 
|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
|
|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
194.214.158.221
Hello,During the last European Triode Festival, aside the "transformers shootout" , I could perform measurements on 30 output transformers for Single End 300B amplifiers.
The document with the measurements is downloadable at the bottom of the page:
European Triode Festival - ETF2009 Shootout
http://www.triodefestival.net/index.php?page=etf2009-shootout
look for: "see the measurements here".
direct URL:
http://forums.melaudia.net/attachment.php?aid=1228
Best regards from Paris, France
Jean-Michel Le Cléac'h
Edits: 12/11/09 12/11/09Follow Ups:
...in the end it was nothing more than a little fun.
The amplifier in question is known for it frequency going down very fast after 15k? IMO it would have been better if an amplifier that was more "wideband" was chosen.
The transformers were chosen by knockout. I.e. 2 are put up against each other winner goes through. Etc.
How long did each listening test take? Etc. I could go on and on. In the end one is left with more questions than answers.
To infinity and beyond!!!
"inconvenient truth" to most posting here ,hehehe
I can only assume you are saying that in jest.
Because the Audio Note tested really bad.
The Sun Audio and Silk had the best waves.
Lundahl and AE Europe had the best response.
DanL
Yes , and not ;) I had a custom opt which measured perfect but sounded like a POS. Poor AN OPT sounded like a revelation next to it. Funny, that you mention mesurements as a qualifying factor on SET forum.
Response & 10KHz squarewave are the ones
that interest me as far as measurements go.
Response flatness and squarewave ringing
are what concern me the most as they are
what I hear that botheres me the most.
DanL
Dear Dan,
Since the Williamson amplifier arrived in the late 1940's the audio industry has increasingly been indulging in single parameter evaluation with focus shifting to whatever parameter is fashionable at any given time, needless to say not much real progress is made this way, as it is a one step forward, one step to the side and one step back approach which never yields any genuine improvement or perhaps more importantly any deeper understanding of what governs the relationship between sound quality and what we measure, something that is demonstrated by the fact that the Partridge TK4519 (which is essentially a 1950/1960's design) performed the measured tests as well as or better than anything in the test, thus indicating that transformer design was fully understood at the time it was designed and not much has happened since, yet it did not win the sonic shoot out.
Over the years the "single parameter" advocates, of whichever persuasion, like yourself have sat on the sidelines and sniped at any attempts to try to balance multiple parameters as this approach always results in any one parameter being "disadvantaged", a fact which is always ignored and therefore generally never properly discussed and thus the critique is always based on single measurements alone, it is therefore refreshing to see a test like this which with all its potential shortcomings considered (no square wave measurements for nearly half the entrants including the winning design, the use of perhaps not the best sounding circuit as the "platform" and perhaps also systems matching issues) still, at least highlighted aspects of transformer design - measurements - sound that allowed a more comprehensive comparison which gave sound a fair chance on its own terms rather than just being interpreted through measurements.
Our transformer designs have evolved over a 15 -17 year period and have come to be designed for "wide band dynamic signal transfer" rather than any singular measured parameter, we always look at what happens as the signal enters (starts) as it is at this ultra low level that the signal signature is formed, loudness just amplifies this signature it does not change it to any significant extent. A key part of the design process involves a great deal of time listening to various winding - core material configurations to establish how they sound in relation to different circuit/power supply combinations.
Over time this work has given us an overview of most aspects of design versus sound and allowed Andy Grove to custom design a modeling program (based on Matlab) which balances all the parameters to favour what we have learned during the many hours spent designing and listening, this work is ongoing.
A good indication of this work is the measured performance comparison between the "small and Old Audio Note" which I suspect is either a TRANS-144 or TRANS-150 and the TRANS-300 which won the shoot out, both the 144 and the 150 designs are 20 year or so old and in spite of this they measure very similar to the 300 but the 300 is much better than the other two sonically, which only goes to show how much measurements are worth when it comes to judging ultimate sound quality.
It is worth remembering also that the TRANS-300 is at the bottom of this range of transformers and better still is in progress.
I think that was it.
Sincerely,
Peter Qvortrup
Hello,
The frequency response of a SE 300B amplifer based on the WE91 schematic is highly dependant on the output transformer.
I see no reason why the schematic in itself (WE310 + WE300B) apart the OPT, should be a limitation for the frequency response.
I guess that for the shoot out major limitation was the loudspeaker itself. (My own choice will never be that loudspeaker... much more detail and frequency extension would have been brought with such loudspeaker as Jean Hiraga's one we could listen at the end of the festival...)
