Philips 50PF9431D/37 Chassis BJ2-5UPA

[PlainBill]

LONG!! Feel free to skip to "AT LAST, The problem:"

This is yet another Craigslist special. When power is applied the relays click several times and the error LED starts flashing error code 12 - problem with the 12 volt supply. After pulling out over 50 screws, I was able to verify the 12V supply came up to only 6.8 volts. Time to pull the power supply.

This supply should be a testament to bad design. Not only are there caps adjacent to heat sinks, there are caps UNDER heat sinks. There are even a few that would require removing the heat sink to remove the cap. Of course, schematics are not available.

After tracing the 12V output back to it's source, I found it was a fairly simple design. Rectified and filtered line voltage is fed to a SMPS which drops it to about 13.5 volts (Vcc). Vcc supplies three loads - the Ambilight inverters (not present on this chassis), a bucking regulator to produce a 5.6V supply, and a low drop-out linear regulator to produce the 12 Volts. All caps in the Vcc supply and the 12V supply test good (ESR less than .05 ohm).

The 12 Volt regulator is a KIA378R12PI - a 12 volt, 3 amp regulator. This crosses to a Fairchild KA378R12C. This regulator is rated to work with as little as a .5V drop between input and output. Maximum power dissipation is 15 watts with a heat sink. The typical load in this set is 1.7 amps. It is used only on the small signal board as a supply for bucking regulators.

AT LAST, the problem: There are two possible explanations for the low output voltage - a failure in the regulator itself, or excessive load. There are no convenient jumpers to lift so the current draw can be checked. Does anyone have any suggestions for measuring the load? The regulator itself is available from Mouser at a very reasonable price (plus an unreasonable amount for shipping).

PlainBill

[Wizard]

Re: Philips 50PF9431D/37 Chassis BJ2-5UPA

Linear regulators fail by low voltage output as symptom. Is the supply voltage good, then replace the regulator. Ohm the load you should find this is not a issue.

Cheers, Wizard

[Krankshaft]

Re: Philips 50PF9431D/37 Chassis BJ2-5UPA

In this day an age why are they using linear regs anyway? It's understandable for Ma loads but when you get in the Amp range they're very wasteful of energy and produce alot of heat.

How much would it have really cost to wind an extra secondary on the switching transformer?

They even manufacture switching voltage regulators today which are much more efficient however they require a storage inductor which raises the part count and cost.

Then again as you mention if the engineers thought putting caps under a heatsink was a good idea this is just another crappy rebrand. Praying on the buyers perception of the original brands reputation.

If I have a suspect V reg I just remove it and hook it up to a variable bench supply dial in the original supply voltage and measure the regulated output. If I need a load I just install a resistor across it. Probably going to need at least 15 volts on a 12 volt V reg.

[PlainBill]

Re: Philips 50PF9431D/37 Chassis BJ2-5UPA

Linear regulators fail by low voltage output as symptom. Is the supply voltage good, then replace the regulator. Ohm the load you should find this is not a issue.

Cheers, Wizard

Thanks for mentioning the failure mode of linear regulators. You just focused my thoughts. Here's the game plan:
1. Hook up the scope and verify the input to the linear regulator is stable. if it is,
2. Pull the regulator and hook it up to a 15 volt supply. Throw a 10 ohm load across the output. If the output is anything other than 12V, replace the regulator.

I agree, the load is not likely to be an issue. According to the service manual, the regulators on the small signal board are disabled until the 12V is stable.

PlainBill

[PlainBill]

Re: Philips 50PF9431D/37 Chassis BJ2-5UPA

In this day an age why are they using linear regs anyway? It's understandable for Ma loads but when you get in the Amp range they're very wasteful of energy and produce alot of heat.

How much would it have really cost to wind an extra secondary on the switching transformer?

They even manufacture switching voltage regulators today which are much more efficient however they require a storage inductor which raises the part count and cost.

