We all know how cheap and cruddy these things are--But Has anyone tried doing some proper upgrading throughout the entire set to improve reliability?
I have one of these things, and picture wise--it aint bad. Its a PS-42C450, not the most reliable beast in the world from the number Ive seen and heard about!
My plans for upgrades are as follows.
Replace the BN44-00329B PSU with one from a 50" set,--The BN44-00330B.
--Its pin for pin/plug-n-play compatible. The 50" PSU has a larger Transformer, and better heatsinks fitted so will be less stressed in comparison to the 40" original--which runs--Pretty Darned Hot!
Before fitting however--That one needs some work. Replace all those nasty cheap caps with Rubycon/Chemicon/Nichicon, of the correct series for their jobs.
Re-work the standby supply regulator to remove the issue whereby the 30V rail derived from the switching transformer by two 15V zeners--In Series, the regulator chip runs from--kills the 32V rated chip!
--Seen a few of these--so something is obviously causing this problem, giving a symptom of completely dead--no life at all-set.
Sanken the maker of the chip specify 'Absolute Maximum VCC of 32V' guessing so as not to stress out the internal 5V reg the chip has. The chip's minimum running voltage is 8.5V
The Max power-output this chip can supply (assuming the transformer was rated for that amount) is 90W. Its actually supplying only 40 odd watts when the set is on, and just a couple in stby. Stress due to output power overload isn't the cause of the problem.
--Similar PSU's like say--the '329 using a similar chip with similar ratings from same maker, run it from 18V using a single zener diode as its reference.
Oh--and replace the original STR-W6053S switching-regulator chip for a new unstressed one.
Up-value some specific caps, particularly those of the Vs, Va, 5V and 15V rails, add some Os-Cons to the 5V and 15V rails.
The Vs decoupling cap of the '330B PSU is larger anyway at 220uF instead of the 150uF of the '329B, prolly go up to 330uF for that one.
Replace all the nasty leccy-lytics of Y-Main and X-Main, and uprate the 150uF 200V Vs de-coupling to 200uF 220V Rubycon BXC-P caps.
(You can almost hear the bean-counters saying--Capacitance costs money boys--Cut it down to bare minimum! hence my increasing of some in the set.)
Increase cooling of the Y-Main and X-Main by adding heatsinks to the existing and a small fan to assist in cooling.
Add a case fan to the set to extract air from the top--as certain Panasonic sets do. I have some excellent virtually silent fans culled from a wine-chiller that will come in handy for this....
Increase the size of the heatsink fitted to the main-board's processor, that is known to cause problems due to temperature issues.
Anything else you guys think would be a good idea?
I have now Three complete sets of working boards for this set, (and 4 PSU's) so all would not be lost if it went titties upwards!
--I can re-work the PSU and fully bench-test it under lamp-loads at full power, set the rails up etc run it for some hours before fitting to the set, --one board then I know will be fine!
I have one of these things, and picture wise--it aint bad. Its a PS-42C450, not the most reliable beast in the world from the number Ive seen and heard about!
My plans for upgrades are as follows.
Replace the BN44-00329B PSU with one from a 50" set,--The BN44-00330B.
--Its pin for pin/plug-n-play compatible. The 50" PSU has a larger Transformer, and better heatsinks fitted so will be less stressed in comparison to the 40" original--which runs--Pretty Darned Hot!
Before fitting however--That one needs some work. Replace all those nasty cheap caps with Rubycon/Chemicon/Nichicon, of the correct series for their jobs.
Re-work the standby supply regulator to remove the issue whereby the 30V rail derived from the switching transformer by two 15V zeners--In Series, the regulator chip runs from--kills the 32V rated chip!
--Seen a few of these--so something is obviously causing this problem, giving a symptom of completely dead--no life at all-set.
Sanken the maker of the chip specify 'Absolute Maximum VCC of 32V' guessing so as not to stress out the internal 5V reg the chip has. The chip's minimum running voltage is 8.5V
The Max power-output this chip can supply (assuming the transformer was rated for that amount) is 90W. Its actually supplying only 40 odd watts when the set is on, and just a couple in stby. Stress due to output power overload isn't the cause of the problem.
--Similar PSU's like say--the '329 using a similar chip with similar ratings from same maker, run it from 18V using a single zener diode as its reference.
Oh--and replace the original STR-W6053S switching-regulator chip for a new unstressed one.
Up-value some specific caps, particularly those of the Vs, Va, 5V and 15V rails, add some Os-Cons to the 5V and 15V rails.
The Vs decoupling cap of the '330B PSU is larger anyway at 220uF instead of the 150uF of the '329B, prolly go up to 330uF for that one.
Replace all the nasty leccy-lytics of Y-Main and X-Main, and uprate the 150uF 200V Vs de-coupling to 200uF 220V Rubycon BXC-P caps.
(You can almost hear the bean-counters saying--Capacitance costs money boys--Cut it down to bare minimum! hence my increasing of some in the set.)
Increase cooling of the Y-Main and X-Main by adding heatsinks to the existing and a small fan to assist in cooling.
Add a case fan to the set to extract air from the top--as certain Panasonic sets do. I have some excellent virtually silent fans culled from a wine-chiller that will come in handy for this....
Increase the size of the heatsink fitted to the main-board's processor, that is known to cause problems due to temperature issues.
Anything else you guys think would be a good idea?
I have now Three complete sets of working boards for this set, (and 4 PSU's) so all would not be lost if it went titties upwards!
--I can re-work the PSU and fully bench-test it under lamp-loads at full power, set the rails up etc run it for some hours before fitting to the set, --one board then I know will be fine!
Comment