Hi all,
Just learned something that may be helpful to others and didn't see it discussed here before so...
I recently repaired a BN44-00192A power board from a Samsung LNT3253HX/XAA (this board and it's similar brethren are common to many smaller Samsung LCD TVs).
After correcting the conductive/burnt glue that was the main problem, and replacing all the large electrolytic caps on the cold side (a couple were bulging/obviously bad), the power supply and the TV were working fine again, all voltages were within spec, it started up every time, etc. However there was an annoying whistling sound coming from the power supply when the TV was on that was just loud enough to be disturbing.
I used an app on my iPhone called "iAnalyzer Lite" to narrow down the origin of the whistle to the 13/12/5 V output area on the "upper right" of the board (in the vicinity of LM851, the toroidal inductor) and to identify the frequency as about 3500Hz. I then used my O'scope to look for the corresponding ripple, and found it on the 5.3V output, about 100mV pp, at 3.5kHz.
I had already replaced CM861, the 1000uF filter cap on the 5.3V output, so I didn't think that was likely to be the problem, and decided to look back at the schematic (downloaded from some thread in this forum) for other likely culprits.
On the schematic, RM864, which is part of the feedback loop for the MC33167 single chip DC-DC converter, has a value of 15k. (Notably, in the schematic pdf file, RM864 is circled in red as a notation, but without any further explanation.) On two other good working BN44-00192B power boards I have here, that resistor is indeed a 15k. But on the BN44-00192A board that was whistling, RM864 had a value of only 2.7k. I looked at the spec sheet for the MC33167, and the sample circuits had even higher values for that feedback resistor.
So to make a long story short, I swapped out RM864 (a surface mount resistor on the solder side) for an 18k resistor I scavenged and the whistle is gone.
So I am guessing that the value of RM864 was one of the things that changed going from the "A" to the "B" version of this board, and that the larger value is needed to keep the feedback loop stable on the 5.3V supply as the components age. Note also that the additional ripple on the 5.3V supply might also be in part responsible for the premature failure of CM861, which otherwise doesn't see much ripple.
I hope this helps someone someday.
And I recommend you check out the "iAnalyzer Lite" app. If you are nerdy enough to be reading this post, you will probably have fun with a real time audio spectrum analyzer. (And no, I have no connection whatsoever to the author/publisher of that app. I just thought it was cool and well done.)
John
Just learned something that may be helpful to others and didn't see it discussed here before so...
I recently repaired a BN44-00192A power board from a Samsung LNT3253HX/XAA (this board and it's similar brethren are common to many smaller Samsung LCD TVs).
After correcting the conductive/burnt glue that was the main problem, and replacing all the large electrolytic caps on the cold side (a couple were bulging/obviously bad), the power supply and the TV were working fine again, all voltages were within spec, it started up every time, etc. However there was an annoying whistling sound coming from the power supply when the TV was on that was just loud enough to be disturbing.
I used an app on my iPhone called "iAnalyzer Lite" to narrow down the origin of the whistle to the 13/12/5 V output area on the "upper right" of the board (in the vicinity of LM851, the toroidal inductor) and to identify the frequency as about 3500Hz. I then used my O'scope to look for the corresponding ripple, and found it on the 5.3V output, about 100mV pp, at 3.5kHz.
I had already replaced CM861, the 1000uF filter cap on the 5.3V output, so I didn't think that was likely to be the problem, and decided to look back at the schematic (downloaded from some thread in this forum) for other likely culprits.
On the schematic, RM864, which is part of the feedback loop for the MC33167 single chip DC-DC converter, has a value of 15k. (Notably, in the schematic pdf file, RM864 is circled in red as a notation, but without any further explanation.) On two other good working BN44-00192B power boards I have here, that resistor is indeed a 15k. But on the BN44-00192A board that was whistling, RM864 had a value of only 2.7k. I looked at the spec sheet for the MC33167, and the sample circuits had even higher values for that feedback resistor.
So to make a long story short, I swapped out RM864 (a surface mount resistor on the solder side) for an 18k resistor I scavenged and the whistle is gone.
So I am guessing that the value of RM864 was one of the things that changed going from the "A" to the "B" version of this board, and that the larger value is needed to keep the feedback loop stable on the 5.3V supply as the components age. Note also that the additional ripple on the 5.3V supply might also be in part responsible for the premature failure of CM861, which otherwise doesn't see much ripple.
I hope this helps someone someday.
And I recommend you check out the "iAnalyzer Lite" app. If you are nerdy enough to be reading this post, you will probably have fun with a real time audio spectrum analyzer. (And no, I have no connection whatsoever to the author/publisher of that app. I just thought it was cool and well done.)
John
Comment