Gigabyte Z390 Aorus Master

[evgenij]

Hello folks,

unfortunately my Gigabyte Z390 Aorus Master died last weekend.
The PSU's short circuit protection triggers when the ATX12V_2x4 connectors are plugged in, all 8x2 pins have a voltage drop of 0V in diode mode and the multimeter beeps on every pin when I probe to ground.
Can anyone help me diagnose this?

Many thanks in advance!
Evgenij

[McWhitePink]

when you are testing the board in diode mode, you should disconnect the PSU from it. otherwise you are measuring the PSU, which is supposed to be low.

[evgenij]

I was testing the board only, without PSU connection, with no GPU/RAM or CPU installed.

[McWhitePink]

I was testing the board only, without PSU connection, with no GPU/RAM or CPU installed.

Hard to believe all the power domains are shorted. but if you have DC power supply you can apply voltage to each pin, and see which component is getting hot.

[evgenij]

Hi and thanks for your reply.

Lisa Sue provided a boardview file, and all 12+ pins are interconnected if I see it correctly.
So I tried to measure the voltage drop in diode mode, starting with connector and going step by step to next components.
Between +12V and the first three caps, there is a voltage drop of 0.024V, which appears to be fine for me. But from 12V to the first coil there is a voltage drop of 0V and I'm not sure if this is normal (probably yes, it's just a coiled wire)?


Also resistance between 12V and ground is 24,5Ohm(but it beeps). So it's not shorted?

[McWhitePink]

Hi and thanks for your reply.

Lisa Sue provided a boardview file, and all 12+ pins are interconnected if I see it correctly.
So I tried to measure the voltage drop in diode mode, starting with connector and going step by step to next components.
Between +12V and the first three caps, there is a voltage drop of 0.024V, which appears to be fine for me. But from 12V to the first coil there is a voltage drop of 0V and I'm not sure if this is normal (probably yes, it's just a coiled wire)?


Also resistance between 12V and ground is 24,5Ohm(but it beeps). So it's not shorted?

so you mean the 12V power rail is shorted, not other powers? 24.5ohm is a short, even 245ohm is abnormal. You should connect your bench top power supply to find out the hot spot. this could be a one shot kill.

[McWhitePink]

so you mean the 12V power rail is shorted, not other powers? 24.5ohm is a short, even 245ohm is abnormal. You should connect your bench top power supply to find out the hot spot. this could be a one shot kill.

BTW, the picture you are showing is the CPU power connector, all the 4 pins are 12V. DAL1 is a inductor, you can solder a wire on one end of the inductor to connect your DC supply. Don't foget to set current limit when you are applying voltage on it.

[evgenij]

Thanks, the desktop power supply has been ordered and will probably arrive either tomorrow or next week.
What current limit should I set? 0.5A@12V should be enough, right? Does it matter which pin of the inductor I use for the voltage injection? Would it be correct to do it the way I measured the voltage drop? Or should I wire my desktop power supply with connector pins?
Since I don't have a rosin atomizer yet or thermal camera, I will use isopropanol for hotspot detection.

Sorry for all the stupid questions, but I don't want to make the situation worse. This is the first time I'm trying to repair a PC motherboard.

[McWhitePink]

you can set it to 0.5A first and adjust it later, the voltage will be clamped since you have set the current limit, if you cannot find the hot spot then you can increase the current limit. usually you should be able to see the hot spot below 1W power consumption, higher may cause something burnt out.

[evgenij]

Hello, it seems to be the IR35201. I had to supply up to 0.5A@8V to get it visibly/noticeably warm, but after removing it the short circuit is gone.. I have a tl866II+ for programming, but is the firmware available for download somewhere? Is there anything else to consider when replacing it?

PS: Merry Christmas to you all!

[evgenij]

Hello everyone,

maybe someone can help me and tell me if:
- it is really necessary to program the IC, or is the "default configuration" sufficient?
- if the default config has to be adapted:
-- where can I get the necessary parameters/firmware?
-- if the default config has to be adapted: can it only be done with PowIRCenter?

[DynaxSC]

Hi my Friend

I have some experience with Renesas programmable VRM controllers ISL and RAA. Renesas has a software called Power Configurator allowing preparation of an configutation file, but the issue is this works only one way, so you can prepare the config file and program it to the VRM chip, but you can't read any configuration file from the chip. However it is possible to compare if the prepared configuration is same as in the vrm chip, but you get only a Yes/No information.
The configuration setup is quite difficult, there are some default settings in the software, but they do not work, when the chip is programmed with it.
Also if you solder an unprogrammed chip to the board it will just do nothing, there is no "default configuration" inside, the chip is just blank.

I don't know if in case of International Rectifier (nowadays Infineon) it is similar, but I strongly suspect this is the same way.

The easiest way to repair your board is to buy another demaged board, preferably with a known demage not in the VRM area, and transplant the chip to your motherboard.
But before you do this, check all mosfets on the board, and also the drivers if exist, as the demage of the vrm controller must not be the only one demage. Just measure if any of the mosfets has short between source and drain, and if the gate has no short to ground nor to the 12V. The impedance on the gate should be several kohms till even mohms, and should be similar on all mosfets. Aslo if you measure the impedance of the vcore net, when you start the measuring, should start from several ohms till probably to severel hunded's of ohms, but may also stop at 100 ohms, this is OK then. Don't forget to check the IGPU phase/phases the same way (they are most probably also controlled by the same vrm chip).

Another thing you can try is to contact infineon, if they can supply you some default configuration file, eg. for a development board or so, maybe also there is a forum there, you can register and look for help.
Gigabyte will most probably not help you, as they treat this content as intellectual property, but who knows, this is a quite old board.

If you want to go the way of defining a new configuration you will need to setup the VRM phases and the VRM layout, the IMVP version (the communication standard between CPU and VRM allowing the CPU to control the VRM chip), the chip address on the bus, and many many more. In case of ISL and RAA chips they can be programmed in-circuit with a special device (can be bought at Aliexpress and works but is not quite cheap, a broken board will be probably cheaper), further the number of programming cycles allowed per chip is limited, so it might happen you need more chips to have success. Hint - the IMVP version used on a board can sometimes be found in the pdf mobo schematics. Also if the IR software has some default parameters set, you might count on the lazyness of the Gigabyte engineers, and that they changed anly some basic parameters not getting into detail parameters, so if that is true, you might have success with the defining of the configuration.

The vrm firmware potentially could also be included in some bios or firmware update file, so it is probable but not sure. To find it could also be not easy, you would need to have some sample bytes of such a config file which are same for all firmware versions, so potentially this is a way how to do this, but I never tried this method. You could try to generate some dummy firmware files with the IR soft and compare them, what is not changing, and how long they are, then look in the bios/firmare for that patterns.

Most difficult thing could be to try to read the config file from the vrm chip (anyhow here you need to buy a demaged board to source a programmed one). Maybe it can be done with some open source device and software (the communication is based probably on I2C protocol or a similar) which does not limit the read function. However the chip is probably "read protected" by a hardware or software fuse and it is most probably not possible to get any documentation about the way how to do it. If chip is "read protected" one can try to alter (reduce) the supply voltage of the chip step-by-step. If you have luck the protection circuit may stop to work before the memory read functions and you can read the configuration. But with this method is used in "professional" industry reverse engineering and might be against the law, and it is not very probable to have luck, and you could waste a lot of time to setup all the test environment and do the trials. However if you have a hacker gene this might be a nice task to do.

Another last idea would be to look for some open source software which can read I2C or similar protocol devices on a motherboard under windows or linux, maybe somethig like this exists, never investigated this ? The question here is also if the motherboard itself has a hardware device which is capable to do this job, maybe the chipset or the Super I/O, you cloud look into the schematics or boardview to check what is connected to the vrm programming bus ?
You might also ask Infineon/IR for some programming specification for the vrm chip, but this might be difficult, as even motherboard vendors get only the software tools for config definition and mass production programming.


So good luck with your board.