• 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.

What tubes for a OTL tube amp?

If you go back a lot of pages - about 40, Ralph had a different way of calculating that load impedance, which as I recall, came out much lower for 8 6AS7 triodes in PP (against another 8 triodes). I think it came out between 8 and 16 ohms, but I have to go back to lacing, so I can find it later. It is generally accepted that a higher impedance speaker ~ 16 ohms would be much better than 4 ohms for a match and power out. At this point, for me, it is kind of moot.
 
wattsman,

Are you talking about 15 series connected "4 Ohm"drivers? (60 Ohms load).

I expect the drivers have at least 1 Ohm DCR per driver; 1 x 15 = 15 Ohm total DCR.
15 Ohms DCR does reduce the effective electrical self damping of the speaker system woofers, Right?

And even if the series 60 Ohms impedance is the load line, how linear is a 6080 driving that?
(Before negative feedback is applied)?
 
Markus, running 4 6AS7s per channel seems to be in the 20-25 watt range.

There are several ways to determine output impedance, apparently based on what you believe in. I'm not a fan of what I call the Voltage Paradigm as it tends to lead to the use of negative feedback in a lame attempt to get the amp to look good on paper.

Gentlemen, what looks good on paper has little to do with how we perceive sound. The bench specs measure things that are not important and for the most part ignores the things that are.

So I measure output impedance by seeing what its actual impedance really is. Most techniques used by the Voltage Paradigm are actually putting a number on the servo gain of the feedback loop. Feedback enhances the odd orders used by the human ear/brain system to measure the volume of a sound, if we tamper with that the system will have a false loudness and a brightness. If you want it to sound like real music feedback has to be avoided. Consequently the method of measuring the output impedance looses is value as soon as the feedback is reduced or eliminated.

We have used this formula from the Radiotron:

2Rp/u + 2
where Rp is the plate resistance of all the power tubes in the circuit, as they are in parallel.

In the case of 4 power tubes this comes to 33.75/4 or 8.44 ohms. In practice our measurements come well within 10% of this formula.

If you use the Voltage Paradigm rules the 33.75 ohm value is what you will get. FWIW as I write this I am listening to an amp that only has 3 power tubes per channel, and its driving our test speakers (High Emotion Audio Bellas) quite well. Obviously a high damping factor is not all its been cracked up to be. I think that myth was started to sell the Voltage Paradigm to the public, and they bought it hook line and sinker. We all paid a price for that too but now I'm way too far down the rabbit hole...


I said I would find this, so here it is. Not saying which method is better but I am saying that I have been running a Futterman for almost a year now with no issues except that it sounds better than all of the commercial tube amps I have. I was driving the original build into AB for sure (again by meter indication) it is 4 6AS7s per channel. The newer one is 6 per and the circlotron is 8 (160V on the Futtermans and 140 on the M-60 based unit).

A friend of mind noted that it is harder to get good clean power at half a watt compared to 5 or 10 watts from most amps. One other thing I have noted is that the dynamic range seems to be extended in the OTLs compared to conventional designs. When that helicopter appears in Pink Floyd, The Wall you would think it was landing just to the left of you.

For me, there is no going back, but I could go forward if somebody has a better design, which is why I spent the last few months building a reasonable facsimile of Ralph's amp. I wanted to compare it to the Futterman, which I have yet to do with my primary speakers.

I will eventually note the mods I have added to the M-60 like metering the power rails, changing the fixed 37.? K resistor to a 50K variable (10w), 2 inverters per channel (suggested by Ralph) and a lot more. Most of this was to give me more precise control of individual tubes which can now be adjusted in pairs. BUT I can see why many of these mods might be bad in a commercially built unit offered to the public. For a DIY guy with a lot of experience, they are great.

My evaluation has been with my ears, of which I have the only two, so your results may be different, but I won't be going back to the Fisher, HH Scott, Heathkit, Dynaco, Sherwood - - - any time soon. But I could use some help moving the OTLs around.

If you haven't heard an OTL, find somebody that has one and give it a listen. The Old Radio Shop is always open to visitors by appointment (clubs also welcome) at Russ old radios .com (P.S. we don't sell anything)
 
That seems like that took some digging!

