ASRock Z68 Extreme 4 can't boot

[senderj]

I am trying to fix this board (3 years old) which cannot boot. While measuring voltage at various points, I encountered another problem.

Original problem first. If I leave it untouched for a few days, it boot for 3-4 sec, the 2-digit LED lighted up and changing values, then it went off and only the CPU fan and chasis fan turning, nothing else works. If then I shut it down and restart the power again, only the fans turn and LED never light up.

After studying for months, I now can figure out the VRMs and most of the large components on the board. So I check the connectivity without the power. For components that I can identify their pinouts, they look good. Then I power it up and measure the actual voltage of the points I know. Chips Vcc seems ok. Power MOSFET has their reasonable difference between D & S. Vtt is 1.05V. VRM for CPU display (i7 2600K) is only 0.91V which I don't know if it's ok or not. When I measure Vcore I got 0V, input to VRM all 12V. But then only when I saw the psu fan stopped I realize there was no power to the board.

Here comes the second problem. I then tried to power it up again, but it never would. Even the fans didn't turn any more. Pressing the power button just trigger nothing. Checked the +5VSB on ATX power plug is 5.01V. The SIO chip (NCT6776F) also has its 3.3V SB.

I don't know how to proceed further. But I know I have to solve the second problem first before resuming the frist problem. Can anyone help me?

[momaka]

Re: ASRock Z68 Extreme 4 can't boot

Post some pictures of your board here.

The CPU VRM needs to be fixed first. Looks like there is something shorting it. If you disconnect the CPU 12V power cable, does the motherboard turn on at least? (It should).

[senderj]

Re: ASRock Z68 Extreme 4 can't boot

Thanks for the reply. Tried power up without CPU & its 12V. Still couldn't turn it on. Measuring the SIO chip, 3VSB ok, PSIN & PSOUT go low in response to the power button, but PSON (to ATX psu) stady at +4.46V. Not able to continue checking the CPU VRM if I can power up the board.

Here is the link to the SIO spec. Page 36, 260 & 261 are where I am learning but not totally understand.
http://webshop.atlantikelektronik.de...sheet_V1_2.pdf

Here are some photos of the board.

[momaka]

Re: ASRock Z68 Extreme 4 can't boot

Since the PSU still doesn't start even with the 12V CPU connector removed, then you might have a short somewhere else on the motherboard.

With a multimeter on resistance setting (choose 200 Ohms or lowest setting if it's a manual one), measure the resistance between each rail on the 24-pin ATX connector and ground. That is, check resistance between 3.3V rail and ground, 5V and ground, and 12V and ground. Post here what resistances you get.

Here's a link to help you determine where each rail is on the motherboard's ATX connector:
http://en.wikipedia.org/wiki/ATX#Power_supply
Keep in mind, though, that the connector diagram above is for a power supply, not a motherboard. For a motherboard, everything is flipped horizontally (i.e. mirrored so that pins 1 through 12 actually appear on the right side and 13 to 24 on the left side).

Also, forgot to mention, do the above tests with the power supply DISCONNECTED from the motherboard. Also remove the CMOS battery and other peripherals. CPU can stay if you want.

[senderj]

Re: ASRock Z68 Extreme 4 can't boot

The board was designed such that same voltage are shorted. My measurement are like this: On the 24-pin ATX, all are open except +3.3V is 161 ohm to gnd with or without CPU. On the 8-pin CPU power socket, +12V is 91 ohm to gnd without CPU and 4.8 ohm to gnd with CPU.

Are they ok?

[momaka]

Re: ASRock Z68 Extreme 4 can't boot

The board was designed such that same voltage are shorted.

That's normal. I actually haven't seen a motherboard yet where this doesn't apply. 12V rail on ATX connector and on the CPU 12V connector is separate, though.

My measurement are like this: On the 24-pin ATX, all are open except +3.3V is 161 ohm to gnd with or without CPU.

3.3V resistance seems a bit low, but it may be okay, so let's just note it but leave it alone for now.

On the 8-pin CPU power socket, +12V is 91 ohm to gnd without CPU and 4.8 ohm to gnd with CPU.

That may be a problem there.
Without CPU in the CPU socket, you should get open circuit on your multimeter when you leave the probes connected to the board for a few seconds. If the reading is steady at 91 Ohms for more than a few seconds, you likely have an "upper" shorted MOSFET in the VRM.

