KOTriode wrote:

The Melquiades give me inspiration to build this 6C33C amplifier. It's on a stereo chassis that has few unique features - Vacuum tube rectifier separately for Input/Driver and Output tubes. - Regulator using the same tube types as the amplifier (for same sonic signature). - Direct coupling without capacitor in the signal chain (from input to output). I prefer this method over interstage transformer. - No negative Feedback. - Can easily be reduced to two stages.

Interning, KOTriode, thanks for sharing. Would you like to publicize the amp security?

The Vacuum tube rectification was greatly debated during the creation of Melquiades. My contusion was that if the soft-switching SS rectification is use and the most important if the input choke filters used with large last cap then there is no need in tube rectification. In addition, I was not able to get bass that I was after with tube rectification.

Regulator. This is a very interesting subject. There was a LOT of debate in Russia of regulator has to be used form 6C33C. Some people feel that it is very much must and others feel opposite.  I did single experiment regulating my 6C33C with 3x807 triodes. I was not impressed with the result but I do not insist that my result was definitive. The Jury about the need of regulation is still out in my view. Of cause the most powerful statement of the need of regulation for this tube would be the use of regulation in the best commercial amp built about this tube – Lamm ML-2.0. Still it was a commercial amp and it was “the Vladimir” – there are 2353 other sinister reasons why he would use more tubes in his design.  The reality is that I would be very interested to learn how he 6C33C behaves regulated and not regulated. The reason I did not go into the depth of this subject at the time when I worked on Melquiades was my modeling of regulation with my phonostage. I had built a very good regulator around 105V gas voltage reference tube, the 7788 pentode as Arial amplifier and 6C19P triode as the regulator switch. It was working fine (still have it), sounded fine but I detected that it a bit hold the Sound back and did not let to play with “hair down” if you know what I meant. At that time that slightly more polite and overly courteous  then I would like it to be sound of my Lamms ML2.0 power amps was exactly what I would like to defeat and it lead me to  abandon the regulation as the possible reasons for the “softness”. I do not insist that it was THE reason and I do not mind to try the regulation on 6C33C again. Sure, I will not be doing it now but I am still observing the commentary of others on the subject.

The direct coupling without capacitors or interstage transformers in the signal chain is fine. I juts a bit afraid of this 6C33C tube. It is not type 45 tube and not 2A3 that will run for 45 year with no sign of any problems. This is Russian 6C33C with own stupid behavior and ability crash. I do not know what kind driver you use but it get to be very reliable as if your driver in direct coupling application go down then your 6C33C will fry not only your speaker but half of your house. Talk to the people who use 6C33C for OTLs… BTW, when I saw you use 2 driver tubes per channel I though you use some kind SRPP to drive 6C33C…

Anyhow, it looks like nice amp. The design and assembling techniques are very pleasant. What driver and what kind output transformer you use? Did you find while you was working with 6C33C anything that I did not cover at my site? The most interning: how do you find your amp sound? Can you elaborate about the sound of the amp in greater details? Is anything that you learned, discover, realized from the sound of your amp?

KOTriode wrote:

Thank you for the comment on this amp, this is my first time of using the 6C33C tube, I decided to build this amp after reading about the Melquiades, this amp is just a quick experiment on the 6C33C sound, since I have a single chassis intended for a phono/preamp laying around for year now, normally I would build a mono block. I picked up 4 these tubes for less than $50 on my trip to Asia last year, they are harder to find now over there, why? nobody care of these cheap tubes, so nobody stock them any more.

I am surprised to hear it. The 6C33C is incredibly widely available tube and it is VERY cheap. In the Milq time I was getting them from Russians for under $5 shipped to Boston. I think nowadays they are a bit more expensive but still then are in the range if 12-15 bucks for tube in you buy a quantity of them. They shall be widely available on eBay and everywhere else.

