I sent an email to Brian asking... also about how far down in frequency their largest mumetal core could go.

I see why you want those tap options now - essentially you want to be able to load down the 6e5p but still use all the secondary sections. There may be some frequency response variations in the ultrasonic range by playing the tap game, but it shouldn't make all that much difference to what you hear.

A while ago, prices were of the order of GBP 100 each for the mumetal OPTs. Assuming they haven't changed much, that's about $400 plus carriage/import duty etc for the pair.

For the HF channel transformer; well, air core would be big, with very high capacitances. You really want a very low Z tube to drive an aircore.

However, something has occurred to me. The RAAL already has a matching transformer on a small core to which you are cap-coupling; I wonder if Aleks might wind something with a much higher turns ratio and cap couple directly to the 6E5p anode on the S2 amp. Fly in the ointment: he would need to ensure the insulation between windings could handle 400+v. Where the RAAL takes over, the inductance of the S2 should be high enough so the 6E5 isn't loaded down too much.

Having said that, do you know what the DCR of the ribbon element is? I have a pair of 90:1 aircores in London designed for driving raw ribbons direct from a SET... there would be some rolloff at 20kHz, but would well be worth trying if I can get hold of them in the next few months...

Paul S wrote:

One thing I noticed right away was a certain "image density", not unlike the ML2.

This is very interesting subject. I do not know what core the ML2.0 used but I would like to have ML2 a little bit  “amorphousised”…. To me it one of the things that I feel the ML2 was missing – it has dynamically limited and was able to play wild and violent when Music is called upon. It was in was over polite and overly gracious. Mostly it was desirable but when the opposite was vital then ML2 faced it’s limitations.

Paul S wrote:

Another interesting quality is absolutely unflappable dynamics, with no hint of saturation, and no harmonic shifting, compression, smearing, or thinning with speed, SPL or  complexity; in other words, pretty much what you said.

cv wrote:

I see why you want those tap options now - essentially you want to be able to load down the 6e5p but still use all the secondary sections. There may be some frequency response variations in the ultrasonic range by playing the tap game, but it shouldn't make all that much difference to what you hear.

Actually it is important to name properly- it is not the taps but just different connection of sections. I do not like taps and I would not go for it under any circumstances. Also, Chirrs, may I ask you subjectively what you feel is special in the Mumetal core Sound, besides the fact that the core “sings” loudly.  BTW, doe the Sofwter make amorphous core transformers?

cv wrote:

However, something has occurred to me. The RAAL already has a matching transformer on a small core to which you are cap-coupling; I wonder if Aleks might wind something with a much higher turns ratio and cap couple directly to the 6E5p anode on the S2 amp. Fly in the ointment: he would need to ensure the insulation between windings could handle 400+v. Where the RAAL takes over, the inductance of the S2 should be high enough so the 6E5 isn't loaded down too much. Having said that, do you know what the DCR of the ribbon element is? I have a pair of 90:1 aircores in London designed for driving raw ribbons direct from a SET... there would be some rolloff at 20kHz, but would well be worth trying if I can get hold of them in the next few months...

I certainly considered it and spoke with Alex. Since the amps will be sitting 10” from tweeter then the secondary will not have too high impedance. Alex is wiling to take this project but he needs to do some research and some thinking and he dose not know what might be an ultimate solution. I am planning to send him the amp (a tube with two resistors!!!) and let thin to think himself.  If he come up with one transformer (powder or amorphous) for HF channel then would serve his tweeter the bets then would be it. I like the fact that Alex will be able to think about the problem not in scope of juts a transformer but in composite scope of the entire plate-trans-ribbon interface…

op.9 wrote:

I now run my 'tweeter' amp 1st order high pass at 2200hz (driving BMS 4540nd in 1500hz leCleach horn) so my impressions might have some bearing on Romy's new amps. I was playing around with operating points and components in my single 5842 amp when the sowter mumetals arrived. Immediately all my energy went into listening to music again. First impression were of real power and grip! A sort of palpable sound - always my holy grail (maybe this is analogous to romy's 'connectivity'?) My BMS4540 nd drivers had never sounded like this before - it has taken a few months to realise quite how good they are. Only later did I notice a huge increase in delicacy and detail. Its not a 'flashy' sound at all, in fact now I'm happy to take my tweeter amp for granted, and forget about it. Interestingly, I was using a single 5842 for a long while until I realised that I didn't have enough system gain. Now I'm using a 5842 RC driving an EL84 triode-connected and I have to say that 98% of all the good things about the 5842 are in control of the amp. The 2% EL84 adds a nice 'sexy shimmer' to the sound. I'm not at all surprised that the Milq's driver stage sounds so close to the original and I'm very interested to see if a fully optimised single-stage could outdo the full Milq. . .

Another consideration where I am planning to make the single-stage Milq different will be the power supply for the thing.  The amp will have my typical SS fast-soft rectification with mandatory “gentle enter” LCRC filter, however I consider it to make it different from where I was before.

Usually I have quadruple (or more) inductance of critical inductance followed with a very small cap, then a decoupling resistor and as big ass last cap as I can get. In the regular full range Milq I use 10.000-15.000uF and it dose job. It is not juts about the ripples – they are irrelevant with this capacitance, but rather is it very beneficial for bass to have a huge capacitance tank.

However, I also noted that a large capacitance in PS’s filters make HF harder and less “connected” (red my post: http://www.GoodSoundClub.com/TreeItem.aspx?PostID=4307). So, my thought is to keep the PS capacities as low as possible for HF channels. From my two stage Milq experience I know that the amps has no nose at 620uF, so I am wiling to keep slightly above this level (probably 1000uF) and to go higher only for the upper bass channel. Frankly speaking I do not care about ripples as I do not see any relations between their size and quality of sound. I also did not see that the presents of theoretical un-auditable LF hum in any way modulate something auditable in auditable range. However, the minimum capacitate for HF channels I feel is very beneficial for space and “connectivity”…

op.9 wrote:

I chanced upon an interesting PSU arrangement for my 5842. I run a valve regulated powersupply at 210v and then drop down to 160v with an un-bypassed 8uf oilcap. Any more capacitance or bypasses seemed to be a step backwards in terms of pleasure. A nice side-effect of the regulator is much more consistency of sound from day to day. (I run my 5842 with a 2v lead acid cell in place of a cathode resistor)

I’m kind of debating how I will organize the PS. I am considering going for LCRLC with a lot of inductance and small capacitance to make the tube stage to “see” Bessel curve in PS…  Nevertheless, I do not know at this point… The LCRLC chain will be driving 4 RC parallel filers for each channel. (To keep the shot pass to ground for each stage). I do not know for now how it all will be organized from sonic perspective… In fact, considering my demands for compact size and the dimensions of the parts it become quite complex to built everything inside the Macondo’s frame.

I know that there are some people who would like to hear form me about my findings regarding the Singe-Stage Milq. I do not talk about it for a while but not because I do nothing in that direction. The decision was made to convert my 3 stages Super Milq into 6-chanls Super Milq featuring 3 full two-stages Milqs and 3 one-stage Milqs under one roof.

I have found all voltages, filaments all supporting thing – it do-able and it is all almost have done with one amp (you can see 6 driver tubes). There in nothing is connected yet internally as I have no out transformers for one-stage amp. The transformer are ordered and they will be ready within a month of so. The special attraction will be given to MF transformer the will be driving the main S2. I have recruited 4 transformer manufactures who will be making their versions of OPT according to my specification but utilizing this own implementation techniques. I will report which transformer will turn out the best (I will not report on the contestants). The budge that you see atop the Super Milq is the home for 3 new output transformers, currents meters and all supporting things. It should be very interesting…

op.9 wrote:

Thanks, look forward to hearing about your results. I am tempted to make some similar experiments myself - but all I have to hand is some 17.5:1's. I think I'll give it a go anyway..

