grünschnabel !
Exactly ! Still wet behind the ears
As you are on your 224th Edition, I'd be happy to cite you in the Chapter: Dirty tricks - How to get 2nd harmonic to 0.023% and 3rd harmonic to 0,004% by using artistically, carefully unmatched semisouth or ZenMods famous ZenPot
With CCS at 20mA and some semisouth source-resistor tweaking the 2nd and 3rd harmonics looked fine, when getting back to 10mA CCS the values are far too low and the amp sounds too thin and uninvolving.
I need some help... easy instructions....for dummies like me.....
Last edited:
the problem is not as easy...
my friend dsdjoy succeeded in cloning the THD spectra 98%, even Nelson got a nice picture of the similar spectra of clone and original and accepted that dsdjoy had done a good job.
But the original still sounds better...... !
and the master is laughing......or..... or....
my friend dsdjoy succeeded in cloning the THD spectra 98%, even Nelson got a nice picture of the similar spectra of clone and original and accepted that dsdjoy had done a good job.
But the original still sounds better...... !
and the master is laughing......or..... or....
Could it be some other part you are overlooking that is effecting the sound of your J2?
Maybe it was some other choice in part(s) and not your circuit design that's having, to your ears, a negative effect or less than perfect result.
What about polarity at the speakers? What about break-in time? What about grounding scheme? Power supply? etc, etc.
Maybe you are dead on target, but just overlooking one small piece. Just a thought.
Hope you find the missing piece. Really do!
Vince
if you boys got it 98% of THD spectra , rest is in your head
I had short circ between two headphones too many times , so I know how's that
So we will do what is to do to hear no difference any more....
Could it be some other part you are overlooking that is effecting the sound of your J2?
Maybe it was some other choice in part(s) and not your circuit design that's having, to your ears, a negative effect or less than perfect result.
What about polarity at the speakers? What about break-in time? What about grounding scheme? Power supply? etc, etc.
Maybe you are dead on target, but just overlooking one small piece. Just a thought.
Hope you find the missing piece. Really do!
Vince
Thank you Vince!
Copycats are looking in the direction you suggested....
if you boys got it 98% of THD spectra , rest is in your head
I had short circ between two headphones too many times , so I know how's that
Sonically after having 98% the same Spectra my J2 inspired amp sounded almost identical to the original. Without a direct A/B comparison they would be almost indistinguishable.
But what am still after is - it is only a nuance - following:
The original J2 has a bit more silkyness, while my J2-inspired version is a tiny bit more on the hard, edgy side.
As said this is a nuance, and you can argue, it is a short-circuit between two headphones
However, I tend to believe that it is due to my usage of 20mA CCS....
Why, did I use 20mA then ?
Well, first I applied the mod you suggested:"fold input LTP (inner drain) to lower SJEP source".
GOOD - it brought THD from 0.015% to 0.022% - right direction.
2nd harmonic looked almost fine, but 3rd harmonic was still too low.
I found that raising CCS from 10mA to 20mA brings up 3rd harmonic - Good - right direction !
Then I began fine-tuning K2/K3 relationship by using LARGE ZenPot (in wrong place,but working for me):
When I was happy with K2 vs K3 relationship, I replaced the pot with a
dsdjoy Butterfly(TM) resistor.
A tiny bit adjustment of feedback to raise overall K2 & K3 and
I got to 98% same Spectra.
The overall process took many many days of iterations, measurements and listening sessions (repeat similar procedure for 2nd channel !).
This is where I am. Only things that bores me (translate: does not let me sleep) is that J2 sounds a bit more silky/relaxed vs. my J2-inspired amp sounds a nuance more on hard/dynamic side.
Is it the 20mA CCS that was a wrong choice ?
A bit like in a known board-game I fear I have to start all-over: Go Back to Start, Do Not Pass.
Well, first I applied the mod you suggested:"fold input LTP (inner drain) to lower SJEP source".
GOOD - it brought THD from 0.015% to 0.022% - right direction.
2nd harmonic looked almost fine, but 3rd harmonic was still too low.
I found that raising CCS from 10mA to 20mA brings up 3rd harmonic - Good - right direction !
Then I began fine-tuning K2/K3 relationship by using LARGE ZenPot (in wrong place,but working for me)
:
When I was happy with K2 vs K3 relationship, I replaced the pot with a
dsdjoy Butterfly(TM) resistor.
