Hi,
I think that what you saw was a SEPP stage.
That'll depend on the plate voltage you use but in general a pair of 6C33-Cs can yield about 25W in PP.
Cheers,😉
I think that what you saw was a SEPP stage.
That'll depend on the plate voltage you use but in general a pair of 6C33-Cs can yield about 25W in PP.
Cheers,😉
Hi, all -
This may not be exactly the right place to post about my neither fish nor fowl hybrid amp, but after a hiatus, I have got back onto it and have hooked up a beefed up version of my filament regulator circuit to the 6K11 amp prototype run as only a single long tailed differential voltage gain stage driving a cathode follower. The early results are encouraging as far as dc offset stability go (It's a dc coupled design). After several hours of warmup after originally having set the offset to be within five millivolts, on at least two power cycles with several hours off, then on again, the dc offset has settled at less than 10 millivolts each time. Of course, it takes between 10 minutes and half an hour to settle in this range, but it's still 30-40 mv or better within a minute or so of powering up.
This looks like 'good enough' stability to proceed with for a practical design, but I'll cycle it a few more times just to decrease the chances I'm 'missing something'.🙂
This may not be exactly the right place to post about my neither fish nor fowl hybrid amp, but after a hiatus, I have got back onto it and have hooked up a beefed up version of my filament regulator circuit to the 6K11 amp prototype run as only a single long tailed differential voltage gain stage driving a cathode follower. The early results are encouraging as far as dc offset stability go (It's a dc coupled design). After several hours of warmup after originally having set the offset to be within five millivolts, on at least two power cycles with several hours off, then on again, the dc offset has settled at less than 10 millivolts each time. Of course, it takes between 10 minutes and half an hour to settle in this range, but it's still 30-40 mv or better within a minute or so of powering up.
This looks like 'good enough' stability to proceed with for a practical design, but I'll cycle it a few more times just to decrease the chances I'm 'missing something'.🙂
Hi,
If I'm reading you correctly you're direct coupling a 12AX7 LTP into a 1/2 12AU7A CF...
Which is than cap coupled to the input of a FET amp or something similar, I assume.
Either way I'm happy to see it seems to work fine but what's the heater supply to do with it?
Cheers,😉
If I'm reading you correctly you're direct coupling a 12AX7 LTP into a 1/2 12AU7A CF...
Which is than cap coupled to the input of a FET amp or something similar, I assume.
Either way I'm happy to see it seems to work fine but what's the heater supply to do with it?
Cheers,😉
I am getting closer to building a little (5w) otl and would like a little feedback (no pun intended) I am thinking about just using
one 6c33c-b as a cathode follower using a 70 ohm resistor. If I
set the bias at 350 ma shouldn`t this be ok for about 5 watts?
Thanks,
Woody
one 6c33c-b as a cathode follower using a 70 ohm resistor. If I
set the bias at 350 ma shouldn`t this be ok for about 5 watts?
Thanks,
Woody
I am thinking about only using this as a midrange amp driving a
pair of serse connected 16 ohm jbl 2123j`s on a open baffel.
So the combined re. would be about 17 ohms.
Thanks,
Woody
pair of serse connected 16 ohm jbl 2123j`s on a open baffel.
So the combined re. would be about 17 ohms.
Thanks,
Woody
If your load impedance is 16 ohm as you mention I think your output power would be max ~1W. As you are using a single tube CF the max peak current can be same as the idle current, in your case 350mA so power output will be Ipk^2*RL/2, remember a single tube can only go in class A. In reality you get a lot of distorsion even at 1W so useable power is even lower, class A is not the way to go for an OTL amp, for higher power any OTL need to go in class AB.
In my OTL the idle current is ~200mA but the peak current is ~2.5A so the output power is 2.5^2*8/2 = 25W in 8ohm.
