Page 1 of 2 12 LastLast
Results 1 to 35 of 68

Thread: Fault Tracing a Carver M-500t

  1. #1
    Better Tone thru Mathematics bob p's Avatar
    Join Date
    May 2006
    Location
    Chicago, USA
    Posts
    2,988

    Fault Tracing a Carver M-500t

    I've got another dead Carver M-500t on my hands, and this one is a little harder than the last one. Its been "serviced" by someone before, and God only knows if they repaired it or not. I suspect there's a strong possibility that its been half-repaired, so that some of the telltale evidence of what's gone wrong is now missing. I think that the amp is now FUBAR'd, which has made diagnosing the problem rather difficult. Please read on if you think you can help.

    First the easy part:


    This amp seems to have a normal PSU. It powers on and idles without popping fuses. All of the voltage rails have appropriate voltages on them. I've used the PSU trimpot to calibrate the PSU output voltage to 74 VDC on the amp board's test point (R179), as listed on the amp's calibration sheet. I've also biased both channels to idle at 6 mA across each channel's TPs at the OT ballast resistors. When the amp powers on, I get the relay clicking on after about 3 seconds as you'd expect. But here's a funny observation: the relay isolates the PT from the speaker outputs when the amp is off (Z measures in megohms), but instead of having the Z between the OT and the speaker posts drop to zero, it drops to about 500 Kohm on each channel when the relay is activated. OK, let's put the relay on our list of things to be worried about.

    Examination of the heatsinks shows that all of the OTs and drivers appear to be un-shorted. That is to say, they all pass my diode chirp tests. Its hard to tell if the OTs have been changed or not. The solder joints may seem a little shiny and new, but they work looks old enough to have been an old repair if the OTs were ever replaced. All of the OTs are OEM Toshibas, and they all have matching lot numbers. If they were replaced, they were all replaced. So far so good.

    Examination of the amp board shows some obvious problems. There are three blatantly obvious areas on the board where the board has been discolored from green to brown due to excessive heat. There are two more subtle areas of minor discoloration which are harder to appreciate. None of the traces are actually burned -- only the green color of the board has been turned to dark brown/black. Here's how the discolored areas look:

    First the 3 blatantly obvious areas:


    1. Scorched earth around R182 and R183. These are 6k8, 2W metal oxide power resistors that attach to Pins 4 and 8 of the TL072 input OpAmp, IC101. Pin 4 is the Vcc- pin, and Pin 8 is the Vcc+ pin. It looks like R182 and R183 had LOTS of current passing through them, as the board is blackened at both ends where the leads mount to the board, though the solder joints remain intact. R182 is mounted close to C103 and scorched its outside. C103 connects to one of the opamp inputs. R183 is mounted close to R179 and looks like its charred the spaghetti insulation off of one of the R179 leads. In this section of the board, I'm assuming that R182, R183, C103 and R179 are all bad from thermal damage. the TL072 is also bad, as are other parts yet to be determined.

    2. Scorched earth around R212 and Q207. R212 is a 270R 1/2W resistor between ground and Q207E. Its burned so badly that its body is fractured. There is a small scorched area around the solder connection between R212 and Q207E that's about the size of a dime. No other parts appear to be physically damaged, though the following parts are physically close enough to be on the suspect list: R211, R218, D201, D202, D203. Tracing the circuit path, it looks like the fault current probably came toward Q207 and went through R212 to ground. Where did it come from? Maybe D203 and the relay? Things are starting to add up.

    3. Scorched earth around IC601 and R601, R602. IC601 is the IC that monitors the amp's outputs to control the power meters. There is a huge darkened area around IC601 and the two 2W metal oxide power resistors: R601 1k8 2W and R602 3k8 2W. The scorched area on the board extends from IC601 to R601/R602, to R612/R613, and to the two meter trimpots. ITs a big scorched area, maybe 2-inches in diameter.

    There are also two areas near the inputs that are subtly discolored, though not so seriously scorched: Q105/Q107 on one side, Q106/Q108 on the other.

    Now that I've got the physical observations out of the way, I'll put the preliminary test results in the next post.

    Edit: Almost forgot: Hacked up ballast resistors: The OT ballast resistors have additional resistors soldered atop their terminals, as if they had all faulted open and someone was trying to jump replacement resistors atop their carcasses. The amp has a total of eight 0.47R, 2W resistors mounted atop the bodies of the ballasts.

    more to follow...
    Last edited by bob p; 10-20-2007 at 08:21 AM.
    "Stand back, I'm holding a calculator." - chinrest

  2. #2
    Better Tone thru Mathematics bob p's Avatar
    Join Date
    May 2006
    Location
    Chicago, USA
    Posts
    2,988
    Some measurements:
    Code:
    Trans    E       C       B
    Q101     1.7    72.4   2.2
    Q102     1.7    72.6   2.2
    Q103    -1.7   -72.7  -2.2
    Q104    -1.7   -73    -2.2
    Q105     73.0   1.7   72.5
    Q106     72.7   1.7   73.2
    Q107    -73.4  -1.7  -72.9
    Q108    -73.6  -1.7  -73.1
    Code:
    Trans    E       C       B
    Q201    43.1   -0.59     43.1
    Q202    .007    12.1     .014
    Q203    12     .007       6.6
    Q204    .014    6.5      .007
    Q205    11.5    43.2     12.1
    Q206    0       12.1     -0.584
    Q207    10.2    18.9     10.8
    Code:
    Q301   .0005   +74.1   .0007
    Q302   .0007   +74.1   .0008
    Q303   .0007   -78.7   .0005
    Q304   .0007   -78.4   0
    "Stand back, I'm holding a calculator." - chinrest

  3. #3
    Better Tone thru Mathematics bob p's Avatar
    Join Date
    May 2006
    Location
    Chicago, USA
    Posts
    2,988
    Code:
    D201   .597 / .495
    D202   .597 / .480
    D203   .184 / .184 *
    R218  10K  (way off)
    Last edited by bob p; 10-20-2007 at 08:26 AM.
    "Stand back, I'm holding a calculator." - chinrest

  4. #4
    Better Tone thru Mathematics bob p's Avatar
    Join Date
    May 2006
    Location
    Chicago, USA
    Posts
    2,988
    Oh yes, the relay housing seems a bit yellower than it should be.

    I can see how a fault current that might have originated in the power meter circuit might have passed backwards through the relay's snubber diode to the protection circuit Q207 to ground. That is to say, at least it looks like there's a path through there somewhere. But Q207 doesn't test with an obvious E-C short.

    I'm still at a loss to explain how these two areas are related to the problem with the TL072 op amp at the input, or to the darkened area around the pre-drivers Q105-Q108. There are some very funny measurements on those 100 series transistors.

    I'm also having trouble explaining how all of this comes together, and then there are those odd resistors soldered atop the ballasts.

    ideas? i'm wondering if this might have been a lightning strike.

    thanks!
    Last edited by bob p; 10-20-2007 at 08:48 AM.
    "Stand back, I'm holding a calculator." - chinrest

  5. #5
    Senior Member Enzo's Avatar
    Join Date
    May 2006
    Location
    Lansing, Michigan, USA
    Posts
    27,077
    Easy part. Either the relay contacts close or they don't. MAke sure you didn't measure resistance from output A to speaker terminal B and vice versa. Also, Measure continuity between the speaker terminals and the relay output side, and also measure resistance from the amplifier circuits - the ballasts perhaps - to the hot side of the relay. Lack of continutiy from amp to terminals is not necessarily the relay. Of course it could be the relay, and if it can be opened, burnish its contacts and verify freedom of motion inside.

    The two 6k8 resistors are dropping resitors and I expect them to get hot. They drop the 74v rails down to the 15v zeners. I would not be surprised to see browning of the board normally.

    What would cause them to OVERheat? SHorted zener or cap, but not likely both zeners are shorted, besides, you can measure them. If the IC burns up, it can short one rail to the other. Neither would short to ground - after all, the chip has no ground conection - but they can and do short together. And that would max the current. Pop the chip out and the rails would come back, assuming the 6k8s are OK .

    What does the TL072 measure between pins 4 and 8? I mentioned elsewhere that the chip was suspect.

