Update: 4/27/97

Quick little FAQ on the 6C33-C-B


Please note this page is provided for historical purposes only.. I long ago abandoned OTL's in favor of SE amplifiers except for headphone amplifiers. I will not publish, nor share the schematics to this design as they are flawed and not representative of a completed work. Please don't ask.. --- Kevin K.

The 6C33-C-B is a low mu dual power triode built originally for the Russian military. Its primary application was (is) as a high perveance low voltage series pass regulator in MIG aircraft. The construction is mechanically rugged to withstand 10G of acceleration and up to 100G of shock - I think not operating in this case. Service life is rated as 1000 hours, but in civilian use as a power amplifier life could be as great as 10K hours. The tube utilizes two separate filaments, and may be operated with one or both simultaneously. One filament operation reduces plate dissipation to 75% of maximum rating.


Translation of Russian Language Spec Sheet.

EMPIRICALLY DETERMINED DATA

mu is approximately 5 - 6

power dissipation should be limited to 40W

Vplate (Eb) should be limited to less than 140V in OTL applications (also per Ralph Karsten of Atmosphere)

Must ventilate well as total dissipation including filaments is typically 80W

PUBLISHED DATA

Filaments: 6.3V @ 6.6A or 12.6V @ 3.3A (Dual Filaments)

Plate Resistance: 80 - 120 ohms, typically less than 100

Plate dissipation: 60W, see warnings

Absolute Maximum Plate Voltage: 400Vdc

Maximum Grid Circuit Resistance: 200K ohms

Basing: Septar


My Amplifier Project

I am currently designing an amplifier around a sextet (6) of these tubes for each channel in a configuration commonly known as the Wiggins Circlotron Circuit popularized over the last decade or so by Ralph Karsten. The design has the advantage that it looks a lot like a cathode follower to the previous stages and is source impedance and gain symmetrical unlike the Futterman designs, but as it is essentially a single-ended bridge amplifier it produces twice the output voltage and half the output current of a Futterman with the same number of tubes. Symmetry and simpler drive circuitry are the benefit, and loop feedback isn't necessary for linear operation again unlike the Futterman. My best estimate of output power is around 110W into a 5 ohm load. Toroids will probably be used in the power supplies, but they may be too expensive for my budget... Driver circuitry will be fully differential, and I currently have two favored topologies I have used in conventional transformer coupled pushpull amplifiers. I will start building the final prototypes within the next month, and will periodically update this article.

UPDATE1:

The power transformers are in! I chose to use EI types because of the extremely high cost of toroids. The chassis are being fabricated, and within a few weeks I will actually start building the amplifiers. At this time I will probably use high current augmented cathode followers to drive these tubes. If performance seems reasonable I will probably operate the amplifiers as a zero feedback design. Driver circuitry will be self balancing fully differential, and will have the option of single-ended or balanced inputs with just a simple modification or two...

UPDATE2:

The chassis are finally ready for assembly, and the driver, power supply, and output stage design is complete. The topology utilizes a wiggins circlotron output stage driven by a high current augmented cathode follower. The bias is applied to the cathode follower and the follower is dc coupled to the grids of the 6C33's through individual grid stopper resistors. Bias is simultaneously adjusted for both halves of the output stage with a separate control provided for dc balance in the outputs, the bias adjustment circuitry is designed to fail safe in the event of wiper failure in the pots. Although not currently planned dc servo circuitry could be added at a later date if dc/thermal stability prove to be a problem. The cathode follower is capacitively coupled to a diff amp stage which is in turn direct coupled to two differential input amplifiers that provide for independant feedback paths for each half of the output stage. Input drive can be either balanced signals or single ended as desired. Driver circuitry supplies and bias supplies are fully regulated. I expect to start the debug on these amplifiers within a month or so. Stay tuned for update3!

