AMB Laboratories DIY Audio

[tenorsaxophone]

Hi,
I have Q3 and Q4, and now the output offset is reversed!
Before the swapping it was 60mV V3 completely counterclockwise turned.
Now, after the swapping it result -38mV V3 completely clockwise turned.
I think this is really a unlucky Q1,Q2,Q3,Q4 combination.
But a lest, changing from a positive to a negative voltage, it shows that the cause is really here, can you confirm it?
I have foud a vendor that sells a quad JFET! 2SK170BL / 2SJ74BL matched +/-3% Idss.
Do you suggest to me to try this way or there is a better solution?
Many thanks for all. Reagards.
Davide.

[amb]

Your results obviously shows that there is an imbalance that swapping the JFETs could cure. Perhaps now you should try swapping Q1 and Q2 to see if you can get things down to within the adjustable range. If that still not enough, a new matched JFET quad is worthy, it's up to you where to buy it from.

[tenorsaxophone]

Hi,
now, swapping Q1-Q2 too, voltage output is -58mV (V3 completely clockwise turned)!!
I am very frustrated.
Now I must replace Q1-Q4....
But there is a problem:
JFets Q1-Q4 that I am using now, I had bought them from your Shop with matching service included. So, if I need a better Idss matched quad, this is not simple.

What is the the Q1-Q4 Idss mA tollerance that could be right?

What do you suggest to me?

Many thanks for all. Regards.
Davide.

[amb]

Typically the matching tolerance from me is more than tight enough, but whether the amp will still have a DC offset that cannot be trimmed to 0 may also be influenced by the other transistors. BC550C and BC560Cs are usually fairly consistent without matching, but perhaps you're unlucky and got one that's really "off". Try swapping one pair of Q9/Q10 or Q11/Q12 to see if a pattern emerges. You maybe able to either get within an adjustable range this way, or narrow down to a specific pair that does not match well.

Failing that, you could go to the next solution which is to add diodes (as described in the Initial setup > Troubleshooting section of the β22 website).

[tenorsaxophone]

Hi,
I own a large amount of BC550c/BC560C BJTs.
I could replace Q9-Q12 with a good matched quad, and to see if something changes....

[tenorsaxophone]

Hi,
I have measured the Hfe of Q9-Q10 that I have unmount from the PCB and I have found a big difference: 511 and 591 (Q11-Q12 are 543 both)! Do you think this could be the cause of that offset?
I have tryed to match my own BC560c, but I the result matching has not be good.
Could be the same using BC556-557-558c as replacement or is it a worse chosen?
Thanks again for your help.
Davide.

[amb]

Hi,
I have measured the Hfe of Q9-Q10 that I have unmount from the PCB and I have found a big difference: 511 and 591 (Q11-Q12 are 543 both)! Do you think this could be the cause of that offset?

Maybe. The design of the amp is such that Q1-Q4, Q9-Q10, Q21 and Q22 all "hang" in balance, adjustable to some degree with VR2. If any of them is far off from the others, could pull the balance out of the adjustment range.

I have tryed to match my own BC560c, but I the result matching has not be good.
Could be the same using BC556-557-558c as replacement or is it a worse chosen?

I would not recommend changing to a different transistor model, because they have quite different characteristics.

[tenorsaxophone]

I finally identified the problem!!
Replacing JFets Q1-Q4 with a matched quad 1% tollerance, the output offset can be eliminated through VR2 trimot.
Now I have completed the amp and it sounds great!!!
Many thanks for all. Best regards.
Davide.