op.9 wrote:

A 20w tripath and a 5w El84 set per channel is a sensible and cheap combination. The tripath and line level crossover section could be thrown in for free!

Ah, the EL84! That brings so many memories…. I was 9-10 years old and was running a PP with Russian 6N14P. Somebody told me that there was a “capitalist” equivalent for this tube and it had a mysterious name EL84. I found that EL84 tube (how that hell we did it at that time without eBay?) and leaned that my amps begun to sound better. I was very surprised….

To the broader decoud’s subject about the DSETs – it unquestionably the direction to go as it deliver better result for less cost – I know, I know – to controversial for high end audio’s objectives. A right selection of tubes and good architectural design would be a very useful in this direction. The tubes should have the “same” sound and be able to be driver by the same driver stage. It would be nice to have a design where the LF output tube would be driver by voltage double from the MF output tube or something like this. For instance the 2A3/45 for MF and GM70/211 for LF – juts hypothetic…

BTW, pay attention to the universality of 6C33C. It does sound very fine for MF and allowed to be used at “half anode” but at the same time it has remarkable low plate impedance and can do a very good 15-20W LF tube. They might still need two separate PS most likely…. The best would be a tune with 3-4 plates what the plated would be connected according to loading needs…

I went today to Srajan’s site check the news and saw the following:

“April 08 - New HalfNote Audio brand: The latest addition to Mr. Barnum's growing stable of import/export products are the $70,000/pr 50-watt parallel 845 Grail SET monos by Chalice Audio which are driven from a 300B with a 5681 input triode. At 150 lbs for the pair, the Grails pack serious iron and at 400 watts consume stout juice to become the latest entry into the high-power SET sweepstakes driven up in price and expectation by Vladimir Lamm's most recent GM70 monoblocks. A current Positive Feedback Online review comments on sonics of the Chalice Audio Grails.”

Surely the session for high-price, high power SET idiocy is opened, but how about we apply my DSET rule: “Any SET that cost more than $5K is west of design objective and it could be easily defeated by a DSET”

The $70,000 for the parallel 845? Why the 845 need to be parallel to begin with? Because they need more power at LF? So, we have compromised parallel MF amplification juts because bass need more power? In any SET amplifier bass region is distinctively different animal with the rest of the region and need own handling.

I hope people embrace DSET notion as the only negative effect of a properly implemented DSET is that it drivers the prices significantly down - that is not a good idea for high-end audio at all….

decoud wrote:

Romy, can you be persuaded to post a sketch of what you think the optimal 2 channel Melquiades circuit would look like?

Decoud, I do not think that there is such a thing as optimal DSET. The entire concept of DSET is a DEDICATED SET or the amps that dedicated to serve the specific purposes. If you look at the schematics of my 6-chennal Melquiades DSET then it gives a very good blueprint how the 2 channel Melquiades might be organized. Still it would be all depends what you would like to accomplish and what are the condition of your acoustic system/room.

Very generally, let make a mental exercises. In order to make a proper 2 channel DSET (doe sake of simplicity let look for identical output tubes) you need to have properly sounding full-range amp and confirm that it does what you feel it right. The amp must have enough power to driver your system.  Then you make two amps identical amps for LF and HF and define what you would like to have a crossover point between them. For the LF amp you’re built a large inductance out transformer and do not worry about the capacitance of that transformer. It will be large but cheap. Target your bass transformer to roll off at 2-3 octaves above your crossover point. For HF amp you make a high quality low capacitance transformer but do not worry about inductance and the necessity to care spare DC current. In fact that transformer should have absolutely minimum inductance, target 3-4 octaves under crossover point. The coupling capacitor of the HF amps acts as the high pass filters, preventing the low inductance HF OPT to be stressed by full-range signal.

So, not you have a perfect DSET configuration and you need to optimize your loading for each channel. You might (or might not) discover that you would like to drive bass with very different operation point then you would need to make charnel in LF’s amp power supplies.

Now, what you have two perfectly operating narrow range SETs you look at them and try to remove a common denominator from them. You might combine drivers, power supplies, servile elements, chassis, controls, filaments, bias… whatever you found is combinable. It might be that 90% of both DSETs would be combinable.  In some case you can have one SET with two out tubes and two output transformer, driven from the same driver stage… it all depends…

In some case you might chose less powerful tube for FH and more powerful tube for LF. In some cases you might have two-stage DSET for FH and but for LF to use a big transmission tube and to put some kind of cathode follower between the HF’s driver tube and the LF output tube, making the LF to see 3 active elements. That would be the same DSET idea…

I do not think so. There is no such a sing as optimal generic DSET but there is a concept of optimal DSET for a given situation. So what I think would be nice if some kind of company come up with a Reconfigurable DSET . The Reconfigurable DSET, or RDSET from not and on, is a two way commercial DSET that has 4 parameters interchangeable depends of the customer demands. It might be not infinitely interchangeable but to be preset for 3 positions for instance.

1)      Low pass fitter in LF stage + 2 octaves LF transformer roll off

2)      High pass filter + 3 octaves HF transformer roll off

3)      Volume between the LF and HF channels (it might be an steeped attenuator)

4)      Power, loading and output impedance of LF channel

The idea to have custom not DIY opportunity id wonderful – in fact it is how it should be. Look at those Thailand guys – they built dozen custom amps. Unfortunately there is no in audio a protocol known to me according to which the proper requirements might be supplely. The boulders would build according to their understanding of what is right; they might or not might be what you appreciate. Even if you develop a detailed circuitry and provide a very detailed and specific specification then there is still a lot of room what you might get. You need to have very in-depth and very intrusive control over what people (if you do not know people) do in order to have the results exactly as you expect.

I think the best is to develop and scope a project and then to recruit people paying them by time. Usually it turns out to be more expensive to yourself. Also I personally never was too successful with this approach as I am pain in ass customer. I also not necessary always know what I need and have more unanswered frustrations in my “designs” then the solid solution to those frustrations. The success is to found a person who is cable to render your sonic frustration into audio implementation… good luck with it…

Anyhow, I think audio people will come eventually to RDSET and tailored acoustic systems. If you traveled and visited different playbacks around the Wold and around your country then you know that then most interesting installations are those where owner have custom-tailored playback for his needs and for his few of sound, not the industry’s pre-caned audio irrelevancy .