BR Hans
Hi, Frank -
It seems that regulating the heater supply voltage decreases the offset drift with this circuit that would otherwise result from mains supply variations. I've already come up with an unregulated plate supply topology that does pretty well in this regard, with a 20% drop in mains voltage only causing about a 25mV offset change. I'm trying by these means to achieve decent dc stability with low feedback and a not fully balanced circuit design. As far as the mosfets go, I plan to dc couple them as followers for an amp that is unity gain at 0 hz. The eventual goal is to have six 150 watt channels like this in the final amplifier. I've already ordered and received the Profusion PLC 8A 200V types, so soon I'll be attaching the output stage (and output mute circuit) to this prototype.
transformer coupling (again)
Say, I just came across the following interesting article on distortion in tubes that rang true with my own experiences:
http://www.nutshellhifi.com/library/FindingCG.html
In the article, Lynn Olson measures distortion components in a 6SN7 tube connected in various SRPP and transformer coupled modes. Quite interestingly, transformer coupling gave average lower levels of harmonic distortion and, what's more, gave, in Olson's opinion, a more sonically pleasing distortion spectrum (ratio of various harmonics), resistor/capacitor coupling giving the worst. Olson's thoughts match those I read many years ago in an article suggesting our brain filters or masks distortion if the next higher harmonic is some percentage lower than the previous. The article suggested the harmonic spectrum produced by a tube in a class A circuit (declining curve) most closely approximates the distortion spectrum the ear finds most pleasing (ie, the spectrum that allows the ear/brain mechanism to effectively mask distortion), which evidently is the same spectrum the ear and air actually produce in the listening, transmission process. This particular ear/brain theory explains why tube amplifiers resolve low level details better than SS varieties despite having a hundred or thousand or more times greater distortion.
The point seems relevant, to me, to choosing a phase splitter in an OTL amplifier and why transformer coupling might be an attractive option. I'd be interested to know if anyone has actually *heard* a well-implemented such variety.
Say, I just came across the following interesting article on distortion in tubes that rang true with my own experiences:
http://www.nutshellhifi.com/library/FindingCG.html
In the article, Lynn Olson measures distortion components in a 6SN7 tube connected in various SRPP and transformer coupled modes. Quite interestingly, transformer coupling gave average lower levels of harmonic distortion and, what's more, gave, in Olson's opinion, a more sonically pleasing distortion spectrum (ratio of various harmonics), resistor/capacitor coupling giving the worst. Olson's thoughts match those I read many years ago in an article suggesting our brain filters or masks distortion if the next higher harmonic is some percentage lower than the previous. The article suggested the harmonic spectrum produced by a tube in a class A circuit (declining curve) most closely approximates the distortion spectrum the ear finds most pleasing (ie, the spectrum that allows the ear/brain mechanism to effectively mask distortion), which evidently is the same spectrum the ear and air actually produce in the listening, transmission process. This particular ear/brain theory explains why tube amplifiers resolve low level details better than SS varieties despite having a hundred or thousand or more times greater distortion.
The point seems relevant, to me, to choosing a phase splitter in an OTL amplifier and why transformer coupling might be an attractive option. I'd be interested to know if anyone has actually *heard* a well-implemented such variety.
Hi,
Interesting you mention those theories today, this afternoon I reread Joe Roberts article for Sound Practices wherein he describes the virtues of the SE amp giving more or less the same reasoning as Olson and yourself suggest.
It has a lot to with harmonic correctness and the way we perceive this as natural sound production.
As for the implentation of a xformer as a phasesplitter in an OTL, this certainly is something worth trying even though I can't think of anyone that's actually tried this out.
One could of course argue that it wouldn't be an OTL anymore but that's another issue...
Finding a splitter xformer capable of faithfully passing, say 5Hz upto 300kHz is yet another.
Very interesting indeed,😉
Interesting you mention those theories today, this afternoon I reread Joe Roberts article for Sound Practices wherein he describes the virtues of the SE amp giving more or less the same reasoning as Olson and yourself suggest.