Here's how we can verify. I'll be using this picture as a reference:
https://www.badcaps.net/forum/attach...1&d=1406595795
Let's take MOSFET Q7493, since I can see its lead connections easily (in reality, you can take some of the other "upper" MOSFETs, like Q7491, Q7489, and etc.). It's located on the left side of that picture above, where the 5 gray "box" components are (these are actually coils).

Select resistance setting on your multimeter and measure the resistance between the tab lead on that MOSFET and any of its three lower pins. Let me know what resistance you get.
If this is unclear, I can edit your picture and re-attach. Also, I have a fairly well-written post explain how VRMs work and how the "lower" and "upper" MOSFETs are connected. Just have to find it.

Anyways, try getting that resistance measurment and post back what you get.

[senderj]

Re: ASRock Z68 Extreme 4 can't boot

Vcore upper mosfet are circled with red, 8 phrases. All 8 have their D-S shorted. 7 have their D-G & S-G opened but the Q7493 (marked with **) has D-G S-G at 19.4 ohm. Lower mosfet all have D-S 91.6 ohm and D-G S-G opened.

Other VRM next to them are for Vtt and CPU display. Their leads are all opened.

Looks like the Q7493 causes the problem. And my CPU is likely toasted. The mosfet is labeled "9130AL PBm". I think it's from NXP but I am not able to find it in digi-key nor RS. Where else I can look for it? Since I don't know why the mosfet failed, even if I can find a replacement how can I be sure that it won't break again before I put on a new CPU?

[momaka]

Re: ASRock Z68 Extreme 4 can't boot

Vcore upper mosfet are circled with red, 8 phrases. All 8 have their D-S shorted.

Well, probably only one has D-S shorted, but it will appear as if the others have D-S shorted too because they are all *pretty much* connected in parallel (actually they are connected through the output coils, but these coils have a low DC resistance, so for all practical purposes, this will make your multimeter *think* that they are in parallel)

7 have their D-G & S-G opened but the Q7493 (marked with **) has D-G S-G at 19.4 ohm.

That's probably the bad one. Remove it and then check D-S on the upper MOSFETs again. Hopefully, the short circuit should be gone.

Lower mosfet all have D-S 91.6 ohm and D-G S-G opened.

That's not necessarily bad, even without the CPU. It is possible that the Southbridge or some other chip could be sharing power with the CPU (although, that's very rare).

But let's take it one step at a time. First remove Q7493 for now.

And my CPU is likely toasted.

You would be surprised.
This is a multi-phase VRM, so if the voltage goes too high due to a shorted upper MOSFET, the lower MOSFETs start to turn on for a very long time and short out the PSU, thus triggering the short-circuit protection.
I've had a graphics card survive a shorted upper MOSFET before, as well as a Xbox 360 GPU.

If it's a high-power CPU (i.e. if it has high TDP), you may not want to use it for testing. A low-power CPU would be better for testing.

The mosfet is labeled "9130AL PBm". I think it's from NXP but I am not able to find it in digi-key nor RS. Where else I can look for it?

I can't even find a data sheet for it. Do you know what this 4-pin package is called? All I can see is that Gate seems to be on the right-most pin rather than the left.

If no replacement can be found, just get a standard TO-252 N-Channel MOSFET, and do a ghetto repair. TO-252 can be found on just about any hardware. Something like 60N02 or 85N02 should be fine.

Since I don't know why the mosfet failed, even if I can find a replacement how can I be sure that it won't break again before I put on a new CPU?

After you remove the shorted MOSFET, measure resistance between ground and the pad where G on the MOSFET used to connect, then also measure resistance between 12V rail (both ATX and CPU-12V connector) and the G pad. Any low resistance on these measurements would mean a shorted driver IC (but highly unlikely bad PWM controller). If the driver is shorted, the new MOSFET will likely short out again. Beware that there may also be a resistor between the Gate pad for the MOSFET and the driver IC's output pin. In such case, also check the resistance of the resistor and make sure it agrees with the resistance of the other upper MOSFET circuits in the VRM.

If you do have a shorted driver and you don't feel like doing this much work to get the board repaired, you can remove the associated lower MOSFET(s) for that shorted upper MOSFET (should be just Q7494) as well as the output coil (should be LT713). With these removed the CPU VRM should still work, though now you will no longer be able to do the highest power CPUs that this board supports.

As far as why the original MOSFET shorted - could be static discharge, power surge, overheating (either bad board design or not enough cooling on the VRM - often the case with water cooling and vertical coolers), or just a marginal MOSFET from the factory.

[senderj]

Re: ASRock Z68 Extreme 4 can't boot

Tried to remove the mosfet yesterday and this morning. The solder didn't even melt under my 20/200W iron. I ended up cutting off the 4 leads, checked to ensure no connection to the pad (except the D), and then measure all D-S and they are now open. CPU +12V to Gnd is open too.

Still trying to remove the mosfet completely. The D has a large pad at the back and soldered on the board where my iron cannot access. I tried to heat up the D lead with my iron switched to High for 2 minutes, the enclosure of the mosfet went off but the pad still has no sign of melting. I can see the board seems melting at the edges of the remains. Search the web for how others do it, finally removed it by adding more solder onto the remains (heat transfer).

Continue with your advise. G pad to Gnd before removal is 110 ohm, now is open. G pad to CPU +12V before is 19.4 ohm, now is open. Resistance between driver ICs (8510) output (hi gate) and G pads are all 2.4 ohm before and after removal. So I guess the driver is Ok.

I'll try to find a replacement. But in case I can't, just to check my understanding: Removing the lower mosfet and coil of that phrase, the cpu can still work with the remaining 7 phrases? But there will be a time slot when the cpu has no power. Won't it interrupt the processing?

What can I expect if I power it up now (i.e. with or without cpu, lower mosfet and coil remain)?

[momaka]

Re: ASRock Z68 Extreme 4 can't boot

Excellent work!

and then measure all D-S and they are now open. CPU +12V to Gnd is open too.

Good.

Still trying to remove the mosfet completely. The D has a large pad at the back and soldered on the board where my iron cannot access. I tried to heat up the D lead with my iron switched to High for 2 minutes, the enclosure of the mosfet went off but the pad still has no sign of melting. I can see the board seems melting at the edges of the remains. Search the web for how others do it, finally removed it by adding more solder onto the remains (heat transfer).

A 60W plain iron should be enough for such work. What you really need is separate flux (preferably "no-clean" type).
What I normally do is I apply flux on the part I want to remove. Next, I put lots of solder on the tip of the (hot) iron (enough to cover all of the pins of the entire part), and then apply it on top of the component. If the board is extra-tough, pre-heat it. There are many ways to do this. setting the board in front of a portable space heater is one way (though not the fastest or most efficient). A hair dryier or a heat gun (on low setting) work great, on the other hand. You could also place the board on something a few inches above your stove's burner (if it is an electric one only). Once the board is heated (heat as much as you can that still alows you to hold it without burning yourself), then you can try the above with the flux and the solder

Continue with your advise. G pad to Gnd before removal is 110 ohm, now is open. G pad to CPU +12V before is 19.4 ohm, now is open. Resistance between driver ICs (8510) output (hi gate) and G pads are all 2.4 ohm before and after removal. So I guess the driver is Ok.

Yes, looks good so far.

I'll try to find a replacement. But in case I can't, just to check my understanding: Removing the lower mosfet and coil of that phrase, the cpu can still work with the remaining 7 phrases? But there will be a time slot when the cpu has no power. Won't it interrupt the processing?

It shouldn't as I don't think the phases are synchronised in time slots. In case that they are, the 7 working phases should just "pick-up the work" for the non-working phase.

What can I expect if I power it up now (i.e. with or without cpu, lower mosfet and coil remain)?

Do NOT do that.
With the upper MOSFET removed and the lower MOSFET still connected in that faulty phase, you will get reverse current in its output coil, and that can kill the lower MOSFETs, the drivers, supervisory IC, and who knows what else.

If removing the MOSFETs is hard for you, just remove the output coil on the faulty phase, and that should be enough. In case the motherboard refuses to start, though, you may also have to remove the lower MOSFET.

Post another picture after removing the output coil and before powering the board, just to double-check everything.

The output coil may also be pretty hard to remove, so pre-heating the board is recommended.

[senderj]

Re: ASRock Z68 Extreme 4 can't boot

Confirmed that my CPU was toasted.

Just received my mosfet replacement and about to solder it on the board. Are there anything I should check before I power up the board with a new CPU?

[senderj]

Re: ASRock Z68 Extreme 4 can't boot

OK. I solder the replacement in, measure the DGS of the mosfet and they are all open. I put in the CPU and RAM, power it up and it works. Everything looked ok and I played around with the UEFI for about 2-3 minutes and believe that I might have fixed it. I put my hand on the VRM heat sink to feel if the temp is normal. At this moment it suddenly reboot and failed. The symptom is exactly the same as the problem I tried to fix before encountering the VRM problem (see Post 1 above): when pressing power on, power on LED ok, cpu fan and case fan turning but diagnostic LED is off and never complete the POST.