    R212. Only thing can burn that up is excess current through Q207. Is 207 short? I'd replace it anyway, if enough juice flowed through it to burn the R. I think we covered how this whole circuit works last time, so refer to that of needed. Why failure? Remember, we don't know what was done to this amp, it may not be in as-failed condition. In any case, the current path is through the relay from the 40v rail, through Q207, and through R212. SInce the relay is designed for continuous on, that means it is OK for Q207 to stay on. Maybe not for the speakers, but Q207 has no trouble holding the relay on all day. So even if Q207 shorts, it is just the same as being turned on as far as the relay and resistor are concerned. SO even if it is bad, Q207 did no harm the resistor. That leaves the relay and the diode across it. Either of those appear to have been soldered recently? A shorted diode D203 means Q207 connects R212 across the 40v rail, right? And that would be 6 watts to dissipate.

    Failure current would not have come through Q207 base, i would have had to come through R208, and that is 1k.

    There is another possibility I discuss below.

    3. IC601 is the meter driver. Two channels. Without looking it up, I would not be surprised if it is a headphone driver IC in real life. But it coulod be specifically made for VU drive. In any case R601,602 are current limit resistors - or voltage dropping if you want to think of it that way - from the +/-40v rails to power the IC. Both are 1k8. The 3k3 is R612. R612 is the dropping resistor that continues on to provide power to U301 with ZD602. There is a companion R613 - it is over further left - for the other U301 supply zener.

    Looks to me like U601 shorted, and just like U101 could have shorted the two rails together. I kinda doubt the two meters themselves were involved, and there is no voltage coming out to the speaker terminals that the circuit was not designed to handle, after all, the peak voltage is the same as the rails. Look for U601 rails shorted together.

    Here is a theory. SO far there are two failures that are independently related to 40v rails. WHat if somehow the 74v rail was shorted to the 40v rail? Now the various parts we have been discussing have an extra 34v to dissipate. How would that happen? SHorted isolation diodes D401-404. ANy one of those shorts, and whenever the 74v rail is commutated on, it sends 74v back into the 40v rail. Just a thought. And if I read it right, you'd never know until you cranked the amp high enough to invoke the higher rails.

    That is some serious sleuthing if that turns out to be the case.

    Funny ballests? Clear it out, and make it stock.


    D201 .597 / .495
    D202 .597 / .480
    D203 .184 / .184 *
    R218 10K (way off)
    What are those, readings each direction?
    D201 is reverse parallel Q205 BE, so naturally we get a junction drop each way. Same thig D202 and Q206. D203 has the relay coil across it with its low resistance. That masks any junction drop readings. It isn't shorted, that is what counts.


    The burnt up U101 may not be related to the burn up 200 and 600 circuits. COuld be but don't assume it is.

    Q301 .0005 +74.1 .0007
    Q302 .0007 +74.1 .0008
    Q303 .0007 -78.7 .0005
    Q304 .0007 -78.4 0
    Looks OK. These run the commutators. The collectors are refernced to the 74v rails and will ne th esource fo curren twhen they turn on. They only turn on when U301 tells them to in response to signal peaks.

    Trans E C B
    Q201 43.1 -0.59 43.1
    Q202 .007 12.1 .014
    Q203 12 .007 6.6
    Q204 .014 6.5 .007
    Q205 11.5 43.2 12.1
    Q206 0 12.1 -0.584
    Q207 10.2 18.9 10.8
    We went through expectations here last time, refer to it. But if the relay clicks on after power up, then these things would seem to be working.

    And all those Q1xx readings? Let's wait and see what the amp is doing. Both channels are reading the same and symmetrical as well. Fix the burnt stuff, fix U101. If the amp sits there politely at idle with no DC output skew, I'd say we are making progress.

  6. #6
    Better Tone thru Mathematics bob p's Avatar
    Join Date
    May 2006
    Location
    Chicago, USA
    Posts
    2,988
    Quote Originally Posted by Enzo View Post
    Easy part. Either the relay contacts close or they don't. MAke sure you didn't measure resistance from output A to speaker terminal B and vice versa. Also, Measure continuity between the speaker terminals and the relay output side, and also measure resistance from the amplifier circuits - the ballasts perhaps - to the hot side of the relay. Lack of continutiy from amp to terminals is not necessarily the relay. Of course it could be the relay, and if it can be opened, burnish its contacts and verify freedom of motion inside.
    I measured from the relay's inputs to the output solder points on the relay where the red and blue channel output wires take off to go to the speaker terminals. I'm fairly certain that I measured across the relay and nothing else. I'll double check that. Insofar as one of the channels is inverted WRT the other, maybe I need to check the blue wire on one channel and the black one on the other channel instead of the red wire. (or vice versa). Its very tough to get the relay package to open up. I stop out of fear as I pull on the relay's case when I hear the PCB flexing and creaking.

    The two 6k8 resistors are dropping resitors and I expect them to get hot. They drop the 74v rails down to the 15v zeners. I would not be surprised to see browning of the board normally.
    FWIW, the other M-500t that I am using for comparision has no discoloration whatsoever around these two resistors. It also does not have the plastic sheath covering C103 charred away, and it does not have the spaghetti that covers the leads of R179 charred away. To me, it looks like one amp has a blackened PCB and several burned parts, while one amp has no signs of excess heat whatsoever. My take on this is that the discoloration on the PCB and the charred parts aren't normal wear and tear, and they're indicative of a more serious current/heat problem. My suspicion is that these signs of heat are all attributable to the failure mode, and not to normal operation.

    What would cause them to OVERheat? SHorted zener or cap, but not likely both zeners are shorted, besides, you can measure them.
    Which zeners and caps are you referring to? And as a silly question -- how do you test a zener, compared to a regular diode? I don't think that my diode tester is going to take the zener diode to avalanche, so are we just looking for voltage drops and the absence of shorts and open circuits?

    If the IC burns up, it can short one rail to the other. Neither would short to ground - after all, the chip has no ground conection - but they can and do short together. And that would max the current. Pop the chip out and the rails would come back, assuming the 6k8s are OK .

    What does the TL072 measure between pins 4 and 8? I mentioned elsewhere that the chip was suspect.
    measured Z across pins 4-8 starts off at several kohms and keeps on climbing with the passage of time. i stopped watching after it climbed to 15k.

    R212. Only thing can burn that up is excess current through Q207. Is 207 short? I'd replace it anyway, if enough juice flowed through it to burn the R. I think we covered how this whole circuit works last time, so refer to that of needed.
    no short in Q207. is it safe to replace Q207 and R212 now (before the meter and relay stuff gets fixed), or are we just going to submit the new Q207 and R212 to being blown-up?
    "Stand back, I'm holding a calculator." - chinrest

  7. #7
    Better Tone thru Mathematics bob p's Avatar
    Join Date
    May 2006
    Location
    Chicago, USA
    Posts
    2,988
    Quote Originally Posted by Enzo View Post
    Why failure? Remember, we don't know what was done to this amp, it may not be in as-failed condition. In any case, the current path is through the relay from the 40v rail, through Q207, and through R212.
    i'm not following the failure path to the 40v rail. tracing backwards from ground, we have R212, Q207, and RL101+D203. Then I'm assuming that the power passed through those two thermal switches at the top of the diagram (points 5, 1, 2 and 38 on the schematic), passed by R613 and Q205, to a point labeled "39" on the schematic, to point 27 which meets the 40V rail. Is that the path you're thinking of?

    If that's the case I'm wondering about if I have to worry about other parts like R613 and Q205 (which isn't labeled as Q205).
    "Stand back, I'm holding a calculator." - chinrest

  8. #8
    Better Tone thru Mathematics bob p's Avatar
    Join Date
    May 2006
    Location
    Chicago, USA
    Posts
    2,988
    sorry about the multiple posts. my browser is acting up and i keep loosing text, so i have to commit the text to a post before i lose it.

