I have been trying to find the original Fletcher and Cooke OTL article online with no success. Anyone have a pointer for me? It is public domain is it not?
Talk of transconductance amplifiers and the sometimes bad behavior of OPTs at low frequencies got me to thinking about the suitability of early OTL designs as transconductance amplifiers. It seems that one of the biggest problems with early OTL attempts is that it was difficult to get high damping. No problem if you want a gm amp.
It seems like direct coupled PPP CF output stages might be just the ticket for open baffle woofers with Qts lower than generally considered proper for OB.
Talk of transconductance amplifiers and the sometimes bad behavior of OPTs at low frequencies got me to thinking about the suitability of early OTL designs as transconductance amplifiers. It seems that one of the biggest problems with early OTL attempts is that it was difficult to get high damping. No problem if you want a gm amp.
It seems like direct coupled PPP CF output stages might be just the ticket for open baffle woofers with Qts lower than generally considered proper for OB.
Figures that I would stumble on it looking for something else. 
http://www.clarisonus.com/Archives/Amp_Design/Cathode_Follower_Amp.pdf
Output Transformerless is not mentioned in the title so that is why my searches failed.
http://www.clarisonus.com/Archives/Amp_Design/Cathode_Follower_Amp.pdf
Output Transformerless is not mentioned in the title so that is why my searches failed.
Here is a discusion of different OTL circuites. One circuite just like this is described
using 8 6as7g tubes to provide 6.32W @ 16 ohms.
6C33C-B OTL Amplifier - Background and OTL Circuits
using 8 6as7g tubes to provide 6.32W @ 16 ohms.
6C33C-B OTL Amplifier - Background and OTL Circuits
Thanks Woody. That looks like a nice tube for the application as it looks like it has plenty of current capacity which is the other problem with this topology. Of course if the plate dissipation exceeds 30W then the current capacity isn't that much greater than the KT88. However if this is designed to operate at such low plate voltages it might be possible to get an operating point where one could idle at 400mA or so. If so a quad of them might be able to make decent power into 8 or 16 ohms.
The problem with hollow state OTL is three-fold:
There are no "P-Channel" VTs.
This remains a high voltage, low current device. The cathode follower can't magically make it into what you really want: a low voltage, high current device. It's still the same loadline, and VTs don't like operating into nearly vertical loadlines.
Your choices for OTL power finals is strictly limited. Not many types were designed to source heavy currents. About your only choices are 6AS7s and their variants, or that Russian, triple-nipple job made for the MiG.
That Fletcher/Cook design is just plain ridiculous: less than 7W into 16R. Don't forget: eight 6AS7s is really 16 triodes, since the 6AS7 is a double triode. Cathode followers are terribly inefficient when operating into nearly vertical loadlines.
You'd do much better with a SEPP variation since this provides active pull-up and active pull-down (much like the output stage of a TTL gate). That means a Futterman, Inverted Futterman, or Circlotron design. (All of which are fundamentally interchangable.) Eight 6AS7s in one of those topologies could get you at least 20W, and more if you really spec-bust on the max current limitation of 125mAdc.
If you go with the 6C33C, these guys definitely need pre-conditioning before you put them into active service. I've seen a couple of suggestions for that:
1) Run the heaters only for 3 days, with all electrodes grounded.
2) Run then at a plate voltage not more than 100Vdc, while drawing enough current for near-max plate dissipation. Watch the cathodes until you no longer see any white spots.
In operation, it's imperative that you preheat before hitting them with the high voltage. The spec sheets for both types specifically call for this. Remember: these were designed for series pass work, and the heater insulation is a good deal thicker to stand up to higher voltages, so the cathodes are slow to warm up.
When mounting, make very sure you have good ventillation. The A Number One fault with this Soviet-era tech is that the pins are too thin. You definitely need to wire the sockets with silver solder, and supply adequate air flow past the socket to prevent the deterioration of the contacts. (MiG techs routinely replaced sockets every time they replaced the 6C33Cs.) Poor contacts lead to poofage, and your speeks will likely
There are no "P-Channel" VTs.
This remains a high voltage, low current device. The cathode follower can't magically make it into what you really want: a low voltage, high current device. It's still the same loadline, and VTs don't like operating into nearly vertical loadlines.
Your choices for OTL power finals is strictly limited. Not many types were designed to source heavy currents. About your only choices are 6AS7s and their variants, or that Russian, triple-nipple job made for the MiG.
That Fletcher/Cook design is just plain ridiculous: less than 7W into 16R. Don't forget: eight 6AS7s is really 16 triodes, since the 6AS7 is a double triode. Cathode followers are terribly inefficient when operating into nearly vertical loadlines.
