Rank: Member
Groups: Member
Joined: 3/9/2009(UTC)
Posts: 65
Location: seoul, korea
|
My multi meter does not measure the c-values. I only checked shorts across caps and no shorts. Then, can I narrow down the problem to bad IC's?
|
|
|
|
|
|
Rank: Member
Groups: Member
Joined: 3/9/2009(UTC)
Posts: 65
Location: seoul, korea
|
If the problem is with IC, is there any way to determin which one is bad?
|
|
|
|
|
|
Rank: Administration
Groups: Administration, Customer
Joined: 10/24/2006(UTC)
Posts: 3,979
Location: Nashville, TN
Thanks: 25 times
Was thanked: 89 time(s) in 83 post(s)
|
Most likely it would be the LM4780.
What are the rail voltages measuring?
|
|
|
|
|
|
Rank: Member
Groups: Member
Joined: 3/9/2009(UTC)
Posts: 65
Location: seoul, korea
|
Russ White wrote:Most likely it would be the LM4780.
What are the rail voltages measuring? Where to measure it? is it btw pin27 and gnd? Edited by user Friday, July 9, 2010 7:11:08 AM(UTC)
| Reason: Not specified
|
|
|
|
|
|
Rank: Administration
Groups: Administration, Customer
Joined: 10/24/2006(UTC)
Posts: 3,979
Location: Nashville, TN
Thanks: 25 times
Was thanked: 89 time(s) in 83 post(s)
|
You can measure between ground and each of the positive and negative rails. There are several places to do that. Refer to the schematic.
|
|
|
|
|
|
Rank: Member
Groups: Member
Joined: 3/9/2009(UTC)
Posts: 65
Location: seoul, korea
|
Russ White wrote:You can measure between ground and each of the positive and negative rails. There are several places to do that. Refer to the schematic. +26v and -26v
|
|
|
|
|
|
Rank: Administration
Groups: Administration, Customer
Joined: 10/24/2006(UTC)
Posts: 3,979
Location: Nashville, TN
Thanks: 25 times
Was thanked: 89 time(s) in 83 post(s)
|
I really can only guess at the problem but my suspicion is that somehow the LM4780 itself was damaged.
There is no way to test each IC individually while they are all on the board.
The LM4780 damage could happen a number of different ways but my first guess is it could have happened when attaching to the heat sink. You have to be very careful not to over tighten. I have done this myself accidentally and not even realized it.
It is possible to remove the LM4780 by very carefully cutting all the leads then desoldering each pin carefully.
You would need to take a lot of care to be sure not to damage the PCB.
You could then install a new LM4780.
|
|
|
|
|
|
Rank: Member
Groups: Member
Joined: 3/9/2009(UTC)
Posts: 65
Location: seoul, korea
|
|
|
|
|
|
|
Rank: Member
Groups: Member
Joined: 3/9/2009(UTC)
Posts: 65
Location: seoul, korea
|
Russ White wrote:I really can only guess at the problem but my suspicion is that somehow the LM4780 itself was damaged.
There is no way to test each IC individually while they are all on the board.
The LM4780 damage could happen a number of different ways but my first guess is it could have happened when attaching to the heat sink. You have to be very careful not to over tighten. I have done this myself accidentally and not even realized it.
It is possible to remove the LM4780 by very carefully cutting all the leads then desoldering each pin carefully.
You would need to take a lot of care to be sure not to damage the PCB.
You could then install a new LM4780.
