Hi !
Want to repair an EMX212s amp module from Yamaha...
Replaced obvious dead 2SA1386A (Sanken) with 2SA1386 icssemi, Chinese mfgr. -
but still have problems with stability.
Has someone experience replacing that transistor.
Is it a good idea to look for the original one (Sanken ) ?
rgds.
fritz
Want to repair an EMX212s amp module from Yamaha...
Replaced obvious dead 2SA1386A (Sanken) with 2SA1386 icssemi, Chinese mfgr. -
but still have problems with stability.
Has someone experience replacing that transistor.
Is it a good idea to look for the original one (Sanken ) ?
rgds.
fritz
Did you repair the fault that caused it ... if not it will quickly fail again.
I would always fit original parts. Yamaha have some in stock as do many other trusted suppliers.
I would always fit original parts. Yamaha have some in stock as do many other trusted suppliers.
I agree. I remember troubles with 2SA1302 fakes. They had smaller chip, thus they were faster and forced the amp to oscillations. The smaller chip also resulted in lower usable maximum Ic.
The fault was probably overheating - the dead transistor was mounted at the least ventilated position.
Replaced part is still alive... because of matching reasons - I replaced both 2SA1386A of this channel.
Everything else looked/measured ok
I already ordered original parts and will replace again...
Output of original stuffed channel stabilizes immediatly at DC zero after turning on - measured with oscilloscope.
Output of repaired channel seems to oscillate.
The output voltage of both channels is combined and filtered + fed back to PSU unit - which finally shuts down the PSU with an optocoupler.
But it looks like the shutting down is primarily initiated by +/- 15V supply goes down due to overloading of the +/-60 V supply,
Now I ordered 2 complementary pairs 2SA1386A / 2SC3519A...
Does it makes sense to change the complete pairs ?
I think its ok to wait for the original replacement before starting advanced exorcissm.
rgds.
Replaced part is still alive... because of matching reasons - I replaced both 2SA1386A of this channel.
Everything else looked/measured ok
I already ordered original parts and will replace again...
Output of original stuffed channel stabilizes immediatly at DC zero after turning on - measured with oscilloscope.
Output of repaired channel seems to oscillate.
The output voltage of both channels is combined and filtered + fed back to PSU unit - which finally shuts down the PSU with an optocoupler.
But it looks like the shutting down is primarily initiated by +/- 15V supply goes down due to overloading of the +/-60 V supply,
Now I ordered 2 complementary pairs 2SA1386A / 2SC3519A...
Does it makes sense to change the complete pairs ?
I think its ok to wait for the original replacement before starting advanced exorcissm.
rgds.
The mica isolation matts broke on removal - thats why I fitted new silicon matte.
This could influence a bit due to different parasitic capacitance and different thermic resistance.
Do you think its a good idea to replace the mica matte with silicon or should I revert to mica ?
rgds.
This could influence a bit due to different parasitic capacitance and different thermic resistance.
Do you think its a good idea to replace the mica matte with silicon or should I revert to mica ?
rgds.
I replaced with silicon... but there is a lengthy thread regarding that topic on this board anyway ;-)
The post from PMA makes lots of sense for me. If the "fake" part has a smaller chip size and higher bandwidth the frequency compensation might not be adequate... and the beast will oscillate... will see if the original parts solve the problem.
Thanks for your feedback...
The post from PMA makes lots of sense for me. If the "fake" part has a smaller chip size and higher bandwidth the frequency compensation might not be adequate... and the beast will oscillate... will see if the original parts solve the problem.
Thanks for your feedback...
Actually the fake parts have poorer bandwidth. Leading to decreased loop stability margins and massive problems with local oscillations, in output stages tailored to high performance parts.
A link to a service manual, too large for attachment: https://elektrotanya.com/yamaha_emx212s_sm.pdf/download.html#dl
I agree. If the fake part was faster then there would be less phase shift and that wouldn't be a problem.
Faster parts are also more expensive. They put 2N3055-oids in fake flatpack and TO-3 packages because they’re cheap.
ISC’s are actually better than that - the die is full size (but no heat spreader) but it is just an MJ15003/4-ish type which is still slow and causes oscillations if a 20+ MHz part was expected. In more tolerant applications they will actually work fine. Maybe not as good a THD20k figure, but no instant explosion.
ISC’s are actually better than that - the die is full size (but no heat spreader) but it is just an MJ15003/4-ish type which is still slow and causes oscillations if a 20+ MHz part was expected. In more tolerant applications they will actually work fine. Maybe not as good a THD20k figure, but no instant explosion.
Last edited: