EEVblog Electronics Community Forum
Electronics => Beginners => Topic started by: tcz on April 09, 2021, 08:02:55 pm
-
Hello,
I'm trying to repair my Dell G2410T LCD monitor. The power light turns on green but the screen is completely black. Sometimes it's just the backlight and you still see a dark image but it didn't seem to be the case.
So far I opened up the monitor and took the electronics out.
I found:
-A larger power supply board roughly 6 inch * 6 inch
-A smaller board with all the video outputs on
-A even smaller board which I believe drives the backlit... forgot what those were called.
At first sight everything looks good, nothing is burned and all the caps have a flat top.
Using a TC1 tester (one of those cheap lcr esr tester) I tried to test the Caps in circuit... when that failed :palm: I undersold all the caps on the power supply and tested them individually.
I got the specs from the manufacturer to compare the results but I must admit: I don't quite know what to make of my readings, when are the values too far off that they should be replaced?
In the attached image the first COLUMN is the manufacturer's specifications, the others are individual caps grouped by Type.
[attach=1]
I could use some pointers.
Thank you for reading.
-
For some reason I'm unable to modify my post so here is some extra info as a reply:
Turns out the small PCB is the Inverter, one of the stickers on the back has yellowed. Maybe it got slow cooked over time, not sure if it indicates a problem.
I attached extra pictures of the FRONT and BACK of each boards.
Here is where I took the specs for the caps: https://su-scon.com/wp-content/uploads/2021/01/Su-scon-Catalog_2021.pdf
Thank you.
-
A couple of those capacitors look a little marginal, but nothing really jumps out at me as bad. Did you check the voltages of all of the power supply outputs? Often the correct voltage is printed on the board by the connector, but sometimes you have to guess.
-
No that was my next step, should have been the first... |O
I will re-solder everything and test voltages.
-
Back light Driver boards are usually the first things to go on LCD and LED displays.
The small board is the LED driver using the OZ9967 OZ99XX (https://datasheetspdf.com/pdf-file/731978/O2Micro/OZ9967/1) . See page 28. Check the power on the 7 pin connector with 6 wires. Black and possibly Brown will be common grounds and/or reference grounds . Red and possibly Orange will likely be power wires. Yellow and Green are likely to be signal wires Dim and power On/Off . Check those voltages . If those look good check the six transistors on the driver board . I can't see the numbers on those . Check for voltage at the Base of those transistors . If no voltage then the Driver is either not turning on or broken . Trace pin 19 on OZ9967 back to the 7 pin connector and check for signal voltage when power on and power off . If that DOES NOT work the problem is likely on the control board. If you get a signal on the on/off then the driver OZ9967 is not functioning. Careful not to short any pins or connections .
Also check F101 (fuse) top left black resistor looking thing with a 0 .
If the driver and everything else seems fine then check the Mosfet Q101 for possible failure.
Fixed spelling and an error! Please reread
-
I will test all of this over the weekend.
Thank you very much for all the info!
-
I tested the voltages going from the power supply to the inverter and the video output board.
All the 3.3v, 5V and 18V pins are at the proper voltage.
On those connectors there were pins labeled "off", "on/off" and "BRI". Not sure how to test those.
I will now try to troubleshoot the inverter board as described by Jwillis.
-
Here is a a close up of the OZ99xx chip and the transistors.
Are those holes on the transistors??? They all look the same except for one that doesn't seem to have a hole in the "dot" of the marking.
-
2T (http://www.taitroncomponents.com/Catalog/Datasheet/MMBT4403.pdf) General Purpose Transistor (PNP)
Can you make a diagram of the readings you got on the connector.
Pin 19 on the driver either goes high or low on turn on . Connect your DMM to the appropriate wire on connector that traces to pin 19 and see if it goes high or low on turn on . That's the On/Off . The data sheet doesn't show any voltages for that chip. I'm guessing it should go high when referenced from ground. Not familiar with that driver chip.
The remaining wire should trace to pin 23. I believe that is the brightness.This will control the output duty cycle of the driver.
Getting voltage readings on those transistors would be easier on a scope .
Q101 is the driver MOSFET and the six transistors sink current to ground via the LED strings. If that MOSFET fails or if there is no signal to the gate there will be no supply current to the six LED strings. If there is no signal to the gate then its the driver chip or the inductor has failed.
Basically work your way from problem (no lights) back towards the driver chip.
It's to bad the manufacture of that driver chip is so secretive and won't show typical operating values. Unless some one else has a better data sheet kicking around.
-
2T (http://www.taitroncomponents.com/Catalog/Datasheet/MMBT4403.pdf) General Purpose Transistor (PNP)
Can you make a diagram of the readings you got on the connector.
Pin 19 on the driver either goes high or low on turn on . Connect your DMM to the appropriate wire on connector that traces to pin 19 and see if it goes high or low on turn on . That's the On/Off . The data sheet doesn't show any voltages for that chip. I'm guessing it should go high when referenced from ground. Not familiar with that driver chip.
The remaining wire should trace to pin 23. I believe that is the brightness.This will control the output duty cycle of the driver.
Getting voltage readings on those transistors would be easier on a scope .
Q101 is the driver MOSFET and the six transistors sink current to ground via the LED strings. If that MOSFET fails or if there is no signal to the gate there will be no supply current to the six LED strings. If there is no signal to the gate then its the driver chip or the inductor has failed.
Basically work your way from problem (no lights) back towards the driver chip.
It's to bad the manufacture of that driver chip is so secretive and won't show typical operating values. Unless some one else has a better data sheet kicking around.
I will work on it.
In the mean time I connected my T1 tester to the 6 transistors and got those readings:
(They were all identified as BJT-PNP)
hFE= 39 39 40 40 39 39
Ube= 656mV 654mV 656mV 652mV 653mV 653mV
Ic= 6.0mA 6.0mA 6.0mA 6.0mA 6.0mA 6.0mA
I'm not sure what to compare those values up against, am I correct to assume I wouldn't get a reading if they were fried?
-
Can you make a diagram of the readings you got on the connector.
If you mean the connector for the cable coming from the power supply (6 wires, 7 pin connector):
Black 18v
Brown 18v
Empty
Red GND
Orange GND
Yellow BRI
Green On/Off
Both 18v test at the proper voltage (I was also surprised about the usual GND and Voltage colors been reversed but the markings on the PCB itself confirm this)
Getting voltage readings on those transistors would be easier on a scope
Unfortunately all I have is a 24mhz 8ch Logic Analyzer, a TC1 tester and few Multi-meters.
It's to bad the manufacture of that driver chip is so secretive and won't show typical operating values. Unless some one else has a better data sheet kicking around.
A repair manual for a Dell G2410T Monitor may also have valuable info but I couldn't find one of those either.
-
Well I am not sure (even surprised) WHY the folks
are calling a simple LED driver an "inverter" ..
Actually it is not.
Inverters are used on CCFL devices - not LEDs
LED based monitor drivers are just BUCK converters
with some protections.
Usually a peripheral component aged or dead shorted.
LEDs are mostly the culprits.
Why they call that INVERTER? no clue
Paul
-
I'm sure you can get help. But please start your own thread or it will be a mess in here.
-
You are totally right, but if you look at the pictures even Dell is calling it an "Inverter board".
It goes back to the first LCDs that used cathodes as their back light.
"Usually a peripheral component aged or dead shorted.
LEDs are mostly the culprits."
You mean some components on the "back light driver" board (aka inverter board) ?
Also, are you implying the LEDs in my LCD panels may be to blame?
This monitor has seen maybe a year or two of use only (even though it's an old model).
-
LED based monitor drivers are just BUCK converters
with some protections.
Boost converter usually.
-
Usually a peripheral component aged or dead shorted.
LEDs are mostly the culprits.
There are 6 individual LED strings . Although not impossible , it's unlikely all 6 strings failed at the same time . The strings can be tested simply by providing 18V in series with a resistor (1K should work) to the Drain of the MOSFET then grounding each string at the Pins on P103 connector momentarily . It would only need to have one LED in a string for that string to fail .
For the sake of simplicity and to minimize confusion I will call it a LED driver because that is what the main component does .
Both 18v test at the proper voltage (I was also surprised about the usual GND and Voltage colors been reversed but the markings on the PCB itself confirm this)
Good thing the colour of the wire won't make the electrons flow in a different direction.
The transistor readings you gave look fine . They're not shorted or open . MMBT4403L Data Sheet (https://www.onsemi.com/pdf/datasheet/mmbt4403lt1-d.pdf)
I found a schematic for a Toshiba . It's not exact but gives a general idea of whats this driver does . The Schematic shows different voltages , values and has to 2 MOSFET's .
The green wire should go to high to turn on the OZ9967 . The exact value I can't say but if you attach a DMM to the green wire and turn on the display it should give a different value than when the display is turned off . If that works we can rule out that part of the control board .
Then check for a voltage change at any one of the bases of the 6 transistors . If there is a voltage change between the ON state and the OFF state then that part of the driver chip is OK .
Check the MOSFET NIKOS P8008BD Data sheet (https://datasheetspdf.com/pdf-file/1094487/UNIKC/P8008BD/1)
If that is fine we need to go back to the Brightness. If the Modulation to pin 23 is not working then the duty cycle to the LED's will be at the lowest power or dark screen .That would be an issue with the control board . If that is working then check the signal to the gate of the MOSFET . Difficult to do with just a DMM because it's PWM (Pulse Width Modulation) .Easier if you check for a current change if you know how with out shorting the MOSFET out .
Trace the yellow wire on the Driver Board and the Power board . I can't see where it goes in the photos
The Attachment below is for A Toshiba display but the concept is the same .
-
Hello,
Good thing the color of the wire won't make the electrons flow in a different direction.
You never know, the wave function could collapses differently in another part of reality. Thankfully it doesn't on our blue marble ;D
Nice find for the schematics! I tried getting info from the manufacturer but they were quick in replying I didn't sign the NDA and asking were I got my chips.
So I have tested the ON signal using my DMM and Logic analyzer and there doesn't appear to be a signal. I used the red wire as ground (labeled GND on the connection going to the LED driver).
On the logic Analyzer I see no change, just a flat line all the way.
On the DMM I get 0V when the power is disconnected (a little higher actually but it drains to 0v after a few seconds), a steady 0.06v with power connected and when pressing the power button there is NO change.
So I guess I need to take a look at the control board? (That's the one with the video outputs right?)
I will start by probing around the cable coming from the power button assembly.
BTW thank you so much for your help navigating this!
-
Just tested for the ON signal on the wires coming from the control board (between the purple (on/off) and the green (GND) wires)
And I'm pretty much getting 0v all the way. I was expecting to get 0.06v but no.
Not being sure which GROUND I should be using (although what I know tells me all grounds should do) I measured again using other GND wires and the clearly labeled ground connection next to the power connection on the power supply.
Same results.
EDIT
I just found the signal for the power button on the ribbon cable going to the control board.
I attached an image showing the pinout.
When pressing the power button the blue pin gets shorted to the green pins which are both grounds on the control board. See image.
I'm now in the process of tracing the path of the "On" button signal to see were it goes, but there is a ton of tiny VIA in this area that makes things complicated.
-
Glad your progressing . Unfortunately your starting to get in the realm of digital chips. I can't even get a CD4013 flip-flop to work right . :-DD
All I can say is try to get the data sheets for the chips on board . Maybe they can give some direction of what's gone wrong . There has to be a pin on one of them that is for back light power on .
Wish I could help more .
-
I'm trying to repair my Dell G2410T LCD monitor. The power light turns on green but the screen is completely black. Sometimes it's just the backlight and you still see a dark image but it didn't seem to be the case.
if sometimes you got background light and no picture, then your problem is video signal path
no need to dig the LED driver for background light
green LED means the processor detected VGA signal. re-test connecting/disconnecting VGA cable or simply leave the PC in standby and move the mouse or shut dowa/start the PC, the LED should change state
IF it change state (the LED), that means processor board have at least some +5V voltage and detects VGA signal
normal repair means :
- verify voltages (at least with DMM)
- if voltages are OK and you got no scope, cut the apple in half by establishing if:
- it's the backlight LED driver defective or
- the LED background matrix
- the control board (ON/OFF from start button works, VGA signal detection works, VGA signal is correctly processed --> MAKE SURE you display on that LCD some supported
resolution/refresh rate, this is very important, you gotta setup output resolution of your device like a laptop first then connect your monitor, you can be out of monitor supported modes
range
you are a little lost in this steps, so re-focus like i proposed and rething the approach to diagnose
this 'organigram' flow works for every monitor from 1980 to today
-
Glad your progressing . Unfortunately your starting to get in the realm of digital chips. I can't even get a CD4013 flip-flop to work right . :-DD
All I can say is try to get the data sheets for the chips on board . Maybe they can give some direction of what's gone wrong . There has to be a pin on one of them that is for back light power on .
