DIY Sony VFET Builders thread

Kapton tape (generic name: polyimide tape) attaches the thermocouple's "bead" to the transistor case or to the heatsink. Then a special DVM with special thermocouple I/O connectors, reads the temperature directly.


_

OK I think it’s safe to say an infrared thermometer even set properly is not the right tool to use to measure amp temperature. I can put my finger on this same spot for almost 8 to 10 seconds.

Can anyone duplicate my findings if they have the same type of meter and set it to .2 and check it about an hour after runtime.
 

Attachments

  • 16D70308-F173-46B5-BD9B-9382F350D743.jpg
    16D70308-F173-46B5-BD9B-9382F350D743.jpg
    978.1 KB · Views: 280
ClaudeG,

Doing some testing to see how the transformer stock front end works.

When the output is 1w into 8 ohms I see .4% THD. The stock front end hooked up driving the output makes .02% THD.

So, it seams like, as you said, the output follower is the dominant distortion. Seem to me this is what you want.

At the Nelson recommend 20v you get the .4%. Best for average speakers. Have efficient speaker? Increase it to 22v and you are at .26%.

It’s a fun Amp to play with. I’m sure there will be many tweaks coming.

🙂
 
No it’s actually the other way around the two laser dots show the sample area that’s being measured as the dots grow further apart the accuracy of the measurement begins to diminish. so a 12 feet it sample area is 1 foot and at 6 inch its 1/2 inch sample area. Closer is more accurate.

The laser pointer and IR reading sensor only meet at the same place if object is further apart from the meter.
In very close proximity, error where laser points and sensor reads, is biggest.
That's what instructions for mine said.
 
Thanks Mark 🙂

That is indeed what I suspected (lowest bandwith...), but taking the "Edcor suspect" out of the equation didn't really speed things up. Doesn't though indeed mean that a faster FE wouldn't "help" things overall: the OS speed being what it is... but a very fast FE would make things faster overall. Not a neutral FE then either, more a pushy one, but some kind of compensation then... I still keep my ability to swap FEs very quickly, "because" of you.

I meanwhile have found some compensation, and I am nearly there... so this Edcor somewhere still fits right here despite being just a tranny...

Enjoy music and thanks again

Claude
 
Hi Botte

You are opening an entire new field!

Indeed, I didn't even thought of daring to deviate from the 20.00V setting I have!

Thanks for your figures, and yes indeed, the key part for me of this amp is the OS board, as that is what makes this amp so special to my ears and eyes.

Thanks again for sharing your interesting findings!

Claude
 
No it’s actually the other way around the two laser dots show the sample area that’s being measured as the dots grow further apart the accuracy of the measurement begins to diminish. so a 12 feet it sample area is 1 foot and at 6 inch its 1/2 inch sample area. Closer is more accurate.

My manual says the most accurate measurements are where the dots overlap. If they don't, it's measuring the area between them.
 

Attachments

  • Screenshot_20210712-174845_Drive.jpg
    Screenshot_20210712-174845_Drive.jpg
    618.3 KB · Views: 238
Moving on to tweak 3...

As documented in another post, this is the very simple addition of a low ESR quality 1500uF/50V cap on each OS board, between the +V and G feeds. Full credit goes to Tungstenaudio as he was I believe first to mention it.

This mod seems to work fine with caps from 470uF as other report. I don't know if it helps the FE, the OS or both, but helping it does. Oh, and our amp started several time as expected without any issue (hiccup modi etc.)

We found the mod addressed nicely the areas where this amp was less outstanding in our set up. That is:
- rising edge, speed, slam... notes, when quickly played, simply start much faster... when needed
- composture, maturity, while not losing the joyfull kid inside the unit. The amp is simply less confused on complex music with many instruments, it sounds clearer, can follow all instruments distinctively
- the addictive cinemascope soundstage is now nailed. On complex music, it could lose on spatial precision, or become a bit stretched with focus on the two LS sources. Now it still fills the room while remaining incredibly consistant in density and spatiality in all its dimensions.
- Control over the music, like a better damping factor. The groove factor is still there: it is groovy, but less boomy. That is across all frequencies, perhaps a tad less increased in the bass though, not a problem.

We couldn't hear any negatives or any change in what makes the VFET so addictive re sound quality. It is the same, relaxed, just with more torque when needed, "standing up" in a much quicker way on the "rising front" - that's the way it is perceived by us. I would strongly recommend anyone to try that mod, unless you like it very relaxed and soft all the time. For those worrying: I haven't lost any of the relaxness of the amp, it is just it sounds faster... only when needed, and not say soft all the time and somewhat confused on complex music. Also, it has gained in maturity and neutrality, but is still a very joyfull kid full of colours, nothing is lost on that account either.

An excellent amp on many aspects now gets even better, becoming more consistant across its various qualities: the excellent ones are untouched, the good ones get the very good label. I have some more bits to try, but TBH now - as before and ever - I could live with it as is 🙂

All to our ears for whatever it is worth

Enjoy music

Claude
 
Pix...

Yes the cables run underneath, +V and G went through and stand proud half an inch or so to make an easy connection to caps (fitting and removing without damaging the boards)
 

Attachments

  • IMG_20210712_123213.jpg
    IMG_20210712_123213.jpg
    51.2 KB · Views: 330
  • IMG_20210712_123219.jpg
    IMG_20210712_123219.jpg
    52.9 KB · Views: 327
  • IMG_20210712_123228.jpg
    IMG_20210712_123228.jpg
    53.4 KB · Views: 316
Last edited:
now, find biggest possible motor run cap and tuck it inside

use hot glue gun, if needed, to horror of all Puritans :clown:

or, even better - put one on each channel rail, and second one as bypass for each output cap

it can be - either broader Green and nicer sleep, or just one more Audiophool moment

:devily:
 
Nah, have enough to do with my 2 SMPS filters 🙂

Given my recent failures, I am more likely to perform a "quick and dirty offboard test", with long wires and lid off, just to see if it goes into the right direction before looking into any permanent installation. Papa has provided this function already, so...

All that is likely to wait for after the summer 🙂

Enjoy music

Claude
 
The slowest piece of the VFET amp is its Edcor transformer. To speed up the amp you need to find a way to speed up the Edcor. Or else use a front end card which has no Edcor.

Hmm, daylight is back... I come again on this and see perhaps a hidden message.

"Relentless" sounds like a nice program... a very mysterious one that is likely to remain so, but I can spot... an EDCOR transformer on it! I believe you said you liked that FE a lot. And there seem to be quite a few of us liking the iron sound, neutral or not (and if it is not, it isn't by a mile and remains pleasant anyway). So...

May I humbly ask for the sake of learning what kind of means can be used to "speed up" a transformer?

Learning everyday, many thanks in advance

Claude
 
now, find biggest possible motor run cap and tuck it inside

use hot glue gun, if needed, to horror of all Puritans :clown:

or, even better - put one on each channel rail, and second one as bypass for each output cap

it can be - either broader Green and nicer sleep, or just one more Audiophool moment

:devily:

Love ZM's humour. Me say no thanks. I like originality just how pa intended.

For your viewing pleasure. Look at the bottom ma! I'm not hiding any wires :spin:

K2cQBFj.jpg