Pairing of store jfet DIYAUDIO

[uchu007]

Hello,
I just got some Jfets from the DIYAUDIO store that was supposed to be paired but when I tested them I noticed a very big difference between the two.I bought a few months ago on the bay a pair a punkydawgs recommended by Nelson and it was really identical.
Are the Diyaudio store pair really well matched? I’m really disappointed

[Zen Mod]

I believe Store JFets are matched simply by Idss at fixed Uds

even if slightly dependable of their role /sort of circuit where you're going to use them , I wouldn't loose my sleep over difference I see in your graph , if they're matched by Idss

edit: what I wrote applies for same sex one - exact matching by Idss

for N against P matching ...... it really depends what I want to achieve and where I'm going to use them ..... but there I'm having even more relaxed criteria

for instance - F5 - you'll never same conditions of work for N and P Jfets , considering that needed voltage sag across drain resistors must be different , due to different treshold (biasing ) voltages of N and P mosfets

absolute symmetry is overrated

[uchu007]

Originally Posted by Zen Mod

I believe Store JFets are matched simply by Idss at fixed Uds

even if slightly dependable of their role /sort of circuit where you're going to use them , I wouldn't loose my sleep over difference I see in your graph , if they're matched by Idss

edit: what I wrote applies for same sex one - exact matching by Idss

for N against P matching ...... it really depends what I want to achieve and where I'm going to use them ..... but there I'm having even more relaxed criteria

for instance - F5 - you'll never same conditions of work for N and P Jfets , considering that needed voltage sag across drain resistors must be different , due to different treshold (biasing ) voltages of N and P mosfets

absolute symmetry is overrated

Yes but this is really disappointing when have expected a pair close enough. This is for an F5T V2 .
Here are the graphs of those bought a punkydawgs on Ebay it’s already much better , I think use this especially that have bought them around the same price

[Zen Mod]

on higher voltages they're asymptotic closer and closer

so , you're using them on which voltage ?

just chill ....... even 1mA of difference between N and P will not spoil your sleep

take care to use P3 in construction and to set THD by your liking and that's it

[woofertester]

Having curved traced hundreds of transistors, I can tell you that N and P parts for both JFETs and BJTs behave differently even if you match gain at some Vds or Vce at some Id or Ic.

Try plotting the gain curves. For JFETs, plot Id vs Vgs. For a BJT, the curve is Ic vs Ib. After plotting those curves, you calculate the derivative of these curves (gain) and plot that derivative vs Ic or Id.

Even when gain is matched for an N/P pair at a particular voltage and bias current, I have seen that the way in which the gain changes with current is different for the N than it is for the P.

It appears that you like the Toshiba parts more than you like the Linear Systems parts. Vote with your wallet.

[uchu007]

Originally Posted by Zen Mod

on higher voltages they're asymptotic closer and closer

so , you're using them on which voltage ?

just chill ....... even 1mA of difference between N and P will not spoil your sleep

take care to use P3 in construction and to set THD by your liking and that's it

I used it in cascode as in the 6L6 guide. So with a transfo of 2x25v I should have rails of about 32v and 32/ (10k +4.75k) * 4.75k = 10.30 v.
So they should be used in the 10.3v.

[Zen Mod]

as I said , no worries

if you want , test them at 10Vsomething ....

[uchu007]

Originally Posted by Zen Mod

as I said , no worries

if you want , test them at 10Vsomething ....

Ok thank you Zenmod

[Nelson Pass]

At $20 each or so, Punkydawgs does a fine job.

However it is unrealistic to expect complete matching between P and N,
which is why there are three pots to adjust bias, offset, and harmonic
structure on the F5.

Even if you have perfectly symmetric Jfets, good luck finding power Mosfets
with comparable matching.

I recommend 1 mA matching on these parts, and I don't see much point
in paying for better in this application.

BTW, 1 mA is the standard for matching in the store if you are buying
the matched parts. Apparently one customer was disappointed when
buying unmatched parts and found that they were a bit outside this
standard.

There has been some discussion about offering a 0.1 mA at extra
expense. I imagine they will sell like hot cakes.

[uchu007]

Originally Posted by Nelson Pass

At $20 each or so, Punkydawgs does a fine job.

However it is unrealistic to expect complete matching between P and N,
which is why there are three pots to adjust bias, offset, and harmonic
structure on the F5.

Even if you have perfectly symmetric Jfets, good luck finding power Mosfets
with comparable matching.

I recommend 1 mA matching on these parts, and I don't see much point
in paying for better in this application.

BTW, 1 mA is the standard for matching in the store if you are buying
the matched parts. Apparently one customer was disappointed when
buying unmatched parts and found that they were a bit outside this
standard.

There has been some discussion about offering a 0.1 mA at extra
expense. I imagine they will sell like hot cakes.

Thanks Nelson , yes but with a deviation as much I do not call it an pairing especially the Idss of the LSk170 are 5,14 and 5,23 well below the 6ma that may not be a problem but still.There are also gaps of more than 1ma on the curves. It is true that Punkydawgs does excellent work because they are very well matched