Another HannsG HG281D

[budm]

Re: Another HannsG HG281D

So the resistance between Source pin and the Power pin 4 (5.05V pin) is about 1.5 Ohms? I doubt that the T-CON for that 28" LCD panel will be running at 5V supply.

[Capt. Cap]

Re: Another HannsG HG281D

I figured out why I was having such a hard time following the +12V to the Q14 source pins:
Q14 source is connected to pin 4 of the power supply header, which is 5.05V and goes through FB14
1.5Ω from Q14 source to power pin 4.

Also I think I had a short between the logic board and a screw terminal the last time I took voltage measurements - I had placed an antistatic bag behind it for insulation but just noticed that it slips around bit. I hope that didn't do any damage. I'm now seeing 4.88V at Q14 source, 4.86V at drain and 0V at gate.

Apologies for the bad info guys, I'll work on getting the TCON numbers now...

[budm]

Re: Another HannsG HG281D

So at this point what Voltage do you have on the Source pin of Q14, 0V?
0V also on the two FB?
There should be big wide copper trace from the POWER 12V pin on the connector all the way to the Source pin of the MOSFET or to the inductor first before getting to the Source pin.

[Capt. Cap]

Re: Another HannsG HG281D

That's a problem. I have never seen small ferrite beads like that read open-circuit. Some had resistance as high as 4-Ohms due to high inductance. But never open.

Sorry, I should have stated that clearer, by no resistance I meant 0Ω, not open-circuit. Or in this case, 0.8Ω with my HFT special

I bothered mentioning because, just like you said, they checked out bad at first due to a layer of crust over the solder. They're really finicky to check, the only reliable place to probe seems to be the very skinny top metal edge of the bead. I hate drilling the probe tips into the solder, makes me feel like I'm going to break these tiny SMD components, but it seems like the only way to get a reliable reading most of the time.

If it was like that, then yes. Grounding the Gate of a MOSFET essentially turns it into a diode (due to the body-diode junction in MOSFETs).

Ah thanks, I didn't see why they wouldn't just use a diode there, but it sounds like this is common practice, so not the issue.

Could be. Though I don't see why the power going to the T-con would go through two switching devices. (Read: more losses) But there could still be a valid reason for that.

I'm trying to figure out where the Q14 source pins' supply is comes from, but the traces are pretty hard to follow on this board. It seems like the nearby FB7 & FB8 only supply Q15, I'm not sure where to check next - should I power it on and just probe around looking for the same voltage as is at the source pins?

Now if this was AC noise/ripple... yeah 40-50 mV is the threshold about where things can start to act up on lower voltage rails.

I'll try my new scope probes later today, but the reading was very steady, so hopefully not noise.

I started with a really cheap meter (not HFT, though) when I got into electronics some 15+ years ago. Went through a lot of headaches during troubleshooting until I realized that 1) my probes were messing everything up and 2) meter dial contacts also caused issues to some extent (though it's a lot worse on HFT meters.) So when I started getting coupons in the mail for Harbor Freight stuff and got those meters for free (like you ), I immediately knew what to expect. One of the first things I did when I got that multimeter is stuff the probes with hot glue around where the wires exit out of the plastic handles. Believe it or not, that made the probes survive over 3 years with medium use and no problems. This only left me with the dial contacts being bad once in a while (due to high humidity here, they oxidize). Spinning the dial cleans them for a while and everything is well for a little bit. Opening the meter and cleaning everything with IPA sometimes helps for a bit longer... but I don't bother anymore.

FWIW, those HFT multimeters are not too bad for low-voltage, low-power circuits. I wouldn't trust them for high power stuff, though. There are also a few thread here on these HFT meters (and one by me, where I explain in more detail how to glue the probes and why.)

Thanks for pointer to your hot glue thread, I'll have to try that out!
I've never had oxidation issues on the dials, but I tend to fry or lose them before they get too old.
Another issue I've come across with these multimeters is that the probe plugs don't fit into the multimeter body tightly enough giving a bad connection there; I've found that pushing a knife into the plug slats to spread them out a little yields a nice tight fit.

[budm]

Re: Another HannsG HG281D

FB7 & FB8, the inrush current through those two inductors is probably 5A or more when T-CON first getting the power, they use two of them in parallel but they may be marginally rated at best.

[momaka]

Re: Another HannsG HG281D

FB7 & FB8
No resistance across each of them

That's a problem. I have never seen small ferrite beads like that read open-circuit. Some had resistance as high as 4-Ohms due to high inductance. But never open.

Before concluding they are bad, try moving and poking harder at them with your multimeter probes. Sometimes SMD components have a layer of flux on top of them after being soldered at the factory. This can make readings appear open-circuit or high resistance until you poke and dig a little harder with your multimeter probes through the flux layer. Once you get through it, you should get an accurate reading. If they are still open circuit, they are definitely bad.

Is gate being grounded on Q15 normal?

If it was like that, then yes. Grounding the Gate of a MOSFET essentially turns it into a diode (due to the body-diode junction in MOSFETs).

Maybe the inductors feed Q15 which in turn feeds Q14?

Could be. Though I don't see why the power going to the T-con would go through two switching devices. (Read: more losses) But there could still be a valid reason for that.

see my reply to budm for voltages.

They look normal.
And I agree with budm about the 44 mV drop on ground cables - it's nothing to worry about. Even up to 1V should still technically not cause any problems. Reason why is because it's DC offset/loss. Now if this was AC noise/ripple... yeah 40-50 mV is the threshold about where things can start to act up on lower voltage rails.

Lol, how did you know?
Yep, it's a cheap (free) red Harbor Freight multimeter, in this case the problem turned out to be a loose wire in the ground probe - jiggling it causes the resistance to jump around.

I started with a really cheap meter (not HFT, though) when I got into electronics some 15+ years ago. Went through a lot of headaches during troubleshooting until I realized that 1) my probes were messing everything up and 2) meter dial contacts also caused issues to some extent (though it's a lot worse on HFT meters.) So when I started getting coupons in the mail for Harbor Freight stuff and got those meters for free (like you ), I immediately knew what to expect. One of the first things I did when I got that multimeter is stuff the probes with hot glue around where the wires exit out of the plastic handles. Believe it or not, that made the probes survive over 3 years with medium use and no problems. This only left me with the dial contacts being bad once in a while (due to high humidity here, they oxidize). Spinning the dial cleans them for a while and everything is well for a little bit. Opening the meter and cleaning everything with IPA sometimes helps for a bit longer... but I don't bother anymore.

FWIW, those HFT multimeters are not too bad for low-voltage, low-power circuits. I wouldn't trust them for high power stuff, though. There are also a few thread here on these HFT meters (and one by me, where I explain in more detail how to glue the probes and why.)

Can't have too many multimeters.

Completely agree. I got some coupons ready for those HFT meters again, next time I go in the area near the store for whatever reason.

[Capt. Cap]

Re: Another HannsG HG281D

Logic board PM549DA3 M06 VER6.0 images attached

Q14 is a 9435GM P-channel enhancement-mode power MOSFET
Q15 is a 2N70002K N-channel enhancement mode MOSFET
Pins 1-3 are source, 4 is gate, 5-8 are drain

I poked around some more with the multimeter while I had the logic board completely detached for the photo session.

For Q14,
LVDS pins marked 1-3 on the logic board (and go to F400 on the TCON) go to this drain
drain to source is 558mV in diode mode
No pins are shorted to ground, though gate sometimes looks like a capacitor charging (resistance quickly rises to infinity over a second or two)

Q15
gate is grounded, source to drain is 577mV in diode mode on the multimeter

10.1kΩ between Q14 drain and source pins (1-3) of Q15
There's > 2MΩ resistance between the +12V supply line and all pins of Q14 and Q15

FB7 & FB8
No resistance across each of them
I get infinite resistance from each FB to source on Q14
FB7 and FB8 both have 509Ω resistance to both source on Q15 and ground

Q8 (obscured by C69 in the photo) is unpopulated


Is gate being grounded on Q15 normal?
Maybe the inductors feed Q15 which in turn feeds Q14?

[Capt. Cap]

Re: Another HannsG HG281D

Sorry for the delay in response. My laptop from which I usually visit BCN decided to get corrupted HDD / Windows boot sector.

No worries, I hope you didn't lose anything Sounds like my typical week.

Anyways.
Yes, if T-con is not getting power, then we need to look back on the logic board. T-con usually runs on either 3.3V or 5V. Follow the power pins on the t-con board going back to the logic board. Using this picture as a reference...
https://www.badcaps.net/forum/attach...2&d=1538959176
... you can see that there are 4 pins on the right side of the LVDs connector. After the first one, the next three have a trace that goes to fuse F400. Color-wise, those wires appear to be the just orange and red with the 3rd pin being unused. So follow these two orange and red wires to the logic-board.

On the logic board, there is likely going to be a PNP transistor or MOSFET that switches power to the T-con board. Using this picture uploaded by user Mishannya...
https://www.badcaps.net/forum/attach...3&d=1285531627
... and assuming your logic board is exactly the same (pictures of yours still wouldn't be a bad idea, just so we can verify), it appears that Q14 (8-pin SOIC chip) may be what switches power to the t-con. Because I can't see if that is 100% true, use your multimeter's continuity/lowest resistance setting to find what switching component these two orange and red wires on the LVDS connector connect to. If it's a PNP transistor, it may have gone open-circuit. If it's a MOSFET, it may have a Gate shorted to ground, which won't allow it to turn on.

I'll check resistances on Q14 tonight, see my reply to budm for voltages.
I have v6.0 of the board, it's mostly the same but seems to have less parts populated, pictures are pending.

The issue could also be that this switching part is not getting a signal from the main controller to turn on. In that case, we will have to investigate what exactly controls this t-con power switch and why it is not getting a turn on signal (for example, it could be that the bad caps on the t-con caused too much noise and corrupted the flash chip, which then doesn't allow the main controller to boot and run everything else... though that's unlikely.)

The CCFL switches on and off at the appropriate times when the computer starts / stops sending video to the display, so something is getting signalling from somewhere, though I don't know how much logic board is involved in kicking the CCFL on and off.

All of my cheap red Harbor Freight multimeters are like this. I still use them a lot, though.

If that's not the issue, you may need new probes then.

Lol, how did you know?
Yep, it's a cheap (free) red Harbor Freight multimeter, in this case the problem turned out to be a loose wire in the ground probe - jiggling it causes the resistance to jump around.

I just used another I had in a drawer for recent measurements, probe-to-probe resistance on that one was 0.8 Ω, F400 looks totally fine when using it 👍

Can't have too many multimeters.

[budm]

Re: Another HannsG HG281D

Do not worrying about 44mV potential diff (you will always have resistance between point A and Point B for Voltage to be developed, beside, 44mV is way too low to affect logic level), concentrate on as to why you are not getting 12VDC to the T-CON by checking Voltage on that Q14 pins base on data sheet so you can see which pin is the feeding pin., it is more likely to be P-Channel MOSFET, so the 12V should be fed to the Source pin may be via some kind of inductor.

[Capt. Cap]

Re: Another HannsG HG281D

BTW, when you are testing the DCV at the fuse of the T-CON board, you have the monitor connected to the running PC?

Yes, connected, sending signal over HDMI, made sure it's not asleep, and the CCFL is powered on.

The CCFL does power on / off with connecting and disconnecting HDMI, and when the computer goes to sleep, so there is some sort of signalling still working there.

Did you have 12VDC feeding the logic board?

Yes, also checked all of the power input pins, they're as follows:

1 yellow 12.03
2 orange -0.04
3 red 4.89
4 green 5.05
5 brown -0.04
6 black 3.14
(ground reference is an on-board screw hole pad)

+12V measured on the power supply board is 12.08V

I apparently have a steady 44mV potential difference between the power supply and the logic board grounds - the logic board is not screwed into the chassis right now. I thought the ground wires running between the two boards would be good enough, I should probably do something about that. Though that could also just be providing an easier path for something shorted to ground on the logic board?

And yes, Q14 is the switched MOSFET so some of the pins should have 12V feeding Voltage on the pins.

Q14 voltages are the following (where pin 1 is the top rightmost in this pic and pin 4 is the top leftmost):
pin# voltage
1-3 1.30 - 1.56V
4 0V
the bottom 4 pins were all 0.20 - 0.35V

[budm]

Re: Another HannsG HG281D

BTW, when you are testing the DCV at the fuse of the T-CON board, you have the monitor connected to the running PC?
Di you have 12VDC feeding the logic board?
And yes, Q14 is the switched MOSFET so some of the pins should have 12V feeding Voltage on the pins.

[momaka]

Re: Another HannsG HG281D

Sorry for the delay in response. My laptop from which I usually visit BCN decided to get corrupted HDD / Windows boot sector.

Anyways.
Yes, if T-con is not getting power, then we need to look back on the logic board. T-con usually runs on either 3.3V or 5V. Follow the power pins on the t-con board going back to the logic board. Using this picture as a reference...
https://www.badcaps.net/forum/attach...2&d=1538959176
... you can see that there are 4 pins on the right side of the LVDs connector. After the first one, the next three have a trace that goes to fuse F400. Color-wise, those wires appear to be the just orange and red with the 3rd pin being unused. So follow these two orange and red wires to the logic-board.

On the logic board, there is likely going to be a PNP transistor or MOSFET that switches power to the T-con board. Using this picture uploaded by user Mishannya...
https://www.badcaps.net/forum/attach...3&d=1285531627
... and assuming your logic board is exactly the same (pictures of yours still wouldn't be a bad idea, just so we can verify), it appears that Q14 (8-pin SOIC chip) may be what switches power to the t-con. Because I can't see if that is 100% true, use your multimeter's continuity/lowest resistance setting to find what switching component these two orange and red wires on the LVDS connector connect to. If it's a PNP transistor, it may have gone open-circuit. If it's a MOSFET, it may have a Gate shorted to ground, which won't allow it to turn on.

The issue could also be that this switching part is not getting a signal from the main controller to turn on. In that case, we will have to investigate what exactly controls this t-con power switch and why it is not getting a turn on signal (for example, it could be that the bad caps on the t-con caused too much noise and corrupted the flash chip, which then doesn't allow the main controller to boot and run everything else... though that's unlikely.)

Also, I checked F400, was getting some resistance there (~35 ohms) but I get that just touching the tips of the probes together.

Some cheap multimeters will do that when they have bad contacts in the dial.
One way to remedy this, temporarily, is to push down on the dial and rotate it quickly several times through the resistance scales. Then check the resistance again. Should be much lower now. Also, cheap meters have bad switch positions sometimes, so getting the lowest possible resistance on the lowest resistance scale may require to play with the dial a bit by pushing left or right on it (without changing to another scale) to find where it makes better contacts. All of my cheap red Harbor Freight multimeters are like this. I still use them a lot, though.

If that's not the issue, you may need new probes then.

[Capt. Cap]

Re: Another HannsG HG281D

I started checking resistances and voltages on the TCON, the voltages were all way off (VGL, VGH, and VGH_M are all 0V) and eventually got back to F400 which has 0.84V on both sides.

Not sure what the supply voltage is supposed to be to this TCON, but it seems U404 and U406 are not going to be cranking out VCC and VDD+1.8v respectively from 0.84V.

So back to the logic board, V18S is 1.784V and V33S is 3.30V, but the pin 1 of the TCON LVDS (which carries power to the TCON) was only 816mV (same as was on the TCON at the time, it fluctuates a bit).

Any ideas which components are responsible for the supply voltage to the TCON?

[Capt. Cap]

Re: Another HannsG HG281D

Also, I checked F400, was getting some resistance there (~35 ohms) but I get that just touching the tips of the probes together.
I need to dig up my good multimeter, but in the meantime I think F400 is ok.

[Capt. Cap]

Re: Another HannsG HG281D

Pictures of the TCON are attached.
Let me know if there are any others I can take that would be of help.
Thanks!

[momaka]

Re: Another HannsG HG281D

What should I check next??

T-con board.
That's the one permanently attached to the TFT screen and usually located at the top of the monitor, under a metal shield. It's possible the badcaps may have cause damage to the t-con board.

Take some pictures of the t-con board and post them here. The are of most interest is where around the LVDS connector (from the LVDS cable that comes from the logic board) and also there should be a round inductor and a fuse on the t-con. The round inductor is where the main boosted rail and other secondary charge-pump rails get produced. The fuse should be near the LVDS connector. Start with checking the fuse.

Post pictures and results here.

[Capt. Cap]

Re: Another HannsG HG281D

I went ahead with replacing all of the caps on the logic board.
Unfortunately, I'm still seeing the same issue (all white screen).

I was pretty hopeful this would work as the 100μF caps near the regulators were really bad, as were some of those 10μF caps.
It could be that I did a crappy job of replacing them - I used some new solder and just wasn't getting very good flow-out on the pads on the top side of the board, it could also be that they were oxidized from age and not taking the solder well.
But seeing as how the symptoms are exactly the same, I'm thinking it could be something else.

What should I check next??

[Capt. Cap]

Re: Another HannsG HG281D

When you are measuring PSU noise with a scope, You need to add two small capacitors to the osciloscope's probes. I forgot what value these were supposed to be and how to place them.... but in esscence, without those two caps, your oscilloscope's probes will act like an antenna and pick up more noise than there is.

Regarding the noise measurements, I checked ground from one side of the logic board to ground on the other side of the logic board and got 14mV of noise - I figured anything much greater than that would be legitimate noise. I also got some real probes with the capacitance built-in since then, but haven't really used them yet.

[momaka]

Re: Another HannsG HG281D

When you are measuring PSU noise with a scope, You need to add two small capacitors to the osciloscope's probes. I forgot what value these were supposed to be and how to place them.... but in esscence, without those two caps, your oscilloscope's probes will act like an antenna and pick up more noise than there is. Granted none of your measured noise above appears to be too terrible (132 mV on the 24V line is not that great, though). Still, the 3.3V and 1.8V rails are linearly regulated, so the noise should be a lot lower than you measured. Not sure if that's due to picked up noise (as mentioned above) or if the caps on the outputs of those regulators really bad. But with linear regs, often times you will see less than 10 mV noise.

[Capt. Cap]

Re: Another HannsG HG281D

I opened up the monitor and broke out my multimeter and oscilloscope as I wanted to get an idea of the difference any modifications make.

Apologies for the _s, I couldn't get it to align correctly with spaces.
Here are the measurements I took:



line_V______%off_______Vpp__interval_f
------------------------------------------
1.8V_1.767V_(-1.83%)_|_17mV_200ms____7.5Hz
3.3V_3.27V__(-1%)____|_38mV_200ms_____16Hz

VSB5V_4.96V_(-0.80%)_|_35mV__200ms____8Hz
+5V___5.09V_(+1.80%)_|_22mV__200ms____9Hz
+12V_12.10V_(+0.83%)_|_48mV__200ms____5Hz
+24V_23.7V__(-1.25%)_|_132mV_200ms____5Hz


Vpp should represent noise on the line, the interval is for the sample rate on the DSO138, and frequency is what I was showing more or less for the noise oscillations at that sample rate. Not really sure what is normal there, but these don't look like they're ridiculously out of line. The +5V line had some weird resonance thing going on (frequency on the noise would sometimes slowly speed up and then suddenly slow down again). All of this was measured as AC at the 10mV sensitivity with the CCFL powered on. I don't think the DSO138 is sensitive enough for measuring noise in DC mode. Also, there were oscillations going on at 60/120/240Hz but that's probably normal right?

I can replace those regulator caps on the logic board tonight, was wondering if the above info changes anything though?