Thought on OTL tube amps

I had a pair of Fourier Panthere monoblocks with 8 6C33C Russian military triodes each that powered a pair of Essence Super Gem and later Wilson Sophia speakers through an autoformer.  When things were right, the sound was glorious.  Not the lean, hyper-detailed, flat-soundstage, micro-image sound you hear so often, but big, transparent, harmonically-rich sound that washed over you and left you mesmerized.  The problem was:  you never knew what you would get when you turned these beasts on and watched the lights in the house dim as the amps doubled as space heaters.  Sometimes everything went right and on others bass authority seemed light and a check of the amps indicated one or more blown tube fuses.  Sometimes in the middle of the music something seemed askew and I noticed one channel had popped its little red circuit breaker.  After I took the amp apart and cleaned and tensioned the tube sockets, things seemed more stable, but finally I realized that I wanted to escape the world's problems by listening to music, not add to them with a problematic amp.  Perhaps the newer OTL designs are more stable, such as Atma-Sphere, but with older designs listening often became a project.  Perhaps this is why almost every manufacturer of these designs went out of business.

Upon re-reading my previous post, to be more precise it occurs to me that I should re-word the second paragraph as follows:

The point, though, is that you and Realdeal appear to be envisioning (incorrectly) that autoformers behave as if they are inductors, and nothing more. But while autoformers operate based on inductive principles, they behave differently than inductors, assuming that a load is present. Just as the primary winding of a two-winding transformer behaves differently than an inductor, assuming that a load is present. And as I indicated, an autoformer can transform voltages and impedances just as a transformer can.

Regards,
-- Al

Xonex77, as someone with multiple degrees in electrical engineering and multiple decades of experience designing and managing design of analog and digital circuits for advanced defense electronics, I can assure you that I am well aware that placing an impedance in parallel with some other impedance (of similar type, i.e., not an L in parallel with a C) results in the total impedance being less than either of the two individual impedances, at a given frequency.

The point, though, is that you and Realdeal appear to be envisioning autoformers as behaving as if they were inductors, and they are not inductors. While autoformers (as well as transformers) possess some degree of parasitic (undesired) inductance (and also resistance), they are different animals. And as I indicated, an autoformer can transform voltages and impedances just as a transformer can.

See the section in this Wikipedia writeup on the equivalent circuit of a real-world transformer, and this Wikipedia writeup on autoformers.

Some brief excerpts from the latter writeup:

... The voltage and current ratio of autotransformers can be formulated the same as other two-winding transformers.... As in a two-winding transformer, the ratio of secondary to primary voltages is equal to the ratio of the number of turns of the winding they connect to.... In audio applications, tapped autotransformers are used to adapt speakers to constant-voltage audio distribution systems, and FOR IMPEDANCE MATCHING [emphasis added] such as between a low-impedance microphone and a high-impedance amplifier input.

Surely you don’t deny that a transformer can transform impedances (in proportion to the square of the turns ratio)? Well, an autoformer ("auto-transformer") can also.

Regards,
-- Al

Please review your your electrical circuits about series /parallel of resistance,inductance and capacitance,I don't like to trash anybody but there is no point in  continuing  the discussion if the  theory is not fully  understood.

Realdeal, as I indicated in my response to the very similar comment that was made by Xonex77, that is simply not correct. A suitably designed autoformer can transform voltages and impedances no differently than a two-winding transformer can. Although there are obviously other differences between the capabilities of the two kinds of devices, most notably the obvious fact that an autoformer can’t provide electrical isolation.

You (and Xonex77) may want to read up on autoformers a bit more. You might also want to reflect upon the various McIntosh solid state amplifiers which use autoformers at their outputs. Which by presenting the output stages with the same load impedance when a 2 ohm speaker is connected to the 2 ohm tap as when a 4 ohm speaker is connected to the 4 ohm tap as when an 8 ohm speaker is connected to the 8 ohm tap, enable those amplifiers to have identical power ratings into 2, 4, and 8 ohms.

Regards,
-- Al

The Zero autotransformer will not make  amplifier output impedance and the speaker input impedance higher as  Paul Speltz claim,It only decreases impedance on both input and output but it increase output  power due to less impedance and more current.There's a lot of Paul explanation about the Zero  are wrong.

realdeal, you are truly misinformed. I have run three different levels of the glorious Atma-Sphere OTL amplifiers S-30, M-60, MA-1 Silver on numerous loudspeakers with no issue whatsoever! All were played without autoformers, all were played through 8-16 ohm nominal impedance loads. It is abundantly clear that you have no proper, real life experience with these amplifiers. You really should do yourself (and the rest of us) a favor, and go out and get a good solid audition of these babies and then come back with an honest opinion.

Happy Listening!

realdeal- You are incorrect.  Many, many people use Atma OTLs direct to speakers that have the appropriate impedance characteristics, including myself, and they sound terrific.  Have you ever heard any of Ralph's amps driving the appropriate speakers? 

Sorry to say that the atmas are not stand alone OTL tube amp but need output transformer  plain and simple.

Most OTL design  require high high impedance speakers,flaws in the design that these amps should work without these impedance matching,but not all OTL have these inherent disadvantage,I have an OTL using KT120 or VV52/300B output tubes  that welcomes 4 or 8 ohms  and a   full-range speakers.But if I insert  the ZERO, it  kills every thing.  

Ralph (Atmasphere), I think that when Xonex77 refers to "increasing the impedance of a speaker" in the second sentence of his latest post, he is referring to increasing the load impedance seen by the amplifier, for a given speaker. But more significantly, in the first sentence of that post he appears to be envisioning that a Zero, when inserted between an amp and a speaker, would act as an inductor (rather than acting as an autoformer), with much of its inductance placed in parallel with the speaker's impedance. Resulting in a lower overall impedance, as seen by the amp, than if no Zero were present. Which as you'll certainly agree is not correct.

Best regards,
-- Al
 

It not exactly clear what you are talking about, but if I read this right you are incorrect about increasing the impedance of a speaker. One way to do it is to put drivers in series rather than parallel; of course that is a design issue.

There are a good number of 16 ohm speakers made these days, on account of the fact that 16 ohms works extremely well with tubes. The speakers we use at our shop are 16 ohms. They have two woofers which are in series.

Paralleling a choke/inductor/coil in the amps 4 ohms output will  be  less than 4 ohms and same thing to a voice coil of 4 ohms will  be  less than 4 ohms due to ohms law,then sound will be louder due to less resistance and more current.The only way to increase the impedance of the speaker is to put a coil/inductor   in series with the voice coil,which of course it kills dynamics.

How do you categorize amplifier with an output transformer?OTL?Are we going in circle?or we might just go back to SE or PP.The mean reason why these amp were design is that they will work without OPT.

Again, what we found after getting into the business is that some people wanted to use our amps with really low impedance speakers. The combination was non-functional. Like many tube amps, OTLs don't like low impedance loads. So what to do? If the customer really wants to use the OTL with that kind of speaker then you make it possible by using a set of ZEROs.

In some cases its just the woofer section that might have the low impedance. In such cases you can install the ZEROs so the amp is driving the mids and highs directly while using the ZEROs to deal with the lower impedance of the woofer array.

The ZERO is intended as a problem solver plain and simple.

Al,That's their selling point plus the exotic wooden case, which has nothing to do with sound or else  zero will not be in bizznes,I already ask him,the autotransformer maker  about if it kills dymanics and  he said decide it yourself.Auto transformer works two ways either step up which makes the sound louder or step down which make the sound lower.

The output can not be equal to input always,How do you categorize amplifier with an output transformer?OTL?Are we going in circle?or we might just go back to SE or PP.The mean reason why  these amp were design is that they will work without OPT.

The ZERO will not kill dynamics. The ZERO is a problem solver, used when someone wants to drive a speaker that is lower impedance, perhaps too low for the OTL to do on its own. All you have to do if you don't want to use the ZEROs is get a speaker of higher impedance (8 ohms minimum usually works with our stuff) or get a bigger OTL.

The ZEROs usually get used with our smaller OTLs, an example being our M-60s used with Magnaplanar 3.7s or the like. The result of using the ZEROs is that the output power of the amp is doubled (from 40 watts to 80 watts), the distortion is reduced and the tubes run cooler as the power they are making is dissipated in the load rather than in the output section. MA-1 amplifiers drive the Magnaplanars just fine without the ZEROs.

The ZERO would not be possible if the amp did not already have a lower output impedance and did not make any DC voltage at its output. The result is that the ZERO then has actually greater bandwidth than the amps- going from 2Hz to 2MHz. So it does not limit bandwidth of the amp. The resulting combination is impressive- a tube power amp with full power from 2Hz to about 100KHz or so, which can drive 2 ohms or more. You can't do that with a regular tube amp!

Using the Zero in OTL, kill dynamics and will not be OTL anymore.Going back to  PP.

If the Berning folks might be stretching it a bit to call their amp an OTL, what about companies, like Tenor (I believe), that put out a solid state amp that they described as OTL?  This was absolutely true, but, aside from a few solid state amps like models produced by McIntosh, all solid state amps are OTLs (low output impedance of transistors makes transformers unnecessary).

I remember when Allied used to write up the receivers in their catalog back in the early 70s. They referred to their circuits as ITL/OTL and maybe with good reason at the time. A lot of solid state amps back in those days employed input transformers as phase splitters to the output section!

These days though OTL generally refers to a tube amp, not a solid state amp as its no big shakes with solid state in that regard.

By the way, all of my amps are FCL (flux-capacitor-less).

I suspect that they'll perform best with contemporary music  :-)

Chris (Cal3713), thanks for the nice comment.

Best regards,
-- Al
 

If the Berning folks might be stretching it a bit to call their amp an OTL, what about companies, like Tenor (I believe), that put out a solid state amp that they described as OTL?  This was absolutely true, but, aside from a few solid state amps like models produced by McIntosh, all solid state amps are OTLs (low output impedance of transistors makes transformers unnecessary).

By the way, all of my amps are FCL (flux-capacitor-less).

Thanks for posting Al.  I should just start reading every post of yours so that I continue learning the science behind the music.

Using the Zero nullify an OTL status.You already get rid of transformer then put new Transformer?Whats happening? going in circle?

Xonex77, your question is a logical one, but there is a good answer to it.  From the FAQ at the website for the Zero:

What is the difference between the ZEROs autoformer and a typical tube amplifier's transformer?

The ZEROs have a few advantages over "typical" tube amp transformers:

A) The music comes out on the same winding wire that it goes in on so the music does not have to pass from a primary winding to a secondary winding.

B) There is no DC current to contend with. When a transformer is made to accommodate a DC field, its audio transparency is compromised.

C) The impedance ratio is very small (16 ohms to 4 ohms, compared to a few thousand ohms to 4 ohms). This simply means that it is much easier to achieve things like, frequency response extremes, than with a "typical" tube amp transformer. The ZEROs sport a frequency response of 2 Hz to 2 MHz.

All transformers (and autoformers) are not created equal!

Regards,
-- Al

I use the Classic Audio Loudspeakers model T-3.3 at home. It goes down to 20Hz. A set of our M-60s has no troubles driving them to very high volume levels with very satisfying bass (shakes the house). M-60s are full power down to 2Hz. 

Like **any** other amp, if you want to hear what its really capable of, its a good idea to put it on a speaker to which it is well-suited.

bsimpson

Perhaps I should put my $0.02 here from an end-user perspective.

I have owned at least seven (7) different OTL tube amps, some are big brands, some are DIYs from experienced (10+ yrs) tube amp hobbyists with extensive engineering experience.

IMHO, OTL tube amps are reliable if you understand their operation limitation first. I found that:

1) OTL tube amps do not work well under 4 ohm or less speakers. Even some transient passages at 2 or 3 ohms may blown the tubes. This is not specific to the tube type.

2) OTL tube amps are much less tolerant to wide impedance swing. The more stable the speaker's impedance, the better.

The best sound I got from OTL amps is when I use them to drive impedance-friendly monitors, with the aide of quality subwoofers covering 180Hz-250Hz and below.

Hope this helps.

This is a good summary bsimpson. And that's quite the experience with so many different OTL's, although you don't list the brands amps/speakers. Thanks for sharing. I particularly take note of your last comment - aide of quality subwoofers 180hz - 250hz. Very difficult to do going this high in the frequency range, even in a dedicated space. 

I'm convinced…I'm immediately removing the output transformer from my amp…those wire cutters are around here somewhere...

Using the Zero nullify an OTL status.You already get rid of transformer then put new Transformer?Whats happening? going in circle?

Perhaps I should put my $0.02 here from an end-user perspective.