While ordinarily I would agree with you, this is a LDO regulator - full load regulation is maintained with a voltage drop as low as 0.5 volts. In this case, the output current is 1.7 A, the voltage drop across the regulator is about 1.5 volts. Total dissipation less than 3 watts.

One of the other loads on the SMPS is the inverters for the Ambilight package. As a matter of fact, this regulator seems to be slipped into the circuit to isolate the 12V supply from the Ambilight load.

Then again as you mention if the engineers thought putting caps under a heatsink was a good idea this is just another crappy rebrand. Praying on the buyers perception of the original brands reputation.

Well, in this case, the panel and power supply are both Samsung products.

If I have a suspect V reg I just remove it and hook it up to a variable bench supply dial in the original supply voltage and measure the regulated output. If I need a load I just install a resistor across it. Probably going to need at least 15 volts on a 12 volt V reg.

Great minds think alike. I'd already come to the conclusion that it was going to be necessary to pull the regulator; why not test it just to give myself a 'warm fuzzy feeling'.

PlainBill

[PlainBill]

Re: Philips 50PF9431D/37 Chassis BJ2-5UPA

Well, THAT didn't work out!!! In a rush to get this working, I checked the input voltage to the regulator with my scope. It looked like a clean 14 volts, so I ordered a new regulator WITHOUT pulling and testing the old one. It came in yesterday and I put it in, crossed my fingers, and plugged in the power. Same problem, 6.6 volts out for about 1 second, power supply turned off, red LED flashes error code 12!!!

Next I take the old regulator and hook it up to a bench supply with a load resistor. Sure enough, the output is about .5 volts below the input until the input goes over 12.5 volts, then the output stays at 12.0 volts.

Next step, check the voltage on the new regulator. What!!! Output goes to 12V momentarily, then shuts down. Error code is now 8 - no 1.2 volts. Remove Small Signal Board cover, sure enough, fuse 1U01 is now open.

Now to figure out if the problem is the 1.2V, 2.5V, or 3.3V supply. At least, they appear to have current sensing resistors, and there is a chart in the service manual showing expected load on each of them.

PlainBill

[RodmanG]

Re: Philips 50PF9431D/37 Chassis BJ2-5UPA

Well, THAT didn't work out!!! In a rush to get this working, I checked the input voltage to the regulator with my scope. It looked like a clean 14 volts, so I ordered a new regulator WITHOUT pulling and testing the old one. It came in yesterday and I put it in, crossed my fingers, and plugged in the power. Same problem, 6.6 volts out for about 1 second, power supply turned off, red LED flashes error code 12!!!

Next I take the old regulator and hook it up to a bench supply with a load resistor. Sure enough, the output is about .5 volts below the input until the input goes over 12.5 volts, then the output stays at 12.0 volts.

Next step, check the voltage on the new regulator. What!!! Output goes to 12V momentarily, then shuts down. Error code is now 8 - no 1.2 volts. Remove Small Signal Board cover, sure enough, fuse 1U01 is now open.

Now to figure out if the problem is the 1.2V, 2.5V, or 3.3V supply. At least, they appear to have current sensing resistors, and there is a chart in the service manual showing expected load on each of them.

PlainBill

Hi IM new here, But I do know what the problem is with that tv. There two 3300uf 10v 110c caps in that power supply. Here is a link, to video that will show you right where they are. Let me know if this works for you.

http://www.youtube.com/watch?v=OsJSz...eature=related

[PlainBill]

Re: Philips 50PF9431D/37 Chassis BJ2-5UPA

Hi IM new here, But I do know what the problem is with that tv. There two 3300uf 10v 110c caps in that power supply. Here is a link, to video that will show you right where they are. Let me know if this works for you.

http://www.youtube.com/watch?v=OsJSz...eature=related

Actually, you are wrong. There are no 3300uF, 10V caps. This is a later generation design of the power supply. In this case, the power supply is fine, the problem is on the Small Signal Board (SSB).