Hi Russ, Is it possible to test LM4780 when it is separated from the board? I would like to make sure the old LM4780 is bad before I put a new one in. Edited by user Monday, July 12, 2010 4:32:57 AM(UTC)
| Reason: Not specified
|
|
|
|
|
|
Rank: Administration
Groups: Administration, Customer
Joined: 10/24/2006(UTC)
Posts: 3,979
Location: Nashville, TN
Thanks: 25 times
Was thanked: 89 time(s) in 83 post(s)
|
No, I really can't think of a good way to test LM4780 without putting into a working circuit. What you can do is test the rest of your amplifier with the LM4780 out. You would do that by putting temporary 2K or so resistors (value not critical) from the output of the OP1632 to close the feedback loop. This would be easily accomplished by placing those resistors across C1 and C2. As always test with the amplifier inputs grounded. Once you have those in place the output of the opa1632 should very close to 0V. Any problem with the first two ICs would be exposed this way. Remember to remove those test resistors before you put the LM4780 back in. Edited by user Monday, July 12, 2010 5:20:08 AM(UTC)
| Reason: Not specified
|
|
|
|
|
|
Rank: Member
Groups: Member
Joined: 9/2/2008(UTC)
Posts: 53
Location: Sydney, Australia
|
hkminn,
My output also has significant DC relative to ground. To diagnose the problem, I have made a few voltage measurements at critical points. Could you please test your board at these points and let me know what you get. My results are identical for both boards.
BTW all measurements are relative to ground. I find it easy to connect one probe to CT1 or CT2 (both are connected to GND) and use the other probe to test the various points. Look at the schematic to work out where to get the readings from.
Heatsink = 0V
VCC = 30.7V
VEE = -30.7V
Regulated Rail+ = 13.87V
Regulated Rail- = -13.85V
IN + = 0V
IN - = 0V
IC3.1 OUT = 0V
IC3.2 OUT = 0V
IC1.1- = -2V
IC1.1+ = IC2 VOUT- = 0.257V
IC1.2- = -2V
IC1.2+ = IC2 VOUT+ = -0.217V
OUT+ = -21.8V
OUT- = -22.0V
(using a 22V+22V secondary transformer)
Also, I find it odd that all 4 amps (both mine and hkminn) are doing the same thing. Russ how sensitive is the LM4780? Is it possible that both of us could have damage all 4 of the chips by tightening the screws too much?
Thanks
|
|
|
|
|
|
Rank: Member
Groups: Member
Joined: 3/9/2009(UTC)
Posts: 65
Location: seoul, korea
|
sureshm wrote:hkminn,
My output also has significant DC relative to ground. To diagnose the problem, I have made a few voltage measurements at critical points. Could you please test your board at these points and let me know what you get. My results are identical for both boards.
BTW all measurements are relative to ground. I find it easy to connect one probe to CT1 or CT2 (both are connected to GND) and use the other probe to test the various points. Look at the schematic to work out where to get the readings from.
Heatsink = 0V
VCC = 30.7V
VEE = -30.7V
Regulated Rail+ = 13.87V
Regulated Rail- = -13.85V
IN + = 0V
IN - = 0V
IC3.1 OUT = 0V
IC3.2 OUT = 0V
IC1.1- = -2V
IC1.1+ = IC2 VOUT- = 0.257V
IC1.2- = -2V
IC1.2+ = IC2 VOUT+ = -0.217V
OUT+ = -21.8V
OUT- = -22.0V
(using a 22V+22V secondary transformer)
Also, I find it odd that all 4 amps (both mine and hkminn) are doing the same thing. Russ how sensitive is the LM4780? Is it possible that both of us could have damage all 4 of the chips by tightening the screws too much?
Thanks Hi sureshm, I disassembled everything and I cannot make the measurements at the moment. Have you made the measurements with LM4780 on the board? How hot do your LM4780s get? Mine get so hot even with heatsinks that I decided to wait the measurments until I replace the chips. Also I am trying to do make the measurements Russ suggested but I am still struggling to put the resistors across the caps. I have no experiences with SMD parts and not good enough with irons to work in that small spaces. About the too much pressure on the chips. I also doubt whether I put too much pressure on the chips since I was very careful and even used a tension bar to fix chips on the heatsink. But since Russ have had the same experience I will put new chips and see what happens. For new chips I will try double sided termal tapes to fix chips to heat sinks without screws. I will make the measurements as soon as I get things together with new chips and post the results. I am still waiting for the new chips to arrive.