Since I wrote that I've changed some of my thoughts about feedback. Norman Crowhurst mentions that feedback is a applied to a non-linear point at the input of the amplifier which results in the feedback signal being distorted- so higher ordered harmonics and IMD are generated.

This can be avoided by simply applying the feedback to the input of the amp by mixing it with the incoming audio signal using a divider network; the same technique used in opamps. In our case we apply feedback to both grids of the differential input in the case of the M-60. Because resistors are a lot more linear than tubes, the damage the feedback does is less profound and its more able to do its job. We don't really have enough gain in the M-60 to take proper advantage of this but if another gain stage were added (or the top tube of the differential cascode were a 12AT7) we just might. This would allow the amp to have flatter frequency response on a wider range of speakers but would not do anything about the 4 Ohm problem that OTLs always seem to face.

I still feel that spec sheets are mostly the Emperor's New Clothes since they tell us nothing about how the amp will sound. But we do have the ability now days to make the measurements that confirm what we hear. IMO the two most important aspects of that are the distortion spectra (meaning that a 2nd or 3rd harmonic needs to be dominant at enough level to mask higher ordered harmonics) and that distortion must not rise with frequency. Zero feedback amps do not have rising distortion with frequency. It is that issue that was stopping us adding feedback to our amps for so many years. Its one of the more important reasons most solid state amps tend to sound bright (and harsh). To what degree often has to do with what frequency the distortion begins to rise.
 
So, I felt strong this morning. Picked up and moved the 6 (12) tube Futterman and hooked up the Circlotron in it's place. At first I was sure I had a phase problem. The sound stage was wide, but I could not point at the lead singer. The bass was strong but it seemed that some frequencies were muted - - Lifting OTLs and early mornings seem to affect my golden ears. After lunch I swapped +, - on the rt speaker. That was defiantly wrong. Returning it to normal and retesting proved the issue was with me. More coffee! (or less blood pressure)

I have a lot more experimenting to do , but initially the Futterman and the Circlotron sound very similar. There is no hum in spite of my RCA cables. There is better gain on the M-60 clone(s). I had to add another stage (unused portion of the first triode also changed to 12AU7) on the Futterman to get that kind of level. I think most people would consider that a real plus. The song I was just playing just ended - it is so quiet, which leads me to say - don't forget to turn it off. It's going to be 100 deg. plus and I don't need a space heater.

Question for Ralph: You have a .1 cap and 7.5 ohm resistor across the output. I thought that this may be because the output tended to be a little bright. I swapped out the 7.5R for a 0 to 50 ohm variable (10W, yeah I have a bunch of these). I wanted to see what effect changing that value had. Am I correct in my appraisal of that function or was it more inline with matching impedance?
 
Last edited:
You don't need that cap and resistor. They are there in case a power tube arcs to cut high frequency energy. But in practice tweeters don't go down if a power tube fails so most examples of the M-60 don't have those parts. If a power tube shorts its usually a DC event; so the fuse for the B+ transformer is rated to blow if the amp tries to make more than its rated power. this system has served us pretty well over the last 46 years; the only tricky bit is if someone is using a full range driver that doesn't handle much power. For those I recommend a series speaker protection fuse since some of those drives might only handle 5-7 Watts and could be damaged if you were careless with the volume control!
 
phlogiston,

Thanks! Great!

The ins, outs, and foibles of listening experiences, as well as the experiments to investigate or fix the problem(s), such as reversing one speaker's +/- leads . . .
is one of the best things about Tubes/Valves Threads.
Your Post # 1146 is a good example.

Anybody that thinks the time of day, coffee, near Hernia experiences, etc. does not affect the sound is Either not listening, Or has never changed their listening routine, and never changed their system.
(It is getting dusty under those speakers, and the rug is collapsed, and the rubber feet of the amplifier is fused to the table top, etc.).

Just my $0.03
 
I used a similar cabinet design that separates the hot tubes from the power supply. The top is hinged. The heat shield below is where the bias pots, fuses and other maintenance items reside. I intend to place a screen in the gap, probably chromed or polished, but I need about 51" x 2". Finding 51" is the problem. If a person wanted, this could become a 2-piece construction with the lid separated from the power supply. I had to add a fan to the first Futterman. Even with the tubes on top it was getting too hot inside - I don't like fans. It only had 4 6AS7s per channel.

The Futterman is setting on a 1920s radio cabinet. Every inch off the ground (floor) makes for easier lifting.

The Circlotron cabinet is solid walnut, the Futterman, mahogany. They have the same footprint but the Circlotron is an inch or two shorter and because of the toroids, is a few pounds lighter.

All of the 6AS7s used in the Circlotron are RCAs with the single grid radiator (sink). The Futterman has GE - GAs with no sink. They have been ok at 160V but if you can get the RCAs for the same price, get them. If you can get the STs at the same price, get them (2 sinks).
Circlotron and Futterman.jpg
 
Last edited:
Thanks for your input Ralph.

Here's a question that might have been asked but - - standing here , wondering what I will do with the larger Futterman if/when I decide to keep the circlotron in service:

I believe that I calculated that the Futterman could be bridged in my stereo application for single channel output. Has anybody ever done that with the M-60s? (assuming that you have all of the power supplies polarities set up properly)

I had thought of running the two amps into one channel (left or right) to do a comparison. After all they are just setting there side by side. They, of course, don't have to be bridged for that.
 
Stability is quite often a key issue when Bridging, and/or Paralleling.
Doing that with any pair of amplifiers is not something I will try.
Part of the success or failure depends on the speaker load.

Remember Galloping Gertie, the short lived Tacoma Narrows Bridge.
Stable, until a 15 MPH wind came along.
 
I used an outboard phase inverter to drive a pair of single ended amplifiers. The Se amplifiers input signals were out of phase.
Then, I ungrounded one SE amplifier's output common tap, and connected its common to the other amplifiers 8 Ohm tap; and connected that amplifiers 8 Ohm tap to the the other amplifier's grounded common tap.
This is a form of push pull, but using air gapped SE output transformers.

Caution:
Doubling the amplifiers in your system also doubles the chances of you getting a Hernia.
Check with your surgeon; I had operations for 3 Hernias already.
 
Last edited:
  • Like
Reactions: atmasphere
Air gapped transformers have real trouble making the inductance needed to play the bottom octaves. This causes the load line to be elliptical, which isn't the sort of way you want to operate a power tube. So full power to 20Hz typically requires a very large transformer even if only 7-10 Watts. Such a transformer would have grave difficulties making it to 10KHz.
 
atmasphere,

You are exactly right! . . . Elliptical loads.
I miss all the various shows that I used to attend, and seeing those OTL amplifiers.

Now, consider a typical closed box speaker system's woofer from 20Hz to 160Hz, 3 Octaves.
At 20Hz Z = DCR (4 to 6 Ohms)
At 35Hz, the woofer is perhaps 10 Ohms, primarily Inductive reactance.
At 50Hz, the woofer is perhaps 30 Ohms, primarily Resistance.
A 70 Hz, the woofer is perhaps 10 Ohms, primarily Capacitive reactance.
And . . .
At 160Hz, with the woofer's capacitive reactance connected in with the series crossover coil's inductance, the Hi Q series resonance at 160Hz, causes the impedance to be resistance (4 to 6 Ohms DCR).

That complex and varying load is presented to each and every amplifier that is ever connected to it for a listening session.
If that is not a mix of resistive, capacitive, and inductive load, which is Elliptical at many frequencies, I do not know what it is.

The listening tests usually survive all that, and people enjoy the music.
According to early engineers and physisists, the bumblebee was not capable of flying, but further studies proved that it can fly.
(The bumblebee, a very poor engineer/physicist always knew it could fly, so it always did).

Vinyl records, and the super small stylus contact area, and other factors, should not sound very good.
But, like the bumble bee, it works very well. I stole that from Lynn Olson's book: "The Soul of Sound" Thanks Lynn!

Eduardo B. DeLima gave a presentation at VSAC, and wrote an article for Valve magazine.
It related the ouput transformer's output reactance versus frequency, and the woofers reactance versus the same frequencies.
Sometimes a reactance is not your enemy, it is your friend.
 
Last edited: