I exchanged the voltage driver chip in the amp, because one channel was dead. Original was a STK3042-II, which is now obsolete. So I took STK3042-III. I planned to change all semiconductors in the dead channel, too but didn't do cause I couldn't remember if it was left or right. Stupid me!
So I soldered the new driver chip in and turned it on. Relay switched on after a few seconds, no smoke, no smell. But damn, 2SA1104 and 2SC2579 got really warm on both channels! I switched it off, let it cool down and attached my voltmeter to measure bias. Manual says it should be 6.5mv, but I had 165mv on the next turn-on and 4mV in speaker-protect-mode before switching of relay! I can regulate between 120 and 190, but I wasn't planning in turning that old amp classA
Driver chip remained cold, at least during the few seconds i had it all turned on.
Why do I have such high bias?? Any ideas? All trannies measure ok, no shorts. I didn't measure the points given in the manual, cause i don't want to grill parts longer than necessary...
matthias
The schematic:
www.listmember.de/diyaudio/yamaha_a400_schematic.jpg
(380k)
So I soldered the new driver chip in and turned it on. Relay switched on after a few seconds, no smoke, no smell. But damn, 2SA1104 and 2SC2579 got really warm on both channels! I switched it off, let it cool down and attached my voltmeter to measure bias. Manual says it should be 6.5mv, but I had 165mv on the next turn-on and 4mV in speaker-protect-mode before switching of relay! I can regulate between 120 and 190, but I wasn't planning in turning that old amp classA
Driver chip remained cold, at least during the few seconds i had it all turned on.
Why do I have such high bias?? Any ideas? All trannies measure ok, no shorts. I didn't measure the points given in the manual, cause i don't want to grill parts longer than necessary...
matthias
The schematic:
www.listmember.de/diyaudio/yamaha_a400_schematic.jpg
(380k)
Update:
Bias is ok, when the machine is cold and relatively stable. Driver chip gets warm. However, when I switch on one speaker, even without one attached to the terminal, the bias rushes away. Doesn't matter if pair A or B. Once rushed away, the bias reacts very nervous on tries to regulate it. It will only work again if the machine cools off and then gets turned on again. I connected a speaker. Volume was zero, but i had a massive 50hz-hum through the speaker
What's wrong? I'm clueless!
matthias
Bias is ok, when the machine is cold and relatively stable. Driver chip gets warm. However, when I switch on one speaker, even without one attached to the terminal, the bias rushes away. Doesn't matter if pair A or B. Once rushed away, the bias reacts very nervous on tries to regulate it. It will only work again if the machine cools off and then gets turned on again. I connected a speaker. Volume was zero, but i had a massive 50hz-hum through the speaker
What's wrong? I'm clueless!
matthias
Hmmm, nervous bias current? No wonder – the trim pot is tied up between base and collector of TR201. Any dirt or damage in the trim pot leaves B-C open and TR201 switched off which means high bias current. The sanest thing to do is to reverse the circuit a bit and to tie a trim pot between base and emitter of TR201. In this case, open circuit in that trim pot will force TR201 to open and to reduce the bias current to minimum. If this is hard for you to do, try to replace the original trim pot with the new one – multi-turn is a must. This way, bias current is not going to be so jumpy. Solve this problem and see how it gets. Anyway, be warned – any game with the bias circuit is quite dangerous – in most cases it leads to smoke…
Regards!
Regards!
Yes, but the offset is only about 20mv.... not way too much. And I checked with my scope... no oscillation except about 12.5mV pp at around 2mhz.... at the limit of my old Hameg 10mhz-Scope.
I disconnected everything after the speaker switches, and bias stayed stable at about 8mV. With terminal plugged, it raised up to 230mV
within a second... without any speaker connected!
I connected a speaker before the terminal. Bias stayed stable and low, the tested channel working...
matthias
I disconnected everything after the speaker switches, and bias stayed stable at about 8mV. With terminal plugged, it raised up to 230mV
within a second... without any speaker connected!I connected a speaker before the terminal. Bias stayed stable and low, the tested channel working...