KOTriode wrote:

Vaccum tube rectification/regulator in amplifier: Yes,  there is quite a few debate on these subject. Two years ago after building a 845/WE212E amplifier with tube rectification/regulator (yes, with 1200V regulator), and compare the sound with previous rectification/unregulated 845 amplifier, I would never go back. With this 6C33C amp, I have made an extra step in this experiment, using the same tube in the amplifier for the regulator, so that the regulator tube  do not add it's own tube sonic signature (I normally used 6AS7 in other regulator). Also, with tube regulator, I did not have to use too much capacitance, and do without choke, saving amplifier real estate and weight. Another strong point for regulator: I can adjust B+ anywhere from 160-220V, and with adjustable bias on the tube from 100ma-300ma, I should be able to find its sweet spot.

I am not convinced by “I would never go back” comment. In addition I think that each type of tube, each type of biasing and each specific application might response differently to the benefits of regulation. Your experience with tungsten-coated direct heated tune might be very different with indirect heated tubes. What I am trying to say is that I would find your experience more educational if you use the same 6C33C in the same application with and without regulation. For instants I was able to get much more interesting sound regulating the screen supply of phonostage input pentode (7788). Still, if I regulated with the same regulator the 6C33C grid supply then it was disasters to sound. So figure…

Another very important reason why I do not regulate my 6C33C is that you have mentioned – the regulation does not go with input chokes and large caps. Regulator does not drive them well and mostly I saw after regulator very flimsy filtration. This all make the stage to sound pretty much like the “sound of regulator”. I do not like it. I tend to stabilize current with very big and heavy PS along with strictly input chokes. The adjustable current on the tube from 100ma-300ma and 160-220V? Why do you need it? The experiments with tube operational parameter people do during the design phase and in most of the case the tune is drive from external power supply. I use Fluke 409, half-ampere, 600V PS with two dozen tubes in it. In the operational amp there is no need to change current so much. Also, if you change the current from 100ma-300ma then you change the plate impedance too much to keep the same transformer ratio…

KOTriode wrote:

It's still too early to tell how this amp sound, as of last night the meter showed only 6.6 hours of break in time (you can barely see the hrs above the "Curtis" hour meter. So far I learned that one of the brand new tube failed with properly heater burn in, it needed -125V bias instead of around -70V. I also learned from the tube seller that you need burn-in 24 hrs just for the heater on these virgin tubes before applying any current. This amp run really hot, the whole chassis get hot, I installed an fan under each power tubes and and it helped a lot. I should able to tell the sound of this tube after about 100hrs on the meter.

The hrs meter was a great move. I would like to have it on my amp but it is too late. Post your observation about sound when you feel you are ready. Also, be advised that 6C33C is indirect heated it means it will not work in A2. With your former 845/WE212E if your driver was able to drive them with plate currents then these DHT more or less were operational on the “negative side” the 6C33C will clip VERY sharply. I do not know what you drive with this amp and what the size of your room but make sure that it does not send the 6C33C to A2.  H

http://www.goodsoundclub.com/Forums/ShowPost.aspx?PostID=6057

KOTriode wrote:

Without regulation, this amp would not work correctly due to its design: DC coupled and fixed bias. The high transconductance will make the bias current varied as much as 50-100ma due power line fluctuation without the regulators. It might work if the amp used cathode bias, but I chose fixed bias for various reason. The input stage (with or without the Cathode follower stage) bias the output tube from the plate voltage, fluctuation in unregulated B+ will cause the grid voltage to swing which in turn change the output tube bias current. Transformer with high current output, choke and capacitor would not help much with this design with line voltage fluctuation.Also since this is a setreo amp, good regulator definitely help crosstalk.All my systems are multi-amp and put a heavy load on the power line on the average of 1000W and more, so regulator in this amp is necessary, not an option. Why the adjustable voltage?  this tube to me is an experiment,  the regulator for me is very convenient to understand the bias of this tube with different plate and grid voltage in the amplifier, and how the sound will be affected with different operating point. I am not too crazy building a high voltage amp (the voltage is as high as 520V on this amp) with multiple power supply and wire hanging. This amp is never intended to work in A2, so does my other DHT amp. The amp can be played with or without the Cathode follower, my aim to get the driver as linear and wide bandwidth as possible, hence feedback will be unnecessary. I will try the amp with few different speakers as full range and later in tri-amp system, among them are mini monitor, Stacked Quad ESL, Horn system.... It would be very interesting to compare with the big DHT tube. Into the 10 hrs of break-in with a monitor, I am very happy with the sound so far.

KOTriode, I think you are missing what I am trying to say. It is very much not the debate of the usefulness of regulation but rather me stressing the point that neither yours nor my experience with regulation does not provide a universal answer if the 6C33C might be benefited by regulation.

It is not doubts that your design needs to be regulated to stabilize all your direct coupled operational parameters. However, your regulation does not address the needs of 6C33C but rather serves the environmental purposes. So, even in your design we do not know how 6C33C will sound in the same application with no regulation. BTW, I regulate my power AC supply – here is another way to stabilize operation without use of DC regulators.

If you look closely then you will see that in Melquiades I use regulation on the grid supply of the driver tube. I do have very substantial advantage on sound quality but here is an interesting question: do I have this advantage become of regulation of because some other reason?  The negative regulator in Melquiades drives the driver grid. The grid has own minute grid voltages that run back to the PS. Here on the PS side we have gas regulator – a very soft entry that instead of hard-shorting grid voltages or bouncing them back, the gas regulator very gently absorbs them. So, it is possible that in my case the regulation itself is not important but rather that I use the regulation as some sort of damper that sucks out the reflective voltages from driver grid. If cause if I go with voltage stabilization on negative side on the biasing grid resistor then I need to have one on positive side to lock the bias. BTW, try the same biasing approach on the 6C33C and it will not sound good at all.

What I am trying to say is that so far I do not see any evidence for or against regulation of 6C33C. Neither your use of mandatory for your circuit regulation not my “creatively-intentional” use of bias regulation does not say anything about the regulation and it benefits for sound of 6C33C.

About the A2. I do not say that you intended to be in A2 but driving the 86dB sensitive Quad in a room larger then a dog house your bass will be jumping for a fraction of seconds into A2. The  845/211, beside to having more power will handle this entering into A2 with some distortions but it will still be fine and the signal will be cared on. The 6C33C will not even go into A2 – it will flat cut the signal producing enormous amount of distortions across all bandwidth. All the I a saying is that being indirect heated the 6C33C is much deadlier with grid voltage approaching the bias voltage. If you have 65V on grid and 30-40 times gain in your driver stage then it is not difficult to be “right there” very fast. Sine you use an extra buffer stage, useless people do it to get current drive then I remind you that it would be fine with DHT 845/211  but it will not “help” to any indirect heated output stage.

About the 6C33C vs. 845/211. This might be interesting. Are your 845/211 also regulated and direct coupled with use of the same Tango magnetics?  Are they use the same Tango’s vintage? The Tango 20 years back was very much not the crap that Tango dose nowadays, if you know what I mean…

oxric wrote:

I am looking for someone who might be willing and capable of building an amplifier inspired from Romy's Melquiades schematics for me once my horns are finished.

Romy the Cat wrote:

If I were you then I would not pursue for any amplifier and particularly DSET. If you are building the building a multi-way horn system but build it, drive it from whatever amplifier your have now, get a good sound and good setup of your multi-way in your room. It might take a year or so, then introduce anther, perhaps better amp and begin to evaluate how different amp might or might not benefit the sound of your setup. The Cat

KOTriode wrote:

With 6C33C output run at 40W plate dissipation, this amp run much hotter than an 845 amp with output tube running at 100W.  I measured the hottest point on the bulb and it is about 380F. I guess the due to short bottle , the heat spread on the chassis make it hotter than the tall 845 bottle.  I have to increase fan speed under each tube to cool it down, while on bench , I make some measurement on the amp. Frequency response at 1W: 10Hz @ -1.37db, 59.1Khz @ -3db. 20Hz @ -0.45db, 20Khz @ -0.3db                                10W: 20Hz @ -1.1db, 20Khz @ -0.3db. This is excellent by SE standard and without negative feedback, better than any tube amp I ever built. Picture below is square wave performance @ 1W, 100Hz, 1Khz, 10 Khz. The bottom trace is the amp input, the top trace is the amp output. The last picture , the bottom trace show the grid input of the 6C33C at clipping , the top trace is the output of the amp, the 6C33C can't go into A2 as Romy mentioned.

KOTriode, a few things. The 40W plate dissipation is bit too low for both plates. The well heated 6C33C (at least 30 min) you can would drive at 50-55W. If you do not need power and 40W on plate is enough for you then use one anode on the 6C33C, it would be better for thermal operation of the tube and will be positively reflected on sound.

About the measurements. I personally discard any talk about frequency response if the measurements were no done at fool power. You got 10W but I think that with 40W plate dissipation you shell get around 12W. I “feel” that you might have a bit “small” transformer for 6C33C as fractional power I would anticipate better numbers then 20Hz @ -1.1db, 20Khz @ -0.3db. So, in order to speak with the same language I will present my SET frequency response acquisition guide. I am sure it will be nothing new for you or to anybody but for some reason very few people follow this directive and without the very literal follow on this sequence I feel any conversations about frequency response are worthless.

1) Turn amp and heat it for at least 1Hr

2) Connect you loudspeaker, play music and find what tap or combination of sections on your output transformer produce the more proper balance of harmonics within you playback.

3) Recalculate the ratio of your output transformer and find what impedance the output stage sees.

4) Substitute the speaker with 15-20W resistor of the value that would assure output tube to see the same load after the transformer. Let premise it was 12R for instance.

5) Drive 1000Hz into 12R and increase the input voltage until you see very first clipping, from here ad extra 5-10% of input voltage.

6) Measure the response of the driver and assure that driver is not clipping.

7) Using voltage and current on the output tube try to make the clipping symmetrical from top and bottom of the sinusoid. The top way is voltage, the bottom wave is current (or vice versa, I do not remember already). As the symmetrical will be find try to play with operation point more in order to get lessen symmetrical clipping at higher power.

8) If change of voltage and current will get rid of clipping then add more input voltage to drive the amp into clipping again.

9) Let say that the final numbers with perfect symmetrical clipping against the GIVEN IMPEDANCE would be 190V and 280mA winch gives 53W on plate. Now drive more or less input voltage in order to maintain the best symmetrical clipping as amp distort more or less. The objective is that output tube shall starve from voltage and from current identically.

10) Mark the found generator’s output voltage as the reference voltage.

11) Now, as the proper operation point against given load is found drop the input voltage up to the point as the clipping just smoothed out. Switch timing on the scope to faster to see the precise moment the amp stop distort. You need to find a few last mV where amp clips – this will be the true max power of the amp.

12)  Without touching the output voltage of your generator increase the frequency of your generator unit you see the very first clip. This would require you to adjust the timing of your scope.

13) From the position of your very first clip drop the generator’s output voltage to 3dB and drive frequency of your generator until you see the very first clipping again. Read the frequency from your generator display. This is the true upper frequency response of your amp at full power

14) Return the generator’s output voltage to the reference voltage and drive the frequency all the way down to the first clipping. Repeat the same ceremony for LF and find he true lower frequency response for your amp against given load.

15) Disregard your “hurt” ego as your frequency response will be not as great as the Morons claims out there. No you will not have 10Hz to 120kHz as some fools claims as you did measure your amp properly and at full power.

Generally an average 6C33C output stage with a proper transformer shell have 20Hz-23Hz at bottom and around 19kH -20kH atop. If you have some VERY good output transformer then you might get in full range operation something like 15-17Hz at about and 25kHz -30kHz. The 17Hz-23kHz at the very fill power of around 16W-17W are VERY good numbers for a full range SET.

7) Using voltage and current on the output tube try to make the clipping symmetrical from top and bottom of the sinusoid. The top way is voltage, the bottom wave is current (or vice versa, I do not remember already). As the symmetrical will be find try to play with operation point more in order to get lessen symmetrical clipping at higher power.

8) If change of voltage and current will get rid of clipping then add more input voltage to drive the amp into clipping again