|
|
Post by mmartell on Dec 4, 2018 19:50:13 GMT -5
I found the 5v supply to the hdmi port drops to zero after a minute or so, killing video and audio output. Attached is the relevant portion of the schematic. 5v comes in on the left from the SMPS and goes out on the right to the port. It is switched by the CEC_POWER line from the CPU. The CEC_POWER line is present even when the port goes down. I infer this because putting my finger on the gate of the mosfet brings the pic back after a couple of seconds.
At the moment I only have access to the Q1058, the three parallel resistors and the two caps, the rest are on the bottom side of the board. Something is pulling up the voltage on the gate. The circuit is mostly a mystery to me I know it's simple for some of you so any help is appreciated.
|
|
|
|
Post by mmartell on Dec 5, 2018 15:54:12 GMT -5
What is the reason for not being able to get a good reading from the gate pin ? Putting a meter probe on it seems to immediately start draining it towards ground.
|
|
|
|
Post by mastertech on Dec 5, 2018 19:07:58 GMT -5
I need to see more of that schematic. Or the manual.
|
|
|
|
Post by mastertech on Dec 5, 2018 19:11:04 GMT -5
Never mind. I just grabbed a copy. Give me a sec to look things over.
|
|
|
|
Post by mastertech on Dec 5, 2018 19:15:51 GMT -5
So you are saying when you loose video you have 5v at pin 4 of Q1058 but you do not have 5v at pins 1,2,5,6 of the same Q1058?
|
|
|
|
Post by mmartell on Dec 5, 2018 19:51:50 GMT -5
Yes that is correct. I can't measure the gate properly at that point my meter says around 2v but decreases quickly then the video/audio comes back.
|
|
|
|
Post by mastertech on Dec 5, 2018 21:39:42 GMT -5
What I am seeing is this:
When the CEC line to the base of Q1057 is low then the line at pin 3 of Q1058 should be high.
When the CEC line to the base of Q1057 is high then the line at pin 3 of Q1058 should be low.
Can you test at pin 3 of Q1058 and see what you have when there is video?
|
|
|
|
Post by mmartell on Dec 5, 2018 23:04:34 GMT -5
You have it right. Except something is charging up the Gate of Q1058 such that after a minute or so the Gate is high again and video cuts out. Grounding my finger on the gate brings the video back for another minute or so. I can do that all day long if I'd like. Same thing with a meter probe - the reading is high but falling fast, then video returns.
|
|
|
|
Post by mastertech on Dec 6, 2018 11:17:17 GMT -5
Well I see 2 possibilities for this problem.
Q1057 is responsible for pulling that line low so if it is breaking down when turned on it is not properly holding that line low.
Or, Q1058 has some internal zeners that could be leaking causing the line to rise some.
To verify I would probably jump the emitter and collector of Q1057. You could just use a solder blob to jump or a piece of wire.
If the problem remains the same it is most likely Q1058 as the problem.
If the problem goes away then it is most likely Q1057.
This is providing that the CEC line remains high at all times like you said it does.
|
|
|
|
Post by mmartell on Dec 6, 2018 14:52:48 GMT -5
If Q1058 is the problem and I remove it, its gate pin should remain low at all times when the CEC line is high, right ?
I think I would rather try that first since it's topside and if it turns out to be the culprit then no disassembly would be required.
That's if I can get a good reason on the gate. Is there a reason you can think of why measuring the gate simply grounds it ?
|
|
|
|
Post by mastertech on Dec 6, 2018 15:05:23 GMT -5
Before I answer that lets clarify something. I think you are saying that Q1057 is on the bottom of the board that you do not have access to. If that is correct then how do you know the CEC at the base of Q1057 is still high when you loose video?
|
|
|
|
Post by mmartell on Dec 6, 2018 15:14:16 GMT -5
Yes only Q1058, the three parallel resistors and the two caps are topside.
I measured CEC_POWER at the CPU which is also topside, pin 137. It never wavers, always 3.3v possibly because passthrough is enabled.
|
|
|
|
Post by mastertech on Dec 6, 2018 15:25:04 GMT -5
Ok, I understand where you are coming from. But that does not confirm that you have a good CEC signal at Q1057 so you cannot rule that out as a possibility.
To first answer part of your question, whatever part is causing leakage is probably just enough voltage to cause the mosfet to go into shutoff and when you touch pin 3 with your meter or finger is causes just enough voltage loss to turn the mosfet back on again. That is not uncommon.
Here are 2 things you can try before you have to pull out that board.
If you have a resistor around 10K solder 1 leg to pin 3 of Q1058 and the other leg to DGND. This will essentially bypass 1057. See what happens.
The other thing would be, if you are real good at desoldering and soldering, swipe the mosfet from Q1063 or Q1067 or Q1061 or Q1074 and swap it into your Q1058 and see if anything changes on that circuit.
|
|
|
|
Post by mmartell on Dec 6, 2018 15:44:32 GMT -5
I expect soldering a resistor would have the same effect as keeping my finger on the gate, no ?
Your second suggestion is more interesting. Let me see if I can find where these other mosfets are located.
|
|
|
|
Post by mmartell on Dec 6, 2018 15:53:40 GMT -5
Wait. If the leakage is occuring in Q1058 then I should have no problem reading the gate value after its removal. That should tell me right there if the problem is elsewhere ?
|
|
