DIY Sony VFET Builders thread

BEST MOD EVER

Sooo... We finaly went the trouble to try Mark's excellent "small" SMPS filters with the VFET.

As posted 2 months ago, in order to try to address 200kHz and harmonics garbage produced by the SMPS I thought wise to add 2 filters after the existing / original SMPS filter and the OS PS input (which already have a great tweak, the addition of PS caps). For the details, reference or the filters, reasoning behind and some calculations please refer to my previous post.

Special thenks to Mark for his excellent filters, eventhough I somewhat derouted their initial purpose probably, and very very big thanks to Gilles who spent an entire day assisting me (or was it me assisting him) with the builds and also confirming listening impressions on his own system.

Instead of posting pictures of inside the unit, where you are likely to see... nothing (or so), I decided to post pictures of our test rig and then pix of the small casing we built to house the 2 filters we retained. That casing fits very neatly under the OS boards...

First things first, listening impressions. As you can see we went for a very rough test equipment. Inded, my last try to find a potential weakness, bypassing entirely the FE boards, ended with a failure as I couldn't hear any improvement. Master Papa had done it right again, as ever. And he had already make provision for a SMPS filter, so I had rather cold feet on that possibly last mod. Therefore, I decided to fit everything off board of the amp, so that direct comparisons were easy (just a matter of a few seconds to undo a connection externaly), while also enabling the possibility to add more than 2 of Mark's SMPS filters in an eye wink. I have 12, so not a problem.

That how it looks like - horrible, with very long cables and even a potential ground loops when you consider the way we rerouted the grounds...

Oh, the small gaffer tape is just an extra insulation in case of, of course there is more than that...
 

Attachments

  • IMG_20211026_155515.jpg
    IMG_20211026_155515.jpg
    15.2 KB · Views: 431
  • IMG_20211028_122914.jpg
    IMG_20211028_122914.jpg
    46.4 KB · Views: 436
Last edited:
All I can say is... WHAOU!

This is by far the best mod I did so far to the VFET. Let's face it, my FE bypass was a failure (no audible difference), the caps bypasses brought some marginal improvement - kind of thing you have to listen to very carefully. The additional PS caps on the contrary brought an immediate plus on many accounts and could be directly assessed. Well, the additional filters are even greater in magnitude! Of course they bring different things than the PS caps, in fact welcome things that are complementary to the additional PS caps.

We could switch from direct, to 1 additional SMPS filter per OS board, to 2 additional SMPS filters per OS board. The good news is the Meanwell started several times without any problem despite the increased capacity. OK, Mark's filters include some R, my long cable saswell, and the coils help somewhat probably, but at the end with 4 filters and additional PS caps we are talking of a total of nearly 8000uF...

"Direct" was our baseline. TBH nothing to complain with, since I fitted the additional PS caps the sound is fascinating. With a total of 2 filters, the sound became... from another planet! With 4 filters, we had something between direct and 2 filters, a bit too soft, loss of dynamic & Co and we clearly found 1 filter per OS board was the golden number.

Gilles and I assessed the gains, and without consulting each other we consolidated them. That wasn't difficult: this tweak has the greatest sonic magnitude change, difficult to miss it, and we converged on all findings and magnitudes of the changes. With a total of 2 filters we found no negative and only positive, despite trying hard:
- mids are unchanged
- bass gains definition, more depth, groove and precision (YES, I know...), flesh, some marginal speed and slam but that was already there thanks to PS caps
- treble gains definition, precision and smoothness (YES, I know...), holds longer in time without fading badly
- overall flow was incredible, music integration at its highest, regardless the material
- composure on complex music was enhenced, nothing collapsed whereas the original VFET kit wasn't at its forte there
- all registers benefitted from a faster rising front / attack on notes, additional definition (like a higher res pix that adds though no harshness at all) and most enjoyable, all notes are flowing longer, are hold longer in the time until their extinction sounding incredible present and realistic... musci at its best
- and now the big surprise... soundstage! Not that the intial soundstage was lacking anyhow, but now all instruments are nailed and nearly palpable, while incredibly soundstage depth and height have increased a lot, everything sounding so much more integrated and consistant, more realistic rather than just a (very) nice illusion / trick.


If you have the same first VFET (can't comment on the second edition), then this mod is a MUST DO IMHO. It cost nearly nothing and with the additional PS caps it move the VFET to an entire new level.

The initial unit sounded marvelous, but soft and not always well defined on complex passages. But sooo nice. Once the additional PS caps came in, smash and speed moved to an acceptable level. Perhaps some magic was though somewhat gone, the unit sounding less "charmeur" and more mature, more precise and realistic also. Not a big change, and for long time listening perhaps the better compromise. More realistic, but a tad less special. With the SMPS filters all the charm is back and some more, but this time not like a slightly blurred trick, but rather due to an incredible level of realistic sounding musicality regardless the music material thrown at it.

