I dont think you will hear a difference in sound if offset is 100 mV or 0V, but in generall the AB class push-pull designs doesnt play correctly because of crosstalk.
It's essentially "class A biasing". Some topologies may benefit from it, others won't. But by the time it has any noticeable effect, you're well into dangerous territory. Offset should always be within about +/ 100 mV. Amps with protection circuitry usually turn off at about this point.
The 3020 has much lower DC offset because like in many other amps, gain is reduced to unity at DC. This is fairly standard practice. Otherwise it would be very difficult to get offset down in such a single-ended input topology or when using non-matched FETs in a differential input. The Rotel does not follow this practice (although gain is reduced to ~22 dB), giving an offset adjustment instead.
The 3020 has much lower DC offset because like in many other amps, gain is reduced to unity at DC. This is fairly standard practice. Otherwise it would be very difficult to get offset down in such a single-ended input topology or when using non-matched FETs in a differential input. The Rotel does not follow this practice (although gain is reduced to ~22 dB), giving an offset adjustment instead.
Last edited:
Thanks for the info, did you see this http://www.diyaudio.com/forums/solid-state/206338-rotel-ra-820bx3-not-working-10.html#post2911673 ?
Depends. Can be good, can be crap too. Their 2SC5200/2SA1943s seem to be fine, others may not be.
I'd think one should be able to compensate for any variation in hFE via the offset adjust, though it can't hurt to select them.
Restoring thermal tracking is reasonably important though (it keeps these transistors at equal temperature and thus reduces offset drift). I could imagine several ways of doing this. Taping them together with some thermal compound (Wärmeleitpaste) in between might work, or maybe even using some thermal glue (-kleber).
Installing the resulting 6-legged critter could be a bit tricky, so you may want to remove the original output transistors first and bend the legs on the new ones as necessary.
I hope you do have a decent soldering iron / station, desoldering braid (Entlötlitze), oldschool leaded solder and flux (Flußmittel), along with the usual isopropyl alcohol to clean up the board afterwards. Soldering with flux-core solder or solder + flux is a breeze (comparatively speaking), with no flux at all it's a pain.
Last edited:
Transistor voltages:
Q603 (old/new)
Base 93/113 mv
Emitter 0,68/0,67 volt
Collector -9,5/-9,8 volt
Q605 (old/new)
Base 98/117 mv
Emitter 0,68/0,67 volt
Collector -9,6/-10,5 volt
Q601
Base 31,7 volt
Emitter 32,6 volt
Collector 0,71 volt
Q607
Base 31,8 volt
Emitter 32,3 volt
Collector -0,96 volt
Q609
Base -10,8 volt
Emitter -10,3 volt
Collector -31,5 volt
Q611
Base -10,9 volt
Emitter -10,4 volt
Collector -31,9 volt
Q613
Base 31,9 volt
Emitter 32,4 volt
Collector 31,9 volt
Q615
Base 31,9 volt
Emitter 32,6 volt
Collector 1,75 volt
Q621
Base 31,9 volt
Emitter 32,6 volt
Collector 1,75 volt
Q621
Base -135 mv
Emitter -0,48 volt
Collector 1,83 volt
Q623
Base 1,79 volt
Emitter 1,18 volt
Collector 34 volt
Q625
Base -0,5 volt
Emitter 101 mv
Collector 33,6 volt
They all look good, hmm?
Also its not possible to adjust DC offset anymore, thats funny to me...
Last edited:
1.79V On the base of Q623 is too much. Check R635, R683 (but maybe you did it before)
No! Voltage at Q623 base is too high because of the DC offset. Q625 base voltage is high by the very same amount, that's normal. Their difference should be about 2.3 V. Q627/633 bases should be about 1.1 V apart. (And looking at the values measured, they are indeed.)
Q627/629 bases are one B-E voltage drop up from the output, and Q623 two B-E voltage drops. If the output shows DC offset, it will obviously show up there.
I think R607 and R609 have already be checked, haven't they? Maybe check the bias adjust pot.
Checking C623 for shorts may be worth doing, at least a dead short would be visible.
I still think a semi-dead Q603 is the problem though. Inspecting its solder joints might still be worth a shot.
If you ever feel like modifying the amp after this whole ordeal is over, R601, R603 and R615 (+ Rch counterparts) could be swapped for metal film. May reduce noise a little.
Next time you have a problem like that, you could also come here btw. Should be a lot easier on the Engrish.
Q627/629 bases are one B-E voltage drop up from the output, and Q623 two B-E voltage drops. If the output shows DC offset, it will obviously show up there.
I think R607 and R609 have already be checked, haven't they? Maybe check the bias adjust pot.
Checking C623 for shorts may be worth doing, at least a dead short would be visible.
I still think a semi-dead Q603 is the problem though. Inspecting its solder joints might still be worth a shot.
If you ever feel like modifying the amp after this whole ordeal is over, R601, R603 and R615 (+ Rch counterparts) could be swapped for metal film. May reduce noise a little.
Next time you have a problem like that, you could also come here btw. Should be a lot easier on the Engrish.
Last edited:
R667, R637, Q621 are ok.
Q621 measuring Hfe showed 265, the new ones are only around 160. Can you comment on this?
Q615 was bad! After exchange the DC offset dropped from 200 mv to around -10mv
I still can't adjust the idle current
Now I have to look other channel. It has DC Offset around 600mv. It will get more difficult now, cause there is no schematic with corresponding values to it, shall I start to measure the drivers?
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.