Wish I could help more .
You helped me a lot thank you. If I can be sure that the issue is the control board I can get a refurbished one for 26$, it may be worth it. Will look at schematics first if I can find it.
you are a little lost in this steps, so re-focus like i proposed and rething the approach to diagnose
this 'organigram' flow works for every monitor from 1980 to today
Thank you will do but FYI:
The power led is green even with no video cable connected.
It's not a computer resolution problem
I cannot get in the monitor's setup menu.
The ON/OFF from start button works but an ON signal never gets sent to the other boards.
Basically it wont turn on! |O
While I'm here I will post an updated picture of my trace investigation and some close up.
-
yep, more clear now
change the quartz or try reflowing or just press a little with your finger the proc and try on button then
i'm sure the proc gets on/off command, he's dead or stuck
if you're thorough in explaining what you did, we can diagnose, if you transmit half of your diagnose, we can't point the defect component
-
Not sure if this will help but I found a service manual for the LCD Monitor 2209WA which uses the TSUMU58EHJ-LF-2 . Not sure how much difference there would be between the chip versions . They are both QFP-128 packages https://www.manualslib.com/manual/794321/Dell-2209wa.html (https://www.manualslib.com/manual/794321/Dell-2209wa.html) free down load .
Anyway if you can trace the power on button to the chip it says that it should set the pin low when button is pressed . That turns on the TSUMU58EHJ-LF . Power pins are on page 27 not sure if they are the same . Doubt it but you never know .
According to the trouble shoot on page 47 if power to the chip is OK but no back light then the TSUMU58EHJ-LF needs to be changed which usually means changing the board .
Not sure if you can figure out which are the power pins .
-
Hope you know that monitor can be bought refurbished with a warranty for about US$100 from NewEgg and others. Now, if you want to work on it for fun and experience … great! Many times, I’ve spent hours fixing something where the parts alone cost more than a brand new one … learned and enjoyed, and very satisfying.
Mike in California
-
@Jwillis Thank you very much for the link to that service manual. I tried to set the pin low but it didn't work. BUT I'm going go through the rest of the manual as i'm sure there's a lot of similarity in how dell builds monitors.
@calzap Yes I know :) I really want to learn how to fix this thing myself.
I'm also looking into the possibility of using one of those "universal LVDS" controllers but it's hard to get straight answers from Chinese sellers.
They do have some available on eBay for my panel (M240HW01 V1) but their picture show a kit with an inverter board for cathode back light instead of an LED driver.
-
I have a confession to make...
10 years ago, I sold a 100mhz HP Agilent Oscilloscope for 100$ or maybe it was 150$ either way I thought I would never learn electronics enough for it to ever be useful.
I am so very angry with myself. |O
-
I was just able to confirm that the LED back light AND the LED driver board are working by turning them on.
I did it by shorting the 3.3 volt pin to the ON/OFF pin on the cable that connects the power supply to the control board.
So the Interface board definitely IS the culprit.
I also did another potentially "stupid" thing; since I couldn't find proper diagram for TSUMU58EHJ-LF-3 and setting pin 27 low did nothing, I tried setting low other pins, tried all of them one by one. Maybe I fried it who knows.
I was able to find 3 pins that send the monitor in standby and then off, but I cant turn it on again until I reset the monitor (dis/reconnect power), then I can do it again as many time as I want.
Could this indicate the TSUMU58EHJ has some life in it? and from this can we imply that the crystal is fine? (otherwise the chip would not run at all right?)
I'll be back later, I gotta go clear my head by performing various voodoo rituals around the board... :scared:
-
Here is a a close up of the OZ99xx chip and the transistors.
Are those holes on the transistors??? They all look the same except for one that doesn't seem to have a hole in the "dot" of the marking.
Nobody has apparently commented on this yet -- but yes, that does look like holes burnt into the SOT-23 package to me. While I have not personally seen this in SOT-23 transistors, I have encountered SOIC packages with very similar-looking small holes after they had let out the magic smoke. Given the fact that more than one transistor is affected, some other root cause has probably triggered their failure.
Or am I misinterpreting this? Has anyone come across similar "holes" which were benign?
-
I was just able to confirm that the LED back light AND the LED driver board are working by turning them on.
I did it by shorting the 3.3 volt pin to the ON/OFF pin on the cable that connects the power supply to the control board.
So the Interface board definitely IS the culprit.
I also did another potentially "stupid" thing; since I couldn't find proper diagram for TSUMU58EHJ-LF-3 and setting pin 27 low did nothing, I tried setting low other pins, tried all of them one by one. Maybe I fried it who knows.
I was able to find 3 pins that send the monitor in standby and then off, but I cant turn it on again until I reset the monitor (dis/reconnect power), then I can do it again as many time as I want.
Could this indicate the TSUMU58EHJ has some life in it? and from this can we imply that the crystal is fine? (otherwise the chip would not run at all right?)
I'll be back later, I gotta go clear my head by performing various voodoo rituals around the board... :scared:
That's awesome ! I would say that the main board is the problem as well. I looked into those universal boards and if you search the on board chip NT68676 you'll find various info on it. Unfortunately I didn't come across any data sheet .
I found this for you M.NT68676.2A (https://www.mikrocontroller.net/attachment/380200/User_s_guide_of_M.NT68676.2A_controller_board_V1.1.pdf)
You can find just the board in different versions for your LCD display. But if your back light is working you won't need the CCFL inverter board .
Here is a a close up of the OZ99xx chip and the transistors.
Are those holes on the transistors??? They all look the same except for one that doesn't seem to have a hole in the "dot" of the marking.
Nobody has apparently commented on this yet -- but yes, that does look like holes burnt into the SOT-23 package to me. While I have not personally seen this in SOT-23 transistors, I have encountered SOIC packages with very similar-looking small holes after they had let out the magic smoke. Given the fact that more than one transistor is affected, some other root cause has probably triggered their failure.
Or am I misinterpreting this? Has anyone come across similar "holes" which were benign?
The LED driver board was confirmed to be operational and the transistors were confirmed to be working as well .
-
Here is a a close up of the OZ99xx chip and the transistors.
Are those holes on the transistors??? They all look the same except for one that doesn't seem to have a hole in the "dot" of the marking.
Nobody has apparently commented on this yet -- but yes, that does look like holes burnt into the SOT-23 package to me. While I have not personally seen this in SOT-23 transistors, I have encountered SOIC packages with very similar-looking small holes after they had let out the magic smoke. Given the fact that more than one transistor is affected, some other root cause has probably triggered their failure.
Or am I misinterpreting this? Has anyone come across similar "holes" which were benign?
The LED driver board was confirmed to be operational and the transistors were confirmed to be working as well .
I'm glad I'm not the only one wondering about theses. Gotta admit it's strange to find what looks like holes on top of transistors. BUT it's true that the driver works, I was able to light the whole panel in a very bright white and Jwillis says the transistors reading I took looked good so I guess it's either normal or some form of aging that is not critical yet?
About the M.NT68676.2A controller, I was looking at the exact same one. I was finally able to find a seller that understood what I was talking about and he sent me a document (attached) with specifications, schematics and pin out info.
He has a compatible LED driver for me but their controller board requires an external 12v 4A power source and I would rather keep it internal.
If I find 12v on the current power supply can I just connect to it to power their controller?
If not is it OK to power it using a step down circuit from eBay on the 18v line to the driver board?
I also wonder if their LED driver will be as bright as the current one.
Preserving the current PS would make it easier to use the original driver, using an external PS would require stepping up the 12v to 18v for the driver.
Also, are we sure the "DIM" signal from the new controller is the same as the original "BRI" signal? Same thing for the ON/OFF signal.
Back to some more :scared:
-
One thing to consider is after removing all of those caps I’d be very reluctant to put them back in. By doing so you have just put them through another soldering thermal cycle plus the mechanical stress of disassembly. It just is t something I would not do on a power supply. More importantly I wouldn’t have removed them before checking the supply.
As for the backlight I’d consider the driver in the same way. Check it’s output. That is if you can easily, backlights are becoming increasingly more complicated. In the simple ones anyways you have inputs and outputs that should be easy to figure out.