Touching the heat sink may or may not related to the problem. It seems related this time but when this problem first happened the board was inside the case and nobody touched anything. I don't know how to proceed since last time I try to diagnose it by measuing voltage here and there, the VRM failed and toasted my CPU. Can anyone tell me what to do next?

[RJARRRPCGP]

Re: ASRock Z68 Extreme 4 can't boot

OK. I solder the replacement in, measure the DGS of the mosfet and they are all open. I put in the CPU and RAM, power it up and it works. Everything looked ok and I played around with the UEFI for about 2-3 minutes and believe that I might have fixed it. I put my hand on the VRM heat sink to feel if the temp is normal. At this moment it suddenly reboot and failed. The symptom is exactly the same as the problem I tried to fix before encountering the VRM problem (see Post 1 above): when pressing power on, power on LED ok, cpu fan and case fan turning but diagnostic LED is off and never complete the POST.

Touching the heat sink may or may not related to the problem. It seems related this time but when this problem first happened the board was inside the case and nobody touched anything. I don't know how to proceed since last time I try to diagnose it by measuing voltage here and there, the VRM failed and toasted my CPU. Can anyone tell me what to do next?

May be a bad socket. Darn LGA!

[momaka]

Re: ASRock Z68 Extreme 4 can't boot

The symptom is exactly the same as the problem I tried to fix before encountering the VRM problem (see Post 1 above): when pressing power on, power on LED ok, cpu fan and case fan turning but diagnostic LED is off and never complete the POST.

Touching the heat sink may or may not related to the problem. It seems related this time but when this problem first happened the board was inside the case and nobody touched anything. I don't know how to proceed since last time I try to diagnose it by measuing voltage here and there, the VRM failed and toasted my CPU. Can anyone tell me what to do next?

Is the motherboard CPU heatsink braket bending or flxing the motherboard? Can you post a picture of it? Although many boards work and continue to work without getting affected from a CPU bracket that bends the board too much, I wouldn't be surprised if this problem is highly undocummented, causing many unknown problems. Try turning ON the motherboard with the CPU heatsink sitting on top of the CPU but not bolted/fastened to the bracket.
.....

There are also these simple things I'd like to ask you to check:

First, what brand and model of power supply are you using to test this motherboard? (Sorry if I asked this before. I looked through the thread again, but I couldn't see that I asked.)

Next up, you can try taking measurements of the board voltages, but I suggest doing it one at a time and cycling the motherboard ON and OFF.
That is, let's say you want to measure voltage on the CPU V_core rail. With the board OFF, put your probe on the lead of a a component that is connected to CPU V_core. Next, turn on the motherboard with the switch and watch the voltages. Finally turn OFF the motherboard and write down your measurements. After this, you can measure another voltage on the board, following the same procedure above.

[senderj]

Re: ASRock Z68 Extreme 4 can't boot

Sorry, the heatsink I am talking about is the mosfet heatsink, not the CPU heatsink.

The PSU I use is CoolerMaster Silent Pro M600.

Will follow the method to measure V of various points and report here. But it will take sometimes since I have to work OT this week. Will work on the board this weekend.

[momaka]

Re: ASRock Z68 Extreme 4 can't boot

Sorry, the heatsink I am talking about is the mosfet heatsink, not the CPU heatsink.

Right.
But I was saying, try the CPU with the heatsink on it but without clamping it down on the motherboard. That way, there would be no pressure on the board. We are just trying to see if this may be a BGA problem.

The PSU I use is CoolerMaster Silent Pro M600.

Not sure who makes those.
In the review below, however, you can see some Teapo caps in the pictures. Yours may or may not have Teapo caps, but I suggest you open it and check.
Review here:
http://www.hardwaresecrets.com/artic...y-Review/929/2

But it will take sometimes since I have to work OT this week. Will work on the board this weekend.

No hurry. I myself am on and off badcaps at random. Been here for many years, so will see your post whenever you reply.

[senderj]

Re: ASRock Z68 Extreme 4 can't boot

I missed your post yesterday. Will try the CPU heatsink method tonight if I can. Thanks for the info on the PSU. I had tried using another PSU but the board still failed to boot. But I'll open up the PSU to check the caps. Afterall, I learnt a lot up to now so repairing the board is no longer the only thing I am after.

