Head-Fi.org › Forums › Misc.-Category Forums › DIY (Do-It-Yourself) Discussions › eXStatA Build Thread II
New Posts  All Forums:Forum Nav:

eXStatA Build Thread II - Page 3

post #31 of 567

Hello again,

 

I am now matching J271s prior to populating my amp board.  Unfortunately, it is not yet obvious to me how to test for IDSS from a schematic and so, in my search for a guide, came across Tangent’s  instructions at  http://tangentsoft.net/audio/opamp-bias.html  ( bottom of page), which I followed.   Below is a pic of my matching setup.  I am using a brand new Energizer A23 12V battery with the positive side connected to the jumper.  When I measure, I connect the positive lead of my DMM to the JFET’s drain and the negative lead of my DMM to the negative side of the battery.  I am certain that my DMM leads are connected properly to the meter and that the meter's settings are correct.  However, I’m telling you all of this because I'm surprised at my results and so wonder if I've done this right. Out of 20 J271s, I only have 6 whose IDSS measured above 10.  Their readings are:  20.75, 14.18, 12.76, 12.72, 11.13 and 10.48.  (I subsequently used the second new battery that came in the pack to do another reading of all of JFETS.  The results were all very close to the first readings --  generally within  5 to 15 one hundredths of a milliamp.) 

 

So, gents, am I good to go with 12.76 and 12.72 on one side and 11.13 and 10.48 on the other?  Or have I done something wrong to get so many low readings?

 

Thanks.

 

Matching.jpg

post #32 of 567

that seems ok. datasheet says the idss range is 6 to 50 ma. also when you test the jfet the gate must be conneced to the source. dont know if you have done that properly. just be careful not to blow them because it s your only quad from what i can tell.

 

also there is a very good website i use sometimes, full of informations:

 

http://www.cavalliaudio.com/exstata/main.php?page=matching

post #33 of 567
Thread Starter 

The fets need to be matched better if you want the two channels to have similar gain. Not absolutely necessary, but would help unless you are using two separate volume controls.

 

It is useless, of course, to say that generally with 20 fets I get one good set. But, statistically, your result has some probability.

Dr. Cavalli gained notoriety with his first DIY amplifier projects. His success has blossomed into Cavalli Audio, a world leader in amplifier design.
Reply
post #34 of 567

Thanks.  I will buy more to get a better match.

post #35 of 567

My additional J271s will arrive tomorrow.  In the meantime, I have a couple of unrelated transistor questions pertaining to my amp board:

 

(1)  On the board, the sides of the outline for Q3 and Q4 show a taper.  The parts also have an ever so slight taper to them.  If I use this as my guide to insert these transistors, the printing on them is on the opposite side of the wider white line that indicates the width on the outline on the board.  Yes?  No?

 

(2)  Opening up my package of Q11-12-13-14s woke me up faster than my cup of coffee this morning.  The Mouser P/N printed on the envelope is identical to what is specified on the BOM:  512-KSC5042FYDTU.  Fairchild’s P/N, also printed on the envelope, says KSC5042FYDTU.  But – these transistors do not have their three pins aligned in a row as the board indicates.  Rather, when you look up at them from below, you see that the pins form a triangle:  all three pins come out of the body the same way, but then the center pin bends out and then down.  The printing on the transistor says C5042F and then J11 below that.

 

Finally, since I’m posting (I’ll try to shorten these things in the future), let me ask one more question, please:  Does the heatsink wiring image on the eXStatA website apply to my hybrid amp?  If yes, am I right in saying that you simply connect two wires to the heatsink pins from underneath the board and run them to the ground of the terminal connector?  If this isn’t right, what should I do to ground my heatsinks?

 

Many thanks.  

post #36 of 567

Nearly an hour on the phone with Mouser has answered my question #2 above.  KSC5042FYDTU came to me from Mouser in the shape I described because a customer ordered 5,000 in that configuration and Mouser has some extras that they are passing on to a guy like me to get rid of their inventory.  The reason I didn’t get the straight-pinned version is because the part is now obsolete in all of its forms: http://www.fairchildsemi.com/obsolete/obs_result.jsp

 

Mouser shows limited stock on hand:  67 pieces of KSC5042FTU (not the version specified on the BOM but the same transistor) and 56 pieces of the KSC5042FYDTU that I got.  The tech I spoke to (whose comments and manner inspired confidence in me) is not sure whether the “FTU” also has the triangular configuration on the pins.  (TU just refers to the packaging =  “tube”), but he is certain that both are the right transistor and was confident that if I were willing to straighten out that center pin to resemble the shape of its neighbors I’d be good to go.

 

I apologize for not having checked this out before posting the above.  Hope no one has been inconvenienced.   (Still need an answer to my other two questions, guys.)

 

post #37 of 567

My mother used to tell me, “Look before you leap!”  Apparently she knew me very well.  I have found among the photos I have collected one with sufficient resolution (by the designer, no less) that clearly shows that the orientation of the Q3 and Q4 transistors do indeed go on the board the way I asked about above.  Sorry again, gents.  Guess I only need the answer to my heatsink question.  

post #38 of 567

#1 - Yes. The tapers should match up.

 

I grounded my heatsinks exactly as you described.

 

That's interesting about the other transistors. Sounds like they will run out soon then...
 

Quote:
Originally Posted by rbarth View Post

My additional J271s will arrive tomorrow.  In the meantime, I have a couple of unrelated transistor questions pertaining to my amp board:

 

(1)  On the board, the sides of the outline for Q3 and Q4 show a taper.  The parts also have an ever so slight taper to them.  If I use this as my guide to insert these transistors, the printing on them is on the opposite side of the wider white line that indicates the width on the outline on the board.  Yes?  No?

 

(2)  Opening up my package of Q11-12-13-14s woke me up faster than my cup of coffee this morning.  The Mouser P/N printed on the envelope is identical to what is specified on the BOM:  512-KSC5042FYDTU.  Fairchild’s P/N, also printed on the envelope, says KSC5042FYDTU.  But – these transistors do not have their three pins aligned in a row as the board indicates.  Rather, when you look up at them from below, you see that the pins form a triangle:  all three pins come out of the body the same way, but then the center pin bends out and then down.  The printing on the transistor says C5042F and then J11 below that.

 

Finally, since I’m posting (I’ll try to shorten these things in the future), let me ask one more question, please:  Does the heatsink wiring image on the eXStatA website apply to my hybrid amp?  If yes, am I right in saying that you simply connect two wires to the heatsink pins from underneath the board and run them to the ground of the terminal connector?  If this isn’t right, what should I do to ground my heatsinks?

 

Many thanks.  

post #39 of 567
Thread Starter 

Yes, the triangular pad shape is standard now, but it is very easy to straighten the middle pin into alignment with the other two. I think most everyone has had to do this and all works ok.

 

I probably should have changed the pad spacing on this new rev but I didn't think of doing so.

 

The heatsinks that you want to ground are the ones where Q11/Q12 are attached as in the diagram.

 

On the SS board, even though the ground traces are built-in, you still have to solder the heatsink pins to their pads to make this work right.

Dr. Cavalli gained notoriety with his first DIY amplifier projects. His success has blossomed into Cavalli Audio, a world leader in amplifier design.
Reply
post #40 of 567

Here are the results/groupings of my “good” IDSS measurements after getting my second shipment of J271s.  (My son will be building an SS eXStatA and, in light of the poor results I got from my first set of 20, I went ahead and ordered an additional 50 to ensure we would both have really good matches.)

 

10.00, 10.11, 10.29, 10.34, 10.39, 10.40, 10.51, 10.52, 10.56

11.09, 11.11, 11.12, 11.36, 11.98

12.40, 12.72, 12.84, 13.27

20.34, 20.39, 20.41, 20.75

 

So, how do I pick from these?  Is a higher IDSS better?  Or, are those J271s whose IDSS values vary the least across four pieces better?   Also, are there any guidelines or rules for how to select which should be Q1L, Q1R, Q2L and Q2R once the quad is chosen?

post #41 of 567

Many excellent potential quads in there. I would go for the 10.29 through 10.40 and the 11.09 through 11.36.

post #42 of 567

Just one amp?  I'd grab the quad of 10.39, 10.40 / 10.51, 10.52 and pair them that way. 

 

Sell off a couple of the other close quads like the 11.09, 11.11, 11.12, 11.36.  I know a peep who could use them.  I'll have him send you a PM. 

post #43 of 567
Thread Starter 

If it were me, I would take the ones in the 20s and then the 12s.

Dr. Cavalli gained notoriety with his first DIY amplifier projects. His success has blossomed into Cavalli Audio, a world leader in amplifier design.
Reply
post #44 of 567

I would like to learn as well as just build this amp.  Why the 20s and then the 12s?

post #45 of 567

Well, Alex can chime in here with his reasoning, but I think it's because the jfets are the principal component providing gain in this design. Again, I think the better the idss of the jfet, the better the transconductance. On the matching page Alex recommends 10mA is the absolute minimum.

New Posts  All Forums:Forum Nav:
  Return Home
Head-Fi.org › Forums › Misc.-Category Forums › DIY (Do-It-Yourself) Discussions › eXStatA Build Thread II