In my searches of OTL schematics, I have yet to come across any examples of single triode cathode follower output stages. They are all multi triode push-pull cathode follower, White cathode follower, etc.
This site indicates that it can be done with a single triode section: http://members.aol.com/aria3/otlpaper/otlhist.htm
I understand that the power output would be so small for such a stage as to make hardly it useable, say only as either a headphone amp or driving some 100+ dB/W/m speaker or tweeter. Other than that, is there anything bad about something like a single 6C33C cathode follower OTL headphone amp?
This site indicates that it can be done with a single triode section: http://members.aol.com/aria3/otlpaper/otlhist.htm
I understand that the power output would be so small for such a stage as to make hardly it useable, say only as either a headphone amp or driving some 100+ dB/W/m speaker or tweeter. Other than that, is there anything bad about something like a single 6C33C cathode follower OTL headphone amp?
Hi,
The main problems with a single tube OTL is to get low enough output impedance and useable output power. A 6C33C which is one of the most suitable tubes available have a ra of ~80 ohm and a mu of ~2.5 this give an output impedance of ~22ohm as a cathode follower, this can be lowered by parallelling tubes even though 22 ohm should be good enough for a headphone amp.
A cathode follower OTL will always be inefficient and you will always have the problem providing a good cathode load be it a inductor or some kind of CCS.
Most people building OTL amps want to have a reasonable output power and the easiest way to get that is by using 2 tubes in some kind of push-pull circuit. The inverted Futterman circuit always provide the lowest output impedance for the same tubes, thereby eliminating the need for parallell connceting a large number of tubes.
Why this interest of single tube OTL? what are you trying to achieve?
Regards Hans
The main problems with a single tube OTL is to get low enough output impedance and useable output power. A 6C33C which is one of the most suitable tubes available have a ra of ~80 ohm and a mu of ~2.5 this give an output impedance of ~22ohm as a cathode follower, this can be lowered by parallelling tubes even though 22 ohm should be good enough for a headphone amp.
A cathode follower OTL will always be inefficient and you will always have the problem providing a good cathode load be it a inductor or some kind of CCS.
Most people building OTL amps want to have a reasonable output power and the easiest way to get that is by using 2 tubes in some kind of push-pull circuit. The inverted Futterman circuit always provide the lowest output impedance for the same tubes, thereby eliminating the need for parallell connceting a large number of tubes.
Why this interest of single tube OTL? what are you trying to achieve?
Regards Hans
tubetvr said:
Thanks for your detailed reply. Basically, I want to build as inexpensive an OTL as I can for a learning experience. A single tube cathode follower appeals to me since it can be powered from a single rail, the same rail as the input stage. I have looked at OTL schematics and they are far from inexpensive, since the cost of the OPT is more than offset by the extra cost of multiple rails, extra filtering to fight bad PSRR, etc.
The first OTL I built was a SEPP without phase splitter using 2 pentodes, it was similar to the one described as "Philips original OTL" here http://www.tubetvr.com/otl.html it has the disadvantage of needing a output capacitor and it also give quite a lot of 2nd order distortion because of the lack of symmetry but the one I built sounded so good that I after that always wanted to build a real OTL.
I can recommend this as a first project and it should work quite well as headphone amp.
Regards Hans
I can recommend this as a first project and it should work quite well as headphone amp.
Regards Hans
Is that the same as this one: http://www.bonavolta.ch/hobby/en/audio/el86.htm ?
Wasn't that built for 600 ohm speakers? Will it work as a headphone amp into 32 ohm headphones?
Wasn't that built for 600 ohm speakers? Will it work as a headphone amp into 32 ohm headphones?
Single output tube otl
It is possible!!
I made a one output tube otl using one 6AS7 for each channel
The resulting output is about 60 Volts into an 800 Ohms speaker
Output impedance is about 200 Ohms
Freq. response -1dB is 40 Hz to 200+ kHz
I use Philips 9710AM speakers. Also sold under the name Norelco iirc.
Lately i found some nice 800 Ohms speakers in a Philicorda organ, haven't tested them yet though
Sound is sweet and fast, unlike any transformer coupled design i have heard so far.
I build it with parts out of the bin, to keep costs low.
See att. schematic
It is possible!!
