Hi, Frank -
As far as output stage regulation goes, I'm allergic to silicon, so I'll pass on that approach, but if I had time, a swinging choke input filter or ferroresonant supply approach sounds like it might perhaps have possibilities. Right now, all the supplies more or less track🙂
I have a 7500 uF 200Vdc capacitance followed by a handwound choke (only a few millihenries using 13ga magnet wire), then a 12,000 uF 200Vdc for each supply polarity paralleled with 80uF polypropylenes for each output rail of the dc coupled OTL. All these 'lytics are 3" dia x 5 1/8" tall, btw. With today's new high CV product lytic's, I could probably find something with two or three times the capacitance in the same sized package if I looked, but I'm afraid that I'd upset the amp's turn on or turn off characteristics and I don't have enough room to scale the input and driver stage supplies filter capacitances by a similar ratio and stay with the polypropylenes in oil that I'm currently using.
Speaking of new 'lytics, I managed to snag 4 47,000uF 63Vdc 'lytics for my hybrid 6 channel amplifier, and these babies are only 2" dia x 4 1/4" tall. Pretty impressive CV product for the size.
As far as output stage regulation goes, I'm allergic to silicon, so I'll pass on that approach, but if I had time, a swinging choke input filter or ferroresonant supply approach sounds like it might perhaps have possibilities. Right now, all the supplies more or less track🙂
I have a 7500 uF 200Vdc capacitance followed by a handwound choke (only a few millihenries using 13ga magnet wire), then a 12,000 uF 200Vdc for each supply polarity paralleled with 80uF polypropylenes for each output rail of the dc coupled OTL. All these 'lytics are 3" dia x 5 1/8" tall, btw. With today's new high CV product lytic's, I could probably find something with two or three times the capacitance in the same sized package if I looked, but I'm afraid that I'd upset the amp's turn on or turn off characteristics and I don't have enough room to scale the input and driver stage supplies filter capacitances by a similar ratio and stay with the polypropylenes in oil that I'm currently using.
Speaking of new 'lytics, I managed to snag 4 47,000uF 63Vdc 'lytics for my hybrid 6 channel amplifier, and these babies are only 2" dia x 4 1/4" tall. Pretty impressive CV product for the size.
fdegrove said:
Walking the regulation path seems to have a certain persuasive logic about it: if power supply regulation improves the sound of preamps (name a preamp lacking power supply regulation) (oops, name a good sounding preamp lacking power supply regulation), it seems not a large leap to think power supply regulation would improve the sound of an output stage. But perhaps that puts one, then, into Krell territory? Perhaps we need PWM-based output power supply regulation? How would that be for combining modern and traditional! Call it the Futterman Super Switcher OTL.
Maybe it is just me but for me the idea of designing an OTL amp is to get rid of the limitations due to output transformers in traditional tube amps. Introducing interstage transformers in such an amp in places where they are not necessary seem to somewhat defy the purpose of designing the OTL amp in the first place.
Yes, I agree that you can get rid of some components by using a interstage transformer as phase splitter but I dont agree that this justify the use of a transformer with all its limitations. Using an interstage transformer in an OTL amp will severly limit the frequency response and also limit the amount of feedback that can be applied without affecting stability, I cant see how this is something that can be justified.
What are in your mind the big advantages with using interstage transformers, (with their well known limitations) that would justify its use when it is quite easy to design a well performing OTL amp completely without them?
Regards Hans
My OTL
Thanks for your deep thoughts on the tube choices. I remember reading about a person who has kept the heater on of a 6c33c for a week non stop to make the bias stabilise, due to the high transconductance of the tube. Perhaps judging by the very low cost and few tubes required to build a decent OTL with 6C33cs the latter would be a good choice.
As far as the design is concerned, the driver section can be found in Audio Reality and is patented and copyrighted therefore I cannot present the design as my own, nor attach it without the consent of the author. But basically I will do away with the biasing scheme of Rozenblit, follow Frank's advise of individual biasing of the 6c33c as well as the bias supply design offered by Ciuffoli.
The power section of the tubes will be exactly the same as the Ciuffoli 8.2 ie the caps the chokes (4KG each) the power transormer 1KVA. For capacitance I will use 8 5900 uf 200v (bloody huge). Furthermore I would add some choke filtering in the B+ section of the driver and use twice as much capacitance in this section to aim a C L C circuit as it is a voltage doubler scheme.
Rozenblit states in his book that voltage regulation has no obvious advantages there will do away with that.
Basically I am aiming for the best parts of the two designs since definitely it would be a shame to avoid using a patented driver design which empirically is superior.
Thanks for your deep thoughts on the tube choices. I remember reading about a person who has kept the heater on of a 6c33c for a week non stop to make the bias stabilise, due to the high transconductance of the tube. Perhaps judging by the very low cost and few tubes required to build a decent OTL with 6C33cs the latter would be a good choice.
As far as the design is concerned, the driver section can be found in Audio Reality and is patented and copyrighted therefore I cannot present the design as my own, nor attach it without the consent of the author. But basically I will do away with the biasing scheme of Rozenblit, follow Frank's advise of individual biasing of the 6c33c as well as the bias supply design offered by Ciuffoli.
The power section of the tubes will be exactly the same as the Ciuffoli 8.2 ie the caps the chokes (4KG each) the power transormer 1KVA. For capacitance I will use 8 5900 uf 200v (bloody huge). Furthermore I would add some choke filtering in the B+ section of the driver and use twice as much capacitance in this section to aim a C L C circuit as it is a voltage doubler scheme.
Rozenblit states in his book that voltage regulation has no obvious advantages there will do away with that.
Basically I am aiming for the best parts of the two designs since definitely it would be a shame to avoid using a patented driver design which empirically is superior.
I can only think that if bias take this long to stabilise there is a problem with the tube sockets, (high contact resistance affecting the heater supply).
I have always initially used heater only for about 30 minutes and then anode voltage with 200mA current for 1 hour, after that I didn't have any problems with bias drift, however as the transconductance is high it is important that the bias supply itself is stable otherwise the current will of course drift.
It can be argued, (and has been by for instance John Broskie) if actually the Bruce R solution solve the problem itself, (or what he describe as a problem) a patent by itself is not a guarantee of superior function.
Regards Hans
Hans,
You are right that a patent is no guarantee of a superior design nor sound quality but the reviews of the T8 have been outstanding and it is a safer bet than Ciuffoli's design who happens to be an IT expert and not an engineer.
Bruce's patent was filed with 6c33cs not EL509 and his comments on the tube (6c33c) have been to the effect that the tube is heavily unreliable hence my original plea for assistance to everybody!
Furthermore Bruce's driver design gives the flexibility in case 6c33c proves unsuitable to switch to a EL 509s with minimum hassle (or 6c19pi or 6AS7 or ...).
Well there is one way of finding out, and I'll keep you posted of the results. Furthermore, what is the joy of DIY if one cannot experiment...
You are right that a patent is no guarantee of a superior design nor sound quality but the reviews of the T8 have been outstanding and it is a safer bet than Ciuffoli's design who happens to be an IT expert and not an engineer.
Bruce's patent was filed with 6c33cs not EL509 and his comments on the tube (6c33c) have been to the effect that the tube is heavily unreliable hence my original plea for assistance to everybody!
Furthermore Bruce's driver design gives the flexibility in case 6c33c proves unsuitable to switch to a EL 509s with minimum hassle (or 6c19pi or 6AS7 or ...).
Well there is one way of finding out, and I'll keep you posted of the results. Furthermore, what is the joy of DIY if one cannot experiment...
Definitive Schematic – Frank, Hans?
HI,
I've been following this thread for some while hoping someone would upload what they considered a great sounding circuit. Although I've had a stash of 6C33's for some years, my interest in circlotron OTL really took of after reading about the Tenor amps.
Any links?
Mike
HI,
I've been following this thread for some while hoping someone would upload what they considered a great sounding circuit. Although I've had a stash of 6C33's for some years, my interest in circlotron OTL really took of after reading about the Tenor amps.
Any links?
Mike
tubetvr said:
The purpose of designing any amp is to get the best sound possible. Ruling out any design element from any given stage, for what appear in this case to be almost religious reasons, is generally counterproductive.
A decent (unfortunately, read EXPENSIVE) IT should be flat to well above 50kHz. Using an IT doesn't limit the ammount of feedback one can use, only the ammount of global feedback, which in my experience should be avoided anyway.
Depends on the rest of the design. It's quite difficult to design a phase splitter that balances as well as a good IT. I've never heard a coupling cap that is as transparent as even a cheap IT, but I've never heard an IT that's as transparent as a good direct coupling implementation. Of course, a properly balanced direct coupled phase splitter and output stage is going to make the power supply get a bit more interesting.
Bottom line here is that you need to try things out before you comment on whether they're good or bad. Saying that an amp using (fill in the blank) is bad merely because it uses (fill in the blank) is just plain foolish. There are good amps out there based on a wide range of techniques, and an even greater number of so-so at best amps based on each of those techniques. It all comes down to how you implement those techniques...
Peace
Re: My OTL
Nice experimentation, Nicos. I have heard and appreciate the positive qualities of a beefy power supply. I've lately been taking a more careful glance at the downsides of high supply capacitance, especially on a capacitor input supply. With higher capacitance comes higher power supply voltage, and therefore a higher charging current. With a higher charging current comes, among other effects, a lower power factor (never good) and a higher voltage drop across resistive elements, not to mention stress-related distortion products. The voltage drop, of course, causes amusical harmonic distortion that affects not only the amplifier in question, but other components connected to the same AC source.
The above considerations perhaps place upon or highlight a hard limit of C-input design, at least in an unregulated supply. One solution would be to filter the incoming AC in such a way as to reduce AC supply harmonics, but that's an entire other discussion.
Nicos said:
Nice experimentation, Nicos. I have heard and appreciate the positive qualities of a beefy power supply. I've lately been taking a more careful glance at the downsides of high supply capacitance, especially on a capacitor input supply. With higher capacitance comes higher power supply voltage, and therefore a higher charging current. With a higher charging current comes, among other effects, a lower power factor (never good) and a higher voltage drop across resistive elements, not to mention stress-related distortion products. The voltage drop, of course, causes amusical harmonic distortion that affects not only the amplifier in question, but other components connected to the same AC source.
The above considerations perhaps place upon or highlight a hard limit of C-input design, at least in an unregulated supply. One solution would be to filter the incoming AC in such a way as to reduce AC supply harmonics, but that's an entire other discussion.
Serengetiplains
you forgot to mention the weight of it all. Had I known that I would get so much iron I would not have gone through this avenue...
Just the transformers weigh 30KG, you can guess about the total
you forgot to mention the weight of it all. Had I known that I would get so much iron I would not have gone through this avenue...
Just the transformers weigh 30KG, you can guess about the total
I think what you say can easily be twisted around, it is difficult to explain the recent popularity of interstage transformers maybe this is due to religous reasons?
It seems that our experiences differ, I have no problem with global feedback as long as I can make the amplifier behave well in open loop, it is well known that for instance Futtermans OTL amplifiers used heavy global feedback and still have a reputation of very good sound. Global feedback is almost a prerequisite when designing an OTL amp in order to reduce the output impedance to reasonable values without an unnecessary amount of output tubes
I dont agree with you regarding the difficulty to design a phase splitter with good balance, a simple split load phase inverter is as balanced as you need it especially if the loading is the same on the 2 outputs, even a transformer is sensitive for loading.
I have never heard or measured an IT that can be called transparent, are you not mixing up transparency and euphonic distorsion?
Thaknk you for the advice but it is not needed, my road to OTL amps have been long, (>35 years) and I have went trough many different amp technologies before I have reached what for me is the best sounding technology, I have tried building transformer coupled amps with and without interstage transformer coupling, single ended and push-pull and I have listened to more tube amps here in Tokyo and elsewhere then I care to remember.
I dont think I have been saying that an amp is bad just because it use transformers in the signal path, what I said is that it is possible to design better amps without these. There are many amplifiers that sound good both with and without transformers the question is which sounds better and why.
It seems that i stepped in on a sensitive issue, (as I expected) when I started to question the use of interstage transformers and chokes in the signalpath for me it would be interesting to hear othet peoples experiences of inductive components and what advantages or disadvantages they have from an objective point of view.
Regards Hans
Hi,
Tubetvr,
I entirely agree with you on this.
There's no iron like no iron just as there's no cap like no cap.
While the iron is easy enough to avoid, avoiding coupling caps, be that directly or indirectly, is another chapter altogether.
I don't think I can find any iron that is as broadband and phase linear as I can make an OTL and it shows up in the listening experience as well.
Just like yourself I've dwelled on just about anything containing a few tubes, even some transistor based designs and finally settled for what sounded most neutral to me: an OTL amp.
That this isn't to everybody's taste I can easily understand but I see little point in going OTL to litter it with iron in the signal path...
Cheers,😉
Tubetvr,
I entirely agree with you on this.
There's no iron like no iron just as there's no cap like no cap.
While the iron is easy enough to avoid, avoiding coupling caps, be that directly or indirectly, is another chapter altogether.
I don't think I can find any iron that is as broadband and phase linear as I can make an OTL and it shows up in the listening experience as well.
Just like yourself I've dwelled on just about anything containing a few tubes, even some transistor based designs and finally settled for what sounded most neutral to me: an OTL amp.
That this isn't to everybody's taste I can easily understand but I see little point in going OTL to litter it with iron in the signal path...
Cheers,😉
You can see the principal schematic on my amp here http://www2.gol.com/users/tube/otl.html
it is not a circlotron, I believe that the Futterman variation give better result.
For more details send me an E-mail
Regards Hans
Btw, my dc coupled OTL driver circuit resembles none of those in the previous link at all. And my design is not a circlotron, either. In fact, I've never seen a driver circuit like the one I've used in my dc coupled OTL design for the last 16 years used anywhere else.😀
I don't want to say much more about it at this time, because it may be worth patenting, since it provides both balanced drive and dc coupling to a totem pole output stage.
I don't want to say much more about it at this time, because it may be worth patenting, since it provides both balanced drive and dc coupling to a totem pole output stage.
Would a White follower work as an OTL SE output stage?
http://www.tubecad.com/october99/page4.html
I have this idea that the lower triode could be substituted by a power Mosfet. I´ve seen it done somewhere in a SE mosfet amp.
What do you guys think about, say, two 6AS7 (four sections) in //
Iq ~ 400mA , B+ 80-100V.
The negative rail (not on the schematic, but necessary if we want DC coupled output) could be lower, say -40V or so if a Mosfet is used.
Should be good for a couple of watt or so if my maths are correct.
http://www.tubecad.com/october99/page4.html
I have this idea that the lower triode could be substituted by a power Mosfet. I´ve seen it done somewhere in a SE mosfet amp.
What do you guys think about, say, two 6AS7 (four sections) in //
Iq ~ 400mA , B+ 80-100V.
The negative rail (not on the schematic, but necessary if we want DC coupled output) could be lower, say -40V or so if a Mosfet is used.
Should be good for a couple of watt or so if my maths are correct.
Flutterman vs Circlotron?
"I believe that the Futterman variation give better result. "
Thanks Hans,
Did you arrive at this opinion by trying both circuits? I am curious because I haven't seen much discussion in this thread about which one is generally preferred. I had gotten the impression that circlotron was considered better.
Mike
"I believe that the Futterman variation give better result. "
Thanks Hans,
Did you arrive at this opinion by trying both circuits? I am curious because I haven't seen much discussion in this thread about which one is generally preferred. I had gotten the impression that circlotron was considered better.
Mike
Yes I have tried the circlotron but only as an experiment. Actually it is not so big difference between a series connected output stage and a circlotron it only differs in the way where the voltage sources are located in the circuit.
The attractive function of a circlorton for many is that is balanced but balance to a good enough degree can be achieved in other ways and for me other parameters are more important then to achieve balance in a circuit just for its own sake. I reach equal or better values of distorsion with the output stage I have chosen then for the same tubes in a circlotron coupling so for me the circlotron doesn't offer me any advantages.
In order to achieve high reliability in an OTL tube amp I believe it is important to minimise the number of tubes used, if low output impedance is a goal the inverted Futterman circuit then offers lowest possible practical output impedance or roughly half of the Circlotron for the same number of tubes, that for me that is a very important parameter.
there are other disadvantages with the circlotron circuit as the need for 2 floating power supplies or the fact that the speaker terminals are live but people make different choices based on different thinking.
Regards Hans
Thanks Hans,
I admit that the possibility of DC on the speaker terminals with the corclotron was not an inviting possibility.
I did a search of your other posts and found the schematic with your circuit values on it - am I correct in assuming that these measurements are with the 12AT7 instead of the E188CC?
Also, noting that you say that you are very satisfied with the sound, what amps have you compared your version of 8.2 to?
Mike
I admit that the possibility of DC on the speaker terminals with the corclotron was not an inviting possibility.
I did a search of your other posts and found the schematic with your circuit values on it - am I correct in assuming that these measurements are with the 12AT7 instead of the E188CC?
Also, noting that you say that you are very satisfied with the sound, what amps have you compared your version of 8.2 to?
Mike
Yes, it is possible to get something out of it but it is not really well made for higher output power as its function depend on the use of an anode resistor that will waste power, (it is possible to get rid of the cathode resistor but not the anode resistor). The other disadvantage is that it only works in class A which seriously limits the output power.
As far as I know no one has used a White cathode follower in a commercial OTL, (even if Morgan Jones in his book mention these as an alternative for OTL's but I dont really think he has thought a lot about how to build practical OTL amps)
Regards Hans
I have never used 12AT7 or E188CC in my OTL but I have used 12AU7 and 12BH7 as driver tubes.
My amp is not as a version of the ciufioli 8.2 the only similarity is that is uses a output stage connected the same way, the performance of these 2 amps are quite different as what I understand.
I have compared my amp to several amps both tube and SS, SE and push-pull, for me I cant find any thing another amp do better but that is what I think.
Regards Hans
- Status
- Not open for further replies.