sharp lc32sb24u bulb / inverter problem

[cashkennedy]

I picked up this sharp lc-32sb24u and it has a flashing green light seems to be an even pace of 1 flash per second. Holding down the input and volume down while plugging it in resets the error and it will start and run for about 3 seconds.

I consulted the manual for this model which says to just read the manual for lc-32d44u. So in that manual it says that 1 green light flash is an error with the bulb.

Normally after knowing the bulb / inverters are probably the issue i would see if i could disconect each bulb seperately and test the tv with 1 disconected at a time, but if you look at the picture of this unit the inverter is screwed directly into connectors for the bulbs, so im not sure if there is any easy way to disconect 1 at a time.

Second thing to check is if the windings have the same resistance on each inverter, which im going to try and check, but im not sur eif the secondarys have any exposed pins considering the way they go through the bottom of the board and connect straight to the bulb connectors.

[cashkennedy]

Well actually i just looked at it, to try to test the inverters, and they are infact quite easy to get to, since they are the spools of copper wire next to the black connectors.

So there are 5 inverters, at first when i measured the secondaries all were around 2.58 kohm, although one is 2.65 so slightly above the rest. But then after testign the primaries (which all have around 0.5 ohms resistance) i retested the secondaries and the bottom most inverter reads open. Occasionally id get a few seconds of a reading, but no matter how good of a connection i tried to make i couldnt get it to consistently read 2.58 ohms. So im guessing the problem is with that inverter.

I tried flexing (the board has a LOT of flex ability) the board so that it wouldnt be connected to the bottom bulb, but all times i try it like that the tv wont clear the flashing green error, so i guess it isnt dumb enough to start with a bulb unplugged.

So what do yall think? Bad inverter?

Oh also, ive never worked on bulb inverters before, so im not sure if im choosign the right places to read the windings from, but these ones have 5 poles on each side of each transformer. when i measure the primary poles 1, and 2 on the top, and 1 and 2 on the bottom are all 0.5 ohms from each other. Poles 3 on top and bottom, are open to all the other poles so i think are not being used. Pole 5 on top and bottom is where i measured the secondardy on each one and got 2.58 kohms on most.

[budm]

Can we see closed up pictures of the transformer top of the board and bottom of the board?

[cashkennedy]

Yep here are some pics, tell me if other would help.

I numbered the pins from left to right when looking at the front of the board (yellow side). And just from looking at the backside i can see now that the pin 3 is neither the primary nor the secondary but appears to have some sort of management function, as well as pin 4 appears to possibly be what monitors the ouput and reports the bulb failures.

Pins 5 (top and bottom), which i was testing as the secondary, are what i got 2.58 kohms for all but the 1 that for the most part is open.

And pins 1, 2 on top and bottom of each have 0.5 ohms (so almost no resistance) (i tested 1 top to 1 bottom, 1 top to 2 bottom, 1 top to 2 top and all were the same )

[cashkennedy]

Any idea where to look for a replacement transformer / inverter like the one in the picture (z1358 08086).

Theres a almost exact match for the whole power inverter board for 19.99 on ebay, so i might just check if that guys board has the right inverters. (The cheapest exact match for the power inverter board is 60 though and from a bad seller)

[budm]

Try www.lcdparts.net, but if you can get the new board for $20, that will be a lot better to have spare transformers and other parts. One transformer plus shipping will be $12~20.

[retiredcaps]

Originally Posted by budm

Try www.lcdparts.net, but if you can get the new board for $20, that will be a lot better to have spare transformers and other parts. One transformer plus shipping will be $12~20.

We had one member buy an entire inverter board (iirc, around $20) just to get 2 replacement transformers because it was cheaper than buying 2 separate transformers ($24.99 each) from other vendors.

[cashkennedy]

I received the used / non working similar inverter and it did infact have the same transformers (it appears every sharp for like a year used basically the same mainboard and very similar power supply / inverter layouts.

I realized the bad transformer just had the winding broken off at the pin, so fixed my existing transformer and it tested as 2.68 kohm, which is a tiny bit high but within 4% of the others now. So i didnt really need to buy the other inverter from ebay.

Anyway it didnt work still after fixing that. As a recap after a error reset the screen comes on for 2 seconds and i can see the channel 2 display at top. Then it turns off and flashes a green light once per second (which means inverter problems).

My current question : Im trying to test the DC/AC IC that controls the FET's for the backlight (it might also control the error reporting?). The part number was melted on the one on my existing board (not a good sign), so luckily i had the other board to read the part number off of, and it is BD9883.
http://www.alldatasheet.com/datashee.../BD9883AF.html
Is there any good method to test this IC? I did verify that all the pins went to the parts described and all seem to have good contact and go where they are meant to go.

As a side note, with the other "untested / not working" inverter, it doesnt turn on the backlights and just flashes 2 green lights (some other error)

[cashkennedy]

Ive attached 3 pictures. The first is a view of the whole back of the board. The largest IC with 3 pins a regulator, the 10 pin is the AC/DC converter IC, and im not sure what the small IC does. The design of this inverter doesnt lend itself to testing very well since all the transformers are in parallel.