Late last night I steal sometime to measure the voltage. This time I didn't touch any heatsink. I power up the board with a new CPU, it booted ok and entered the UEFI for about 17-18 sec. and then stopped and then reboot itself. Then PSU and CPU fans turning but no diagnostic LED no UEFI. I then turn it off, wait for few mins.

Then I power it up again and quickly measure the Vcore reading 1.02V. After about 7 sec. it reboot itself again before I can measure the Vtt and Vdisplay (for CPU's HD graphic). This reboot also failed with fans turning. I then measure Vcore again and it read 0V, Vtt is 1.05V and Vdisplay is 0.92V. Will work on other parts tonight if possible.

[momaka]

Re: ASRock Z68 Extreme 4 can't boot

Afterall, I learnt a lot up to now so repairing the board is no longer the only thing I am after.

Yeah, that's how I am as well, so I often repair "worthless junk". The knowledge is what's worthwhile to me, though .

Late last night I steal sometime to measure the voltage. This time I didn't touch any heatsink. I power up the board with a new CPU, it booted ok and entered the UEFI for about 17-18 sec. and then stopped and then reboot itself. Then PSU and CPU fans turning but no diagnostic LED no UEFI. I then turn it off, wait for few mins.

Then I power it up again and quickly measure the Vcore reading 1.02V. After about 7 sec. it reboot itself again before I can measure the Vtt and Vdisplay (for CPU's HD graphic). This reboot also failed with fans turning. I then measure Vcore again and it read 0V, Vtt is 1.05V and Vdisplay is 0.92V. Will work on other parts tonight if possible.

Interesting issue.

Check the Southbrdige heatsink and its thermal compound.
The only other thing I can think of is if there is some other fault in the CPU VRM that eventually causes it to overload and thus making it shut down, hence the 0V V_core.
Perhaps a partially-shorted lower MOSFET in CPU VRM? Without a CPU, that 91.6 Ohm reading between D-S on the lower MOSFETs sounds a bit suspicious now. Some (very few) motherboards power their Northbridge from the CPU V_core, but now that I look at your motherboard again, I see that there is no Northbridge to speak of . So where that 91.6 Ohm reading is comming from is a mystery.

Try powering on the motherboard with no heatsinks on the MOSFETs for the CPU VRM (assuming there are any?). Then run your finger through each and see if any one is getting particularly hotter than the others. Of course, this will only work as long as the PC has POSTed and hasn't crashed/restarted yet. With only a few seconds before the PC crashes, I think this may be hard to do, though.

If you do get your hands on a very powerful iron, you can remove all of the CPU VRM coils. That would make it much easier to verify if there is indeed a shorted lower MOSFET and where that 91.6 Ohm reading is comming from.

[senderj]

Re: ASRock Z68 Extreme 4 can't boot

Tried the CPU heatsink without clamping, didn't remove the problem. Tried removing the VRM heatsink, same problem. Tried with another PSU again, the board still didn't boot.

Tried keeping the power on time short and measure various points on the board. The Vcore PWM controller works, its Vcc was 12V, all its PWM output were 3.36V. The Vcc of the VRM driver was 11.6V and the PWM input was 3.36V as expected. However, all VRM drivers had HIGATE and LOGATE at 0V all the time, and so were the G of all upper and lower mosfet. All 8 drivers can't be demaged at the same time!?! I attach the VRM driver spec here if anybody is interested. It's the CHL8510.

Bad news! I might not manage the timing very well. After the above measure, I found that 3 upper mosfet has G-S at 3 ohm while others are opened. I though I might demage more while fixing it. But more interesting was, after lunch I measured the same again, 2 of them "back to normal" (i.e. G-S opened) while the third one was 14ohm. More to come. 3 hours later, it was 22 ohm, now it is 29 ohm. But this seems coincide with the fact I described before: if I store the board for a few days, it boot for 10-15 second before fail. If I try again tomorrow, it only boot for 1-2 second and if I try a few hours later, it never boot.

I leave the probes of my DMM on G-S and expected I can see the resistance raising, but it drops slowly. If I leave it there and come back after a while, it raises, also slowly.

[momaka]

Re: ASRock Z68 Extreme 4 can't boot

I'll have a look at the driver datasheet in a day or two and let you know if I find something. Almost sounds as if there must be some kind of DC offset on the PWM signal, thus eventually pulling all of the lower MOSFETs on.
Also, what is the PWM controller for the VRM?