I have owned at least seven (7) different OTL tube amps, some are big brands, some are DIYs from experienced (10+ yrs) tube amp hobbyists with extensive engineering experience.

IMHO, OTL tube amps are reliable if you understand their operation limitation first. I found that:

1) OTL tube amps do not work well under 4 ohm or less speakers. Even some transient passages at 2 or 3 ohms may blown the tubes. This is not specific to the tube type.

2) OTL tube amps are much less tolerant to wide impedance swing. The more stable the speaker's impedance, the better.

The best sound I got from OTL amps is when I use them to drive impedance-friendly monitors, with the aide of quality subwoofers covering 180Hz-250Hz and below.

Hope this helps.

11-25-15: Psag
... That is why Berning refers to his amp as an OTL. Those who say otherwise are obviously invested in saying otherwise.

Certainly not true in my case, apart perhaps from whatever negligibly small interest I might have in not having my original statement that it is not an OTL contradicted.

Also, Ralph makes a good point that the term "OTL" is normally used just in the context of tube amplifiers, since the vast majority of solid state amps don't have output transformers or autoformers anyway. And since the Berning's output stage utilizes solid state switching devices, it arguably does not even fall within that context.

Regards,
-- Al

BTW, David most certainly does call his amp an OTL. As Atmasphere states above, the transformer in the Berning amp operates at RF frequencies, while the output transformer of a conventional amp operates at audio frequencies. That is why Berning refers to his amp as an OTL. Those who say otherwise are obviously invested in saying otherwise.

Ralph, yes I had looked at the patent, and I completely agree with your comments.

Tubegroover, thanks for the suggestion! I'll be sure to check out those discussions. Prior to this thread I had no awareness of this amp, as I mentioned earlier, so I'm sure I will find it to be of particular interest.

Best regards,
-- Al

10-30-15: Czarivey
....in Minnesota you need lots of heat in the winter. OTL amps partially solve the heat problems.

True story ....

Our living room is on the main floor & houses the furnace thermostat; the room is hardly used any more as the family room is where the guests go at the other end of the house. Hiding from my wife in this living room - a pair of modded Acoustat Model 3's. So one very cold Ontario winter day, I decided to put my OTL's in there, fire them up with the model 3's and do some listening at - low - moderate levels. It was going great. 2 maybe 3 hours had gone by ?? Not sure.. Then my wife opens the door on me and her first words were ...

"why is the house so cold",

quickly followed by ....

"Hey why is so hot in here",

followed by.....

" Hey..Why the hell are you in your underwear !"
-
-

Maybe the acronym OTL should be changed as applied to David Berning's ZOTL amplifiers as the original ILOT design, iron less output transformer.It never ceases to amaze how this topic is continually debated after all these years. Al you REALLY need to check the archives, if you haven't already, I'm sure you'll find it quite interesting particularly the thread where Ralph and David have a lively discussion of this topic, I think you'll enjoy!

I'm not too sure David EVER called it an OTL, I believe it was Harry "Gizmo" Rosenberg that first did but I may be mistaken. One thing for sure after it was forevermore designated OTL by some, it was equally or more declared NOT by others.

Ralph is from Minnesota(I might be wrong) and in Minnesota you need lots of heat in the winter. OTL amps partially solve the heat problems.

I read the patent many years ago. I suspect Al has too.

The transformer is impedance matching. It is air core, as it is operating at RF frequencies to modulate the output of a switching power supply. The output of the switching supply then has its RF component filtered out and the result is the audio signal. So the output transformer and the switching supply work in tandem to do the impedance conversion.

A conventional output transformer does impedance matching as well only it does it at audio frequencies.

Either way there is a transformer that represents the load to the output tube or tubes. It can only be called an OTL when a transformer can no longer be called a transformer. Since these terms have been with us a long time and will continue to be with us for decades to come, the result is that the acronym 'OTL' is misapplied to this amp.

Yes, there's a tranformer in the Berning amp. The real issues are what its doing there, and how one wants to define OTL. The 'traditional' definition of OTL seems to require parallel tubes, and no transformer anywhere near the output. Ok fine, but what is the Berning transformer doing as compared to the traditional output transformer? Without knowing the answer, the question is moot.

The fact that the Berning amplifier (while an outstanding and very innovative design) does not fall within the definition of an OTL is something that has been put to bed many times. While I respect David as one of the top designers in the audio world, I do find it inexplicable that his website makes this claim.

