fascinating new psu technology..

[lt cdr data]

simple, novel, effective, read the technology bit on this reveiw.

why hasn't this been done before?

http://www.never-connected.com/image...ewsmarch04.pdf

[pinkmouse]

There was a big thread a year or so ago discussing just this topic.

[lt cdr data]

where?

[pinkmouse]

Well, if, you'd searched on "never-connected" as the new thread page asks you to do, you would have found this thread for starters, and I'm sure there is another bigger one as well, but I don't have time to find it for you.

[janneman]

From the description, the NC concept seems clear. Remember that in a conventional supply, the transformer (and mains) is disconnected from the amp most of the time because the diodes are non-conducting. Only during the relatively short (a few mSecs) of conduction is there a galvanic path from the transformer secondary (which has all that mains junk) to the supply cap and to the load. That is then precisely the period that mains noise can get into the amp. Unless you are so 'smart' to put caps across the diodes, that gives mains noise a free ride into your amp, of course.

What NC apparently does is precisely during the conduction period, divert the current from the rectifier diodes into another cap, that is NOT connected to the amp. Then, after the conduction period, and when the mains is once more disconnected through non-conduction diodes (or a FET switch), this 'topped-up' separate cap is allowed to bleed into the main filter caps that feed the amp. Therefore, the main filter caps that feed the amp are never connected to the mains.

Is this worthwhile? Depends, as always. There may be a measurable improvement, again can you hear that? Depends on a LOT of factors. Anyway, it is beyond me how anyone, even Martin Colloms, can state that a certain change makes 15% or 7% audible improvement. How on earth can anyone determine that? How do hear 7% improvement in dynamic bass lines??

Jan Didden

[lt cdr data]

Martin Colloms has a history of trying to quantify dubious things like that, its supposed to give his subjective reviews objective credibility, it could be valuable in the sense that it gives one an idea of the rough degree of magnitude difference, but its not to be taken as gospel.

its a fet switch apparently.

interesting that trichord used? to produce a transreactance toroid along with other companies (ben duncan, smartsound) from a localish firm, which apparently offers total isolation from the mains.

so if you have mains 'cleaners' then, what is the point of nc's? hmmmm...

Jan, just as an aside, did you design some super regs. with Jung?

thanks

[janneman]

Originally posted by lt cdr data
[snip]Jan, just as an aside, did you design some super regs. with Jung? thanks

At one time I have collaborated with Walt Jung on a superreg project, yes. It was published as a 4-part series in Audio Express (then Audio Amateur) in 1995, IIRC. There was a 3rd person involved, Gary Galo.

Jan Didden

[jackinnj]

Originally posted by janneman
That is then precisely the period that mains noise can get into the amp. Unless you are so 'smart' to put caps across the diodes, that gives mains noise a free ride into your amp, of course.

Depends upon which problem you are dealing with -- I'll let those more knowledgeable from International Rectifier and On-Semi talk about "putting caps" across diodes -- if you don't put in a resistor in series with the cap, you get no RC time constannt, you just pass along the ringing.

I didn't manufacture the problem illustrated below, just searched through the junk box of transformers to find one which caused a bad ringing owing to the LC tank formed by the transformer, interwinding capacitance and diode capacitance:



I must say that the NOCONNECT must be a pretty snappy switch to deal with the harmonics in the power line:

[janneman]

Originally posted by jackinnj
Depends upon which problem you are dealing with -- I'll let those more knowledgeable from International Rectifier and On-Semi talk about "putting caps" across diodes -- if you don't put in a resistor in series with the cap, you get no RC time constannt, you just pass along the ringing. [snip]

Agree - it's just that I often see those diagrams with just a cap across the diode, apparently in an attemp to kill diode switching impulses.

Originally posted by jackinnj
[snip]I must say that the NOCONNECT must be a pretty snappy switch to deal with the harmonics in the power line:

Jack, as I understand it, this FET switch switches twice per half mains period (assuming double phase rectification) to connect the auxiliary cap to the rectifiers when the sec mains is higher than the aux cap voltage.

Jan Didden