The 12V supply is the source for two DC-DC converters. The converter outputs are 1.25V and 3.3V; the 3.3V converter also is the source for the 2.4V supply. The problem is on the 1.25V bus. The dual FET in the DC-DC converter was shorted; the service manual was quite helpful with identifying that. The problem appears to be a (roughly) 40 ohm short on the 1.2V bus. It's not the 22uF ceramic cap in the DC-DC converter, it may be one of the other two dozen caps on the bus, or one of the ICs may be shorted.

NOW if you have any ideas on identifying the faulty part OR on obtaining a working SSB, I would really appreciate it.

PlainBill

[RodmanG]

Re: Philips 50PF9431D/37 Chassis BJ2-5UPA

Actually, you are wrong. There are no 3300uF, 10V caps. This is a later generation design of the power supply. In this case, the power supply is fine, the problem is on the Small Signal Board (SSB).

The 12V supply is the source for two DC-DC converters. The converter outputs are 1.25V and 3.3V; the 3.3V converter also is the source for the 2.4V supply. The problem is on the 1.25V bus. The dual FET in the DC-DC converter was shorted; the service manual was quite helpful with identifying that. The problem appears to be a (roughly) 40 ohm short on the 1.2V bus. It's not the 22uF ceramic cap in the DC-DC converter, it may be one of the other two dozen caps on the bus, or one of the ICs may be shorted.

NOW if you have any ideas on identifying the faulty part OR on obtaining a working SSB, I would really appreciate it.

PlainBill

well that was what was wrong with mine. I was trying to help.. Mine was a Philips 50PF9731D/37. I looked and saw the two caps were bubbled on the top of them that`s how I found the ones that were bad. Then I found that video on youtube.

[PlainBill]

Re: Philips 50PF9431D/37 Chassis BJ2-5UPA

well that was what was wrong with mine. I was trying to help.. Mine was a Philips 50PF9731D/37. I looked and saw the two caps were bubbled on the top of them that`s how I found the ones that were bad. Then I found that video on youtube.

Interesting. That TV is the deluxe version of the one I am working on. The power supply in this one is an LJ44-00125A. Do you have the LJ44-00118A?

PlainBill

[Wizard]

Re: Philips 50PF9431D/37 Chassis BJ2-5UPA

Is your regulator seperate design (controller and MOSFETs or just all in one IC?) Ones I see usually have all in one buck reguator with controller in one IC for each voltage output or 16 pin regulator controller IC driving team of MOSFETs for different voltages? What is the numbers on the IC and if any MOSFETs? If I find a shorted MOSFETs I tend to replace the controller along with MOSFETs.

Notice I said regulator which means all in one IC. Also notice different meaning for regulator controller and MOSFETs which is seperate pieces.

40ohms on 1.25V is good,
200ohms on 3.3V is also good.

I used ohm law formula to get current then watts, keep in mind current is for STATIC state, current will go up when active. These stuff on mainboard tend to be way under 150mW to 800mW (if has heatsink installed).

Cheers, Wizard

[PlainBill]

Re: Philips 50PF9431D/37 Chassis BJ2-5UPA

Is your regulator seperate design (controller and MOSFETs or just all in one IC?) Ones I see usually have all in one buck reguator with controller in one IC for each voltage output or 16 pin regulator controller IC driving team of MOSFETs for different voltages? What is the numbers on the IC and if any MOSFETs? If I find a shorted MOSFETs I tend to replace the controller along with MOSFETs.

Notice I said regulator which means all in one IC. Also notice different meaning for regulator controller and MOSFETs which is seperate pieces.

40ohms on 1.25V is good,
200ohms on 3.3V is also good.

I used ohm law formula to get current then watts, keep in mind current is for STATIC state, current will go up when active. These stuff on mainboard tend to be way under 150mW to 800mW (if has heatsink installed).

Cheers, Wizard

This design uses a regulator controller (NCP5422AD dual synchronous buck converter) and a pair of 8 pin dual mosfets (SI4944DY). I wasn't sure if the 40 ohm resistance on the 1.25V bus represented a fault or not, and I'm sure I don't want to pull the SSB in the 42PF7220 in my wife's office for comparison purposes (the schematic is identical).

PlainBill

[Wizard]

Re: Philips 50PF9431D/37 Chassis BJ2-5UPA

Years ago when intel had 8088, 286, 386 and chipsets they were under 100mA to 500mA to keep heat envelope within specs. Also another reason to keep cost down, heatsinks costs too much and Philips is too cheap, I hardly see one. And chipsets still have power limitations especially in TVs. This is where I complain about poor A/D quality and processing due to limitation on how powerful chipsets can be to allow quality processing for analog 480i via any types of analog inputs to make a clean capture and display on flat panel LCDs. HD is already compressed in MPEG2 format and it is easy for chipset to uncompress and do very *little* processing.

Like I said, 40 ohm is about right for extremely low voltage applications (1.25V), This is in order to keep current low - therefore less watts. 31.25mA static load, but mA goes up to say 250mA or bit more with active ICs that use 1.25V. I think this 1.25V is used for LVDS voltage probably. But I usually see 1.8V for this. Read on next comments.

Replace NCP54422AD and MOSFETs. If fault craters that dual channel MOSFETs, either a short or bad controller. I had to replace controller and MOSFETs in a notebook.

Other option is take out the MOSFET and supply 3.3V and 1.25V via external supplies and test run the mainboard. I occasionally do this.

Cheers, Wizard

[PlainBill]

Re: Philips 50PF9431D/37 Chassis BJ2-5UPA

Years ago when intel had 8088, 286, 386 and chipsets they were under 100mA to 500mA to keep heat envelope within specs. Also another reason to keep cost down, heatsinks costs too much and Philips is too cheap, I hardly see one. And chipsets still have power limitations especially in TVs. This is where I complain about poor A/D quality and processing due to limitation on how powerful chipsets can be to allow quality processing for analog 480i via any types of analog inputs to make a clean capture and display on flat panel LCDs. HD is already compressed in MPEG2 format and it is easy for chipset to uncompress and do very *little* processing.

Like I said, 40 ohm is about right for extremely low voltage applications (1.25V), This is in order to keep current low - therefore less watts. 31.25mA static load, but mA goes up to say 250mA or bit more with active ICs that use 1.25V. I think this 1.25V is used for LVDS voltage probably. But I usually see 1.8V for this. Read on next comments.

Replace NCP54422AD and MOSFETs. If fault craters that dual channel MOSFETs, either a short or bad controller. I had to replace controller and MOSFETs in a notebook.

Other option is take out the MOSFET and supply 3.3V and 1.25V via external supplies and test run the mainboard. I occasionally do this.

Cheers, Wizard

That's an idea that briefly crossed my mind, but was rejected. According to the service manual, the set pulls 2.9A off the 3.3V supply, plus another 1A off the 2.5V supply which is derived from the 3.3V supply. And the 1.25V supply supplies another 2.6A. I just don't have bench supplies that can put out that kind of current.

PlainBill

[Wizard]

Re: Philips 50PF9431D/37 Chassis BJ2-5UPA

2.9A on 3.3V is just 10W. And 1A on 2.5V is 2.5W and you subtract this from 10W to get actual wattage which is 7.5W for 3.3V and 2.5W for 2.5V loads.

1.25V at 2.6A is 3.25W. Add up watts from 3.3V but not the wattage for 2.5V downstream of 3.3 that draws from this for 2.5V as it is already totalled up on 3.3V wattage. That's 13.25W total at that regulators from one voltage feed and we didn't concern ourselves with losses in the 3 regulators train.

I bet this main board does have heat sinks attached to chip sets this time or multiple chip sets (older design).

13.25W is lot of heat to dissipate. That equals to 3"x3" pin fin heat sink 1" tall to passively cool to get the how much heat had to be radiated off. But this main board have heavy ground & power planes this is direct connected to chip set dies via copper vias and BGA connections in center of each BGA chip sets so whole board get rather warm.

Get this controller and MOSFETs replaced, this should fix the problem. 2.9A on 3.3V would be very low ohms than 200ohms (as active load) which is dynamic ohm of 10ohms. 1.25V ditto, that's dynamic ohm as viewed from power supply load as 3.25 ohms.

Cheers, Wizard

[PlainBill]

Re: Philips 50PF9431D/37 Chassis BJ2-5UPA

2.9A on 3.3V is just 10W. And 1A on 2.5V is 2.5W and you subtract this from 10W to get actual wattage which is 7.5W for 3.3V and 2.5W for 2.5V loads.

1.25V at 2.6A is 3.25W. Add up watts from 3.3V but not the wattage for 2.5V downstream of 3.3 that draws from this for 2.5V as it is already totalled up on 3.3V wattage. That's 13.25W total at that regulators from one voltage feed and we didn't concern ourselves with losses in the 3 regulators train.

I bet this main board does have heat sinks attached to chip sets this time or multiple chip sets (older design).

13.25W is lot of heat to dissipate. That equals to 3"x3" pin fin heat sink 1" tall to passively cool to get the how much heat had to be radiated off. But this main board have heavy ground & power planes this is direct connected to chip set dies via copper vias and BGA connections in center of each BGA chip sets so whole board get rather warm.

Get this controller and MOSFETs replaced, this should fix the problem. 2.9A on 3.3V would be very low ohms than 200ohms (as active load) which is dynamic ohm of 10ohms. 1.25V ditto, that's dynamic ohm as viewed from power supply load as 3.25 ohms.

Cheers, Wizard

Slightly strange things going on here. I had measured the resistance from the 1.25V bus to ground as about 40 ohms; it varied a little. I got the FETs and SMPS controller in today and soldered them onto the board, then checked the resistance again. It was about 30 ohms. The next step was to solder in the 3A fuse, THAT was a bear; they didn't leave much room between it and the inductor, but I finally got it in. Clean up the flux, inspect the board under the magnifier, and gave myself a C- for soldering (I always grade low, it encourages improvement).

So mount the completed (notice, I didn't say repaired), hook it up, double check the wiring, and plug in the power. The relays on the power supply click twice, then nothing. I'm getting discouraged, then the gray cells coalesce. The green LED on the front panel is still on!!! I dig up the remote, hit Menu, and the menu comes up!!! Son of a gun, we did it!!

This is FUN!!!

PlainBill

[tmwalsh]

Re: Philips 50PF9431D/37 Chassis BJ2-5UPA

PlainBill,
Congrats! You don't want those gray cells to coalesce TOO much. Then there wouldn't be room to put any more stuff in there, no?

I had the same 'this is fun!!' moment last week when the set I had worked on since March(?) finally gave it up and started working again. Even though I had caused part of the problem (gray cells DIDn't coalesce??) through ignorance and even more ignorance.
tom

[Wizard]

Re: Philips 50PF9431D/37 Chassis BJ2-5UPA

Yay!

That take awhile but you got it!

Cheers, Wizard

[PlainBill]

Re: Philips 50PF9431D/37 Chassis BJ2-5UPA

Yay!

That take awhile but you got it!

Cheers, Wizard

Wizard, I hope you realize your advice was a big part of getting this working, and I appreciate it.

NOW, one of the little 'What if's' that has been rattling around is a set that I passed on. These Philips sets allow the user to update the software by downloading the latest version from Philips and putting it on a thumb drive. The doofus that was selling it had tried to update the software, but unplugged the thumb drive before the update was complete. The result was a brain dead TV. Have you ever seen a way to recover from that screw-up?

PlainBill

[Wizard]

Re: Philips 50PF9431D/37 Chassis BJ2-5UPA

Welcome!!

Mainboard replacement or find another mainboard from junker board that uses same board (only difference is different panel used and is reset to different panel by remote control (punching in option number followed by 6 digits). Back to this, copy this eeprom contents and flash onto the corrupt eeprom. No fun.

That doofus should WAIT for it to finish.

Cheers, Wizard