OTLs by definition do not have an output transformer; that much should be pretty clear. 'OTL' refers simply to 'Output TransformerLess. Over time (the last 50 or 60 years) the acronym generally has referred to a *tube* amplifier without an output transformer; in that regard having semiconductors in the output section would not qualify the amplifier as an OTL. And of course the Berning amps all include semiconductors as an essential part of the output circuit.

Hi Psag,

I had looked at the writeup at his site, and also at the patent which it references. There is a transformer between the outputs of the power tubes and the output of the amplifier, along with some solid state switching devices. As I read the writeup, what he said is that there is no AUDIO output transformer, but "the high-voltage, low-current tube impedance-plane is re-mapped to the high-current speaker impedance-plane through a special transformer at a constant RF carrier frequency of 250 kHz."

So while his very innovative approach avoids the use of an AUDIO output transformer, the design is not Output Transformerless (OTL).

Best regards,
-- Al

Al, check Berning's website. He clearly states that it is an OTL. Check the picture, you will see that there is no output transformer. Looks like one, sounds like one...

Makes sense. Very heavy duty!

Thanks Al.

10-27-15: Mapman
Chalice Audio Grail amps are interesting to me. I didn't realize a SET amp could produce 50 w/ch into 8 ohm. How do they do that?

Mapman, it uses a parallel pair of type 845 tubes, which I believe were originally developed for transmitting applications, and have much greater power capability than the 300B, 45, and 2A3 tubes you are probably more used to seeing in SET amplifiers. There are a number of other 845-based SETs out there. The 211 is another such high powered tube that is sometimes used in SETs. Both types require about 32 watts just to heat the filaments of each tube(!), and run at very high B+ voltages (I believe often in the vicinity of 1000 volts), as well as handling substantial amounts of current. All of which presents much greater challenge when it comes to design of the output and power transformers than in the case of the lower powered SETs.

BTW, A'gon member Ait has described in past threads how he designed and built for himself a SET providing 200 watts per channel, which utilizes an 833C transmitting triode operated at 2500 volts(!), and having output transformers weighing 62 pounds each. Not an OTL of course, but quite a remarkable tour de force!

Psag, thanks for the reference to the Berning Quadrature Z, which I hadn't been aware of previously. Quite an innovative and remarkable design, providing 200 watts of tube power per 32 pound monoblock. But of course it is neither a SET nor an OTL, although its output transformer operates at low RF frequencies rather than at audio frequencies.

Best regards,
-- Al

Chalice Audio Grail amps are interesting to me. I didn't realize a SET amp could produce 50 w/ch into 8 ohm. How do they do that?

No paralleling of tubes in David Berning's Quadrature. 200 W into 8 ohms.

If you want power, quite often "paralleling tubes is the only option" regardless of OTL or not.

Nobody knows all the knowledge about OTL,Some were copied and modified.Most commercial/on the market OTL are very ineffecient in converting tube output capabilities which paralleling tubes is the only option.

Trelja,May the FORCE be in Simon Vinitsky's soul.

Okay, thank you.

Futterman himself seemed to do a pretty good job at keeping the amps together according to one of our customers, although HP related an event in TAS that suggested that they could fail spectacularly.

I heard rumors about the NYAL stuff being not as stable but never knew anyone that actually had experienced a problem themselves. During the 1990s, that rumor was so prevalent that we came to call it the 'Futterman legacy' since it was our single biggest marketing problem. In essence, because we made an OTL people assumed it was unreliable. The simple fact that we don't have to ship amps back to be repaired if a tube fails is why we are still around after 39 years.

But I have no idea where the rumor got started. I do know that Fourier, who went out of business in the late 1990s, didn't help, but their amps had engineering flaws that had nothing to do with them being OTLs!!

Harvey Rosenburg (founder of the NYAL) was adamant that the amps had no reliability problems, even years after the company failed. But he had a lot of ego tied up in them and was never the most reliable witness so I regard his comments as a bit of a red herring.

So I really can't confirm or deny that rumor; sorry I can't shed more light on it.

Ralph, can you either verify or put to rest the notion that the Futterman OTLs are inherently unstable? I have an H3 in decent shape I'd like to have brought back to life. From what I've found, there seems to be a serious contention that this has been rumored early on and stuck, however erroneously.