A tiny bit adjustment of feedback to raise overall K2 & K3 and
I got to 98% same Spectra.
The overall process took many many days of iterations, measurements and listening sessions (repeat similar procedure for 2nd channel !).
This is where I am. Only things that bores me (translate: does not let me sleep) is that J2 sounds a bit more silky/relaxed vs. my J2-inspired amp sounds a nuance more on hard/dynamic side.
Is it the 20mA CCS that was a wrong choice ?
A bit like in a known board-game I fear I have to start all-over: Go Back to Start, Do Not Pass.
Last edited:
dsdjoy
if you wanna exact copy - you need to dismantle bloody thing and measure Idss of all Jfets ........ I mean - all
then try to repeat that
and then ask Pa what he's doing when he's putting some other Idss grades in ?
OK - say that he choose always (approx.) same current through LTP ;
then it's slightly less believable that he's always choosing same Idss range for LTP itself , simply because 2SJ109 BL's aren't growing in a field (never mind that he's probably having Jfet's inventory as big as all other HiEnd makers together ) ...... so that will be just waste of good wine
same as for output Jfets - it's completely sane thinking that Pa is puttin' there spread of +/-10% Idss ....... meaning 20% of majority
tricks are there ; we are having pics of , say , several J2 specimens ; none of them are showing paralleled source resistors as seen on F5
but - even if we didn't saw them , maybe they're here in some
so - there is another trick ( which I can't share publicly ) to alter/influence Iq .... and that change will also influence OLG , so feedback , so THD spectra ..
whatever - I'm writing too much , most probably contemplating even more ....... but I'm almost positive that Pa isn't so obsessive to have all amps of one breed in same 2% of performance
considering most probably scenarios - that 100 J2 specimens will end on at least 89 different speakers ..... spread of performance will be much much higher than these 2%
conclusion - for us - it's worth chasing understanding of performance , not performance itself
if later is the case - drink glass of red , everything will be 20% better
edit - and yes - I chose 20ma in LTP , along with cascodes
I like them bleed , and shiny , too
if you wanna exact copy - you need to dismantle bloody thing and measure Idss of all Jfets ........ I mean - all
then try to repeat that
and then ask Pa what he's doing when he's putting some other Idss grades in ?
OK - say that he choose always (approx.) same current through LTP ;
then it's slightly less believable that he's always choosing same Idss range for LTP itself , simply because 2SJ109 BL's aren't growing in a field (never mind that he's probably having Jfet's inventory as big as all other HiEnd makers together ) ...... so that will be just waste of good wine
same as for output Jfets - it's completely sane thinking that Pa is puttin' there spread of +/-10% Idss ....... meaning 20% of majority
tricks are there ; we are having pics of , say , several J2 specimens ; none of them are showing paralleled source resistors as seen on F5
but - even if we didn't saw them , maybe they're here in some
so - there is another trick ( which I can't share publicly ) to alter/influence Iq .... and that change will also influence OLG , so feedback , so THD spectra ..
whatever - I'm writing too much , most probably contemplating even more ....... but I'm almost positive that Pa isn't so obsessive to have all amps of one breed in same 2% of performance
considering most probably scenarios - that 100 J2 specimens will end on at least 89 different speakers ..... spread of performance will be much much higher than these 2%
conclusion - for us - it's worth chasing understanding of performance , not performance itself
if later is the case - drink glass of red , everything will be 20% better
edit - and yes - I chose 20ma in LTP , along with cascodes
I like them bleed , and shiny , too
Last edited:
Thanks a lot for good hints !
Nah...I would need to know J2 repair price at FirstWatt facility.
....or better.....J2 will stay untouched.
I found they grow better on nutrient agar...
(matched them for a good friend)
Learning is my main aim here, too. I learned how different THD and different K2 and K3 levels sound. Interesting !
Now junior wizard apprentice likes to learn more (hunting
)
Nah...I would need to know J2 repair price at FirstWatt facility.
....or better.....J2 will stay untouched.
I found they grow better on nutrient agar...
(matched them for a good friend)
conclusion - for us - it's worth chasing understanding of performance , not performance itself
if later is the case - drink glass of red , everything will be 20% better
Learning is my main aim here, too. I learned how different THD and different K2 and K3 levels sound. Interesting !
Now junior wizard apprentice likes to learn more (hunting
)
Last edited:
If you are referring to the input bias CCS, then that does sound like
a bit much. I run about 8 mA.
This sounds like 2 x 2SK270GR's
