I'd appreciate any suggestions!
This is an LED DLP TV with no lamp or color wheel. It has big R, G, and B LEDs for the light engine. It has seen regular use since late 2007.
It started taking longer to warm up before coming on. Later, it began doing glitchy things before warmed up. Sometimes, I'd get audio, but no video. Sometimes, the screen would get scrambled with odd patterns and there'd be a loud buzz from the speakers. It would reset itself after a bit. Other times, the audio would be higher pitched, making human voices like The Chipmunks, then lock up, then reset. Sometimes, it would display video and play sound but be unresponsive to commands. But after warming up for a bit, and after two or three resets, it would work fine and continue to work fine until turned off and cooled down. It reached a point where before warm, it would just sit there with the green LED on the front blinking once per second for a while and then would give up, never displaying a picture.
I read online of many occasions of success replacing bad caps on power boards. I found a row of five bulged caps on the sub power board. I replaced ALL electrolytic caps for both the power boards, over twenty. No difference.
After some troubleshooting, I decided that the mainboard was suspect. It has many electrolytic caps on it in V-chip/SMD packages. I figured that since the TV works when warm, the problem is likely electrolytic caps, since ESR will drop when warm. I also suspected bad solder joints somewhere, maybe BGA. I connected the guts on the bench in my cold basement. In that cold, it wouldn't warm up enough by itself to work. I used a hair dryer to selectively warm parts, shielding others using a cardboard/aluminum foil shield. Confusing results sometimes suggested a group of caps and sometimes the big chip with the heatsink, the Samsung ARM S5H2201A01. I decided to try to reflow the solder balls on that chip. I removed the heat sink and shielded everything but the chip with foil. Since I can't spare much money for this and don't have the proper tools, I used my propane torch, carefully, using a piece of solder as crude temp probe and avoiding putting flame directly on chip.
No difference. So I did some more careful testing with better heat isolation, putting snow in a little metal cup and putting it on the chip while I heated caps, or putting it on caps while heating chip, all while isolating heated area with special shields. This gave more consistent results, strongly suggesting a group of caps near that chip. Using a soldering iron with intermittent touching to try to heat one cap at a time yielded confusing results.
I ordered replacements for all electrolytic caps on that board. By selecting cheap Wurth radial caps, I could replace all 130-ish of them for around $28. I cut the leads as short as possible while allowing my iron room underneath, making little bent feet for the solder pads. Since I was concerned that these caps should probably be stabilized, I used some GE Silicone II caulk to attach groups of them together and some to the board. I waited 48 hours for the caulk to cure and then tested the board. EVERYTHING WORKED PERFECTLY! The TV started right up, even in the cold basement, and stayed on with no problems, repeatably.
I reassembled the TV and put it back in service. It was used for part of a day and then suddenly, it froze up, making a loud buzzing noise through the speakers, and quit. Now it won't start up at all, even with some heat. I don't get any error LED patterns. The green LEDs light up on the mainboard and LED driver board as they should when all is well. The green LED on the front of the TV blinks once per second a number of times and then it just gives up.
Before replacing caps on the mainboard, I seem to remember measuring 69 ohms of resistance between 1.3VB (feeds part of big chip with heatsink) and ground. I thought this seemed low, but decided that I don't know enough and that it is probably normal. Now, across the same, I measure 5.4 ohms. I also measure 17.8 ohms between 3.3VB and ground, another power line that feeds part of the same chip. Do you suppose something is shorted? I have scoured the board with a magnifying glass for solder bridges, scratching with a tool along the gaps around the solder pads, blowing the board off with an air compressor, cleaning it all with isopropyl, all to no avail.
I have begun to desolder the caps I put on to test them to see if one might have failed. I don't know how else to proceed. I am also concerned that maybe that big chip failed as a result of the torching I had done earlier. What do you think?
This is an LED DLP TV with no lamp or color wheel. It has big R, G, and B LEDs for the light engine. It has seen regular use since late 2007.
It started taking longer to warm up before coming on. Later, it began doing glitchy things before warmed up. Sometimes, I'd get audio, but no video. Sometimes, the screen would get scrambled with odd patterns and there'd be a loud buzz from the speakers. It would reset itself after a bit. Other times, the audio would be higher pitched, making human voices like The Chipmunks, then lock up, then reset. Sometimes, it would display video and play sound but be unresponsive to commands. But after warming up for a bit, and after two or three resets, it would work fine and continue to work fine until turned off and cooled down. It reached a point where before warm, it would just sit there with the green LED on the front blinking once per second for a while and then would give up, never displaying a picture.
I read online of many occasions of success replacing bad caps on power boards. I found a row of five bulged caps on the sub power board. I replaced ALL electrolytic caps for both the power boards, over twenty. No difference.
After some troubleshooting, I decided that the mainboard was suspect. It has many electrolytic caps on it in V-chip/SMD packages. I figured that since the TV works when warm, the problem is likely electrolytic caps, since ESR will drop when warm. I also suspected bad solder joints somewhere, maybe BGA. I connected the guts on the bench in my cold basement. In that cold, it wouldn't warm up enough by itself to work. I used a hair dryer to selectively warm parts, shielding others using a cardboard/aluminum foil shield. Confusing results sometimes suggested a group of caps and sometimes the big chip with the heatsink, the Samsung ARM S5H2201A01. I decided to try to reflow the solder balls on that chip. I removed the heat sink and shielded everything but the chip with foil. Since I can't spare much money for this and don't have the proper tools, I used my propane torch, carefully, using a piece of solder as crude temp probe and avoiding putting flame directly on chip.
No difference. So I did some more careful testing with better heat isolation, putting snow in a little metal cup and putting it on the chip while I heated caps, or putting it on caps while heating chip, all while isolating heated area with special shields. This gave more consistent results, strongly suggesting a group of caps near that chip. Using a soldering iron with intermittent touching to try to heat one cap at a time yielded confusing results.
I ordered replacements for all electrolytic caps on that board. By selecting cheap Wurth radial caps, I could replace all 130-ish of them for around $28. I cut the leads as short as possible while allowing my iron room underneath, making little bent feet for the solder pads. Since I was concerned that these caps should probably be stabilized, I used some GE Silicone II caulk to attach groups of them together and some to the board. I waited 48 hours for the caulk to cure and then tested the board. EVERYTHING WORKED PERFECTLY! The TV started right up, even in the cold basement, and stayed on with no problems, repeatably.
I reassembled the TV and put it back in service. It was used for part of a day and then suddenly, it froze up, making a loud buzzing noise through the speakers, and quit. Now it won't start up at all, even with some heat. I don't get any error LED patterns. The green LEDs light up on the mainboard and LED driver board as they should when all is well. The green LED on the front of the TV blinks once per second a number of times and then it just gives up.
Before replacing caps on the mainboard, I seem to remember measuring 69 ohms of resistance between 1.3VB (feeds part of big chip with heatsink) and ground. I thought this seemed low, but decided that I don't know enough and that it is probably normal. Now, across the same, I measure 5.4 ohms. I also measure 17.8 ohms between 3.3VB and ground, another power line that feeds part of the same chip. Do you suppose something is shorted? I have scoured the board with a magnifying glass for solder bridges, scratching with a tool along the gaps around the solder pads, blowing the board off with an air compressor, cleaning it all with isopropyl, all to no avail.
I have begun to desolder the caps I put on to test them to see if one might have failed. I don't know how else to proceed. I am also concerned that maybe that big chip failed as a result of the torching I had done earlier. What do you think?
Comment