    Quote Originally Posted by Enzo View Post
    Why failure? Remember, we don't know what was done to this amp, it may not be in as-failed condition. In any case, the current path is through the relay from the 40v rail, through Q207, and through R212. SInce the relay is designed for continuous on, that means it is OK for Q207 to stay on. Maybe not for the speakers, but Q207 has no trouble holding the relay on all day. So even if Q207 shorts, it is just the same as being turned on as far as the relay and resistor are concerned. SO even if it is bad, Q207 did no harm the resistor. That leaves the relay and the diode across it. Either of those appear to have been soldered recently? A shorted diode D203 means Q207 connects R212 across the 40v rail, right? And that would be 6 watts to dissipate.
    D201 and D202 look like they have original solder joints with red marker dots on them. D203 (relay diode) has no dots, but could be original. D202 and D203 look like 4007. Don't remember if that is stock or not. since the relay appears to be functioning, would you bother lifting D203 to test it?

    more to follow...
    Last edited by bob p; 10-21-2007 at 01:24 AM.
    "Stand back, I'm holding a calculator." - chinrest

  9. #9
    Better Tone thru Mathematics bob p's Avatar
    Join Date
    May 2006
    Location
    Chicago, USA
    Posts
    2,988
    Quote Originally Posted by Enzo View Post
    There is another possibility I discuss below.
    ...
    Looks to me like U601 shorted, and just like U101 could have shorted the two rails together. I kinda doubt the two meters themselves were involved, and there is no voltage coming out to the speaker terminals that the circuit was not designed to handle, after all, the peak voltage is the same as the rails. Look for U601 rails shorted together.
    I looked it up. The meter driver is a TA7318P, which is a dedicated VU meter driver, not an adapted headphone driver. I've attached the data sheet below.

    for use with a bipolar power supply, the voltage rails are at pins 5 and 9. measuring the Z on them, it starts at about 1k, jumps to 10k and keeps climbing, like i'm charging a cap somewhere.

    data sheet below:
    Attached Files Attached Files
    "Stand back, I'm holding a calculator." - chinrest

  10. #10
    Better Tone thru Mathematics bob p's Avatar
    Join Date
    May 2006
    Location
    Chicago, USA
    Posts
    2,988
    Quote Originally Posted by Enzo View Post
    Here is a theory. SO far there are two failures that are independently related to 40v rails. WHat if somehow the 74v rail was shorted to the 40v rail? Now the various parts we have been discussing have an extra 34v to dissipate. How would that happen? SHorted isolation diodes D401-404. ANy one of those shorts, and whenever the 74v rail is commutated on, it sends 74v back into the 40v rail. Just a thought. And if I read it right, you'd never know until you cranked the amp high enough to invoke the higher rails.
    ...
    What are those, readings each direction?
    i don't remember if i mentioned this before, but the voltages on all of the PSU caps are what they're supposed to be. that is to say, the caps that supply the different rails were't at the same potentials. i think that might rule-out our rail overvoltage hypothesis.

    here are those measurements you asked for:

    D401 and D402: 0.384 / climbs to open loop
    D403 and D404: 0.380 / immediate open loop



    The burnt up U101 may not be related to the burn up 200 and 600 circuits. COuld be but don't assume it is.
    putting the 200 and 600 circuit faults together was easy enough. i have no way of explaining how they'd be related to the IC101 circuit faults.
    Last edited by bob p; 10-21-2007 at 02:09 AM.
    "Stand back, I'm holding a calculator." - chinrest

  11. #11
    Senior Member Enzo's Avatar
    Join Date
    May 2006
    Location
    Lansing, Michigan, USA
    Posts
    27,077
    i don't remember if i mentioned this before, but the voltages on all of the PSU caps are what they're supposed to be. that is to say, the caps that supply the different rails were't at the same potentials. i think that might rule-out our rail overvoltage hypothesis.
    Under my theory the voltage would look OK, but when the commutating transistor turns on, THEN and only then the wrong voltage would get places IF the isolation diode was shorted. FOr example if D401 were shorted, then when Q405 turns on, the 40 and 74v rails would be connected together. Then again that alone would be problem enough I guess.

    If they are not shorted, they are not shorted.

    I looked it up. The meter driver is a TA7318P, which is a dedicated VU meter driver, not an adapted headphone driver. I've attached the data sheet below.

    for use with a bipolar power supply, the voltage rails are at pins 5 and 9. measuring the Z on them, it starts at about 1k, jumps to 10k and keeps climbing, like i'm charging a cap somewhere.
    OK. Regardless, I seriously doubt the meters themselves caused any issues. Any faults in teh IC is more likely to burn the little meter movements out than anything the other way around.

    And yes, you are indeed charging a cap - the power supply caps.

    [QUOTE]I measured from the relay's inputs to the output solder points on the relay where the red and blue channel output wires take off to go to the speaker terminals. I'm%2y phase relationshup between channels will have no impact on the relay contacts.

    My previous post was written stream of consciousness, so if the relay pulls in and drops out like it should, then the D203 is not shorted obviously. If the diode were open, Q207 would give up the ghost pretty quick.

    You ZEN moment of the day:
    Put a sigh in front of you that says "Someone else worked on this before me." Never forget that. This is not an amp that failed and now you have it in the as failed condition. Aside from whatever original problem it had, you also have the previous tech's additions to that. Clearly he could not fix it, and we do not know what he might have done to it while trying.

    For example, my theory about the 40 and 74v rails getting shorted together by a bad diode was more related to an as failed amp, but it could easily have been the other guy shorting power rails together with a slipped probe or something. Leaves no physical evidence other than burnt up stuff.

    The fact that several places in the circuit are BURNT, not merely failed, is too convenient for me. I have to think they are related. These are not things that burn up in a normal blown channel type situation. Nothing should harm the meter circuit. Except maybe power supply.

    Well, one thing maybe... lightning. And I don't say that lightly. Lightning damage does not follow schematics, it follows physical proximity. Various things that have NO relation burn up because they are near each other on the board.

    i'm not following the failure path to the 40v rail. tracing backwards from ground, we have R212, Q207, and RL101+D203. Then I'm assuming that the power passed through those two thermal switches at the top of the diagram (points 5, 1, 2 and 38 on the schematic), passed by R613 and Q205, to a point labeled "39" on the schematic, to point 27 which meets the 40V rail. Is that the path you're thinking of?

    If that's the case I'm wondering about if I have to worry about other parts like R613 and Q205 (which isn't labeled as Q205).
    Yes, point 27 and friends are the 40v I refer to.

    The current through any circuit failure or otherwise has to come from somewhere. SO either it is the existing power rail, or it came from the outside world by way of something metal dropping on the board or some such. SO I follow the path you just described up to the relay, and the relay is connected straight to 40v. The fact that the drawing goes up through thermal cutouts and a wandering path does not alter that it is a direct connection to the 40v rail. If the 40v is OK, then other things connected to it are not threatened. SO no, I don't suspect Q205 and R613 since they are connected to 40v normally. SOme other circuit drawing excess current from 40v won't affect them.

    My only potential worry there would be if some other voltage like 74v from a screwdriver short got on the 40v rail. Then they would face the same over voltage the other stuff did. ERmember that is just a theory. Remember that Q205 and the timing cap work together to pull the relay in, so if the relay waits a few seconds and pulls in, then those xstrs are working.

    I believe the other amp does not show signs of heat. So is U101 original? The zeners I refered to are ZD101,102 above and below U101 on th drawing. The caps are the e-caps in paralle with each zener. If a zener or cap shorts, then the 6k8 feeding it will get hotter. Seems unlikely both side zeners would fail together though. A shorted IC would then move up my list.

    Zeners are just diodes, so if they are shorted, they show it. They can get leaky, so a 15v zener can wind up with say 6v across it. But usually, if you just test them for forward drop as a normal diode, they are most of the time OK when in zene mode too. If they are not shorted, act more or less like a real diode frontwards, and result in 15v at teh IC, then they are OK.

    no short in Q207. is it safe to replace Q207 and R212 now (before the meter and relay stuff gets fixed), or are we just going to submit the new Q207 and R212 to being blown-up?
    Q207 and R212 ARE the relay stuff, and I don't think the meter circuit wil get in the way of anything - unless the IC is shorted inside. If you like, remove the meter IC. Now R601,602 no longer go anywhere, and the resistors dividing off the speaker line won't load anything down. Fix everything else, then reinstall the IC and see if it works. If R601 gets hot then, the IC is bad.

    You had burnt traces before, ever determine what they were?

  12. #12
    Better Tone thru Mathematics bob p's Avatar
    Join Date
    May 2006
    Location
    Chicago, USA
    Posts
    2,988
    Quote Originally Posted by Enzo View Post
    Under my theory the voltage would look OK, but when the commutating transistor turns on, THEN and only then the wrong voltage would get places IF the isolation diode was shorted. FOr example if D401 were shorted, then when Q405 turns on, the 40 and 74v rails would be connected together. Then again that alone would be problem enough I guess.

    If they are not shorted, they are not shorted.
    ah, I see your point now. that was pretty insightful. so insightful that i missed it the first time around!


