Welcome! Need support, you got it. Or share your ideas and experiences.
Return to Planar Speaker Asylum
50.132.28.230
In Reply to: RE: the late Al Sekela's choke tweak and True ribbon tweeter posted by Satie on January 14, 2015 at 14:32:55
Inductors are not evil. The problem with the "choke tweak" is the concept, not the component used to implement it.
A transformer will not give you a brick wall filter. Not even close.
Cheers,
Dave.
Follow Ups:
Call it what you will, this FR behavior is functionally close enough for me to be a "brick wall" I know there is a definite technical meaning to the term that these might not quite fit.
http://www.sacthailand.com/Pic/Transformer_InterstageTestFC1.jpg
http://www.sacthailand.com/Pic/Transformer_InterstageTestFC2.jpg
If you don't need their particular function then yes they are evil.
I'm a fan of transformers in that application. They provide three nice benefits. Galvanic isolation.....bandwidth limiting.....and an excellent way to perform balanced-to-unbalanced (or vice-versa) conversion.
However, I'm wondering how you reconcile your possible usage for a particular function where you really don't need them......since your objective can easily be achieved without coupled inductors. :)
Dave.
Actually if I had not heard the approach work so well in the PS Audio Peirfect Wave DSD I would not have considered it. But it is possible to forgo an active output stage and filter circuit with this, so the alternate option performance threshold is lower than that of a coupling cap.
Did you note/read John Atkinson's test report of the PS Audio Perfect Wave DAC in the a recent issue of Stereophile? I thought that and Paul McGowan's response was interesting.For reference:
http://www.stereophile.com/content/ps-audio-perfectwave-directstream-da-processor-measurements
Paul McGowan's response is not shown in the online article, but I found it completely consistent with his usual commentary regarding products.
Dave.
Edits: 01/15/15
I believe the issue was that the output trannies were undersized and were saturating at high bass levels. It is also possible that the PS was sagging there too. It seems that the firmware from the end of last year has compensated for the problem as the bass has firmed and deepened substantially. As a patch solution we tried out lowering the digital volume level to 70% and raising the level at the preamp. That tightened bass but it seemed to need more, we were going to try 50% digital volume as there was still headroom to do so without losing resolution. I don't know if that was part of what they did with the firmware update.
My magnetics friend and I came to a conclusion that sound is best served if the trannie is never operated beyond 1/3 of its rating and preferably under 1/5th. Subsequently he took his tube preamp's filament power supplies off the mains transformer and gave them a separate power supply. He was over the moon with the resulting improvement in dynamics, tone texture and fine detail as well as imaging.
I'm still very surprised at the cavalier approach Robert Duetsch had for this unit and his apparent gullibility regarding information that was forthcoming from PS Audio. I like to read John Atkinson's comments because you can usually read between the lines very easily to gauge his opinion.
A firmware update won't fix the hardware. You would think that for $6000.00 they could install some adequate transformers to handle the signal levels? I hate to mention this in the context of my go'round with JBen....but this seems to be an issue where inadequate testing took place. A "workaround" that requires you to reduce signal levels 6db seems....interesting.
Anyways, getting off topic for this forum. Back to the normal Asylum silliness.
Dave.
I suspect they took the trannie maker's word that what makes the trannie a good filter would also restrict its dynamic capabilities and early saturation. Which is what happens if you use smaller cores. Bigger cores like those in poweer amp output transformers typically outdo 70 khz and better ones can do over 100 khz. The bigger cores and higher nickel materials and fatter wire can extend LF and limit LF distortion and then you can use bad skin effects to limit HF output.
Outside the bass the PW is really a great performer even at that steep price.
To take the Art Dudley side, read his self psychoacoustic analysis piece in the last Stereophile. Perhaps THD should be compiled with a 3-6db discount on even order bands to reflect audibility and how objectionable they are in musical context rather than electrical.
Probably more discount than that. :)
I think Stereophile should turn an editor loose on Art Dudley's articles before they publish them. (It's 90% flowery rhetoric.)
His underlying point is understood though. However, he seems pretty late to the party with his acknowledgment of the subjective aspects of various distortions.
I'm still puzzled by PS Audio letting a product like this to the market though considering the measured performance. However, that assumes they weigh measured performance as much as I do....which they obviously don't. The firmware upgrade though, at least acknowledges an existence of an issue.
Cheers,
Dave.
Transformers are pretty well understood by this point and excellent ones are available off-the-shelf (for not that much money) that could have exhibited better low frequency characteristics. I'm puzzled by that one, but oh well. Distortion correlation to subjective result is an interesting topic. :)
Also interesting to me was the lack of resolution on the unit and the fact that noise completely swamped the standard jitter (J-test) testing scenario. Relatively speaking, that unit did not perform well on JA's bench compared to other DAC's for much less money.
None of this is to say the unit won't sound excellent. Just another puzzling data point in the crazy world of high-end audio. :)
Dave.
In this case, the output trannies were designed for the device by the trannie manufacturer and I am not sure they "got" what was required. The custom trannies are a big portion of the cost of the PW DSD and I suspect it is also the weakest link. They were talking about the difficulties of getting the trannie people to understand the filter function and they were reluctant to make trannies with such "bad" HF performance.
I do actually expect the fuzzy low level 16 bit HF sine traces to sound better than the normal square looking ones.
JA's test results are always an interesting counterpoint to ecstatic reviews of non-NFB SE devices and of "crafted" compression/expansion in some digital equipment that measures as if it were made by a horse...
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: