My brother's SA-220 started smoking the other day and one channel quit working. I am taking a look at it for him, but don't have schematics for this.
U3 (AD42020-1) and U4 (AD736JN) both split in half. The portion of the board above these has smoke damage from this. VR1 and VR2 trimpots are both charred and the board below them is also damaged, though I haven't taken it apart to inspect how much damage there is.
I figured before I did much on this, that I would try to determine whether the output FETs are still good or not. I don't measure any shorts, but they also do not ohm out the same as the good channel.
Any suggestions? My brother would like this fixed if possible, but not looking real good right now.
If anyone has schematics, I would appreciate a copy also. I do have the SA100 schematic that was floating around, but this problem at least so far seems to have started in the thermal tracking portion of the board which isn't on there. Thanks for any help, and I can post pics if they would help.
U3 (AD42020-1) and U4 (AD736JN) both split in half. The portion of the board above these has smoke damage from this. VR1 and VR2 trimpots are both charred and the board below them is also damaged, though I haven't taken it apart to inspect how much damage there is.
I figured before I did much on this, that I would try to determine whether the output FETs are still good or not. I don't measure any shorts, but they also do not ohm out the same as the good channel.
Any suggestions? My brother would like this fixed if possible, but not looking real good right now.
If anyone has schematics, I would appreciate a copy also. I do have the SA100 schematic that was floating around, but this problem at least so far seems to have started in the thermal tracking portion of the board which isn't on there. Thanks for any help, and I can post pics if they would help.
Bump.
Chris (anatek)? You still around?
Any hints on determining if the FETs are still good? I'll have to find the datasheets for these to determine the pinout. If the FETs are good, I would also guess I will need a new PCB, so if anyone has a line on one shoot me a PM or email. I can't find a datasheet (or anything) on Analog Devices site for the AD42020-1 also. I saw someone had a couple for sale here about a year ago or so. Not sure if you can just disable the thermal tracking portion of the board or not. Schematics would again help. Thanks for any help.
Chris (anatek)? You still around?
Any hints on determining if the FETs are still good? I'll have to find the datasheets for these to determine the pinout. If the FETs are good, I would also guess I will need a new PCB, so if anyone has a line on one shoot me a PM or email. I can't find a datasheet (or anything) on Analog Devices site for the AD42020-1 also. I saw someone had a couple for sale here about a year ago or so. Not sure if you can just disable the thermal tracking portion of the board or not. Schematics would again help. Thanks for any help.
Hi there!
The IC's that you said are blown are true RMS-DC converters so are probably used in somekind of protection circuit along with a relay or maybe a VU meter (if there is one on the front panel).
Checking the output FET's would be a good idea (a quick google provides plenty of guides) because if the protection circuit or VU meter IC's have blown up as you described, then it must have been a pretty catastrophic failure.
It may also help if your brother could descibe the kind of failure that he experienced. When it started smoking, did the speaker cone move in any strange ways, did anything actually explode or did it just smoke? This might help us determine the problem further.
Others seem to point towards the AltaVista audio site to find schematics but at $75 a pop I wouldn't bother at all.
Hopefully some more people will see this thread and provide some more answers.
The IC's that you said are blown are true RMS-DC converters so are probably used in somekind of protection circuit along with a relay or maybe a VU meter (if there is one on the front panel).
Checking the output FET's would be a good idea (a quick google provides plenty of guides) because if the protection circuit or VU meter IC's have blown up as you described, then it must have been a pretty catastrophic failure.
It may also help if your brother could descibe the kind of failure that he experienced. When it started smoking, did the speaker cone move in any strange ways, did anything actually explode or did it just smoke? This might help us determine the problem further.
Others seem to point towards the AltaVista audio site to find schematics but at $75 a pop I wouldn't bother at all.
Hopefully some more people will see this thread and provide some more answers.
His wife was home at the time, and was going to listen to something so turned the amp on. He usually leaves it on all the time (I've warned him about these before) but noticed a funny smell the night before so had turned it off. Apparently it just started smoking; she thought it was the preamp, so I'm not sure if she turned the amp off or not.
It looks like I will probably have to pull the output devices to check them; it they are toast, most likely game over on this amp. Ohming them out in circuit isn't telling me much other than they aren't shorted. I'm not sure I'm up to designing a replacement output section using the Exutech (???) lateral FETs or something else for this. AltaVista audio is up to $400 or so for replacement pulls, and it would likely need an entire output PCB as well. The $2k or whatever he is charging for the "Basic" upgrade isn't an option either.
It looks like I will probably have to pull the output devices to check them; it they are toast, most likely game over on this amp. Ohming them out in circuit isn't telling me much other than they aren't shorted. I'm not sure I'm up to designing a replacement output section using the Exutech (???) lateral FETs or something else for this. AltaVista audio is up to $400 or so for replacement pulls, and it would likely need an entire output PCB as well. The $2k or whatever he is charging for the "Basic" upgrade isn't an option either.
Strange, the protection must have kicked in properly (because there were no reports of abnormal speaker movement) but then because of the extent of the failure, the actual protection IC must have blown too. This will probably have blown the fuse and turned the affected section off.
For the protection IC to have blown like this, the FET's were most likely shorted and supplying the power rails to the protection circuit. Replacing the FET's would be a pig of a job too, firstly because the originals are so hard to find and also the fact that they have to be closely matched.
Short from sending it off to AltaVista and paying some serious money for them to repair it, or to buy one of their diagrams (still mega $$$), I don't think there is much else that can be done apart from maybe trying to contact someone on the forum who knows more about this gear.
For the protection IC to have blown like this, the FET's were most likely shorted and supplying the power rails to the protection circuit. Replacing the FET's would be a pig of a job too, firstly because the originals are so hard to find and also the fact that they have to be closely matched.
Short from sending it off to AltaVista and paying some serious money for them to repair it, or to buy one of their diagrams (still mega $$$), I don't think there is much else that can be done apart from maybe trying to contact someone on the forum who knows more about this gear.
Thanks. From reading previous threads, the protection circuitry has been described as an analog computer which attempts to sense the heatsink temperature. There are protection relays (input signal perhaps?) as well as rail fuses on each of the two channel output boards. I'm still not convinced the output FETs are blown, as I would expect to see some shorts. If they are, then FET replacement with the same type would be out of the question. The speakers were Apogees, so hopefully that side was not damaged. I think my brother has spare ribbons however.
If the amp had the AD736 IC inside though, surely it must have some kind of output protection from DC as well. A true RMS-DC converter probably wouldn't have a place in just sensing heatsink temperature and would instead be used to monitor the output instead (oscillation, DC maybe?). The analog computer that you talk about may be a simple microcontroller that reads the DC output of the converter and acts on this as an input. Depending on their size, the protection relays could be for the inputs (startup/mute??) if they were small but if they are bigger, higher amperage and voltage units, then they are most likely for the outputs (which would make more sense). The protection circuit may also incorporate some kind of heatsink temperature sensing to turn the amp off if it overheats too.
As there is no schematic freely available for the amp, I am making quite a few assumptions here. It may help to get the problem solved though.
As there is no schematic freely available for the amp, I am making quite a few assumptions here. It may help to get the problem solved though.
I'm hoping that Chris (anatek) weighs in here as he used to own/run a Counterpoint service center. I pm'd him, but know he has had health issues the last couple of years or so.
The relays are small signal type, so input would be my guess. The SA100 schematic (not official) floating around also shows input relays (from the tube section to the SS output stage).
The relays are small signal type, so input would be my guess. The SA100 schematic (not official) floating around also shows input relays (from the tube section to the SS output stage).
Board and Mosfet
Pars,
I have a blown SA20 with one bad channel. I planning to try this mod/repair from Denis Vilfort: http://www.diyaudio.com/forums/solid-state/31131-hafler-dh-200-220-mods-92.html
I'll use his output stage with an Aikido configure with a White Cathode Follower.
You can have my SA20 Board and what I believe is the good channel's mosfet.
I'm not sure what is the difference in the circuit board between the SA220 and the SA20.
Let me know if you are interested
QuinnLing
Pars,
I have a blown SA20 with one bad channel. I planning to try this mod/repair from Denis Vilfort: http://www.diyaudio.com/forums/solid-state/31131-hafler-dh-200-220-mods-92.html
I'll use his output stage with an Aikido configure with a White Cathode Follower.
You can have my SA20 Board and what I believe is the good channel's mosfet.
I'm not sure what is the difference in the circuit board between the SA220 and the SA20.
Let me know if you are interested
QuinnLing
The main board from the SA20 is the same as the main board from the SA220. The modifier to this statement is that the main board was in a continuous state of 'development' so while the last production SA20 main board is identical to the first SA220 main board, this is not the case for an early SA20 main board.
Unless your SA220 main board is damaged, then the offer of an SA20 main board is of no use to you. The offer of a set of output MOSFETs is however worth considering if they are a "working pull". The eight MOSFETS are closely matched within the n- and -p channel sets and then the two sets are 'reasonably' well matched from set to set. Although the latter is not so critical as it just affects the limits of the DC offset adjustment.
I'll respond in more detail to our earlier questions a bit later - the weekend family taxi service is about to start.
I assume he was offering the good channel's output board and MOSFETs? AFAIK, the main board in the SA220 is fine, just the left channel went. The bad channel board started charring around VR1/VR2 trimpots on the top edge. The two 8-pin DIPs I mentioned were both split in half. I unbolted the heatsink and lowered the board as far as it would go, but I have not disconnected any of the wiring to the main board yet to get a better look at it.
The analogue computer converts an approximate measurement of the drain-to-source voltage of the output stage and the current flowing in the speaker return lead into the RMS power dissipated in the output stage MOSFETs. This is taken to a trigger circuit that has selectable thresholds. If the threshold is exceeded then transistor clamps short out the MOSFET gate drive to ground. At the same time the normal mute circuit (located on the main board) is tripped, which causes a relay to also short the MOSFET gates to ground. The mute circuit remains active for around 60-90s after the fault is cleared which make it obvious to the use that something untoward has occurred.
The reason for the doubling up of MOSFET gate drive clamps is that the MOSFETs can die in nanoseconds and a relay takes many milliseconds to operate.
There is no output DC or temperature sensing involved in the analogue computer circuit at all.
Temperate sensing is done using two thermal switches that sense the average heat sink temperature and cause the normal mute circuit to trip if the 'sink temp exceeds some threshold - can't remember what that was, maybe 70 degrees Celsius, which is typical for such circuits.
This is complex circuit. If you want to service it you will need a steady hand, DVM and preferably a signal source and oscilloscope. If you have not serviced anything before then this is not the amp to start on...
The reason for the doubling up of MOSFET gate drive clamps is that the MOSFETs can die in nanoseconds and a relay takes many milliseconds to operate.
There is no output DC or temperature sensing involved in the analogue computer circuit at all.
Temperate sensing is done using two thermal switches that sense the average heat sink temperature and cause the normal mute circuit to trip if the 'sink temp exceeds some threshold - can't remember what that was, maybe 70 degrees Celsius, which is typical for such circuits.
This is complex circuit. If you want to service it you will need a steady hand, DVM and preferably a signal source and oscilloscope. If you have not serviced anything before then this is not the amp to start on...
OK, that sounds like either (or both) of the +/- 15VDC supply rails on the APC board have failed and you have +/-70 VDC supplying all the protection circuitry. If so, and you want the APC repaired, then you will be replacing every IC on the board AFTER you have repaired the +/- 15VDC supplies.
The good news is that this damage has probably not affected those pesky matched MOSFETs.
So what WERE your plans for the weekend?
So the design drops 55V across an LM317/337? Good idea alm: I haven't pulled the board down far enough to read the regulators, but that is what they appear to be. Rails are at +/-70V and not +/-53V as in the sa100 schematic floating around? I don't want to fire this thing up until I am sure of everything and what I am doing. I do know from replacing the cap on the 555 timer circuit that it is a PITA to work on with the chassis design
I might be able to wire the pots in assuming some traces are burnt up, but I'll have to see what state the bottom side of the PCB is in. The amp was very dusty inside, might have contributed to this.
Thanks for your input, most appreciated.
alm: I haven't pulled the board down far enough to read the regulators, but that is what they appear to be. Rails are at +/-70V and not +/-53V as in the sa100 schematic floating around? I don't want to fire this thing up until I am sure of everything and what I am doing. I do know from replacing the cap on the 555 timer circuit that it is a PITA to work on with the chassis design I might be able to wire the pots in assuming some traces are burnt up, but I'll have to see what state the bottom side of the PCB is in. The amp was very dusty inside, might have contributed to this.
Thanks for your input, most appreciated.
Most often, when an output stage goes on these amps, the gate protection zeners and/or the gate stopper resistors turn crispy. If those parts look like they have not sustain high temperature damage, there is a good chance your output stage is OK. In that case, you could disconnect the gate leads and ground them out to protect them from ESD while you work on that amp.
As far as devices? Exicon ECF10N16/ECF20N16 with changes to the bias supply to permit adjustment to new threshhold voltages. The other device looks like the dual-die
BUZ901/905P pair from Semelab (Newark sells them).
-- Jim
As far as devices? Exicon ECF10N16/ECF20N16 with changes to the bias supply to permit adjustment to new threshhold voltages. The other device looks like the dual-die
BUZ901/905P pair from Semelab (Newark sells them).
-- Jim
So the design drops 55V across an LM317/337? Good idea
The regulators are 7815 and 7915, but the series pass transistors are MJE340/MJE350 which at 300V are quite capable of dropping 55V each. In normal operation, the three terminal regulators have no more than ~ 20V input-to-output voltage differential which is quite OK.
The nominal rails are +/-70V, but that depends on that does depend on the transformer taps. The SA100 valve amp section is similar to the SA220.
The originals are VMOS devices and substitutes are still available although the matching requirements means that you need to purchase a bunch and do matching. Not for the faint hearted.
The Exicon ECF10N16/ECF20N16 are both lateral MOSFETS. The 160V Vds rating is marginal for these amps.
The BUZ901D and BUZ905D (the P suffix is a TO247 package and therefore unsuitable) is a 200V contender. But again a lateral device.
The difficulty with using lateral MOSFETS is that:
1) The matching requirement is not eliminated
2) The output impedance rises significantly compared to the VMOS devices. If you went down this route then you would need to change both channels...
3) The pin out is different so the PCB gets hacked in major way
Are the gate stopper resistors the ones that span the output devices? They look like Rodersteins encased in vinyl tubing or something? They measure OK (the same as the good channel) and are not crispied. There are two devices (?) encased in rubber or something in the center of the board. Markings on the PCB are D something. Are these the zeners? The diodes around the outside of the board measure OK and do not look crispy. Here are a couple of pics:
