• WARNING: Tube/Valve amplifiers use potentially LETHAL HIGH VOLTAGES.
    Building, troubleshooting and testing of these amplifiers should only be
    performed by someone who is thoroughly familiar with
    the safety precautions around high voltages.

If you were going to build a OTL?

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
If you were going to build a OTL (output-transformerless) amplifier and you knew you had to live with it long term which one would it be.

MOSFET Circlotron final, hollow state front end. The A Number One problem with any hollow state OTL is operating the finals into nearly vertical loadlines. That makes for big time distortion, and it becomes questionable whether that is preferable to just going ahead and using source followers. Just make sure that your source followers stay away from the rails since MOSFET internal capacitances go squirrelly when Vds drops to low voltages.

Besides, a MOSFET Circlotron doesn't need a high current heater transformer, nor does it require knocking a dozen holes in the chassis. There is also the safety question: the MOSFET design won't put bigvolts on the speaker terminals if something poofs.
 
Brit01, as far as protection goes, and if you are using 6AS7Gs, careful B+ transformer fuse selection seems to do the trick. In the last 30 years, I've seen about 9 drivers damaged, but so far not a single one was a legitimate case where a tube failed and damaged the speaker.

In two cases people put in alternate(!) tubes that we didn't recommend. In three others, people had the B+ fuse over-rated. The remaining four had damage unrelated to tube failure or malfunction of the amp- the speakers simply could not take the power.

That's over a period of 30 years. If you have the fuse rated right, and a tube does some sort of arc-over or internal short, the fuse blows, you replace it, deal with the tube and you're back up in no time.

If you are using 6C33s, 7241, 6336, EL509 or other types you will have to fuse the individual tubes. OR- install a protection circuit :)
 
If you were going to build a OTL (output-transformerless) amplifier and you knew you had to live with it long term which one would it be.

None. Tube based OTL amps are inherently unreliable : be prepared for frequent tube replacements,adjustements,troubleshooting and speaker reconing when the amp fail. (and it will,sooner or later). A reliable tube OTL amp still need to be designed. Maybe this is the main reason why OTL amp's remained commercially relatively unsuccessful. Also,the choice of power tubes is limited to a few poor linearity power supply regulator triodes or TV horizontal sweep tubes which requires lots of NFB to help lower the output impedance an linearize the amp. At best (?), you'll end up with a lousy expensive unreliable heat generating amp. If you really need an OTL amp,better to go solid-state.
 
Brit01, as far as protection goes, and if you are using 6AS7Gs, careful B+ transformer fuse selection seems to do the trick. In the last 30 years, I've seen about 9 drivers damaged, but so far not a single one was a legitimate case where a tube failed and damaged the speaker.

I'm looking into a transistor protection circuit that switches off the B+ if the bias fails or falls too low, plus fuses on secondary and primary windings, Better be safe than sorry.
 
Administrator
Joined 2004
Paid Member
My understanding is that the Russian made version of the 6AS7G is the one to use in OTL designs. (The Chinese ones reportedly are OK too.) Long ago I sold 16 NOS RCA 6AS7GA that I had originally purchased for my own OTL project to a friend who reported unfortunately that they had a tendency to arc over and fail.
 
None. Tube based OTL amps are inherently unreliable : be prepared for frequent tube replacements,adjustements,troubleshooting and speaker reconing when the amp fail. (and it will,sooner or later). A reliable tube OTL amp still need to be designed. Maybe this is the main reason why OTL amp's remained commercially relatively unsuccessful. Also,the choice of power tubes is limited to a few poor linearity power supply regulator triodes or TV horizontal sweep tubes which requires lots of NFB to help lower the output impedance an linearize the amp. At best (?), you'll end up with a lousy expensive unreliable heat generating amp. If you really need an OTL amp,better to go solid-state.

So, you are claiming that Atmasphere's products are "unreliable", failure prone, " commercially relatively unsuccessful", "lousy unreliable" heat generators? Wow, I hadn't heard that about them before. Atmashpere's post above yours claims otherwise. Untrue? are you able to document the failure rates of this apparently poor design which uses poor linearity output tubes?

I can understand DIY project failures, but this is news to me.
 
Administrator
Joined 2004
Paid Member
So, you are claiming that Atmasphere's products are "unreliable", failure prone, " commercially relatively unsuccessful", "lousy unreliable" heat generators? Wow, I hadn't heard that about them before. Atmashpere's post above yours claims otherwise. Untrue? are you able to document the failure rates of this apparently poor design which uses poor linearity output tubes?

I can understand DIY project failures, but this is news to me.

I agree, and have known a number of owners of Atmasphere amps and it struck me that these amps were quite reliable, and also sounded very, very good without massive amounts of feedback being employed.

My own 6C33 based amps had lots of problems related mostly to my inexperience with OTL design, but even without global feedback applied they sounded quite good. Today I am sure I could design and build one that would both sound good and be reliable. This does however talk to the need to choose a good, proven design, and to implement it properly - not trivial, but far from impossible either.
 
None. Tube based OTL amps are inherently unreliable : be prepared for frequent tube replacements,adjustements,troubleshooting and speaker reconing when the amp fail. (and it will,sooner or later). A reliable tube OTL amp still need to be designed. Maybe this is the main reason why OTL amp's remained commercially relatively unsuccessful. Also,the choice of power tubes is limited to a few poor linearity power supply regulator triodes or TV horizontal sweep tubes which requires lots of NFB to help lower the output impedance an linearize the amp. At best (?), you'll end up with a lousy expensive unreliable heat generating amp. If you really need an OTL amp,better to go solid-state.
Its been my experience of the last 30 some years that nearly every statement of the above is incorrect. I've seen tubes last as long as 22,000 hours, even with the manual bias not needing adjustment more than maybe every 6 months or so (with continuous use), and we see untested tubes fail as you might expect when you buy 100s at a time but no need for speaker repair...

It might be that we designed the first really reliable OTLs, as the above comments were commonplace when we got started- mostly relating, as far as I can tell, to the Futterman circuit. A lot of people said things to us like 'OTL huh? pretty unstable? How do you drive the top tube?' stuff like that. They were often surprised to find out that there was no 'top tube'...

At shows I would frequently pull power tubes out of the amp while it was playing music- without any reaction from the amp at all. I don't think there is another tube amp or transistor amp where you could pull something like that off!

We also made the first zero feedback OTLs... of course that limited the market since that prevented the amp from having the right voltage response for certain speakers, but more to the point the linearity was there so that the feedback was not needed. IMO/IME, if you can find a speaker that works with a zero feedback amplifier, the result will sound better if that equipment has its ducks in a row, and with OTLs that is possible.

I love seeing the DIY stuff on this site. Its cool to see people excited about what this technology can do, and the vision that drives them to explore!
 
None. Tube based OTL amps are inherently unreliable : be prepared for frequent tube replacements,adjustements,troubleshooting and speaker reconing when the amp fail. (and it will,sooner or later). A reliable tube OTL amp still need to be designed. Maybe this is the main reason why OTL amp's remained commercially relatively unsuccessful. Also,the choice of power tubes is limited to a few poor linearity power supply regulator triodes or TV horizontal sweep tubes which requires lots of NFB to help lower the output impedance an linearize the amp. At best (?), you'll end up with a lousy expensive unreliable heat generating amp. If you really need an OTL amp,better to go solid-state.

Wow, you REALLY do not know what you're talking about. I'm not even going to bother...
 
This thread is turning into a typical bullying session, so I'll just offer my opinion for the thread starter to consider and move on.

Regarding the issue of reliability, of course an OTL typically has lower reliability than a OPT tube amp, it's basic undergraduate engineering. You multiple the operating factor of all the components to arrive at a cumulative operating factor. More tubes = more frequent replacement of a tube in the amp. Note that it is a linear relationship, but this effect exaggerated by some to make OTL's sound unreasonably poor in this regard.

As for design, I would design my own. I don't carry a bias against gNFB or SS that many others here do, so I would probably pursue something with SS followers as the drive stage and CCS loaded front end. For the front end I would probably use something like Morgan Jones' Crystal Palace (cascaded LTP's) or Williamson type like Wavebourn champions. For those who are so against gNFB just read a @#$%^ book.
 
Wow. When someone states something that is outright false, the product of opinion rather than fact, would you prefer that it go unchallenged?

Global negative feedback, while reducing overall THD, actually **adds** distortion to the odd harmonics. Not by much (100ths of a percent), but since our ears use these harmonics to gauge the volume of a sound we are very sensitive to this 'enhancement'. Go read a book indeed: the use of GNFB is thus violating a fundamental rule of human perception- how we perceive volume.

If built right, OTLs are as reliable as any other tube technology. 'If built right'- and IME experience, many are not. One of the primary issues that plagues many OTL designs is the means of biasing the power tubes. Quite often there is a tension between the value of coupling caps used and that of the divider network values in the bias network. Most power tubes used in OTLs need the divider network to be a lower set of values (you aren't going to control the tubes with 1M bias resistors...) and that means larger coupling cap values. The problem is that these amps are very transparent so the artifact introduced by large coupling caps (needed to play bass) is easily audible. So to fudge that, the bias network gets pushed to higher values.

The result is an amp in which bias drifts easily, has poor overload recovery and a variety of other ills, all of which are avoidable. Add to this an excessive amount of feedback to compensate for the distortion of an asymmetrical drive or output, and you have an unstable and unreliable amp.

Design problems are not unique to OTLs; as long as humans are making things, flaws will be found. But at the same time there will be designs that overcome flaws of the prior art; anyone who ignores this fact is at risk of engaging in mythology.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.