(BTW: I am in now way organizer of ETF09 nor of the shoot out... and I have done my mesurement independantly, after having required the authorization to proceed by myself to the organizers with the help of 1 or 2 friends).
Best regards from Paris,
Jean-Michel Le Cléac'h
If I am reading your graphs correctly, most have a rolled off top end (10K up)
-t
Yep, I missed that. Now I wonder if they level off or continue downhill. I guess I am surprised they are not very linear out to 50K
Another question is whether this rolloff is due to leakage inductance, winding capacitance, or a combination of both?
Well done Jean-Michel!
I am surprised to see that the third order distortion dominated the 2nd.
dave
Hello Dave,
Paul reply is correct for the most, IMHO.
The gap of the OPTs having a dominant 2nd order distortion component operates near the knee (in the low linearity part near the saturation) of the BH curve, this surely induces some dynamic compression + H2 dominant distortion that may be euphonic, but...
I may add that it is probably a good thing that 3rd order generally dominates considering that the 2nd order distortion component will be mainly brought by the 300B itself. Higher order harmonics may be due also to intermodulation products and to the input/driver tube(s).
The difficult "art" is then to find the proper mix of all parameters to accord to your loudspeakers (it depends a lot on their efficiency for sure... this means that IMHO there is no universal solution to the question of the better choice of an SE OPT)
Here in France the SE OPT subject lead to very hot discussions between the tenants of the use of the best measuring transformer and the tenants of the "listen and choose" solution.
It is very funny that during ETF2009 shoot-outs I voted very often for the "looser" and that when the results were known I could see that those "loosers" were -for most of them- the best measuring transformers. I could see few "known" audiophiles (known at least from me) that voted the same...
But the majority voted for OPTs having a frequency response having very few bass frequency extension and an extended frequency response in the high frequency range with a rising slope. (those OPTs sounded bright to my ears). All in all this was an interesting experience but more related to anthropology than to audio ;-)
Best regards and an Happy New Year 2010 from Paris.
Jean-Michel Le Cléac'h
You shouldn't be though :^) Transformer hysteresis loops for small signals are symmetrical and pretty independent of the DC bias as long as you are not too close to saturation.
hey,
granted 1/2 a watt is small signal but 70ma is a lot of current to put through a grouping of 31 transformers to only have a few of them show dominant 2nd order distortion.
Does anybody know the average listening power of the actual musical passages?
dave
For what it's worth ... there are a few papers that deal with what studio engineers do. Traditionally, peaks on a VU meter corresponding to 82dB +/- 2dB sound pressure level are said to be normal for a substantial sampling of studio practice. Headroom of 14dB above that (96dB) is typical of tape recording in such practice, with tape saturation limiting further peaks. For THX film work, another 6dB headroom is said to be used; that would correspond to 102dB per speaker, or 105dB peaks with two speakers uncorrelated - which is I believe the THX theater specification.
14dB above 1/2 watt is 8 watts - typical of 300B amps at clipping. For what it's worth.
I keep saying "for what it's worth" because all the above is scientific measurement, with no clear connection to subjective acoustics beyond the fact that it's common practice among professionals. Most professionals have good ears and obviously must meet most people's needs, but how that relates to serious audiophiles is a nearly unanswerable question. This is after all an asylum!
Probably more important in the context of this discussion is, what is the distribution of low frequency energy? If you have full power handling to 60Hz, with saturation limiting lower frequencies, is that good enough? 30Hz? 15? How does the power bandwidth requirement relate to small signal frequency extension requirement? The implications for transformer design (and cost!) are huge.... As Edwards Deming was fond of pointing out, the important numbers are usually unknown and probably unknowable - but they are still important.
Hello PaulThe loudspeaker used for the shoot-out of OPTs at ETF2009 is a replica of Shindo's Latour:
http://www.shindo-laboratory.co.jp/English/speaker.htmlIts efficiency is given as 100dB/1W/1m.
Main limitation for the 0.5watt power used for the measurement of each OPT was the 84volt peak to peak max sinus delivered at the output of the very low distortion and very low output impedance transistor amplifier (a 120W/8ohms Bryston PP120) used at the output of the low frequency generator. To find an amplifer with no distortion able to deliver 150V peak to peak would have been ideal for such measurements but that's a dream I cannot afford. ;-)
Best regards from Paris,
Jean-Michel Le Cléac'h
Edits: 12/15/09
Naz
Post a Followup:
| FAQ |
Post a Message! |
Forgot Password? |
|
||||||||||||||
|
||||||||||||||
This post is made possible by the generous support of people like you and our sponsors:
Top of Page ]
[ Contact Us ]
[ Support/Wish List ]
[ Copyright © 2025, Audio Asylum, all rights reserved ]