UPDATE3:

The amplifiers are almost finished now with just some minor power supply issues to be resolved. The design features full regulation to all stages with the exception of the final cathode follower and the floating supplies for the actual output stages. Due to short sightedness on my part I had to add two additional small filament transformers to each chassis as a consequence of minor design changes in the driver and power supplies. The amplifier has a total of 14 tubes, 4 each in the driver circuitry, and power supply, and of course the 6 output tubes. Wiring is Kimber, except in the output stage where I employed Belden teflon insulated 16 ga wire. Resistors are Holco, Resista, and some Dales, coupling caps are MIT/RTX. The design will operate with overall loop feedback of about 20dB from both halves of the bridge output, and the input topology will allow true balanced operation with only one minor change. Soon I will start the debug process, so stay tuned for part IV!

1998.3.19 UPDATE4:

The amplifiers are basically finished with a significant number of design changes being implemented, and debugging is well under way. They have run into speaker loads for as long as 2 hours without problems so far. The sound quality is not yet quite up to snuff with a small, but obtrusive amount of buzz present on the output. Current hurdles to be overcome include improving the quality of the treble, and overcoming biasing and matching difficulties with the output tubes. Drift is not a big problem, but establishing the desired bias point, and controlling balance in the two halves of the output stage is still a minor challenge.. In addition the driver circuitry adds an additional small current which is not part of the circulating quiescent dc current in the output stage, and some additional circuitry was required to null this out. The cathode follower current in theory causes common mode voltage to appear on the output - not an issue as far as the speaker is concerned, but due to the dc coupling back to the input error amplifiers this becomes an issue as it imposes an undesirable dc offset on the input differential amplifiers.

So far I have had a cathode follower tube which I believe failed due to a defective 6C33 output tube which had several visible hot spots on the cathode surface. The speaker line is fused as well as one of the floating B+ supplies in order to protect the speaker and if fortunate perhaps the output tubes as well, when these types of faults occur.




































1998.3.22 UPDATE5:

Some snow fell this week-end so I spent a fair amount of time working on the amplifiers.. Finally resolved a problem I was having with balancing the two halves of the output stage, and also figured out how to null out the common mode dc on the outputs due to the follower currents I am dumping into the output stage.

I am now succcessfully operating the amplifiers with all six output tubes installed, and have not seen any bias or offset instability worth mentioning, however I have observed that it takes about 15 minutes for the tubes to reach full emission, and therefore the final bias adjustment should not be performed until this time.

Russian quality control leaves a lot to be desired, of the 17 tubes I purchased 2 are clearly defective, (12% failure rate) one previously cited in update4, and the other newly discovered with a warped grid support structure in one section - although this one works some bias drift was evident with it installed in one of the amplifiers, and I was concerned that it would eventually develop a grid to cathode short and make a nasty noise, and ruin another 12SN7... The remaining 6C33 seem to be performing well, and offset and bias points are stable after warm up, performance in this regard seems to improve as the tubes age slightly. I have not yet matched the tubes so to prevent problems I am running them at around 200mA each, and after I match them reasonably well I plan to increase the idle current to roughly 300mA each for even better linearity.

The sound is getting close to what I would expect from these amplifiers, although I am not yet using them in my main system, the workshop system is not bad at all.. I am hearing things with these amplifiers that I never heard on these speakers before (Mission 763i), and the bass extension is beyond incredible.... They sound quick on percussive instruments in a way I have not really heard before in a tube amplifier. Treble is much improved as well, in terms of extension and quality.... I also still hear that darned 120Hz buzz when no music is playing - next thing to fix....

These amplifiers also make excellent room heaters - even runing at 2/3 intended idle dissipation they consume about 1/2KW each and just about all of this is wasted as heat. I don't plan to use these extensively during the summer months if they heat up my basement listening room the way they heat up my lab... They also are quite impressive looking to the initiated as well as the unitiated....

Eventually the plan is to use these to drive my MG1.4's, which are essentially a 5 ohm resistive load. Unfortunately OTL output power goes way down as the load impedance decreases. These amplifiers produce approx 130W into 5 ohms, close to 300W into an 8 ohm load, and a theoretically unbelievable amount into 16 ohms... Running these sorts of tests on OTL output stages seriously threatens the output tube life so I won't being doing any more testing at these power levels....

1998.3.28 UPDATE6:

Spent the better part of this week-end working on the amplifiers, sound quality is greatly improved, buzz is somewhat less - but not yet acceptable, one plenum with fan has been completed and installed on one of the amplifiers. The dc fans I am using are a little too noisy for my tastes, but cog badly (and audibly) when run at voltages below 10Vdc..

Mis-haps continue to plague the project, I discovered to my chagrin that the IR drop across the line to my shop was sufficient to keep the amplifiers happy during inrush and early warm up, however the listening room has very short direct feeds from the main entrance panel, and the amplifier ac line fuses blew instantly, and when replaced I discovered that the + series pass raw supply was 50Vdc higher than the cap rating which proceeded to cause further problems with fried fusing resistors in the high voltage supply....Inspite of these problems they did run reliably for about 10 hours over the week-end.

Did I mention they sound great, (when they are working) actually I am really beginning to understand the sonic allure of this technology.. Lots of detail, speed, unbelievable bass - tightly controlled, and neutrality of an order I have not previously experienced listening at home..

1998.4.27.98 UPDATE7:

A full month has passed since the last update, but plenty has been happening as I work out the kinks in the design. Fundamentally these amplifiers perform best when driven differentially (balanced) - an overly optomistic viewpoint on my part made it appear as though internal balance in the driver circuitry would not be a problem even with SE drive, but it turns out that best overall balance was achieved with 0 applied loop feedback... Poor sonic performance with the desired closed loop gain of 32dB forced serious bench evaluation. The gain reduction brought about as a result of loop feedback application was all occuring in the negative leg of the driver stages! (It actually swamped the desired differential output signal and produced a common mode output signal differing only in amplitude by -6dB/not phase!! to the output stage.) This seriously reduced the available drive voltage to the output stage, reducing the available output power, and the output stage insertion loss was nearly 6dB under worst case (highest feedback) conditions. I will have to remeasure the drive stage performance to ascertain that the available drive is adequate, now that the amplifiers have been modified for balanced inputs...

Interestingly the amplifier sounds very good without any feedback applied, hum however due to ripple on the floating supplies is quite audible with no music playing. Chokes or 120Hz traps in the floating supplies would probably solve this problem. In my case 10dB of loop feedback takes care of the problem handily, as well as assuring flat frequency response into real world loads...

The fan cooling is now operational in both amplifiers and keeps the chassis' cool over long hours of operation. Forced convection was the only way to go here due to the limited space, and a unbridled desire that the expensive components in these amplifiers live for a while. The fans are a little noisy when new, but quiet down after a few hours of operation. The 12Vdc power supply arrangement used to power them is a crude stop-gap measure, and I intend to improve them - perhaps even add temperature dependant variable speed control.

As I mentioned earlier I am now driving the amplifiers with balanced drive, I chose to use T3F and T3M connectors for cables and I/O due to their small size and similarity to their larger XLR brothers... All this means is that I also had to design, build and debug a pre-amplifier to do this. The pre-amplifier features a cross-coupled differential amplifier identical to my 6CK4 amplifier which drives a 5814 based dual cathode follower.. Relays are used for muting, polarity, and source select. Most relays seem to be lousy muting devices - even milli-ohms of contact resistance makes for less than exemplary muting performance, but I hate to add series resistors to improve the attenuation because it raises the output impedance of the pre-amplifier..

I am not satisfied that I am there yet in terms of extracting maximum performance from either the amplifiers or their new cousin - the balanced pre-amplifier. Inspite of the reservation the amplifiers exhibit great detail, excellent bass control, slam, depth, ambience retrieval, a certain palpability and presence, and balanced timbre. Better yet they have run for at least 30 hours without failure, and with only minor bias, common mode and diff. mode offset adjustments. The speakers may be safely left connected during power up and down sequences, making them practical for everyday use. It is very doubtful that I will ever go back to my conventional amplifiers once all the new electronics are complete, and debugged. (I need a new phono stage now as well.)


Back to Homepage