    OK. Regardless, I seriously doubt the meters themselves caused any issues. Any faults in teh IC is more likely to burn the little meter movements out than anything the other way around.
    well, if the meters are dead, this amp might end up having a different fate than if they are not. in that case, given the lack of parts availability, i might consider sending the amp out to one of those guys who does flat rate repairs and just letting him fix everything.
    "Stand back, I'm holding a calculator." - chinrest

  13. #13
    Better Tone thru Mathematics bob p's Avatar
    Join Date
    May 2006
    Location
    Chicago, USA
    Posts
    2,988
    Quote Originally Posted by Enzo View Post
    And yes, you are indeed charging a cap - the power supply caps.
    am i correct in interpreting the test to say that the observed charging of the PSU caps indicates the lack of a short across those pins, and that IC601 tests OK on that test? before pronouncing it OK, how else would you assess IC601?



    The fact that several places in the circuit are BURNT, not merely failed, is too convenient for me. I have to think they are related. These are not things that burn up in a normal blown channel type situation. Nothing should harm the meter circuit. Except maybe power supply.

    Well, one thing maybe... lightning. And I don't say that lightly. Lightning damage does not follow schematics, it follows physical proximity. Various things that have NO relation burn up because they are near each other on the board.
    hmmm. lightning. that could open an entirely new can of worms. what surprises me about this is that the burned areas are on the front (input op amp), middle (relay and protection), and back (meter control) sections of the board. although they're kinda far apart in physical proximity, the middle and back sections (protection relay and meter IC) are close electrically. still trying to figure out the input opamp thing. maybe we had two failures and there's no point in trying to tie it all together.


    I believe the other amp does not show signs of heat. So is U101 original? The zeners I refered to are ZD101,102 above and below U101 on th drawing. The caps are the e-caps in paralle with each zener. If a zener or cap shorts, then the 6k8 feeding it will get hotter. Seems unlikely both side zeners would fail together though. A shorted IC would then move up my list.
    yes, the input TL072 opamp appears to be original. ZD101 and ZD102 aren't shorted:
    Code:
    ZD101: .650/.830
    ZD102: .540/.838
    still having broswer faults, so i'll break up my posts. sorry.
    Last edited by bob p; 10-22-2007 at 04:11 AM.
    "Stand back, I'm holding a calculator." - chinrest

  14. #14
    Better Tone thru Mathematics bob p's Avatar
    Join Date
    May 2006
    Location
    Chicago, USA
    Posts
    2,988
    Quote Originally Posted by Enzo View Post
    Q207 and R212 ARE the relay stuff, and I don't think the meter circuit wil get in the way of anything - unless the IC is shorted inside. If you like, remove the meter IC. Now R601,602 no longer go anywhere, and the resistors dividing off the speaker line won't load anything down. Fix everything else, then reinstall the IC and see if it works. If R601 gets hot then, the IC is bad.

    You had burnt traces before, ever determine what they were?
    the problem i had with PCB traces on that other amp were not actually burned traces from something going wrong in the amp. they were traces that got yanked from the board by some yayhoo with a soldering pencil.



    Quote Originally Posted by bob P
    Quote Originally Posted by Enzo
    What does the TL072 measure between pins 4 and 8? I mentioned elsewhere that the chip was suspect.
    measured Z across pins 4-8 starts off at several kohms and keeps on climbing with the passage of time. i stopped watching after it climbed to 15k.
    it looks like PSU charging again, and the absence of a short in IC101, at least across those pins. Same question as before: what do you think of the status of this IC?

    OK, let me see if I follow your recommendations:

    1. pull out those additional non-stock ballasts. who knows why they are there.
    2. pull out IC601. then replace the stuff that's bad in the relay circuit, Q207 and R212.
    3. pull out IC101 (input opamp), leave it out and recheck voltages on the Q100 series of transistors.

    anything else to recommend at this stage of the game?
    Last edited by bob p; 10-22-2007 at 04:18 AM.
    "Stand back, I'm holding a calculator." - chinrest

  15. #15
    Senior Member Enzo's Avatar
    Join Date
    May 2006
    Location
    Lansing, Michigan, USA
    Posts
    27,077
    Yes, replace the ballast mess, I assume the underneath ones are open? Make them all good.

    I'd likely replace U101 on principle. They cost about a quarter. If you are concerned, install a socket, then you can replace burnt ones easy. I think it should be there. SS amps are HEAVILY feed back for stability. Without the IC, the output stage is free to drift around as it desires. Without that IC, even a good amp could wind up with serious DC offset. Generally the only thing that will damage that chip is something on the input line or real power problems. if the zeners are making 15v then the IC is not at risk.

    Oh, the missing traces were a different amp. Sorry.

    An IC with a dead short can measure that way, but plenty of dead transistors inside it could short the rails and not appear so from the outside. If you place a diode between ohm meter leads, it won't show a short, but apply it across a power supply and it sure will. Ohm meter tells you something is shorted, but can't tell you it is not.

    I don't know about U601, but if you pull it, then it can no longer get in the way of the rest of the work. You can verify the meter movements easy enough without it.

    Yes, fix Q207 etc.

    Input problem and other problems may or not be related, but either way they have to get fixed. A lot of time the tie ins become apparent after all the repairs are done. You get an aha moment and realize this cause that to fail and so on.

    Here is a scenario for lightning. If the lines actually were struck, the thing would be toast. But lightning is a huge curent flow nearby, and as such it induces current into conductors around it - as in wires. SO the cables to the input and the speaker wires from the output could have independently had voltage spikes induced in them by the same nearby strike. The connection may not have been inside the amp.

    AS I recall, your amp would sit there powered up with reasonable currents, so there should be no problem taking hot measurements.

    If U601 is already toast, powering for a few more minutes wont hurt it. DO R601,602 get hot? Is the voltage at the chip ends of R601,602 reasonable for that chip per the spec sheet.?

    Same argument for U101. If it doesn't load down the 15v zener rails, then most everythig is OK, at least OK enough to work with.

    I'd say get the amp stable, get it passing signal, get the relay working right, then get the meters going.

    The TA7318 seems to be in stock for $3 at B&D.

    www.bdent.com

  16. #16
    Better Tone thru Mathematics bob p's Avatar
    Join Date
    May 2006
    Location
    Chicago, USA
    Posts
    2,988
    Quote Originally Posted by Enzo View Post
    I'd likely replace U101 on principle. They cost about a quarter. If you are concerned, install a socket, then you can replace burnt ones easy. I think it should be there. SS amps are HEAVILY feed back for stability. Without the IC, the output stage is free to drift around as it desires. Without that IC, even a good amp could wind up with serious DC offset. Generally the only thing that will damage that chip is something on the input line or real power problems. if the zeners are making 15v then the IC is not at risk.
    I was planning on a socket. They're cheap and I can get the socket and the chip locally at Ratshack. having the socket, if the IC blows up, easy to pop in another.

    not sure what you mean by making 15V on the zeners. i suppose that you're talking about the potential difference across them when the amp is under power? my diode drop tests have been power off, of course.



    I don't know about U601, but if you pull it, then it can no longer get in the way of the rest of the work. You can verify the meter movements easy enough without it.
    any tips for removing those 9-pin linear chips without burning them? the suction bulb method doesn't seem to get all of the solder off on either the first or second passes, there's no way to protect the chip with a clip-on heatsink, and unless you get all of those pins cleaned off, the chip just sits there with one pin holding the whole thing to the board, and you have to reheat it over and over.



    AS I recall, your amp would sit there powered up with reasonable currents, so there should be no problem taking hot measurements.
    yes

    If U601 is already toast, powering for a few more minutes wont hurt it. DO R601,602 get hot? Is the voltage at the chip ends of R601,602 reasonable for that chip per the spec sheet.?
    thanks. then that's the next step.

    Same argument for U101. If it doesn't load down the 15v zener rails, then most everythig is OK, at least OK enough to work with.
    not sure i understood what you meant about the changes to the zener rails. i suppose you're talking about votlages dropping below normal somewhere?
    "Stand back, I'm holding a calculator." - chinrest

  17. #17
    Better Tone thru Mathematics bob p's Avatar
    Join Date
    May 2006
    Location
    Chicago, USA
    Posts
    2,988
    another question -- do you know where to get those big, square 5W dual-element ballast resistors like they use in the M-500t?
    "Stand back, I'm holding a calculator." - chinrest

  18. #18
    Senior Member Enzo's Avatar
    Join Date
    May 2006
    Location
    Lansing, Michigan, USA
    Posts
    27,077
    Yes, I mean the zeners regulate the rails to about 15v when the amp is powered. And if U101 is internally screwy, it could load that voltage down. if the voltage sits there at about 15v, then the zeners are doing there job, and the IC is not trying to thwart them. If the zeners were open, then the voltage there could drift a lot higher than 15v and the IC would suffer. Nothing cosmic.

    The inline ICs are to me the same as DIPs. I think of them as DIPs with only one side. I don't do anything different for them. I use a desoldering station, but if I were to do it by hand, I would use my desolder pump. I do not like the bulbs, I hat ehtem maybe less than I hate braid, but I don't like them. For manual solder exttraction I use one of those culinders where you cock it and press the button and the spring loaded plunger pops back and makes a suction pulse. They come in plastic or metal, and I like the metal ones. They can be had cheap.

    I used to buy them in bulk at $5 each. You could spend a lot more on a name brand one, and buy replacement tips for them, but the replacement tips cost about as much as a whole cheap surplus sucker. So I just buy the surplus suckers to begin with.

    The spring loaded thing can make a stronger vacuum pulse than a bulb.

    When desoldering, if the hole does not clear, don't reheat the hole and try again to get the rest. Resolder the hole and try again with a fully soldered hole. The melting solder transfers heat to the rest of the solder better than a half filled hole can. And make sure to hold heat on it long enough for the hole to thoroughly melt all the way through.

    But your board is single sided, yes? SO there should be no issues with plate through holes retaining solder.

    Rock the IC back and forth a little as you desolder, to try to unstick the pins from the edge of the hole. If they stic, I sometimes get my needle nose in there, grasp the end of the pin, and wiggle it around to crack free of the edge solder.

    But a desoldering station is really the way to go. The iron tip is a hollow tube you put over the pin or end of componenet lead. When teh solder melts, I move the tip in a small circle to move the pin away from all the sides of the hole and then hit the suction.

    SOmeplace like MCM will have dual resistors. 0.22 ohm 5w duals would be like MCM # 28-0540. They don't look like yours but would work.
    www.mcminone.com

    But unless I was ordering already, I would just take two single 0.22 5w cement resistors, set them side by side, and at one end, bend one lead over to the other and wrap it around and solder right next to the body. Now that common wire goes in the center hole, and the two loose ends go into either end hole. Voila.

    I am sure other places that sell parts for consumer electronics repair would have them too. Places like MAT electronics or Electronix come to mind. I don't buy from them, but they send me catalogs all the time.

  19. #19
    Better Tone thru Mathematics bob p's Avatar
    Join Date
    May 2006
    Location
    Chicago, USA
    Posts
    2,988
    Quote Originally Posted by Enzo View Post
    The inline ICs are to me the same as DIPs. I think of them as DIPs with only one side. I don't do anything different for them. I use a desoldering station, but if I were to do it by hand, I would use my desolder pump. I do not like the bulbs, I hat ehtem maybe less than I hate braid, but I don't like them.
    well, all that i have is a couple of non-adjustable soldering pencils and a couple of bulbs. that's all i've ever needed for "macroscopic" projects with PTP wiring of vacuum tube circuits. i'm just starting to do "microscopic" SS work and i don't have a temp controlled soldering station, a desoldering station, rework station, etc. just a cheap soldering pencil and a bulb, as that's gotten me along ok with tube stuff for a loooooong time.

    For manual solder exttraction I use one of those culinders where you cock it and press the button and the spring loaded plunger pops back and makes a suction pulse. They come in plastic or metal, and I like the metal ones. They can be had cheap.
    i think i've seen these things before, but i'll have to look to try to find them.

    When desoldering, if the hole does not clear, don't reheat the hole and try again to get the rest. Resolder the hole and try again with a fully soldered hole. The melting solder transfers heat to the rest of the solder better than a half filled hole can. And make sure to hold heat on it long enough for the hole to thoroughly melt all the way through.
    thanks for the tip. hadn't thought aobut that.


    But a desoldering station is really the way to go. The iron tip is a hollow tube you put over the pin or end of componenet lead. When teh solder melts, I move the tip in a small circle to move the pin away from all the sides of the hole and then hit the suction.
    what exactly is a desoldering station, and what does it do to make it worth several hundred dollars to buy one?

    SOmeplace like MCM will have dual resistors. 0.22 ohm 5w duals would be like MCM # 28-0540. They don't look like yours but would work.
    www.mcminone.com
    those resistors at MCM look like an inline package:



    the resistors that i'm looking for are rectangle shaped (close to square) and they have the 3 leads for the 2 resistors mounted at the corners of an isosceles triangle. not that i couldn't get those to work, but since i have to order anyway...

    But unless I was ordering already, I would just take two single 0.22 5w cement resistors, set them side by side, and at one end, bend one lead over to the other and wrap it around and solder right next to the body. Now that common wire goes in the center hole, and the two loose ends go into either end hole. Voila.
    since i have to order a bunch of stuff anyway, i might as well find the right part and place an order with whoever has it. any other ideas on that? i guess if all else fails i could just roll my own with two individual cement resistors, though i'd prefer a cleaner look if i can get it.
    "Stand back, I'm holding a calculator." - chinrest

  20. #20
    Better Tone thru Mathematics bob p's Avatar
    Join Date
    May 2006
    Location
    Chicago, USA
    Posts
    2,988
    now this is odd -- as i was removing IC101, Q207 and its emitter resistor, I noticed that there are some "extra" resistors soldered to the green side of the PCB. somebody has added 4.7R 1/4W resistors in parallel to each of R174, R175, R176, R177, which are 10R each. those are on the collectors of the Q115-Q118 drivers.

    ???

    it almost looks like they were trying to increase the rail voltage on the drivers, instead of addressing a sagging rail problem.
    "Stand back, I'm holding a calculator." - chinrest

  21. #21
    Senior Member Enzo's Avatar
    Join Date
    May 2006
    Location
    Lansing, Michigan, USA
    Posts
    27,077
    a cheap soldering pencil and a bulb, as that's gotten me along ok with tube stuff for a loooooong time.
    Sure. We use what we have. You are finding that it is not quite as effective clearing small holes on a pc board though. And this is a one sided board. That means the solder is on the sirface. Two sided boards have plate through holes so the solder flows from one side through to the other. This is even more dificult to clear easily.

    Some guys like bulbs, and more power to them. A LOT of guys like braid, though I can't see why, but apparently it works well for them, and you might try a roll of it. If you hate it you are only out a couple of bucks. I like the pumps because they apply more powerful suction to the melted joint.

    MCM 21-8235 is one like I use. Six bucks. CEDist sells a similar one with the Weller name on it for like $20 - I have found the cheapo ones work just fine. I have found them at surplus for under $3.


    Or $4.95 at Electronix stock SD1044


    For chassis work like old Fenders, I often get out my old gun. I have used it for over 50 years now. It sits next to me on the tool cart.



    Who says low tech doesn't work?

    Just wait until you encounter the "joy" of surface mount crap... I mean components.

    what exactly is a desoldering station, and what does it do to make it worth several hundred dollars to buy one?
    A desoldering station is a desoldering iron with a base unit, like a soldering station versus a soldering iron. In the case of desoldering, though, there is usually a vacuum pump of some wort, and that base unit houses that. And a desoldering iron, is what I mentioned, a soldering iron with a hollow tip, and a suction pump connectd to it through a hose. Apply the iron to a soldered joint, and when it melts, hit the vacuum and it sucks the liquid solder out of the joint like water through a straw. SLick. Would I buy one for two ICs a year? No. And when I did field service, I did not carry it with me, I used the hand pump. Pulling an 8 leg IC or a three leg transistor off a board by hand is no big deal.

    But when you sit at a bench working on pc boards all day, the economy of it starts to pay off. It is so much faster and easier than by hand. Especially when I have to change out a 64 leg IC. Two rows of 32. That is a lot more than 9.

    And when you place the bulb on the joint, you have to hold it over the tip of your iron, so the iron is in the way of a good air seal around the bulb tip and the work. The iron is like a foot in your door. The desolder iron IS the bulb. The tip itself covers the work, and the IC leg or the end of a resistor lead, or whatever is sticking through the pc board can fit right up the hole in the iron tip. The solder sucks up around it.

    desolder iron tips:


    You would consider buying one when the time saved and convenience make it worth while. My desolder station sits uner my solder station on the corner of the bench. They are both on while I work, so when something needs unsoldering, I reach for that iron handpiece, and when soldering I reach for the other handpiece. Desoldering a resistor or diode from a board is as quick as vwoop vwoop. (Sound effect of vacuum pump.) It leaves a neat clean hole, the unsoldered part usually falls right out of the hole, and the hole is left clean and ready for a new part.

    I know it is your money I am spending. Unless you were extremely serious, I don't think the average hobbyist would need one. But in a commercial bench, I will never be without it by choice. I use a very basic Pace unit, and I like it well enough. I have not used other brands. Apparently the Hakko is popular. I found my Pace worth the $600. But there are less expensive models. And some are not so fancy.