You'd do much better with a SEPP variation since this provides active pull-up and active pull-down (much like the output stage of a TTL gate). That means a Futterman, Inverted Futterman, or Circlotron design. (All of which are fundamentally interchangable.) Eight 6AS7s in one of those topologies could get you at least 20W, and more if you really spec-bust on the max current limitation of 125mAdc.
If you go with the 6C33C, these guys definitely need pre-conditioning before you put them into active service. I've seen a couple of suggestions for that:
1) Run the heaters only for 3 days, with all electrodes grounded.
2) Run then at a plate voltage not more than 100Vdc, while drawing enough current for near-max plate dissipation. Watch the cathodes until you no longer see any white spots.
In operation, it's imperative that you preheat before hitting them with the high voltage. The spec sheets for both types specifically call for this. Remember: these were designed for series pass work, and the heater insulation is a good deal thicker to stand up to higher voltages, so the cathodes are slow to warm up.
When mounting, make very sure you have good ventillation. The A Number One fault with this Soviet-era tech is that the pins are too thin. You definitely need to wire the sockets with silver solder, and supply adequate air flow past the socket to prevent the deterioration of the contacts. (MiG techs routinely replaced sockets every time they replaced the 6C33Cs.) Poor contacts lead to poofage, and your speeks will likely
This OTL design has been around for a while and the build quality has always impressed me.
Ip +‚h‚“‚‡ Hybrid EL509 SEPP OTL stereo power amplifier
Ip +‚h‚“‚‡ Hybrid EL509 SEPP OTL stereo power amplifier
Given the number of contemporary references I take it that the Circlotron is the preferred option these days for most folks. Will have to do some research on these and see how they deal with the inherent imbalance of the SEPP (which it appears is a near kin to the mu-follower). I also recall seeing somewhere (may have been John B.) one based on PP white cathode followers.
I wonder about Kimmel's sweep tube version scaled down to smaller sweep tubes. I only need about 20 or maybe 30 watts with Zout around 50 ohms. Inverted Futterman huh? Another thing to look up.
I built that, and with a pair of 6C33C at 210V plate supply, I got 56W RMS at just under point of clipping....
I redesigned it to remove the mosfet screen-drive and just used conventional cathode-follower driving the grids of the 6C33 tubes....
Its a complicated amp with those two diff-amps in the 'guts' of it, but does work well....
Inverted Futterman huh? Another thing to look up.
Here you go, Have a look at the schematic at thebottom of this page..
http://www.diyaudio.com/forums/tubes-valves/189282-vacuum-tube-otl-power-amp.html
Nice and simple and will give you around 25W into 8 ohm
The original Futterman design used a cathodyne splitter with the plate connected to the upper tube (cathode follower) on the totem pole. The cathode was connected to the bottom (grounded cathode) half, with the tail of the cathodyne returned to ground through the load, so that it would receive the signal as well.
He originally believed that this feedback would make the lower (GC) half like a cathode follower. What he didn't consider is that the whole cathodyne simply bounced up and down on its tail, making the feedback to the upper (cathode follower) half positive, and therefore like the lower grounded grid half. It accomplished balance, but in the wrong way, and not what Futterman wanted. Two grounded cathode halves on a totem pole makes for a Hi-Z output, not what you want.
The Inverted Futterman reverses the connections: plate of the cathodyne to the grounded cathode half; the cathode to the cathode follower upper half. The bottom of the tail is still returned via the signal output, but the feedback is phased correctly to make the grounded cathode half more like the cathode follower half. This gets the Zo down, and makes for the lowest output impedance of any of the usual OTL topologies: SEPP, Futterman, or Circlotron.
Actually in my case (OB modern woofers with over damped low Qes response) this is what I want within reason (25 to 50 ohms or so). However in a low powered version with only a few output tubes the inverted version may in fact still be high enough.
Ah, OK. I get it. So the PI still bounces around but the resulting feedback is more in line with the normally desired results. Is the stability enhanced by going inverted v.s. the original design? Does the resulting feedback (in either case) actually result in perfect balance?
Interesting stuff this.
No. The Inverted connection drops the Zo. Stability is still up to good design, lay out, and construction techniques.
No such thing as "perfect balance". Components -- passive and active alike -- have normal variances. Tubes age, and even if they start out perfectly balanced (a dubious proposition at best) they won't stay that way for very long. The true electrical mid-point in any practical OPT or PTX is not the same as the center tap, though it will come close, but not coincide since the coupling coeffcient between turns is never 100%. And so it goes...
You always design to specified tolerances because nothing is ever 100% perfect.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.