[j4cbo]

Well, here's what I wound up with for the drive transistors:

Group 1
2SA1358Y: 148, 148, 150, 153; spread 5, average 149.75
2SC3421Y: 148, 148, 148, 149; spread 1, average 148.25

Group 2
2SA1358O: 128, 129, 129, 131: spread 3, average 129.25
2SC3421Y: 128, 130, 132, 133: spread 5, average 130.75

It's not perfect, but pretty good. Started with 20 of each.

 

[Mimas]

Im building KGSS... For input stage instead of 2SK389(since they aren't available anymore) I have bought 30 pcs of 2SK170 for matching (on the right). Idss looks ok, but transistors themselves look a bit weired comparing to the ones on the left(I have 4 of those).
Idss(at 9 volt) looks like this:
for left transistors:
6,22 (mA)
6,26
6,92
9,95
And those on the right:
6,65
6,8
6,82
6,85
6,93
7
7,13
7,18
7,49
7,5
7,56
7,63
7,64
7,7
7,74
7,84
7,93
8,13
8,13
8,81
9,15
9,34
9,4
9,45
9,53
9,63
10,25
10,25
10,43
10,73

Looks ok, but I'am still concerned about them. Did anyone see 2SK170 like those on the right? All pics of s170 i could find on inet look like left ones. Hope those aren't fakes...

 

[dip16amp]

I think they are okay since they all are within BL range of 6-12.
I've had other components that had different lettering from the same manufacturer. Just different batches.

 

[luvdunhill]

when matching transistors on the Millet diamond buffer, I'm trying to first get h_fe of Q6 == Q7, that part is clear. For Q2,Q3,Q4,Q5 am I trying to get them to all have the same h_fe, or Q2 == Q4 and Q3 == Q5?

Schematic is here:

http://www.diamondstar.de/dDB_isb_v022_schem.jpg

 

[MisterX]

http://www.diamondstar.de/dDB_partslist.html

Quote:

[size=small]Note 3: hFE matching of BJT transistors[/size] Transistor matching is not a necessity, most probably it doesn't make an audible and just a measurable difference Since we are using bipolar junction transistors (BJT) matching is done by their hFE values. Matching is done within a single channel only, you don't have to match the channels Match input stage Q6 / Q7 (good chances of success) Match current mirror Q2 / Q4 / Q3 / Q5 (medium chances of success) Match output stage Q8a/b / Q9a/b (poor chances of success, little benefit) The small signal transistors used in the input stage and the ccs mirror (2n50xx ... BC5xx) are cheap (~ 5 cent ea.), get about 30 to 50 of each to find decent complementary pairs among them. For the large transistors used in the output stage it's unlikely to get perfect matching pairs, probably npn and pnp types are more or less off each other. Don't get frustrated if the ranges not even overlap, group them as good as possible but don't buy too many surplus. The easiest way matching transistors is to measure their hFE values with a DMM that has a transistor socket and supports this mode. There is a nice transistor matching thread at head-fi.org and a page about BJT transistor matching on my site if you want to read more about this topic.

 

[luvdunhill]

Quote:

Originally Posted by MisterX /img/forum/go_quote.gif http://www.diamondstar.de/dDB_partslist.html

MisterX:

I don't why that verbiage isn't clear to me

So, I want Q2 = Q4 = Q3 = Q5?

 

[steinchen]

Quote:

Originally Posted by luvdunhill So, I want Q2 = Q4 = Q3 = Q5?

yes

 

[Alfiax]

I've been preparing for the upcoming set of amps that use the 2sj74 and 2sk170 by trying to match up for Idss and have run into a bit of a problem. I ordered 2sj74-BL from B&D and a mixed assortment of 2sk170 from Futurlec which turned out to be mainly BL with some GR. The 2sj74 tested just fine within their listed range (6-12 mA,) but the ~50 2sk170-BL I got from Futurlec are way out of spec, in the range of 15-16 mA. This is near the center of the V range!

I'm trying to figure out what's going on here. I know that these JFETs have a reputation for being off spec, but this seems ridiculous. Is it possible I got re-marked or defective parts, or is this par for the course for these devices?

 

[Pars]

If you are certain that you have the polarities flipped around correctly for matching these (from the 2SJs), I'd send them back as out of spec.

 

[Alfiax]

Yeah, I'm pretty sure the polarities are right. I'm a little loath to send them back as my order did not specify gain class; I'd be in the same position with no ability to complain if these parts were properly labeled as V grade. My big worry is that I got another brand of JFET relabeled as 2sk170. There are lots of rumors but little hard evidence on relabeled parts; while there seems to be a good case for power devices commonly being relabeled i don't see too much about small-signal devices. It would seem like the profit-per-unit would be low enough that it wouldn't be worth it for these, but who knows?

 

[amb]

FWIW, I ordered 25 of each (all BL class) from B&D some time back and was able to get 5 sets of reasonably good 4-way matches (this is for the β22 amp). Maybe I'm lucky...

 

[operafollower]

Hi
I am going to build KGSS or BH.
Must I match its transistors (not first FET) like gilmore dynamic too?
I read headwize so I know KG** must adjust offset with potentiometers.
It can not adjust If I did not match transistors?

 

[luvdunhill]

hi! would someone be so kind as to compare the method of matching BJTs recommended in this thread (using a variable resistor and two meters) versus grouping by collector voltage (the method is outlined here http://www.diamondstar.de/transistor...ing_bjt.html)?

Thanks!

 

[amb]

Quote:

Originally Posted by luvdunhill /img/forum/go_quote.gif hi! would someone be so kind as to compare the method of matching BJTs recommended in this thread (using a variable resistor and two meters) versus grouping by collector voltage (the method is outlined here http://www.diamondstar.de/transistor...ing_bjt.html)?

They are essentially doing the same thing, except that in the "grouping by collector voltage" method, you don't actually know what the Hfe of each device is without doing some computation.

 

[lcp12345]

Does your guy know that Dynahi PCB is still available or not? I have email Dan. But, there is no response yet.

I want to build a Balance Dynhi to drive my KK!