    MCM house brand for $289 72-6340


    Hakko hand held self contained MCM 94-410 $219


    ANd the lowest of tech, MCM house brand desolder iron $16.49. Stock 21-8240. it is a hand held iron combined with my hand pump.


    Such tools may not make sense in your shop, I don't know what sort of volume you do in the sort of things what would need this, but pc boards won't be going away soon. On the other hand, surface mount is becoming more and more common, and I just am on the edge of needing it enough to invest in some new gear. But not quite. A few hundred bucks for a basic hot air rework, and I am not quite there. Likewise you may not quite be there for a desolderer.

    I'll stop selling it now.

    Yep, I know those MCM dual resistors are not laid out the same, you'd have to bend the legs into a tripod, but electrically they are the same, and the inline ones like that are a lot more common than the flat ones like yours these days. I don't know who would have the older type.

    I don't know what the parallel resistors were for. if they look like a serious mod as opposed to the last tech dicking around, leave them. I don't think it was for voltage so much, those resistors maily limit current I believe. The lower resistance would make larger currents available to the drivers. ALthough maybe that is the same thing, lower resistance means less sag under higher current draw.

  22. #22
    Better Tone thru Mathematics bob p's Avatar
    Join Date
    May 2006
    Location
    Chicago, USA
    Posts
    2,988
    Quote Originally Posted by Enzo View Post
    I don't know what the parallel resistors were for. if they look like a serious mod as opposed to the last tech dicking around, leave them. I don't think it was for voltage so much, those resistors maily limit current I believe. The lower resistance would make larger currents available to the drivers. ALthough maybe that is the same thing, lower resistance means less sag under higher current draw.
    too late to save them, they're long gone. i don't think that they were an official Carver mod, i think that they were someone else's doing. the amp had the standard 10R resistors mounted on long leads standing off of the board for cooling, and it looked like somebody just tacked 4.7R resistors across the joints on the solder side of the board as an afterthought.

    when i took them apart, i expected that the 4.7R were just tacked on the underside and would be easy to remove. it turned out that somebody had actually removed the original 10R resistors and replaced them with new 10R resistors on top of the board, and twisted the leads around the leads of those 4.7R resistors on the underside of the board. then the twist was covered with a solder joint. i would think that an OEM mod would have replaced the 10R with a lower value rather than strapping something across it. this definitely looked like a hack fix when I was trying to get everything apart, and AFAIK there's no TSB for this mod. there were also a couple of places where i wasn't confident that the revised solder joints only touched the desired components, so i just pulled everything to return to stock.

    I pulled all of the 4.7R resistors out and got rid of the bent-up 10R resistors and replaced them with new 10R resistors so the amp is back to stock like the other one. perhaps the lower Z does allow more current for the drivers. i'll keep that in mind for later upgrades/improvements, but for the time being i think its a good idea to take the amp all of the way back to stock to get it up and running. i can always reassess the "mods" situation once the amp works properly. i just want to get rid of all of the things that look like amateur band-aid fixes.
    Last edited by bob p; 10-23-2007 at 07:11 AM.
    "Stand back, I'm holding a calculator." - chinrest

  23. #23
    Senior Member Enzo's Avatar
    Join Date
    May 2006
    Location
    Lansing, Michigan, USA
    Posts
    27,077
    Is it just my browser, or is this place really SLOOOOW tonoght? Posts are taking forever to load.

    If they were not on your other M500t, we would suspect the original design works. SO stock is good on the 10 ohms.

    While you are at it, if that U101 is waiting, ANY dual op amp you have will work there - 4558, 4560, 4565, 4580, 5532, 2068, TL082, what else yu got? 5532 and 2068 are low noisers, not that TL072 is noisy.. The 072 is a general purpose type. And if you plan to socket it anyway, the sub would work well enough for you to continue on the rest of it.

  24. #24
    Better Tone thru Mathematics bob p's Avatar
    Join Date
    May 2006
    Location
    Chicago, USA
    Posts
    2,988
    Quote Originally Posted by Enzo View Post
    Yep, I know those MCM dual resistors are not laid out the same, you'd have to bend the legs into a tripod, but electrically they are the same, and the inline ones like that are a lot more common than the flat ones like yours these days. I don't know who would have the older type.
    thanks. i pulled off those extra 0.47R resistors that were sitting atop the ballasts. the ballasts measure a small resistance of less than an ohm across each element with my DMM, so it looks to me like the original ballasts are okay. i have no idea why in the heck somebody added .47R 2W resistors in parallel with the .22R 5W ballasts. like I said before, this amp has a lot of strange things going on.

    i finally took that relay apart. on the back side of the coil the plastic cover was so hot that it had discolored from opaque white to orange, and there was a radial fracture in the case. it had obviously become very hot during that episode that caused Q207 and its emitter resistor to fail. (we're thinking that's related to failure in IC601.)

    i took the relay apart, and measured across the leads, and i was getting incosistent resistance readings from the OT emitters to the speaker terminals, anywhere from low Z to 500k depending upon the relay armature position. using all of the magnficiation i could muster, i thought i saw some charring on the contacts. i used an emory board to clean things up (or to ruin the contacts, depending upon how you look at it), and now the Z across the contacts has dropped by 500k to less than a half-ohm. bad relay.



    so here's where we stand:

    1. relay bad, could stand to be replaced. contacts roughed up with an emory board as a band-aid fix.

    2. TL072 input op amp removed. plan to replace with another TL072 in a socket. i tested the Z across those 2W dropping resistors to the opamp, and it took forever to get up to their 6k8 reading. obviously charging a cap somewhere, but the final results were within tolerance.

    3. IC601 for the meters removed.

    4. Q207 for the relay and its emitter resistor removed.




    we're going to be at a pause for a while, as i need to order parts. just as you'd expect Murphy's Law to prevail, i'm missing a few key pieces and i'll have to order them.

    I'm thinking about replacing some additional components in the areas of PCB discoloration, just to be on the safe side. like C103, C104, C105, C111 and R179 that are located close to R182 and R183 and may have suffered heat damage. anything else in that zone on the board that you would worry about, and order just to be on the safe side? zeners?

    when I was removing IC601, the pin 1 solder joint with R615 was cracked. heat i guess. i'm wondering if there are any other parts in the area that would be partricularly vulnerable to heat damage, like C601 or C602. R601, R602, R612 and R613 obviously took a lot of heat, but being metal oxides, they should be able to take it. I have also thought about replacing C605, C606 and ZD601, ZD602 just to ward off the evil spirits. There is a lot of PCB discoloration around IC601 and its better to be safe than sorry.

    if there are any other parts that you think would be suspects, needing possible replacement because of their location on the board or in the circuit, please do let me know.

    thanks.
    "Stand back, I'm holding a calculator." - chinrest

  25. #25
    Better Tone thru Mathematics bob p's Avatar
    Join Date
    May 2006
    Location
    Chicago, USA
    Posts
    2,988
    Quote Originally Posted by Enzo View Post
    Is it just my browser, or is this place really SLOOOOW tonoght? Posts are taking forever to load.

    If they were not on your other M500t, we would suspect the original design works. SO stock is good on the 10 ohms.

    While you are at it, if that U101 is waiting, ANY dual op amp you have will work there - 4558, 4560, 4565, 4580, 5532, 2068, TL082, what else yu got? 5532 and 2068 are low noisers, not that TL072 is noisy.. The 072 is a general purpose type. And if you plan to socket it anyway, the sub would work well enough for you to continue on the rest of it.
    REALLY SLOW tonight.

    parts on hand? unfortuantely no. i haven't done much SS work, so I'm starting off with no appreciable parts on-hand. i have to order just about evertyhing when it comes to opamps, transistors, low voltage radial caps, etc.

    regarding opamps, is there a lower-noise replacement for the 072 that's pin compatible? if i've got to order something, it might be worth ordering something that's better than plain vanilla.
    "Stand back, I'm holding a calculator." - chinrest

  26. #26
    Better Tone thru Mathematics bob p's Avatar
    Join Date
    May 2006
    Location
    Chicago, USA
    Posts
    2,988
    back to the subject of the desoldering gear -- thanks so much for your detailed explanation. my volume for this sort of stuff is so low right now that i can't justify a reworking station that i'll only use ocasionally. SMD? forget about it! but i really do like the idea of your suction-pen replacement for the bulb. i'll give that a try for sure!
    "Stand back, I'm holding a calculator." - chinrest

  27. #27
    Better Tone thru Mathematics bob p's Avatar
    Join Date
    May 2006
    Location
    Chicago, USA
    Posts
    2,988
    on the subject of ordering a replacement relay: the carver relay doesn't have a particularly helpul part number on it, it just says its a 24V audio relay. It looks like a DPST.

    are all of these sorts of relays designed so that they have a similar footprint and they'll be interchangeable? I'm just trying to get familiar with how i need to spec a replacement. fwiw, here's a photo from an existing thread, though admittedly its not a great one:

    "Stand back, I'm holding a calculator." - chinrest

  28. #28
    Senior Member Enzo's Avatar
    Join Date
    May 2006
    Location
    Lansing, Michigan, USA
    Posts
    27,077
    Your overall failure mode is still a mystery. What if someone plugged the amp into 240VAC somehow? Oh, we may never know.

    Something like the relay burning up just shouldn't happen. It is designed to be on continuously, so even if Q207 shorts, it would be the same to the relay as if it were simply on. The only thing I can think of keeps coming back to somehow the voltage went way high on the 40v rails. R212 does serve to drop voltage so the 24v relay can run off the 40v rail, but resistors don't short as a rule, they burn open. And the common current through the relay would not burn the resistor normally, and a shorted Q207 would no more affect the resistor than it would the relay.

    The relay did have bad contacts. Emery board is a good first step. Then a smoothree nail file or board. Then even smoother, fold a piece of emery cloth or wet/dry paper in half rough out and rub it up and down between the closed contacts. Press on the armature with your finger to apply a little pressure. Maybe 400 grit,and follow with 800 grit if you have it. I spent many years servicing pinball machines in the field, and prior to 1976 they were all relays inside. believe me I have a lot of experience servicing relay contacts. I have a burnisher I use on CLiff type jacks a lot, but I originally got it for relays.

    Look up MCM 22-1235, the burnisher looks like a fine file, but it feels smooth. But it really is coated with fine diamond dust or other hard material so it is a super fine file, so it burnishes (polishes) contacts smooth.


    Relays of that shape often come on a standard frmat, but these are a little different. I have the layout with trace patterns. These are DPST. I know you hate hacks, but in the absence of the real part and need a new one - assuming you cannot restore this one - one method would be to glue a new common type relay to the board upside down, and run wires down to the pads.

    The TL072 is a common general purpose chip. ALL the dual op amps in 8-leg DIPs have the same pinout, they all will work in each others place. I routinely swap 4558/4560/TL072 and others. The TL072 is not a low noise part, but it is also not noisy. I don't think Carver went cheap to save 10 cents on in IC. 5532 is a low noise IC, 2068 is a low noiser too. Oh I think it is NE5532 and NJM2068. The 5532 draw a few more milliamps than the TL072, but since ther is just one chip, it won't matter. Where it does matter is where someone gets the bright idea to low noise his PA mixer so he can "record" with it and decides to change out 50 or 100 chips. NOW all those extra ma-s add up to a ton more current and the power supply can't handle it. No such trouble with one chip here.

    Op amps are a commodity item, and they are the same in all the stuff we service. 4558 was real common through the 70s 80s. Now we see 4560, 4565, 4580 - just later chips in the 4558 series. 2068 and TL072 are also stil common. And TL074 is pretty much stil the universal quad op amp - 14 pin IC. It is the same as two TL072s in one chip.

    I have pretty much discussed what I think might be bad so far. On the other hand, I don't have to order every part I need, I am used to having a pretty extensive parts inventory to draw from. I am not so concerned for nearby parts, but if in your view they look scorched, go ahead and replace them. Little caps and resistors cost little, and in the case where one really was bad, you don't have to go through another order cycle.

    Many time I encounter charcoal situation, but often as not the damage is more superficial than anything. Nearby components are covered in soot. They may not have really been hot, they just got a lot of smoke deposited on them. A Qtip ful of alcohol wipes most of it off. I worry about caps more than resistors. Most resistors don't care how hot their neighbors get unless flames come out.

    I think if I were gong to start a parts kit for general SS repair, I'd stock a few op amps - pick one type - and certainly 15v 1w zeners are common enough. The rectifiers you might use in tube amps will be good enough - 1N4007, a few 1N4148 for the little signal diode spots, higher curent diodes I'd just wait until I needed them, there are many kinds. It is likely premature to duscuss transistor selections and such.

    When I first set up this shop years ago, I already had the Pace desolder station, but I had a partner who had none, so we found a used one online. That is an alternative to consider.

    A quick tour of Mouser did not reveal a relay with your foot[print.

  29. #29
    Better Tone thru Mathematics bob p's Avatar
    Join Date
    May 2006
    Location
    Chicago, USA
    Posts
    2,988
    I got some parts in the mail today, so I replaced the following components:

    removed IC101 (a TL072) and replaced it with a socket and a new TL072
    replaced Q207 and its 270-ohm 1/2 watt emitter resistor
    replaced R179, C103, C104 due to visible external heat damage

    I powered on the amp and the relay is working. After 3 seconds the impedance between the ballasts and the speaker terminals drops to zero. and stays at zero. so far so good.

    i tried injecting a signal into the amp, while monitoring the output of the signal generator on one of my scope channels. the signal disappears as soon as it is connected to the amp's inputs. it looks like the test signal is getting shorted to ground somewhere inside of the amp. The protection circuit doesn't activate.

    I checked the voltages on IC101 and compared them to the reference voltages on the schematic. then I pulled the IC and measured voltages on the socket:

    Code:
    Pin#   Reference  In-circuit   Removed
    1       0           -56mV        6.3V
    2       0             0          26.9v
    3       0             0           0.3mV
    4       -14.7v      -14.6v      -14.8v
    5       0             0           0.4mV
    6       0            -2.0 mV     26.2v
    7       0           -61.5mV       1.29v
    8      +14.7v        14.78v      14.98v
    I haven't had a chance to lift any components to see why the input signal is disappearing. Its late, and I have to put this off until tomorrow.

    Do you find any of the voltage measurements on the IC to be helpful? I haven't yet determined why there are negative potentials on IC pins 1, 6 and 7, or why the voltages are what they are on the empty socket. I'll have to sleep on it.
    "Stand back, I'm holding a calculator." - chinrest

  30. #30
    Senior Member Enzo's Avatar
    Join Date
    May 2006
    Location
    Lansing, Michigan, USA
    Posts
    27,077
    First off, as a rule you cannot measure signal at the input pins of an op amp. Something about them being at "virtual ground" which you might look up. I thhink the National Semiconductor web site has tutorials on op amps - among many other topics. If you are at all interested, you can go through their Analog U. AS in university - training materials for analog electronics. And anything written by Bob Pease there is great stuff. He's my hero.

    If you pull the IC from the socket, you should be able to see signal at the pins to verify it gets there. The main way I test op amps it to look at the output pins. if signal is there, it is a safe bet it was at the input pins too.

    With the IC removed, there is nothing to keep the amp circuit centered on zero. The output bus feeds right back to the input through R113 to pin 2 of the IC. Without the IC, that output feedback is more or less connected to the drive throughR192, R111. And the whole thing will drift to wherever it feels comfortable.

    In circuit, the 072 has readings within millivolts of zero, and it looks OK to me. Rails are fine. Is there signal at pins 1,7? There is no ground connection in the chip, so even if C101 shorted, there would still be a minimum of 1k across the input due to R103. A few millivolts will have one polarity or the other, but it is millivolts, and not many of them.

    Without the chip the amp will drift to whatever, and there will be no op amp to correct for it.

    Since it is socketed, remove the IC and make sure signal gets there.


    And measure resistance to ground across the input sockets just to verify they are not somehow shorted.

  31. #31
    Better Tone thru Mathematics bob p's Avatar
    Join Date
    May 2006
    Location
    Chicago, USA
    Posts
    2,988
    Making some progress. I was getting inconsistent readings to ground on one of the input jacks. Sometimes one channel looked shorted, and sometimes it didn't. (IIRC it was the right but I'm not sure).

    I had already ordered-in all of the small electrolytics needed to replace the 20+ year old 'lytics on the amp board as part of routine PM, so I swapped them all out. I also replaced some metallized film caps that were in the burned spots on the PCB. The short to ground problem on the inputs went away, so it was likely a bad cap somewhere. As we had discussed before, there were a lot of areas on the board that had visible heat damage. With the new lytic caps, a lot less bias current is required to meet the 6 mV spec, so some of them must have been leaky.