I made a one output tube otl using one 6AS7 for each channel
The resulting output is about 60 Volts into an 800 Ohms speaker
Output impedance is about 200 Ohms
Freq. response -1dB is 40 Hz to 200+ kHz
I use Philips 9710AM speakers. Also sold under the name Norelco iirc.
Lately i found some nice 800 Ohms speakers in a Philicorda organ, haven't tested them yet though
Sound is sweet and fast, unlike any transformer coupled design i have heard so far.
I build it with parts out of the bin, to keep costs low.
See att. schematic
Attachments
The trouble with various OTL schemes is not getting output impedance low (as people tend to think), but limited current drive capability. If you are running two sections of 6AS7 in parallel as a Class A cathode follower at 75mA idle per section, the maximum current you will get into the speaker is 150mA, which is just 0.1W into 8ohm load before clipping sets in. Not very efficient. (One can get 5W easily from the same tube in conventional transformer-coupled common-cathode output stage.) To get around this problem, OTL are usually made in push-pull configuration, and run deep into Class B (which is ugly if you ask me).
There is no reason to run an OTL in class B, AB1 is sufficient to get both high power and reasonable efficiency. My own OTL have an idle current of 200mA and the tubes deliver 2.5A in the peaks without drawing grid current.
Output impedance is important in order to get reasonable efficiency, an OTL with very high output impedance like for instance the original EL86 based Philips design need very high anode voltage to get any kind of output power when loaded with low impedance e.g 8 ohm. Output impedance can of course be lowered by feedback but efficiency will still be the same.
Regards Hans
Sorry, when peak output current is twelve times the idle current, it is deep in class B. The usual drawbacks are high distortion, a change in output impedance as signal swings from Class A to Class B region (which leads to very unpleasant waveform distortion), and heavy requirements on power supply. Might as well switch to transistors in output stage. (BTW, if you are willing to put up with Class B, you still could get a lot more power with a tranformer if you just drop it in into your OTL design instead of direct speaker load...)
Andrei,
What do you mean by "deep in class B"? an amplifier operates either in class A, AB or C these are clearly defined.
It seems that you have your own definition of class B, that is OK even though it must be confusing for others.
What is more important is of course the end result you can get, I don't know what experience you have from OTL amps but I recommend you to listen to a good one. When it comes to measurement results I think it is quite difficult to equal the performance in a good OTL using transformers, be it bandwidth or distortion, the remaining problem in OTL amps is low efficiency but that is something you need to live with.
Regards Hans
What do you mean by "deep in class B"? an amplifier operates either in class A, AB or C these are clearly defined.
It seems that you have your own definition of class B, that is OK even though it must be confusing for others.
What is more important is of course the end result you can get, I don't know what experience you have from OTL amps but I recommend you to listen to a good one. When it comes to measurement results I think it is quite difficult to equal the performance in a good OTL using transformers, be it bandwidth or distortion, the remaining problem in OTL amps is low efficiency but that is something you need to live with.
Regards Hans
Leadbelly,
The Philips amp can be built with other tubes, almost any output beam penthode will work, the higher Gm the better as it will affect the output impedance. The capacitor values are not that critical and can be increased if you want to, the output capacitor need to be a very big one as it should pass low frequencies.
If you have the components you should be able to put this together in short time, especially if you build it without a chassis just for test, that is how I did it, it will play very low with ordinary speakers but should work better with headphones.
Regards Hans
The Philips amp can be built with other tubes, almost any output beam penthode will work, the higher Gm the better as it will affect the output impedance. The capacitor values are not that critical and can be increased if you want to, the output capacitor need to be a very big one as it should pass low frequencies.
If you have the components you should be able to put this together in short time, especially if you build it without a chassis just for test, that is how I did it, it will play very low with ordinary speakers but should work better with headphones.
Regards Hans
I have enough experience with listening to OTL transistor amps not to be thrilled with the idea of emulating their design flaws using tubes instead of transistors.
As far as operating point classes go, drawing 2.5A current from stage idling at 200mA means that 95% of the time one of the tubes is in cut-off (assuming sine wave). So your nominal AB is made up of 5% A and 95% B at your rated power. That is no different from Class B for practical purposes. I'm willing to bet that there is a lot of high-order harmonics in the output, even without measuring or simulating the whole thing.
I don't see the point of chasing bandwidth in audio amplifiers. If anything, it should be limited. Do you really want to be amplifying RFI? (When I lived in Toronto, all my solid state audio gear would receive CBC broadcasts on its own and without benefit of having a tuner. True story.)
As far as operating point classes go, drawing 2.5A current from stage idling at 200mA means that 95% of the time one of the tubes is in cut-off (assuming sine wave). So your nominal AB is made up of 5% A and 95% B at your rated power. That is no different from Class B for practical purposes. I'm willing to bet that there is a lot of high-order harmonics in the output, even without measuring or simulating the whole thing.
I don't see the point of chasing bandwidth in audio amplifiers. If anything, it should be limited. Do you really want to be amplifying RFI? (When I lived in Toronto, all my solid state audio gear would receive CBC broadcasts on its own and without benefit of having a tuner. True story.)
If you feel like that then I find it is strange that you take part in a discussion of how to build tube OTL amps, are you trying to discourage people to try?
You still don't seem to understand what is meant by different operating classes, class AB doesn't mean that it switch between A and B at some point, the definition is that current flows in each active element between 180 and 360 degrees of a period.
I think you would find that high-order harmonics still can be made lower in an OTL amp than in a compareable transformer coupled amp.
Agreed that it should be limited but what is the required bandwidth? In order to get high IM caused by high frequencies e.g 19, 20 kHz I want the amplifier to have wide bandwidth before applying feedback and I also want to have minimimum phase shift inside the audio range.
A suitable open loop bandwidth then could be around 200kHz which is not that difficult to achieve in an OTL but very difficult in a transformer coupled amp.
Regards Hans
Impedance
Dear Hans
Let's back to the OTL.
Do you think four 6c19 work together as CF on an inductor
via four 5 ohms resistor as what the SE 1.5W OTL of
Transcendentsound can drive a 8 ohms speaker with good
damping?
How about replaceing the inductor with a CCS as I question
in the other thread.
Regards CHOW
Dear Hans
Let's back to the OTL.
Do you think four 6c19 work together as CF on an inductor
via four 5 ohms resistor as what the SE 1.5W OTL of
Transcendentsound can drive a 8 ohms speaker with good
damping?
How about replaceing the inductor with a CCS as I question
in the other thread.
Regards CHOW
6C19 will work but it is not an ideal OTL tube, the transconductance is only 7.5mA/V and internal resistance is high at 400ohm. Each tube as CF will have an output impedance of about 100 ohm so 4 in parallell will have an output impedance of 25ohm. In order to get reasonably low output impedance you would need quite heavy feedback.
A single 6C33C have a Gm of 40mA/V and can handle the same current as 4 6C19, the output impedance is also about the same or about 22ohm, I think that is a much better choice as you don't get the problems of matching the tubes.
Max output power is determined by how much current you can run continously in the tubes, each 6C19 is specified to max 140mA so 4 of them can run 560mA. As we have a single ended class A stage the output current can never be higher than the idle current so the output power will be theoretically max (R*I^2)/2 or 1.25W, (in reality it will be lower as you can't use the whole available current swing). To get higher output power you either need to run the tubes above ratings or add more tubes.
Replacing the inductor with a CCS doesn't change the calculations, output power and output impedance will be the same but the frequency response and distortion can possibly be better, it depends on the quality of the CCS.
Regards Hans
A single 6C33C have a Gm of 40mA/V and can handle the same current as 4 6C19, the output impedance is also about the same or about 22ohm, I think that is a much better choice as you don't get the problems of matching the tubes.
Max output power is determined by how much current you can run continously in the tubes, each 6C19 is specified to max 140mA so 4 of them can run 560mA. As we have a single ended class A stage the output current can never be higher than the idle current so the output power will be theoretically max (R*I^2)/2 or 1.25W, (in reality it will be lower as you can't use the whole available current swing). To get higher output power you either need to run the tubes above ratings or add more tubes.
Replacing the inductor with a CCS doesn't change the calculations, output power and output impedance will be the same but the frequency response and distortion can possibly be better, it depends on the quality of the CCS.
Regards Hans
No, I'm merely replying to the original question - and that was why there are no single triode OTL power stages. The answer is you can do better.
If you really want OTL, why not a hybrid with MOSFET followers? Even a cheap IRF610 has transconductance 800mA/V, which is an order of magnitude better than 6C33C tube (not to mention size and heater power advantage)...
This is pretty much the standard solid state design philosophy. Global feedback strapped on top of a bunch of non-linear stages. Makes it all better, right?
.- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.