|
|
|
|
|
|
Rank: Administration
Groups: Administration, Customer
Joined: 10/24/2006(UTC)
Posts: 3,979
Location: Nashville, TN
Thanks: 25 times
Was thanked: 89 time(s) in 83 post(s)
|
Usually the problem is tightening unevenly, not so much "hard". Uneven pressure puts stress on the internal die and causes it to crack. The reason I suspect the LM4780s in both cases is that the other voltages are consistent with what I would expect if the LM4780 were being pulled to one rail for some reason. It is very difficult to know how that happened. Edited by user Tuesday, July 13, 2010 4:08:55 AM(UTC)
| Reason: Not specified
|
|
|
|
|
|
Rank: Member
Groups: Member
Joined: 9/2/2008(UTC)
Posts: 53
Location: Sydney, Australia
|
I only ever connected the LM4780 for 30 seconds at a time before it got really hot or I got weird sounds through the speakers. Without heatsink it gets to untouchable hot after 10 seconds.
Instead of trying SMD soldering, you could jumper from LM4780 input to output, effectively taking it out. You could use the holes of the LM4780 once its taken out. Then test the voltage at the vias. You could take it one step further and connect headphones to the output and you would have a headphone amp powered by the OPA1632 with a gain of 20. That way you can test a real signal rather than just ground. Russ correct me if I am wrong.
I will wait for you to try with a new chip and report back. Replacing a 27 pin device is not going to be easy. Let me know if you have any tips for cleaning the 27 holes on the PCB.
|
|
|
|
|
|
Rank: Administration
Groups: Administration, Customer
Joined: 10/24/2006(UTC)
Posts: 3,979
Location: Nashville, TN
Thanks: 25 times
Was thanked: 89 time(s) in 83 post(s)
|
You are right you could just connect the In and Out pads appropriately.
You should never run the LM4780 without a heat sink.
Use a solder sucker to clean out the holes.
|
|
|
|
|
|
Rank: Member
Groups: Member
Joined: 3/9/2009(UTC)
Posts: 65
Location: seoul, korea
|
Hi Russ,
I have received and replaced the chips but I have the same problems as before, -17v on both output terminals relativ to GND and very hot heatsink. This time I used double sided thermal tape to fix the chip to the heatsink, no pressure to the chip at all.
Then, I cannot help suspecting the defects of the board or any of the parts pre-installed on it.
|
|
|
|
|
|
Rank: Member
Groups: Member
Joined: 9/2/2008(UTC)
Posts: 53
Location: Sydney, Australia
|
Russ,
I have also replaced my LM4780 chips and the problem is still there. All measurements are identical to the ones I posted before even after replacing the chips. This time round I took extra care when installing on to heatsink and even used nylon screws for further insulation. I intend on doing some more testing over the next few days.
Brian, could I please request that you check for DC offset on your test board to check if your board has a similar problem.
Let me know if you have any other test in mind.
Thanks
|
|
|
|
|
|
Rank: Administration
Groups: Administration, Customer
Joined: 10/24/2006(UTC)
Posts: 3,979
Location: Nashville, TN
Thanks: 25 times
Was thanked: 89 time(s) in 83 post(s)
|
Did you test the board with no LM4780? What you are seeing suggests that the problem is upstream. Edited by user Wednesday, July 21, 2010 3:58:38 AM(UTC)
| Reason: Not specified
|
|
|
|
|
|
Rank: Member
Groups: Member
Joined: 3/9/2009(UTC)
Posts: 65
Location: seoul, korea
|
No, I did not. I tried to put resistors across SMD caps but I gave up.
I am not confident working with SMD parts in the tiny space and I am worried about further damage to the other parts.
|
|
|
|
|
|
Rank: Administration
Groups: Administration, Customer
Joined: 10/24/2006(UTC)
Posts: 3,979
Location: Nashville, TN
Thanks: 25 times
Was thanked: 89 time(s) in 83 post(s)
|
You don't need to add any resistor(but that is one way to do it), just connect the output pads to the input pads with the LM4780 out as was mentioned before.
|
|
|
|
|
|
Forum Jump
You cannot post new topics in this forum.
You cannot reply to topics in this forum.
You cannot delete your posts in this forum.
You cannot edit your posts in this forum.
You cannot create polls in this forum.
You cannot vote in polls in this forum.