    (As an aside, how easy is it to blow up an opamp? After doing the recapping, the signal generator mistakenly had maybe 5mV of +DC offset because of a knob that wasn't properly zeroed. When the amp turned on, the signal trace at the input looked OK for a short while, and then became very ugly and spikey -- kind of like a tangent function if you remember them from your geometry days. I pulled the IC and swapped in another.)

    Now the Left channel produces a clean sine wave output at the speaker terminals into no load.

    The right channel is causing problems though -- the amp will idle fine with no Right channel input, but the protection circuit keeps triggering when an AC test signal is fed to the R channel input and I scope the speaker terminals. The relay energizes after the 3 second turn-on delay, and on the scope I see a blip of what looks like a tall clipped waveform on the screen for a fleeting instant, then the protection relay triggers again, flattening out the scope trace. The relay continues to cycle in periodic fashion, and I intermittently see a fleeting blip of what sometimes look like either: a) a clipped AC waveform that's transposed upwards with a DC offset, or b) a tall square wave with rounded corners that looks like its all DC. After the waveform appears on the outputs for only an instant, the relay cycles into protection mode. Then the cycle repeats over and over again.

    Now here's what's interesting -- the odd behavior I've mentioned above only appears at the speaker terminals. At the center-taps of the R channel ballast resistors, the sine wave outputs look completely normal throughout the amp's power range. I'm at a loss to explain why the DC appears on the outputs but not on the ballast resistors. The only thing between those two points is ... the relay. But if the DC is coming from the relay and appearing at the speaker terminals, I would think it should appear on the ballast resistors as well. Maybe its coming from somewhere else.
    Last edited by bob p; 11-07-2007 at 12:40 AM.
    "Stand back, I'm holding a calculator." - chinrest

  32. #32
    Better Tone thru Mathematics bob p's Avatar
    Join Date
    May 2006
    Location
    Chicago, USA
    Posts
    2,988
    Well, I was wrong; My DMM tells me that there's 1.25VDC on the R channel ballast resistors at all times (none on the Left), but the DC only seems to trip the protection relay when the scope leads are on the speaker outputs. The relay doesn't seem to be tripping if the scope leads are on the ballasts. I can't explain that. I'm looking at the schem now to try to determine where that low voltage can be coming from.
    "Stand back, I'm holding a calculator." - chinrest

  33. #33
    Senior Member Enzo's Avatar
    Join Date
    May 2006
    Location
    Lansing, Michigan, USA
    Posts
    27,077
    5mv of offset shouldn't harm the op amp.


    What does that waveform do at the ballasts WHILE this happens, not before or after. Oh, and are you viewing the signal at the ballasts with the scope AC or DC coupled? We need DC for this. A DC offset won't appear on an AC coupled display.

    When this happens, when the speaker relay goes OFF, is there any DC remaining on the speaker terminals? Thinking some odd trouble in U601. But with 82k in series, that seems unlikely.

    How abuot this? You interrupt the speaker line while any current is flowing - either signal waveform (at any point other than zero crosssing) or a DC offset - and the inductance of the speaker voice coil will provide an inductive kick. Ever unplug a working speaker and seen the spark at the jack? MAybe this is what the scope sees on the outputs. Back on the amp side of the relay, all this would be invisible, since it was the disconnection FROM the amp that triggered the symptom on the speaker side.

    Then the relay triggering would be a separate problem.

    Is R202 OK? It samples the R channel out into Q202, part of the shut down circuit. SInce the L channel doesn't do this, we can assume Q202 itself and associated parts are OK.

    SInce the R channel seeems to work internally, find out what is triggering the protect relay. WE know that the outputs are sampled through R201, 202 and fed to Q202, and also inverted through Q204 to feed Q203. That way the signal can go pos or neg and still trigger one or the other of Q202,203. SInce those two are in parallel, either one can trigger Q205 and shut it down.

    Q206 can also shut things down via Q201 up there, and it in turn is driven by limiters in each channel - R185/D105 and R184/D106.

    So by monitoring the collector of Q202 and the base of Q206, we might find which path is shutting the relay down, and from that have a clue.

  34. #34
    Senior Member Enzo's Avatar
    Join Date
    May 2006
    Location
    Lansing, Michigan, USA
    Posts
    27,077
    Ah, simulposts.

    DC is enough to cause relay trouble.

    Scope leads? Maybe the output doesn't like being grounded, which the scope probe ground lead would do?

    Well, it is all DC coupled in the amp, and servoed back into itself, so offset anywhere shows up through the whole thing typically. I'd go through and look at the DC levels through the amp, like the test points shown on the L channel. Keep in mind they balance, so where it is supposed to be +1.4 and -1.4, if you find +2 and -0.8, at least we know the parts are still maintaining their separation properly and the whole thing is being pushed to the side.

    And another point, sometimes if a voltage is too positive - like +1.3 instead of zero - it COULD be from something on the + side turned on too far, but it ALSO might be something on the neg side NOT turning on enough, or not conducting well enough. Say a weak xstr or a resistor that upped in value.

  35. #35
    Better Tone thru Mathematics bob p's Avatar
    Join Date
    May 2006
    Location
    Chicago, USA
    Posts
    2,988
    Quote Originally Posted by Enzo View Post
    What does that waveform do at the ballasts WHILE this happens, not before or after. Oh, and are you viewing the signal at the ballasts with the scope AC or DC coupled? We need DC for this. A DC offset won't appear on an AC coupled display.
    Its hard to say -- the problem where the relay cycles repeatedly in and out of protect mode seems to be related to dual-channel hookups to the scope. If I'm simultaneously comparing the output of the ballasts to the output at the speaker terminals, then the amp goes into the protect/reset/protect loop. OTOH, if I disconnect the scope probe that's monitoring the speaker terminals, then the repeated cycling of the relay doesn't seem to happen.

    I never noticed this sort of behavior with my other M-500t, but then I don't think that I ever tried to simultaneously scope the signal at the ballasts with one channel of the scope while scoping the signal at the outputs with the other channel. Perhaps I had the polarity reversed between the probes, and the protection circuit didn't like the way things were grounded.


    When this happens, when the speaker relay goes OFF, is there any DC remaining on the speaker terminals? Thinking some odd trouble in U601. But with 82k in series, that seems unlikely.
    U601 is absent from the amp board. We removed it to simplify troubleshooting of the amp circuit. The original IC is still sitting in my parts box, and I haven't ordered a replacement yet. (Nobody seems to have it except BD, and I'm waiting until I know more to place and order with them.)


    Then the relay triggering would be a separate problem.

    Is R202 OK? It samples the R channel out into Q202, part of the shut down circuit. SInce the L channel doesn't do this, we can assume Q202 itself and associated parts are OK.
    Now its gotten more interesting. I saw the L channel manifest the slow chatter of the relay as the amp entered the "protect/unprotect" loop. This happened when I tried to simultaneously scope the ballasts and the speaker terminals on the LEFT channel. I wonder if this could be a grounding/polarity issue between the scope and the protection circuit. The problem goes away immediately if I disconnect one of the scope channels from the amp. I also noticed -0.9VDC of offset on the L side ballasts, and 0 VDC on the right. The problem seems to be intermittent and/or migratory. :-/

    Now here's the good news: All of these observations were made with NO LOAD other than the scope itself. It turns out that if I scope the output at at the ballasts, and use only one scope channel, I can hook up a DVM and an 8R dummy load to the outputs, and both channels will put out 43+ VAC RMS into the 8R load. That's 231 WRMS, which is pretty close to the amp's rated output of 250 Watts/channel.

    It looks like the problem might be in the protection circuit, rather than in the audio section of the amp itself.
    "Stand back, I'm holding a calculator." - chinrest

Page 1 of 2 12 LastLast

Thread Information

Users Browsing this Thread

There are currently 1 users browsing this thread. (0 members and 1 guests)

Similar Threads

  1. Carver M-1.5t THUMP at Power Off
    By bob p in forum Other Amps
    Replies: 30
    Last Post: 06-23-2009, 03:28 AM
  2. SVT Fault.... fix it?
    By cherin in forum Maintenance, Troubleshooting & Repair
    Replies: 5
    Last Post: 08-07-2007, 06:10 PM
  3. Marshall JCM2000 Fault
    By Ant in forum Music Electronics
    Replies: 3
    Last Post: 11-02-2006, 12:47 PM
  4. Gibson 500T
    By Merlin in forum Pickup Makers
    Replies: 2
    Last Post: 09-17-2006, 10:48 PM

Bookmarks

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •