No the MJE340G is not to be used, that is not optimized for the new spreader. Any of the BD***16S series transistors will work without issue, as JJS said the Vce is low, all 3 are listed just to give options in case there is stock issues
If you check the schematic you'll see the vce for both Q103 and Q104 is well below the DB135 so all those transistors are fine. The reasons for moving away from the MJE340 was explained in @fireanimal post and so those are the transistors that should be used.
Part Numbers as promised.
WOW, I can't say thank you enough for all of your guys time and efforts. I can see, that was a lot of work, but it made an already outstanding amplifier even better. At some point you guys need to stop tinkering, my arms are getting sore taking those mono blocks in and out of my system, LOL.
Thank you very much for this design improvement. I was one who was concerned with the large overcompensation.
I raised this issue in post #488. Repairing a finished amplifier is a bit problematic for me.
Using a teflon washer between Q104 and Q113 I was able to tweak the thermal stability in the desired direction.
@fireanimal Remarkable choice for C112 on the EF3-3 pcb in #2057, this considering your previous opinion on that part. It made me laugh for a moment. Thank you.
@fireanimal Remarkable choice for C112 on the EF3-3 pcb in #2057, this considering your previous opinion on that part. It made me laugh for a moment. Thank you.
While I'm sure that what @OmeEd did make an improvement its not the solution. Many hours has gone into this and you know what they say. If you're going to do it do it right. I encourage you to make the modification outlined by @fireanimal
It will be worth it, but then again it is not urgent. TBH I have 2 Group buy 1 amps in use in my main media room, with 0 mods, or updates from the original BOM and they sound fantastic, and have been running like a champ for almost 1 year now, and they are never shut off.
Tweak is the key word, if you are going to tweak, I recommend just applying the new update.
I put it on to test, and never swapped it out, there was no difference with that cap vs. polyester except it looks nicer lol
I would like to make sure that I understand the mounting of the new transistors. EF3-3 and EF3-4 both have Q104 mounted upright without any heat spreader or heat sink contact.
EF3-3 Q103 is fly wired to the main heat sink between output transistors. For the EF3-4 Q103 is mounted under the board on the main heat sink. The drivers are mount either side, also on the main heat sink. The EF3-4 will no longer use the heat spreader for drivers and Q103.
EF3-3 Q103 is fly wired to the main heat sink between output transistors. For the EF3-4 Q103 is mounted under the board on the main heat sink. The drivers are mount either side, also on the main heat sink. The EF3-4 will no longer use the heat spreader for drivers and Q103.
I make an attempt to keep up with the goings on in the thread, but it's a challenge.
For the current discussion of the new mods, do they apply to the boards from the 1st group buy also?
Is it better to hold off a bit before getting started, since I have not begun construction yet?
Thanks again for continuing to evolve the design!
For the current discussion of the new mods, do they apply to the boards from the 1st group buy also?
Is it better to hold off a bit before getting started, since I have not begun construction yet?
Thanks again for continuing to evolve the design!
Correct, the bias spreader mod applies to ALL generations of Wolverine EF3-3 and EF3-4 boards.
Thanks for this tidbit of info. I have the parts oredered for this change and will wait to do all the work at once when the second twqeak is announced. Like@fireanimal said, "It sounds just fine with no changes".
Nice work (again). Did you obtain accurate temperature coefficients of the MJE340, BD13x16 and maybe other candidates, and is this something you can share? It is hard to find this information in general.
I previously did some testing on the influence of the mains voltage and bias was indeed depending on the power supply voltage. I suspected that the CCS has such a high impedance in this layout that even a large Rcc could not counteract the increase of bias. Changing the value of Rcc did not change much the dependency of bias on supply voltage. Did you measure any improvement on this with the new configuration ?
I previously did some testing on the influence of the mains voltage and bias was indeed depending on the power supply voltage. I suspected that the CCS has such a high impedance in this layout that even a large Rcc could not counteract the increase of bias. Changing the value of Rcc did not change much the dependency of bias on supply voltage. Did you measure any improvement on this with the new configuration ?