Still, I would like in this thread to stay with subject of advantage/disadvantage of DSET approach vs. better/expensive SET: "A DSET is better then an expensive SET" instead of turn it into thread about David Berning. Bernings certainly might be an interesting subject and if you guys are wiling to talks about it then create another thread desisted specifically dedicated to Bernings.

Merlin wrote:

I have heard plenty of traditional SET's from Lamm to Wavac, AN, Kondo and many others. I know I could not live with their failings.

Merlin, in the context of this thread, might I ask you something... When you heard Lamm SET (presumable it was 2.0) then what speakers it drove and what specific problems you found that made you to feel that you can’t live with that sound? I think it would give me some impression what you are talking about and would set certain equalization point for comparing notes..

drdna wrote:

The altered topology of DSET has the potential to have disadvantages.  There is increased overall system complexity and this often seems to be a negative.  Beyond this, it is difficult to expound, for me, on the specific differences because, well is it meaningful?  The way to reflect on this is to listen comparatively to SET and DSET and observe the differences and correlate this to the toplogical changes.  The difficulty is that when one has the DSET the urge is NOT to live with the system as is, but to begin to modify and "optimize" each frequency range, so that it becomes impossible to make a real comparison.

Yes, you are absolutely correct and I very much back up it. When I had my first full-range Milq then I knew exactly what I did and what I was getting as a result (Sound). Moving with Milq into the DSET operation I lost my immediate and clean view of the Melquiades sound. The result is definably better but the reason-consequences patterns are not so easily interpretable, and are absolutely not interpretable when the filters are applied. That is why I absolutely insist that any DIYer or manufacture, if they are willing to “go DSET”, must to verify all operational, construction and performance parameters/characteristics with a conext of a full-range amplifiers and only THEN to apply the success to a DSET.

drdna wrote:

To this end, I would suggest to ROMY that he keep all the parameters the same from DSET as with the original SET setup and then live with the system for a period of time to make a true comparison between the two.  Only in this way can one hope to assess the topological differences' effect on the sound.  It would be nice to do this experiment, since he is in a position to do so.

Well, I understand the objectives and you are right in them but the experiment will not take place as I am already out of the league of the methodological correctness. You see, the 6-chennal of Super Milq is betraying of the DSET idea as the new Super Milq uses one-stage amps. Surely I made all precautions to keep all channels to operate similarly, topologically and operationally, but they still are different amps and the new Supper Milq has basically 2 pairs of 3-chennls DSETs (even despite the HF is not even triode anymore).

So, the Super Milq is not the best illustration of the DSET concept. Also, the Super Milq-type of amps are possible only in context of targeted binding to a specific acoustic system/room. The Super Milq is an irrational aberration, a result of unwell imagination if you wish. My advocating of DSET does not base upon the Super Milq’s “prove”. Also, if we mentally go away from the Super Milq class examples, then we will see that DSET is not necessarily a “remedy” only for DIYes but might have much wider commercial implication. Do not forget that the subject of the thread is not that “DSET is the best amp” but “DSET is better then expensive SET”.

When I said DSET I am very rational. The problem that a regular SET has might be addressed by 2 channels - where one cares LF and another rest of the spectra. THAT is my primary definition of DSET. That DSET is very easily implementable and marketable commercially. We are accustomed in High-End that the speakers are bi-ampable and that the cables are bi-wireable. Why a DSET should not feel the gap where it could. A DSET should not even care a crossover – most of “better” speakers allow bi-amping and their driver-level crossovers are bridged externally… The DSET would be a perfect match in there.

Anyhow, what I would like to see is the industry embraces the DSET idea and begins to educate the participants about the DSET advantages and the DSET economic reasonability. I am sure that the first company and the first marketing cheerleaders who would jump to the DSET wagon and would SPIN IT PROPERLY will make quite a good ride out off it.

Paul S wrote:

I will eventually get another pair of ML2s to slave the bass/save the upper registers, because, as we all know by now, dedication rules.

It wouldn’t be a DSET but rather just a simple bi-amplification. Bi-amplification has own benefits but it is not the same as a channel dedicated to own bandwidth. I was running for a few years two pairs of ML2 but I never considered that I used DSET. Well, I use some very rudimental optimization by driving the bass output tube differently, loading it differently, using different tubes, building filters in the ML2’s feedback but it system advancement not the amplifier advancement and therefore it sound not be consider that it ws a dipping into DSET world.

Romy the Cat wrote:

The result [in the 6-channel Milq] is definably better but the reason-consequences patterns are not so easily interpretable, and are absolutely not interpretable when the filters are applied.

Romy the Cat wrote:

Well, I understand the objectives and you are right in them but the experiment will not take place as I am already out of the league of the methodological correctness. You see, the 6-chennal of Super Milq is betraying of the DSET idea as the new Super Milq uses one-stage amps ... so, the Super Milq is not the best illustration of the DSET concept. Also, the Super Milq-type of amps are possible only in context of targeted binding to a specific acoustic system/room.

Romy the Cat wrote:

The problem that a regular SET has might be addressed by 2 channels - where one cares LF and another rest of the spectra. THAT is my primary definition of DSET. That DSET is very easily implementable and marketable commercially. We are accustomed in High-End that the speakers are bi-ampable and that the cables are bi-wireable. Why a DSET should not feel the gap where it could. A DSET should not even care a crossover – most of “better” speakers allow bi-amping and their driver-level crossovers are bridged externally… The DSET would be a perfect match in there.

drdna wrote:

And the explorations that led you from the original Melquiades to this current set-up are well-documented over a period on months.  Still, it might be instructive (to me at least) to hear concisely stated your dissatisfactions with the original set-up that led you to make the changes.

It might be true but the dissatisfactions did not derive from dissatisfactions about a single SET vs. DSET. If I have 2 channel DSET (LF + HF) then it would be fine with me. I think my desire to make chinches were more driven by me desire to optimize the channels of loudspeakers then the dissatisfactions of amplification.

drdna wrote:

This is true, but it reflects also quite accurately my own situation.  I have the EdgarHorn Titans and no intention of building new horns anytime soon.  However, I am buying up the parts to make the Melqiades amplifier.  My own personal question is whether I should proceed as planned or instead build up DSET versions to drive the four channels of the Titans separately.  I am not keen on building original Melquiades "just as an experiment to compare to DSET I will build afterwards" either because it will end up being a lot of money and time, and I am kind of in the situation where DSET might be appropriate.  However, I ask the advice of someone who has already done this whether it is worthwhile in my situation (this is the argument of course for DSET in general as well).

Well, I would say EdgarHorn Titans do not need DSET and the lowest EdgarHorn channel go down to 100 Hz and a regular full range SET will do juts fine. If you decide to drive the EdgarHorn’s subwoofer then you can consider DSET, still, considering the idiosyncrasies of the bass-horn topology it might be advisable to stay with a regular SET. However, if you feel like to drive each speaker channel with individual amplification channel then it is a totally different game. I presume that it would be better result but it also a considerably higher amount of efforts, both in building and in support.

drdna wrote:

Most people planning to bi-amp, they are told to use the identical amps to do so, so this will build resisitance to the idea of using different amps.  Also, it is easier and cheaper to buy one more amp (or pair of monoblocks) identical to the one you already have than to buy four new monoblocks.

The identical amps but with different power capacity, different damps power supplies, different output stage loading and different amount of inductance in output transformer. It would not violate anything in the speaker or it’s crossover but would enable a cheap and small SET in DSET configuration to drive loudspeaker better.

drdna wrote:

So really the target market demographic is audiophiles who are already bi-amping but who can be convinced that they should be bi-amping with DSETs instead of the paradigm of using the same amps, and whose speaker crossover points fit with the design parameters of the DSETs, AND whose ears and systems can appreciate the relatively subtle changes of going from biamp SET to DSET.  I am going to be honest, I think this is going to be a really small number of people.

It is very seldom that people who bi-amp use the same amps. Most frequently peopel who bi-amp (unfortunalsy) use with different amps. The DSET idea might enable then to use basically the same amps.

el`Ol wrote:

Why not more semi-active speakers with high-sensitivity mid/treble and transistor-powered bass?

Multiamping is using of dedicated amps for different channels. DSET is the use two or more of the SAME AMPLIFIERS (SETs in the given case) optimized for a given frequency range (output transformer in case of DSET). Here is the illustration: 3-channal Super Melquiades was a pure DSET with 3 6E5P-6C33C parallel channels. The late 6-channal Super Melquiades with it’s single-stage amps is a deviation from the DSET evangelism and it is more multiamping. We can make some excuses that the Sound of Milq is not determined by out stage and many other excuses but realty it that it is still not DSET but multiamping. I did spent a substantial amount of efforts and time (~3 month) to make sure that that way how I use the Super Milq channels do not break (or “almost” do not break) the TTH weighing scale but the 6-way Super Milq success is still arguable.

In case of pure DSET – where each channel is IDENTICAL there are many problems not there anymore and with an identical loading of the plates via different transformer of the same type a user is in a very “save” environment.

guy sergeant wrote:

A few months ago, iirc, you linked to a Canadian? amplifier which seemed to be made as or perhaps could be applicable for DSET use but I can't find the link.  Can you remember where its is (or what it was called)?

decoud wrote:

I know you do not want to do it, but it is a shame not to do it: actually manufacture a version of the Macondo with an integrated dset, cut down enough (say three channels) to make it feasible while still demonstrating the advantages of the dset idea. Indeed, was not Jessie thinking of spinning something commercial out of his elegant plaster horns? The point of making it commercial is only to make the dissemination of the idea tangible, not just talk.

rowuk wrote:

I think the problem starts when we short circuit the process. It is easy enough to sell "the Best" of anything and still get mediocre results. If we are looking for a modular approach to DSET/speaker, then the results are also best guess because there is no real formula for success. The "conductor" of the "audio orchestra" needs to earn their baton before real "results" are possible.