Romy the Cat wrote:

So, voicing is a big totally artificial pile of BS and even the conversations and thinking about voicing is BS.

Guys,

Thanks, for your help with all my HF channel frustration.

The saga is not over. The latest update is the following:

The “secured” Cubes 950 (1000pF and 500pF) were placed between the Milq’s HF channel and the result was dissatisfying. The sound was very clean but FH were rolled off. Then 1000pF was placed in input of the amp against of 43K of HF channel (the preamp Rout is 7R). The HF returned back but sound picked HUGE amount of HF noise - anything at HF was resonating. I presume that the brand new Cubes of the low value should work in there for a while before they should be listened….They were juts too new....

I knew that I would have problems to filter HF at the Milq’s level :-(… In the end, I put back between the stages 2uF of broken-in Cube and placed at the speaker level 3uF of Vitamin Q (the configuration that I’ve been using for years). With this configuration the Super Milq HF channel gave out what I needed from it. I have to note that it has not rally anything to do with what I use to and habituated. It was very problem free sound, having which I would not peruse anything else. In fact it was way better sound then I use to as there was no upper bass coil into the upper FH tone.

Anyhow, then, I thought that in order do not overload the low primary inductance of HF channel I should go for lower coupling cup. I took Cube 950 .047uF and placed it there. The funny part that it, although sounded well, it still got some very remote resemblance of the HF noise that 1000pF had when it was placed at input.  I put it to burn for a couple days to see when it would lead me. So far, the 2uF inner-stage and the speaker-level filter performed absolutely and incontestably the best.

Rgs,
Romy

I really do not what is gong on there and why. Thorsten's comment that line-level crossovers do not act like speakers’ level crossover confirms what I feel and what some other people suggested me. Yes, the reactance of the crossover component, the inductive reactance of the driver coils, the lack of harmonic modulations in the amplifier stages when a stage operates within a limited frequency range… all of this certainly impact the final result. So what?

The experiments with the Super Melquiades HF channel are over. I found the optimum filtering configuration (or at least it what I feel presently) and closed the amp up. To my surprise the high-pass filter at the very input of the HF amp, before the first stage, did not sounded well. It had in a way too mechanical sound. I abandoned the frequency calculations and started to use mic-feed RTA on that HF dedicated channel in order to have the identical slopes with diferent filters. i did it  for the sake of more or less objective listening evaluations as evan with the very identical sloped the chanal where I had the filters at it's input sounded very thin and lean. When is puff it up with LF, moving the crossover point lower, the channel picked some “body” but still had sound very artificial and synthetic: It more sounded like “frequency pusher” instead of “normal narrow frequency sound”.

The filter between the Super Melquiades’ stages did not work due to necessity to use very low value caps against the high impedance and consequential HF roll off. However, the sound di not indicated any artificiality and was very natural, although HF channeled. Probably the first channel of the Super Milq should run full frequency. I do not know if that is the specific demand of that specific UHF rube that runs in the first stage or that is a genetic rule or perhaps something else. I know that it certainly turned out to be a rule in my case. Interpreting that in Super Melquiades’ LF channel I do have the lowpass first order filter of 65Hz sitting before the input stage… and I DO LOVE what the bass channel dose. Perhaps the problem that my frequency-limited first stage has is applicable only for HF? Or perhaps whatever the frequency-limited first stage dos at LF made me so love what the Super Melq’s LF channel does? Who knows!

Anyhow, the best, very much the best result that I was able to get form Super Melquiades’ HF channel was with the full range channel (2uF coupling cap) and the speaker level cap in series with the driver. Interestingly that when I began to lower the coupling cap I liked sound form HF channel less and less, although I it was way beyond the crossover point of driver.  I was hesitant to let the output stage to operate full range, due to low inductively of the HF output transformer primary (~1-1.5H). I left between the stages .047uF that makes ~ 35Hz high-pass filter and will let it to run as is for a few weeks to see what happen next. Still, when I go with the full-blown coupling cap of 1-2uF the HF become softer and more natural. Mystery, isn’t it?

Rgs,
The caT

Hm,

since I made my last post I was walking around my room, listing music and a feeling that something was very  still wrong did not live me. It always there was some sense of presence of HF in sound. It was not my typical full-Sound where it is imposable to say if it was enough "frequencies" but it was some kind of Sound + the "frequencies". I tested the  1st stage HF tube. It was fine but I still replaced it. I tested and juts in case replaced the output tube, even tryed thefew different brands of them but the problem did not go away.

Then I said: “OK screw all theory!” I opened the amps and replaced the coupling cap from .047uF  to 2uF. O wonder! It fixed the problem immediately. Then I trued few more caps in there and I learned that when the larger caps were placing between the stages I got the WAY better HF integration . Pretty much anything over .68uF worked very well. I would like to remind that the HF channel works from 3.700Hz electrically....

This all settles down a very important conclusion that for HF channel of multi-amping the presents of the modulative LF harmonics is incredibly important and apparently the HF channel amplifiers MUST run full range.

The 2uF-coupling cup was permanently placed to the place what it should be.

I hope the time I spend to learnt this make some of you to think then you do your multi-amping…

Rgs,
Romy

Morning,
I'm not disputing what you heard, but your conclusion that the LF must modulate the tweeter section can't possibly be right. 0.047uF is a -3db rolloff of 34Hz. If you're saying that over 0.68uF was fine, then by implication 0.47uF is not as good, that means that a cutoffs of 2.3 vs 3.4Hz on the tweeter channel make a difference!

Hence I'm thinking that LF modulation has nothing to do with it.

I think it has to be something else, to me it now sounds like the impedance issue is the most likely. Roughly 1k Zout with the 0.047uF; then again, given that the driver will have a Zout of at least that, can the 6c33 it be that sensitive? Maybe... who knows?

Anyway, you seem happy with it now, so I guess the case is closed and unsolved.

cheers

Chris,

I see where you are coming form and I was thinking about it myself. I would say that anything above 0.68uF was in somewhere listenable but I would not say that it was day and nigh difference. The increase of capacitance GRADUALLY improved sound and the larger cap was there the more natural HF were. I think if I put in there 4uF or 6uF then the result will be even better although the common sense sagest threat the more turns I have in the caps then more then should be prone to the “vibrative” distortions. Yes, the 2uF was better then 0.68uF (quite substantial better) and it still dose not make sense as the difference between them 0.8Hz and 2.3Hz.

I certainly would not insist that the modulations have anything to do with it – although the look for me the most plausible explanations. BTW, I did drop the load of the filter down to 30Hz when I use smaller caps and had no effect. I mean I gained some very minor amount of HF but the nature of Sound did not change.

Rgs,
Romy

Running the fully closed Super Milq for a few days I noted some amassing qualities that took the HF performance of Super Melquiades even further then Melquiades. This quality consists of a complete absent of any king of glitz and very atypical for the 6C33C gentleness. Certainly a new “yellowish” driver in HF channels and using a half of the 6C33C do pay great dividends! The HF channel was a reason of some concern of mine as the HF channel has too high PS capacitance that always makes sound harder and was planning to drop the last cap significantly. Most likely I still will go into this direction but for now I am experimenting with a new output tube for HF channel. I will post my observations about this journey when I will be through…

Rgs,
The Cat

It is kind of pleasant and insulting at the same time. Insulting because I thought I know the Super Melquiades better then anyone else or at least was able to say about the Super Milq Sound whatever might be said. :-)

However, today my ego was severally damaged. I have today a third person in my room who heard the completed Super Melquiades and when I, as usually,  was asking about the problems with my Sound, he made very fascinating comment. The comment was very interesting, educational and it made me to look at the things differently.

He said: “The Super Melquiades makes Sound that allow you to hear very clearly the lowest octaves of the bass instruments but you do not hear.... the actual bass.”

Think about it….
The caT