That is VFET amp freed up, at its best. I am not a SMPS fan but I was told a lesson by Papa and it makes perfect sense with Class A units. However, it requires often quite some play with filtering (SMPS filters, additional caps). I feel the original VFET is hard to beat re FE and OS, so perfect and addictive is the job. I also feel the SMPS is VERY convenient to regulate the power. BUT I also feel as it came with the kit it was more a starting point and to my humble ears and listening preference the amazing music capability of the VFET could be releashed completely by tweaking the PS. Not a criticism, the original unit is amazing... but it can be even more than that IMHO.
 
Now one may say why not going linear regulated? Sure, could work if done properly, but cost an arm, is difficult to implement correctly and IME on Class A devices it is, to my surprise, possible to equal or even beat linear regulated with enhenced SMPS (read often SMPS + caps + filters, all carefully chosen though). Been there on the B1K - lab PS with tweaks vs SMPS enhenced. Having said that, I haven't tried on the VFET whereas some did (or at least one).

But because of my experience, plus the fact I consider the VFET an exceptional gift and an artwork, I would rather keep the mods simple, reversible and keep as much parts and initial spirit as possible. Soooo... I decided to integrate the 2 additonal SMPS filters discretly into the VFET. I won't really go linear, so don't ask for it: based on what I hear I doubt there is much more to be gained, and the sound I now have suits me perfectly.

For those who might be interested in trying that tweak, that's how we did it (again, a final pix wouldn't show much, hence step by step hints). First of all, 2 SMPS filters were integrated side by side into a very small metal housing. Connectors were used to exit the existing entry filter in direction of the 2 SMPS filters. Wiring has been reduced to the minimum, but I must confess I couldn't hear a difference between long and short cables or grounds loops running wide.

Each SMPS filter exits then the small housing to lead to another connector. That's the white one we bolted carefully (read Gilles did it LOL) on top of the small casing. That connector feeds then the usual OS PS cables. In fact, you get 1 SMPS filter per OS board, which can possiblyu be a plus for channel separation. These cables enter as usualy the OS boards, in my specific case at the additional PS caps level.

To make it all nice, the additional very small casing fits underneath the OS boards, right after the original entry / filter board. The casing is somewhat elevated to enable air to circulate under it, but TBH it is so small and centered in the plain floor that its impact on air flow must be very low anyway. There are other places and other ways to do it, many many others once you split the filters or stack them etc. We looked deeply into them but found this was the most logical way leading to the fewest "cable salad" with low signal wires not mixing with PS wires etc., something we deemed essentiel.

Mark's SMPS filters that we used cost a tenner or so at the DIYA shop, the chinese casing a few bucks... you will be very surprised by the VFM ratio of this mod, especialy if like me you thought the VFET couln't really improve any further.

Have fun

Claude
Note: discard the tape and empty casing on some pix, these were just intermediate dry fits to show YOU what we did!
 

Attachments

  • IMG_20211028_155918.jpg
    IMG_20211028_155918.jpg
    36.7 KB · Views: 423
  • IMG_20211028_155927.jpg
    IMG_20211028_155927.jpg
    38.5 KB · Views: 409
  • IMG_20211028_173159.jpg
    IMG_20211028_173159.jpg
    31 KB · Views: 415
  • IMG_20211028_180523.jpg
    IMG_20211028_180523.jpg
    33 KB · Views: 240
Last edited:
Oh, and a last one as my tweak path for the VFET nears its end: again MANY MANY THANKS to PAPA. And one more to have created such a robust amp that survived my many abuses trying to tweak it 🙂

This is an incredible gift... and such a cute looking unit. The VFET goes to my wife and will take place in our living room, where it will enjoy a daily use from the entire family once I find suitable LS partners (on it, need to think hard, not used to high efficiency LS). It sounds so AMAZING, it is very difficult not to fall in love with it!

Enjoy music very much

Claude
 
Thanks Tungsten 🙂

I wanted to close the tweaks on that amp, being very very happy (diminishing returns, other projects etc.)... but never say never! Especialy since given the connectors and mounts I fitted for the FE, changing FE boards would be just a matter of 8 screws per board, so VERY easy. Was done for that purpose initialy... And Mark's projects are always tempting...

Argh!!!
 
ClaudeG , simply judging by your name ( and I'm really having catastrophic memory regarding most things geographical) , I resume there is possibility that your'e either living in France, or you are connected in some way

there are plethora of both Old School and newer ones - speakers made in France, not really being known in other Markets

if you're Boy from Texas and I got it all wrong ....... excuse moi! .......... but even in that case - go and search for Frenchy speakers in Dallas ....... :clown:
 
Hi ZM,

Right guess my friend, I am based in F.

Yep, some nice audiophile LS, some better known Triangle & Co, but for myself I am in love since a few decades with my Renaissance 90. But these magnetostats (for my audio room) obviously need another amp (on it way), so... I need a pair of speakers for the VFET. To my own surprise, I was amazed by the Klipsch RP-8000F, very different from what I had heard from them in the past, and these were the LS I used to test and tweak the VFET as I recommended them to Gilles. As of now, they hold pole position for my VFET in the future, but I always welcome any challenge to these.

BTW, should you be around one day ZM, give me a shout: you are always welcome in F.

Claude
 
Merci Claude 🙂

regarding visiting France ...... for several years now, I have open invite to my Manu's place in old Village in central France (La Celle, by memory), where he's spending summers, recuperating of extreme endurance of being Prof. at Vienna Uni

though, life and work getting in the way .....

Cabasse, Triangle, Audax, Elipson, Rehdeko ........ and more .......... you guys are for speakers what 'Talians are for motorcycles :clown:

I'm in love with more Old School Klipsch-es and - frankly - didn't heard much of newer ones
what I did, eons ago - they were driven with horrible Audiophoolery, I really couldn't hear anything

(hey, but I have pair of 8" FR from Cabasse Dinghy!! :clown: )
 
Yep, I agree older "mainstream / affordable" Klipsches sounded awfull for HIFI.

Hence I didn't expect a lot (or anything TBH!) from the 8000s... and I was (I am still in fact) extremly positively surprised.

Whereas it is the exception or a new trend at Klipsch, I don't know... may want to tackle it further, time permitting.

Bonne fin de semaine

Claude
 
LOL

I have to say I wasn't impressed by the Classic line either, up to the enormous corner Klipsch furnitures - errr, I mean LS - a VG friend of mine has in a dedicated room.

But the few Classics I heard were all built ages ago and the line has been constantly improved, so the 8000s have also the merit to raise my interest for the upmarket Klipsches.

Sadly these (RF 7, Heresy) are rare and very expensive where I leave, used or new making them not very competitive vs other brands. The 8000s sound like a good modern 3000$ speaker TBH, you get something very complete nowadays for that kind of money. We found them new for less than 800$ /pair, half the price so really a bargain. Sadly they don't come across anymore here at that kind of price since a year, but I don't mind paying a bit more for them.

I still need my house to be finished/built first. I have a bit less than a year in front of me to decide what goes in the living room with the VFET, the beauty of these Klipsches being they run since day 1 on Gilles 8000s and I must confess this pairing was made in heaven. So they are on pole position but I keep watching the mirrors!

Have a nice WE

Claude
 
Just to steer this thread back on topic..
I expect to receive my Scourge kit in the mail today. The plan is to install a 1.0 uF polycarbonate cap at C1, and a combination of 100 uF Nichicon ES series + 0.33 uF Wima polypropylene at C2 & C3.

That will follow my findings of using those caps in my Bulwark cards. There are other fun variations to try, similar to what was done with the older M2x Mtn View front end boards. Later.
 
ClaudeG, that is a great read on the SMPS filters... since English isn't my first language, nor (apparently?) yours... just to confirm:

You installed the output of the original PSU filter into A SINGLE SMPS filter and then into the OS board... repeat for the other side? Or did you "daisy chained" the SMPS filters in series?

As I read it, it's the first, but just wanted to confirm.

Also, I believe TungstenAudio is another firm believer of adding extra capacitance to the PSU filter... are those "extra" caps still desirable? Or is this what the SMPS filter is doing? If not, where should that extra capacitance happen? Before or after the filter? Where "exactly" and what values would be the cherry on top 🙂

Thanks for any feedback!
Best,
Rafa.
 
Thanks for the reply, Gilles - busy day! 🙂

Rafa, I only discover this now and I just replied with all details and further recommandation to your question in the filter thread.

Here again, to avoid any confusion...

ClaudeG, that is a great read on the SMPS filters...
=> thanks rafa, appreciated

You installed the output of the original PSU filter into A SINGLE SMPS filter
=> NO. It looks like a single from the outside because at the end, to save space, we used only one casing... but that very casing encloses 2 FILTERS, see pix.
Each filter feeds one channel. So one filter goes to the regular PS feed of the OS right channel, and another does the same for the OS left channel.

and then into the OS board... repeat for the other side? Or did you "daisy chained" the SMPS filters in series?
=> I didn't daisy chain the filters in series in my final build.
I did so though during my experiments, when we thought one filter per channel is great, what about daisy chaining (always per channel, to keep the 1.7A I described for you in the filter thread)? In that very particular experiment, we ended up using 4 filters, in fact 2 daisy chained per channel. That's what you see on some pix. This experiment didn't turn out to be satisfactory, too much of a good thing, so we reversed to 2 filters in total. SO WE DISCARTED DAISY CHAINING ON THAT BUILD.

As said with more details in the other thread, the key point here is that each filter is passed through with roughly 1.7A, so well within specs.

Also, I believe TungstenAudio is another firm believer of adding extra capacitance to the PSU filter... are those "extra" caps still desirable?
=> Ok, that's a new point. Lucky you, I thought on that one already yesterday... great minds LOL
BTW, the credit for additional PS caps on the VFET comes from Tungsten - I followed on this... with success to my ears. Thanks go to him.

Back to your last, interesting, question. First thing is I didn't try the filters in isolation of the PS caps addition so stricto sensu I can't reply 100% to your question. What follows is guesswork...

Based on:
- adding far more PS caps or a bit less ONLY doesn't seem to make any real difference
- the additional PS caps bring other sonic benefits than the filters, so the sonic impacts are quite different and distinguishable

I would say that BOTH mods are worth doing!!!

Detailled consideration, skip if confusing... (as I thought about several points 3 months ago, see my old posts)...

OK, one could still argue that by fitting the filters I did per se add some PS capacity as the filters include some caps aswell at their outputs and they end up adding to the ones I already added, especialy true in the second filtering RLC stage of Mark's filter. BUT, based on the above, and on my experience with the filters with other builds, my bet is that the effect of the filter is solely due to the filtering capability (and probably the coil's performance from 200kHz on in that case) than to any "pure" increased capacity discussion (or even dare I say the resulting lowered Fc of the second stage of the filter by having its output cap in parallel with the one I added on the OS board, hence reducing the cutting frequency of the filter's second stage). Rafa, that rather lowish frequency garbage is probably already more than adequately addressed by Papa's exiting filter...

I exposed my theory and guesswork 3 months ago in a long post with various models to "consider" what things were happening and also why I would try this mod. I was really seeing it as completely complementary to the additional PS caps à la Tungsten and mainly to address "over 190kHz" garabage produced by this very SMPS that comes with the VFET kit - while guessing the original small filter board that came with the kit might be better at filtering lower frequencies than these. Say I trust coils more than cap's ESR at HF, but that's me, and I didn't use a scope.


If not, where should that extra capacitance happen? Before or after the filter? Where "exactly" and what values would be the cherry on top
=> "Cherry on top"... no clue! I just used what I had already fitted, assuming it worked great and was quite separate from the additional filters, despite what I just mentioned indeed above. You could use... what I used re additional PS cap on each OS board. That is the only recommandation I can make and, that is guess work, I doubt varying a bit from it changes massively the sound.

Now, my PS chain is:
1) External SMPS
2) Original small filter board inside (the one by Papa that feeds the blue LED)... YES, I kept it, refer to the old post
3) Split PS for each channel... what follows is for one channel only, same happens to the other one!
4) Mark's filter
5) Additional PS cap at OS PS entry
6) regular OS (which feeds also positive FE board as ever). Read: these mods also help the FE power feed...

I would really keep it that way, that is filters in that order and additional PS caps last at OS level. That is what makes most sense in my mind, please refer to my post 3 months ago where I explained why this sequence would make the most sense (to me!).

No clue if I am wright, no clue if that can be bettered, but all I can say is: GO FOR IT, it works great as it is and IME once a tweak works that great, your are close if not at the limit of the potential gain around its topic, so probably more worth investigating elsewhere now. Read: quite mature that way IMO, topic PS closed. I bet there isn't much into that PS now and a regulated linear (good) one.

Of course I could be wrong, it is just me after all 🙂
Plus there are obviously some guesswork... BUT the result stands, and that means the sound has increased a lot!

Thank for any feedback
=> You are welcome anytime

Claude
 
Last edited:
Yep, it is there already with pix so you can see how to implement them best, saving you some thinking we had LOL.

However, since you asked so nicely and my posts are very exhaustive in all senses 🙂, here we go again in a hopefully clearer format...

Please note this is just what "I did", not necessarly what "must be done" or "what others would do"...

POWER SUPPLY TWEAKS
IMHO MUST DO and both tweaks complement each other very well...

1- PS FILTERS
Addition of two PO89ZB filters, one for each channel, electricaly "between" the existing little filter board (feeding also the LED) and the OS boards.

2- PS CAPS
At the OS PS feed, where the PS cables come in, addition per board of one low ESR Panasonic FC 1500uF/50V cap
Easily done on top of the boards.


BYPASS CAPS IN THE SIGNAL PATH
This is really fine tuning and not a must do, but always good to have...

3) On the OS boards
Bypass the huge 10 000uF output cap with a Nichicon KZ 220uF/50V
Easily done under the board

4) On the original FE boards
Bypass the original 1000uF cap (C2) with a Panasonic ECW 1uF PP cap
I fitted them under the board and used taller stand offs


I ranked all these tweaks more or less in what is likely to be your perceived order of importance regarding sonic impact.

I hope this helps

Claude
 
Last edited: