at high frequencies the large size makes winding geomertries that avoid resonances and early HF rolloff difficult to implement.
If true IT coupling is desired (as opposed to Parafeed IT or actually in essence LCL coupling) using differential driverstages allow interstage transformers with exceptional performance to made. Using a 1:1+1 input transformer, a 5687 or E182CC and a suitable high nickel content core PP Interstage 1CT:1+1 makes for an exceptional driver for either SE or PP Amplifiers, with superb level handling, squarewave performance and sonics.
Sayonara
protos said:
Currently I am not really aware of a truely "no-compromise" design. Closest would be a combination of the SJS & Hoekstra Axiom Design, with two LCLC Supply filter sections.
I am just "specing out" such a design for myself. Intended is the following basic outline:
Medium MU/High GM Triode like WE417A/437A, russian 6S45/6S4, german EC8010/8020 or triode wired High Transconductance Valves with an Anode load choke (50% Mu Metal laminations), ultrapath cathode decoupling with variable cathode resistor and easily changed Valve mounting plates to allow a wide range of valves to be acopmmodated, operating point probably around 10-12mA.
The +B will be switchable around 350V, 280V & 250V for 250/180/150V anode voltage with up to 100V Bias for the Output Valve. PSU will be valve rectified and LCLC using 2-section chokes trimmed for minimum parasitic capacitance within each section and maximum leakage inductance between section to serve as combined differential & common mode chokes, PSU entierley using Film Capacitors with silver mica bypass..
Output Stage direct coupled with the anode current of the driver Valve providing the output stage Bias due to the voltage drop in the drivers anode and using a 5K nominal Output transformer (but with a secondary similar to the Tango X20S, but with an improved secondary winding arrangement set to minimise the unconnected sections of the secondary).
The +B Supply will be adjustable to 450, 400, 350, 300 & 250V to allow a range of Output Valves apart from the 300B to work as well as any of the "super 300B" Valves that work at higfher Voltage & current, again dual section Chokes as above and only Film Capacitors in the Supply and the usual .
The output valve heaters will be selectable 2.5V or 2.5V/5V/7.5V DC with the DC schottky rectified and CLCLC (only electrolytic capacitors in the Amp) smoothed. The two seperate mains transformers will be made with "balanced" primaries and multiple electrostatic screens to maximise the common mode rejection and with a "low flux" design and a small airgap to elimiate susceptibility to DC and potential saturation or high distortion of the core. In both transformers a conventional rectifier heater windiong is used to damp any potential rininging in the transformer/rectifier circuit. The driverstage +B path will include a suitable relais as interlock to switch on the output Mains transformer once the driverstage is operational and the bias voltage is present.
Looks like all Iron for this project will have to be scratch developed ith each item specifically tailored to the job (like first choke as swining input choke, second choke as super wideband filter choke etc.), including a silver secondary and possibly primary on the output transformer and possibly silver winding on the driver's anode load choke.
protos said:
Completely traditonal and not very interesting. Neither Mu-Follower nor White Follower (excellent measurements nonwithstanding) sound all that good and having used active regulation I know how hard it is to get good sound from it.
IF I where to recommend a DIY Design right now it would be a modified Jack Elliano DRD Amplifier from VTV, especially with a more relaxed 300B operating point, a PSU free from electrolytic capacitors and valve rectification. Or build the Marzio/Jelasi direct coupled Amplifier from Sound practices.
protos said:
Not neccesarily, but surely helps. I found a RCRC filter from the main supply often sufficient occasionally even a CR filter, assuming the timconstant is high enough.
protos said:
Often more trouble than it's worth. Fixed bias rarely sounds as good as well implemented self bias scheme due to the clipping behaviour. The advantage of fixed bias is the elimination of the cathode bypass capcitor. If fixed bias is used it is VERY ESSENTIAL to minimise the grid circuits (DC) resistance which increases the demands on the driver or that failing increases the overload recovery of the circuit.
protos said:
Usually completely unacceptable for 5V and higher voltage valves due to excessive Hum and the IMD produced "veil" across the music. DC on the other hand can sound clinical and unpleasant. Solutions are various and will not be detailed here, look at several of muy designs to see what I found to work well.