matthias
The schematic reveals that the speaker switch sub-circuit has several capacitors connected in paralell with the output of the power amplifier. Obviously this translates a small oscillation into a huge one
Check that R257 and L201 are not shorted
Check that C225, C231, C233 are not open
Try to increase R257 to 10 ohms
Try to slightly increase the value of C225, C231, C233
I suspect that the new driver IC has higher gain-bandwidth product than the old one and requires more compensation to achieve stability with this output stage. Note that the new driver IC is not exactly the same model as the old one
I've experienced similar unstability problems from time to time when replacing obsolete or unavailable transistors with others having similar specs but lower capacitances. Adjusting compensation capacitors solved the problem
To check stability you have to look at the amplifier behavior when entering and leaving clipping when driving the rated load at high frequencies. It the circuit is stable when clipping a 10Khz sine wave then it's not likely to oscillate in any other circumstances. Sometimes the oscillation has less than 100mV amplitude so check carefully
Check that R257 and L201 are not shorted
Check that C225, C231, C233 are not open
Try to increase R257 to 10 ohms
Try to slightly increase the value of C225, C231, C233
I suspect that the new driver IC has higher gain-bandwidth product than the old one and requires more compensation to achieve stability with this output stage. Note that the new driver IC is not exactly the same model as the old one
I've experienced similar unstability problems from time to time when replacing obsolete or unavailable transistors with others having similar specs but lower capacitances. Adjusting compensation capacitors solved the problem
To check stability you have to look at the amplifier behavior when entering and leaving clipping when driving the rated load at high frequencies. It the circuit is stable when clipping a 10Khz sine wave then it's not likely to oscillate in any other circumstances. Sometimes the oscillation has less than 100mV amplitude so check carefully
Matthias,
I have had the same issue in the past with one amp I was repairing then let it go because of oscillations. (Still lying somewhere). It was a Sony ESD-110 (110wpc).
It had a blown STK-3042 and I replaced it with the Series III.
I see the amp breaking out into oscillation, I have played around with the compensation caps but no avail.... I do not have the schematics.....
Not sure if it helped, but now you know the problem was caused due to the new driver chip....
I have had the same issue in the past with one amp I was repairing then let it go because of oscillations. (Still lying somewhere). It was a Sony ESD-110 (110wpc).
It had a blown STK-3042 and I replaced it with the Series III.
I see the amp breaking out into oscillation, I have played around with the compensation caps but no avail.... I do not have the schematics.....
Not sure if it helped, but now you know the problem was caused due to the new driver chip....
Hey all, thanks for your hints pointing to oscillation effects. It's my first repair in the signal path of a power amp stage and I would not have thought this small amplitude i have on the outputs might be of any concern.
I will replace the pots with some multiturn types and modify the values of the compensation tomorrow.
@ k-amps: Thanks for this helpfull info. Strange as the specs i have on simple one-sided datasheets are more or less the same. Only recommended operational voltage was derated from 43volts to 36 volts (41volts in the yamaha, Vmax on both revisions 55volts)
@ working at home: Let me know if you can really get hands on a seriesII chip! I might need it in a week or two
14dollars is a fair price. I payed 8.50euros for 3042-III
matthias
I will replace the pots with some multiturn types and modify the values of the compensation tomorrow.
@ k-amps: Thanks for this helpfull info. Strange as the specs i have on simple one-sided datasheets are more or less the same. Only recommended operational voltage was derated from 43volts to 36 volts (41volts in the yamaha, Vmax on both revisions 55volts)
@ working at home: Let me know if you can really get hands on a seriesII chip! I might need it in a week or two
14dollars is a fair price. I payed 8.50euros for 3042-III
matthias
Ok, as it has run 24hours with stable bias of recommended 6.5mV in non-loaded condition, doesn't rush away with 0db 15kHz sine & white noise and now plays music for about 2 hours I assume the Yamaha A400 is ready to be shipped to my friend again.
Thanks for all the help. I wouldn't have done without each of you!!!
Now here's what i had to change in order to get the stk3042-III run instead of -II :
* Changed VR207/208 to 2k 64w multiturns (on adapter pcb)
* R257/258 4.7ohms -> 10ohms -> 10ohms
* C225/226 120p -> 180p -> 330p
* C227/228 5p -> 5p -> 13.2p
* C231/232 82p -> 120p -> 270p
* C233/234 82p -> 120p -> 270p
* decoupling of the STK: each 10n between V+/V- and ground and 10n between V+ and V-
First change meant bias would rush up to 90mV instead formerly 200mV. Second change is now stable, parts in addition soldered on coppered side of board.
Thanks again.
Bye, Matthias
Thanks for all the help. I wouldn't have done without each of you!!!
Now here's what i had to change in order to get the stk3042-III run instead of -II :
* Changed VR207/208 to 2k 64w multiturns (on adapter pcb)
* R257/258 4.7ohms -> 10ohms -> 10ohms
* C225/226 120p -> 180p -> 330p
* C227/228 5p -> 5p -> 13.2p
* C231/232 82p -> 120p -> 270p
* C233/234 82p -> 120p -> 270p
* decoupling of the STK: each 10n between V+/V- and ground and 10n between V+ and V-
First change meant bias would rush up to 90mV instead formerly 200mV. Second change is now stable, parts in addition soldered on coppered side of board.
Thanks again.
Bye, Matthias
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.