It has a lot to with harmonic correctness and the way we perceive this as natural sound production.
As for the implentation of a xformer as a phasesplitter in an OTL, this certainly is something worth trying even though I can't think of anyone that's actually tried this out.
One could of course argue that it wouldn't be an OTL anymore but that's another issue...
Finding a splitter xformer capable of faithfully passing, say 5Hz upto 300kHz is yet another.
Very interesting indeed,😉
You would of course need a dual secondary IT wound very tightly --- an interesting project for our friends at AE, no doubt. "Say, can you wind me a dual secondary 3dB down at 4 and 300KHz?" Another benefit of transformer coupling, also applicable somewhat to the OTL design, is it allows one to tie the grids of the next stage directly to ground, if one so chooses. Grid current flowing in the output stage would then muck up the drive signal to a lesser degree.
The articles to which I referred were written by Scott Frankland and were published in 1996 annd 1997 in Stereophile. They were titled "Single-Ended vs Push-Pull." I personally do not advocate any given topology over another --- single-ended or push-pull or what have you (tube vs ss is another matter entirely) --- and am more interested in the benefits each might offer and in questioning how a given topology might be best employed.
The articles to which I referred were written by Scott Frankland and were published in 1996 annd 1997 in Stereophile. They were titled "Single-Ended vs Push-Pull." I personally do not advocate any given topology over another --- single-ended or push-pull or what have you (tube vs ss is another matter entirely) --- and am more interested in the benefits each might offer and in questioning how a given topology might be best employed.
Hi,
My view on the subject.
I don't see that many benefits with using a transformer as phase splitter, the lower distorsion argument is valid in itself maybe, but in a good OTL design as in any other power amp the distorsion of the driver stages will be much lower then from the power amp and will not affect the total distorsion or the distorsion spectra.
I have seen the article by Lynn Olson and I have always thought it is flawed as it compares apples and oranges. The problem is that in the tests the anode voltage for each tube is not the same and then it is not strange that the circuit that allows highest voltage gives lower distorsion, e.g in the SRPP circuit the volatge over each tube is ~225V and in the transformer coupled circuit ~325V, so the article doesn't give a clear answer which circuit gives the lowest distorsion but maybe the answer to which circuit gives lowest distorsion given a certain supply voltage, which is not the same thing.
I still think that it would be better to build OTL circuits totally without iron, one of the main reasons to build OTL amps is usually to get rid of the output iron due to issues like limited frequency range, distorsion aso, the same arguments also holds for use of iron in any other place inside an OTL amplifier, that is at least my view on the subject.
Regards Hans
My view on the subject.
I don't see that many benefits with using a transformer as phase splitter, the lower distorsion argument is valid in itself maybe, but in a good OTL design as in any other power amp the distorsion of the driver stages will be much lower then from the power amp and will not affect the total distorsion or the distorsion spectra.
I have seen the article by Lynn Olson and I have always thought it is flawed as it compares apples and oranges. The problem is that in the tests the anode voltage for each tube is not the same and then it is not strange that the circuit that allows highest voltage gives lower distorsion, e.g in the SRPP circuit the volatge over each tube is ~225V and in the transformer coupled circuit ~325V, so the article doesn't give a clear answer which circuit gives the lowest distorsion but maybe the answer to which circuit gives lowest distorsion given a certain supply voltage, which is not the same thing.
I still think that it would be better to build OTL circuits totally without iron, one of the main reasons to build OTL amps is usually to get rid of the output iron due to issues like limited frequency range, distorsion aso, the same arguments also holds for use of iron in any other place inside an OTL amplifier, that is at least my view on the subject.
Regards Hans
No, I dont know if anyone even exist, but neither you, I or anyone else have heard all possible amplifier configurations that is possible to build.
Let me guess, maybe you are one of those who believe, (based on listening experiences or other reasons) that transformer coupled or inductor coupled amps have some sonical benefits.
I must say that I don't belong among that group and I base my belief both on extensive listenening experiences as well as circuit theory and measurements. For me transformer coupled amps never give me anything better, (but often worse) then what I get from a amplifier designed either with as good output transformer as possible, (but without interstage transformers) or a good OTL amp.
But as I said before these are my views and others can have other experiences and preferences, one of the great benefits of DIYAUDIO.COM is that you can hear the views of many others, you don't have to agree with everyone.
Regards Hans
I know this post doesn't exactly belong in 'OTL', but the hybrid amp I'm working on sporadically is like a dc coupled OTL, if you ignore the mosfet output stage, which, incidentally, I finally got around to adding to my prototype just this evening. No big surprises here, but I wound up needing three, not two diode drops to bias the single pair of Exicon EC-10N20 and EC-10P20 8A 200V 125W lateral mosfets on the prototype at the moment up to over 100mA.
I also must say that I'm a little disappointed in the fact that Exicon lists only the dc SOA for these devices, yet they label the SOA curve '10mS', which surely must be an error. I really wanted to use the Toshiba 2SK1035 and 2SJ201 for this project, btw, but Toshiba announced they are not going to support these devices in the US any longer which made me a little wary of their possible longevity as supported devices elsewhere in the world. As of a result of Exicons lack of peak current specifications for their parts, I'm seriously considering using no fewer than FIVE pair per channel for my 150 W/channel six channel HT amplifier because I want it to swing to the rails into 4 ohms (and allow two pairs of channels to work properly in bridged mode as a biamp option) yet not risk smoking the output devices.
However, a somewhat offsetting advantage of these parts is supposed to be the low current at which a negative temperature coefficient begins (around 100ma per device), which simplifies thermal protection. However, I don't want the quiescent bias to drop off *too* fast when the amp warms up, so I will likely work up a biasing scheme that is slightly more complex than a string of two or three diodes - probably something using a transistor and also a themistor if necessary on the heatsink with bias adjust and possibly even a tempco trim - still need to work out the details🙂
Anybody else have experience with these parts?
I also must say that I'm a little disappointed in the fact that Exicon lists only the dc SOA for these devices, yet they label the SOA curve '10mS', which surely must be an error. I really wanted to use the Toshiba 2SK1035 and 2SJ201 for this project, btw, but Toshiba announced they are not going to support these devices in the US any longer which made me a little wary of their possible longevity as supported devices elsewhere in the world. As of a result of Exicons lack of peak current specifications for their parts, I'm seriously considering using no fewer than FIVE pair per channel for my 150 W/channel six channel HT amplifier because I want it to swing to the rails into 4 ohms (and allow two pairs of channels to work properly in bridged mode as a biamp option) yet not risk smoking the output devices.
However, a somewhat offsetting advantage of these parts is supposed to be the low current at which a negative temperature coefficient begins (around 100ma per device), which simplifies thermal protection. However, I don't want the quiescent bias to drop off *too* fast when the amp warms up, so I will likely work up a biasing scheme that is slightly more complex than a string of two or three diodes - probably something using a transistor and also a themistor if necessary on the heatsink with bias adjust and possibly even a tempco trim - still need to work out the details🙂
Anybody else have experience with these parts?
Distortion and Human Hearing
I just read Jonathan Valin's review of the Lamm ML2 which Valin considers the best amplifier he's heard ... which I take to be an endorsement, not universally applicable, but perhaps worth investigating. The ML2, for those unfamiliar, is a 6C33C SET with regulated PSU on evidently all stages. What I found interesting about Valin's review is his discussion about the principles the designer, Vladimir Shushurin, used to design the amp. According to Valin, Shushurin wrote his PhD thesis on a mathematical model of human hearing which, after university years, Shushurin used to model parameters (?) of audio equipment. Says Valin:
"There is an important distinction to be made about the way Vladimir 'reads' test results [THD vs frequency vs power etc]. The numerical values [absolute values, I think] aren't what he pays attention to. It is the graphic result---that is, the shape of the curves indicating how, for instance, distortion changes with power and frequency---which tells the story. And once again it tells the story because of the way the curve correlates with the mathematical model of human hearing that Shushurin spent so many years working on in Russia"
I'm not sure what Valin means by "curve" in his last sentence, but I suspect he's speaking of the distortion spectrum or distortion curve. Shushurin, for his part, speaks of the "constancy of harmonic residue shape" as a primary design goal. For an interview of him, see:
http://www.lammindustries.com/intervie/audiorus.html
I would sure like to know more about this subject of distortion spectrum and how a given distortion curve (ie, the relative amplitude of various harmonics), and the linearity of that shape vs power, affects one's auditory perceptions. I suspect both are important, perhaps equally so.
I realise I'm digressing into generalities, but the issues do refer back to earlier threads concerning the utility of transformer coupling in an OTL design.
I just read Jonathan Valin's review of the Lamm ML2 which Valin considers the best amplifier he's heard ... which I take to be an endorsement, not universally applicable, but perhaps worth investigating. The ML2, for those unfamiliar, is a 6C33C SET with regulated PSU on evidently all stages. What I found interesting about Valin's review is his discussion about the principles the designer, Vladimir Shushurin, used to design the amp. According to Valin, Shushurin wrote his PhD thesis on a mathematical model of human hearing which, after university years, Shushurin used to model parameters (?) of audio equipment. Says Valin:
"There is an important distinction to be made about the way Vladimir 'reads' test results [THD vs frequency vs power etc]. The numerical values [absolute values, I think] aren't what he pays attention to. It is the graphic result---that is, the shape of the curves indicating how, for instance, distortion changes with power and frequency---which tells the story. And once again it tells the story because of the way the curve correlates with the mathematical model of human hearing that Shushurin spent so many years working on in Russia"
I'm not sure what Valin means by "curve" in his last sentence, but I suspect he's speaking of the distortion spectrum or distortion curve. Shushurin, for his part, speaks of the "constancy of harmonic residue shape" as a primary design goal. For an interview of him, see:
http://www.lammindustries.com/intervie/audiorus.html
I would sure like to know more about this subject of distortion spectrum and how a given distortion curve (ie, the relative amplitude of various harmonics), and the linearity of that shape vs power, affects one's auditory perceptions. I suspect both are important, perhaps equally so.
I realise I'm digressing into generalities, but the issues do refer back to earlier threads concerning the utility of transformer coupling in an OTL design.
Distortion
Hans, I'm not sure I agree with your last sentence in the quote above. Speakers produce upwards of 5-10% distortion, but one clearly can hear both the level and spectrum of distortion of any electronics driving them. My experience tells me distortion is cumulative both in terms of absolute level and the mixing of various stages' spectra. The output stage amplifies the signal it's fed, which signal will be of a given distorted shape. The output therefore amplifies all distortion components previously introduced into that signal.
Regards,
Tom
tubetvr said:
Hans, I'm not sure I agree with your last sentence in the quote above. Speakers produce upwards of 5-10% distortion, but one clearly can hear both the level and spectrum of distortion of any electronics driving them. My experience tells me distortion is cumulative both in terms of absolute level and the mixing of various stages' spectra. The output stage amplifies the signal it's fed, which signal will be of a given distorted shape. The output therefore amplifies all distortion components previously introduced into that signal.
Regards,
Tom
more Lamm urls
Perhaps also of interest:
http://www.lammindustries.com/faq/expens.html
http://www.lammindustries.com/faq/bhk.html
😀
Perhaps also of interest:
http://www.lammindustries.com/faq/expens.html
http://www.lammindustries.com/faq/bhk.html
😀
- Status
- Not open for further replies.