protos said:
Absolutely. At least unil High Voltage schottky diodes that glow in the dark become cheaper than valve rectifiers.
protos said:
Yes. However, what constitutes the "best" output transformer and what ingrediences are needed is a matter of taste or so it seems.
protos said:
Probably. Mains Transformer design is another often overlooked science. Most commonly found mains transformers are not well suited to real world operation neccesitating extensive (and expensive) external power conditioning to allow good sound. The suceptibility of a given piece of gear to mains cable changes and mains conditioners as well as the "Midnight Hour" phenomenae are strong indicators for ****poor powertransformer design. Of course, a well designed power transformer is larger, more complex to make and subsequently more expensive than an excellent SE Output transformer.
protos said:
Two driver stages invariably lead to a much increased level of higher harmonics and related IMD in the signal input to the Output Valve. It is no problem with modern linelevel sources to use only two stages. With IT's I find that those which require to cope with DC bias tend to be too copmpromised in many areas to perform as intended. At low frequencies they run out of inductance and at high frequencies the large size makes winding geomertries that avoid resonances and early HF rolloff difficult to implement.
If true IT coupling is desired (as opposed to Parafeed IT or actually in essence LCL coupling) using differential driverstages allow interstage transformers with exceptional performance to made. Using a 1:1+1 input transformer, a 5687 or E182CC and a suitable high nickel content core PP Interstage 1CT:1+1 makes for an exceptional driver for either SE or PP Amplifiers, with superb level handling, squarewave performance and sonics.
Sayonara
Kuei
Apologies if this is offtopic but i have recently developed a serious interest towards balanced PP amps as exemplified by Lynn Olson's Aurora, John Chapman's design, Marantz T1 and one or two of your own designs.
As building any of these is a serious undertaking i am still trying to get a handle on the expected sound. Lynn will have us believe it is actually better than SET and sliced bread at the same time, but based on my impressions of the Ariels i have a lingering doubt we may not share the same sonic preferences.
Have you actually built or listened to such a design? Using reasonble, but not sensational iron, as shown in the Aurora's schematics, can one really expect a wide subjective bandwidth? SE type midrange magic?
Interestingly the T1 uses parafeed after the first stage. Why?
Do you reckon the heroic effort of building one of these is really worth it or would a high-voltage high power SET be a better bet.
I'll probably buy the iron and do some tests anyway. Most likely with the mesh plate 300B from Valve Art as it seems reasonably priced.
Thanks.
Apologies if this is offtopic but i have recently developed a serious interest towards balanced PP amps as exemplified by Lynn Olson's Aurora, John Chapman's design, Marantz T1 and one or two of your own designs.
As building any of these is a serious undertaking i am still trying to get a handle on the expected sound. Lynn will have us believe it is actually better than SET and sliced bread at the same time, but based on my impressions of the Ariels i have a lingering doubt we may not share the same sonic preferences.
Have you actually built or listened to such a design? Using reasonble, but not sensational iron, as shown in the Aurora's schematics, can one really expect a wide subjective bandwidth? SE type midrange magic?
Interestingly the T1 uses parafeed after the first stage. Why?
Do you reckon the heroic effort of building one of these is really worth it or would a high-voltage high power SET be a better bet.
I'll probably buy the iron and do some tests anyway. Most likely with the mesh plate 300B from Valve Art as it seems reasonably priced.
Thanks.
What are we talking here? Sowter? S&B? other?
Stuff I consider highly without have real experience are Lynn Olson's ideas...
Valve rectification, (pseudo) choke input supplies, electrostatic screens on transformers. And also some stuff JamesD wrote over at K&K on how he has always tried to optimize each stage and isolate each stage as well as possible from the other (p (This is where an interstage transformer comes into the picture (hope I am understanding James well enough to explain it like this.)
further I also "believe" in wooden chassis...solid core wiring...carbon resistors for signal applications.. star earthing...
Mmmm can't think of anything else..at the moment....
[edit]also posted just after analog_sa and am also interested on the "new" PP renaissance..Lynn, Gary Pimm and Allan Wright being some of the people (each in their own way) that I know of that are pioneering the PP topology...(forgive my ignorance)
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.