DAG7ELMBAC0 motherboard windows crashes.

[Francisco.s]

Hi, have some interesting news.

laptop has been working for a few hours and can confirm that voltage on vcore is too much higher than it should be.

In bios setup screen vcore was 0.8 volts and in windows login screen it was 0.6 volts during that few hours.

So I think it should be a problem with a resistor o capacitor from the feedback circuit.

I have tested some resistor but seems to be good all of them, but without schematic is very difficult to follow the circuit.

Could be an internal cpu voltage sensor problem?

[mcplslg123]

As a matter of last resort,replace the coils and see if the voltage drops a bit on coils. It may be the cpu is overheating and resulting in panic kernel.

[Francisco.s]

Hi, I have replaced the Richtek RT3612EB today with a new one, and still crashes in windows and Linux, but ms-dos works well.

Hi, I have replaced the output mosfets and tantalum capacitor, and still the same. I have not replacement for coils now.

Booting a BlackArch linux iso via Ventoy, it still crashes, but looking at the screen error messages seems that cpu is damaged.

I don't know what messages can say, so I hope someone in the forum can read that messages in the photo attached, and tell what it is about.

Thanks in advance.

[mcplslg123]

1.6V-1.8V is the norm for 10th/11th gen cpu for VCORe/VCCGT. I suggested to replace all the mosfets/drivers/pwm related to VCORE at one go. It wont help if you replace only part of the items suggested.
However, there is no guarantee that it'll solve your problem. have seen success in couple of cases but not always.

[Francisco.s]

No point trying to tweak core voltage as its correct for 10th/11th gen cpu's.

Hi mcplslg123, thanks for your advice, I know that intel haswell and braswell have an internal regulator for vcore, didn't know that 10th and 11th have the same.

Replace the coils on VCORE area with same rating(may sound absurd but sometimes it helps)

The coils are rather unusual, 240 uH. are not frequently used, perhaps there are some in my donor boards.

It has only one tantalum capacitor, (330 uF) I will replace it and see if there are room in the board for another one.

There is another coil near the two of vcore tha seems the GFXCORE coil, but 1.8 volts seems a little high.

I will post results when have new info.

[mcplslg123]

No point trying to tweak core voltage as its correct for 10th/11th gen cpu's.

Replace the coils on VCORE area with same rating(may sound absurd but sometimes it helps)
Replace all the mosfets/drivers in VCORE/GFXCORE area at one go along with VCORE PWM ic.

If this doesnt helps, probe ram circuit thoroughly.

[Francisco.s]

I think the problem with this board is the cpu vcore, being a little higher. There are two phases circuit in the cpu vcore, having 1.6 volts at the output of that circuit.

With the oscilloscope I can measure about 400khz pulses at the input of both coils, but sometimes one of phases doesn't work. (pulses are missing randomly in that phase)

The chip that controls the mosfets of that circuit is a Richtek RT3612EB.

Hi, I have replaced the Richtek RT3612EB today with a new one, and still crashes in windows and Linux, but ms-dos works well.

Still cpu vcore is 1.6 volts, sometimes is 1.8 volts.

I will tray to reduce that voltage tweaking the resistors of feedback circuit, but it is not easy working only with the datasheet.

Do you know if ThrottleStop program or other similar can be run before windows starts?

[Francisco.s]

Hi Tony, seems that bios works, at least like the original one. It will be very estrange that different data on bios behaves the same.

Where did you get the new data for the bios?

Try to boot without any memory in slots and compare. It will be great if you have a lab power supply and could see the current consumption.

By the way, how much voltage do you have in cpu vcore?

Before I can test anything for you, shouldn't my pc be functional ?

You can test diode mode readings, as long as it seems to have cpu vcore, and cpu gets warm.

[tony.knors@live.com]

Hi again Francisco, Sorry again for the delay .. It's quite busy period for me now. Before I can test anything for you, shouldn't my pc be functional ? I just replaced the BIOS chip as we suspected a corrupt BIOS is the cause with a blank chip I flashed myself (still waiting for the pre-flashed I ordered in Taiwan) . I believe the flashing went OK, did some tests. BUT ... it doesn't change anything. When I press the power button, the power LED goes on, ventilator starts spinning, processors get warm (not excessive but warm) but nothing happens. No beep, nothing on screen (also not with external monitor), no response to keyboard presses .... nothing at all ... Memory chips are OK ... No clue what might still be wrong with it .... Any ideas ?

[Francisco.s]

Can you explain diode mode, you mean like for testing a diode ?

Yes, it's like testing a diode. Testing chips is almost the same, but the test have to be done between every pin of the chip and ground.

Red probe from multimeter has to be connected to ground, (ie. negative pole from dc jack or any usb)

Here you have a picture from the chip.

[tony.knors@live.com]

Hi, sorry for the delay as I was out of home for a few days. Can you make a picture of that chip on the board ? Can you explain diode mode, you mean like for testing a diode ?

[Francisco.s]

'm used to flashing chips like Arduino compatible, ESPressif, Atmel ... but never flashed a BIOS yet and not sure if I have the right tools for it.

Hi tony, I know some people use a Raspberry Pi with the flasrom program for flashing bios, but the cheapest way is to buy a CH341A + SOIC8 clip + adapter.

I will tell you how to take the measures and the pines of the Richtek RT3612EB chip.

You should use a multi-meter in diode scale with red probe to gnd and black probe to the pins of the chip.

Most of the pins readings will be in the range of .500 to .600.

Only pins from 6 to 22 are relevant for me. I've attached a screenshot of a board-view from other model for reference.


This chip is located in the same side of the board as the 3 pins smd.

[tony.knors@live.com]

Just ordered the tools ! :-)

[tony.knors@live.com]

I'm used to flashing chips like Arduino compatible, ESPressif, Atmel ... but never flashed a BIOS yet and not sure if I have the right tools for it.

[Francisco.s]

Francisco, no problem but you will have to give me precise instructions what/where to measure. Also please note that currently my mainboard is not functional. It powers up but nothing on screen. I suspect the BIOS is corrupt and I am waiting on a new BIOS chip I ordered in Taiwan.

Hi tony, if your bios is corrupted, you don't need to replace the bios chip.

It only needs to be re-program with new data from Badcaps or HP website.

I will tell you how to take the measures and the pines of the Richtek RT3612EB chip.

[tony.knors@live.com]

Hi Francisco, no problem but you will have to give me precise instructions what/where to measure. Also please note that currently my mainboard is not functional. It powers up but nothing on screen. I suspect the BIOS is corrupt and I am waiting on a new BIOS chip I ordered in Taiwan.

[Francisco.s]

If you need my help, let me know !

It will be great if you can post the vcore voltage in your board. Also the diode mode readings of the pins of RT3612EB.

[Francisco.s]

I'm glad you solve the problem.

[tony.knors@live.com]

UPDATE !!!! I found it ! It was on the other side of the PCB and as it was still mounted in the laptop I didn't see it ! THANK YOU SO MUCH Francisco !!!!!

[tony.knors@live.com]

Hi Francisco ! Really appreciate your efforts to help me out ! If you need my help, let me know ! This is a picture of the area around the HDMI connector. The board revision is : MODEL G7EL - DAG7ELMBAC0 REV: C