Still the final number would depend from:

1) The ACTUAL impedance of your load in the bandpass of the given channel.
2) The plate current
3) The listening distance from the channel and how “live” your room at the bandpass of the given channel.
4) How your other channels are loaded
5) What kind driver are you using and what the driver’s natural harmonics
6) What frequency range you drive
7) How responsive your particular driver for damping
8) How far you are fro axes of the channel if you use horns
9) Reverberation time in your room.
10) How the transformer is made (it might have own 48282 parameters)
11).... esle

Generally I feel that for 6E5P anything from 10:1 to 20:1 into 15R might be useful in one way or another…

I got today my first pair of output transformers for the single-stage Melquiades.  They are 40ma, 5H, 42R primary, 15:1, Amorphous, with remapable secondaries made by Netherlandian Pieter Treurniet. I did not put it in use and heard them – probably it will take another two weeks but I need to say that this transformers made me to think sorry they the will be hosted inside of the Milq chassis, I never thought I would feel this way.

When I spoke with Pieter I asked him do not pot it into a can as it will be inside of amp, Pieter gave me a lot of reasons why he insist to pot it, fi9nishing with something like this: “I know hat I do. It will be potted”.

Well, the sad irony is that when I got those transformers the first thing that very impressed and surprised me was… their potting. It really feels deferens and I did catch myself that I actually experience a pleasure to hold the things in my hands. I am not kidding. It has a very strange quality of very precisely boiled egg, what a few seconds more would make it completely boils but a few second less will make the core too soft. Of cause the transformer has noting inside riddling of liquid but I constantly caught myself that I am wiling to shake it… Very abnormal feeling for me and I usually do not give a damn how the parts look and feel…

Can’t wait to put this baby into the use…

Romy,
It seem what with the tacttile pleasures of boiled egg potting compound sensations you are going a bit Sakuma yourself.

I will say this: I think Pieter is an absolute magician when it comes to iron; he's also a gentleman despite looking like Larry David's Dutch cousin.

I have to say, if there's any way an inanmiate component would capture some of the personality of its maker, it would be the output transformer. All that painstaking winding, layering, core assembly etc. I wonder if it's a coincidence that my fave OPTs all come from folks who are either pals or I get on very well with.

I hope they work well for you.

cheers
cv

The new 6-Channal Milq is up and running, I did manage to calibrate everything in the Macondo’s left channel and was listening it for a while. I might report somehow a controversial result. The MF drive with a single stage is incontestable king – it very clean and it cured the S2 from all its illnesses. Should I rejoice? Well, not real. Along with a phenomenal smoothness and cleanness the MF now exhibits a behavior that I do not like, in fact - hate. That all makes the enter sound absolutely not acceptable. Here is what I found very wrong:

1) With increased of dynamic the channel go thinner, the higher volume the thinner it goes. Extremely ugly!!!
2) The channel now too fast, inappropriately fast, faster then the rest of my channel. It is a way hyper-transient and  harmonically deficient – a typically overly idling tube
3) It sounds OK power vise only with full 8-9W dissipation on the tube – a very full power. Lowering power to 6W and the channel begin to loose in stresses.

Here are possible solutions that I am considering:

1) Heavier loading the tube that should eat some transients and add power.
2) Trying a different output transformer if it will not help
3) Change, S2’s impedance curve after 10K
4) Accept that a single stage amp was a dad idea.

Frankly I do not case about the wrong transient and harmonic balance - I will be able to deal with it by loading the tube it will also shell address the problem with power. However, I am VERY concerned about the increase of  thinness with rise volume… I afraid that it might be a characteristic of a sing-stage amp… I play that the effect will go away what I load the 6E5P much heavier.