There are 2 groups of 2 FETS which provide the 2 AC inputs for the transformers and each group is gated by the AC/DC IC. The FET's appear to convert 13v power to the AC voltage. I have tried testing the AC voltage while the TV is on (for about 5 seconds), but do not get any readings. But the TV does light up, so obviously its an error in how im trying to measure it.

I also included a picture of the connector, because im wondering if theres any way to fool the err pin? (I assume its what reports the error to the main board), Like would cutting the wire work, or does the err pin transmit a steady voltage to the main board to prove everything is working.

[tom66]

Do those error pins, or OFL/OSC show any change in voltage after ~2 secs?

[budm]

You can try lcdparts.net
Does the bulb has any sign of bad indication like blackened ends, pinkish glows?

[cashkennedy]

No it didnt really seem to be any worse then the others (they all have a little bit of black near the edges, maybe 3 mm worth or so).

But since they are in series with another bulb and all the other bulbs are working i figured it was pretty likely that it is passing energy, just not starting. So I would assume it has a good connection and is receiving the proper current.

Earlier when i was getting annoyed at the unit i adjusted a rotating switch that had 1 - 0 - 3 marked on it and it was at 0 to begin with, i set it to 0 and could hear a flickering crackle from the center of the light box (approximately where the bab bulb is), then i tried 3 and it was no different then 0)

[budm]

I have never seen the bulb in series connection used in the TV LCD panel, each one is driven by the dedicated winding of the transformer, the the other end of the lamps should return to ground through a sensing network to tell the inverter IC if the lamp is fired up and stays conducting or if it draws too much current so the protection circuits will kick in.
You may want to check the cold side of the lamp assembly.

[cashkennedy]

I will double check when im infront of the TV again, but the wiring diagram clearly shows the transformers (there are 5) have 2 windings a positive and negative winding which each are connected to 1 bulb (there are 10 bulbs) and then on the other end of the bulbs theres a small crude board (which i didnt examine very cloesly, but might connect them to ground. There is deffinetly no way for them to be monitored where they are grounded, if they are grounded at 1 end.

What is your concern? because i tested all the secondary windings quite in detail and they are identical on all 5 transformers. specifically there are 4 pins on each transformer's secondary (2 per winding). And so if i number them as the service manual does with 7 and 8 on 1 windings and 9 and 10 on the other winding, and 8 and 9 are the pins that go to the bulbs, while 7 and 10 are connected by resistors and are what feeds into the monitoring parts.

measuring 9-10 and 7-8 i get 500 ohms, and 7-10 i get 1500 ohms, and 8 - 9 (bulb to bull) i get 2500 ohms (the sum of the other 3). I get that for all 5 transformers.

I could try swapping 2 bulbs and see if the one that doesnt light continue to be in the same place, or moves with the bulb. Also it would be very hard for me to notice a pink issue with the bulb when they only turn on for 2 seconds each time i reset the errors, and with all the other working bulbs around it.

I think ill order a bulb from http://www.ccflwarehouse.com/ as they seem to have cheaper shipping and many more lengths / sizes then lcdparts. Neither lists the start voltages for the bulbs though which doesnt seem a good sign.

[cashkennedy]

Attached is the page from the manual where they show the bulb connectors are + and - .

[budm]

When you have two lamps in series, if one is not lit, its impedance will be so high just like an open circuits which is not going to let the current to flow and lights up another lamp in series with it. CCFL has negative resistance, resistance goes down as when the lamp conducts and emits the lights.
The sensing can be done on the hot side also. Most uses the cold side since they will be all tie together then it can sense the total current draw of the lamps.
That will be new to me to have lamps in series and one is lit but not the other one. Will chalk one up in my library for this strange behaving that does not go with the electronics theories. That is why we are here to learn.
I wonder if that 1700v is the strike voltage or running voltage.

[cashkennedy]

There are a few samsungs that use U shaped bulbs (which might be a similar design to having 2 bulbs in series. Plus sharp is pretty bizzare with their design choices...

I get what you mean about it being harder to start either bulb if one is burned out. But maybe for the 2 seconds that the other bulb works the inverter is still providing the strike voltage, which has to be enough to overcome the resistance of off bulbs (since the whole point of strike voltage is to overcome them being cold). I have no clue how long the strike voltage is applied though (milliseconds? seconds?)

One website even had for sale stacking U shaped bulbs that each bulb was a different size ... whatever tv maker used that design should be shot though, it was probably toshiba, masters of pointlessly complex designs).

[budm]

Most inverter driver IC uses 1.5 second windows to detect if the lamp conducts or not.
The funny thing on your set is that the other lamps stays on and the dead one did not trip the shutdown circuits. Everyone does things in its own ways.

[cashkennedy]

You were correct that they are grounded on the other side (I confirmed this by removing the bad bulb, and the one i thought it was in series with still worked.

I then swaped the bulbs and the bulb i moved continued to be bad in its new position, so it looks like its just a bad bulb

Ordered one (3mm x 705mm) from ccfl warehouse, cant wait to get it.

[budm]

That is good to know, now that makes sense. Poor Tibimakai, he got the new lamp but he accidentally cracked the screen while putting it back together, poor guy.