[HOWTO] Fix/Repair "Chassis Intruded Fatal Error System Halted" on Asus P5QL

[Th3_uN1Qu3]

Just finished fixing this one. Apparently, this is a common issue with all Asus P5 series boards. I first assumed it's a bad BIOS flash or corrupt CMOS settings, as it would go away once after resetting CMOS, but upon the next reboot it would be there again. Same if i changed any BIOS settings and saved them.

I read that people have flashed BIOSes to no luck, and someone pointed towards a transistor on another (older) board, so i proceeded to look for a hardware issue. And i found it. See attachments for fix.

Instructions:

1. Follow directions in pictures
2. Clear CMOS
3. Power board up, redo all BIOS settings
4. Be happy and have a cold one.

This should apply to *all* boards that exhibit this behavior, obviously the parts may be in a different location. Follow these directions to find it if your board differs:

1. Remove jumper cap from chassis intrusion connector (if you're reading this, you already found it's of no use anyway).
2. With the board powered ON, measure with your voltmeter and find the pin of the chassis intrusion switch connector that has somewhere between 1.7v and 3.3v on it.
3. Disconnect board from PSU.
4. Use continuity tester function of your DMM and probe on the SOT23 transistors next to the super IO controller, you should find it beeps on one of them.
5. Power the board back ON and find a 3.3v supply nearby
6. Power board OFF, remove transistor, solder drain connection (it's gonna be the single pad, just like it looks in the picture) to 3.3v supply.

The rest is the same as above.

I'm also going to tell you what causes the issue to appear in the first place in a follow-up post.

[RJARRRPCGP]

Not getting the issue on my Asus P5QL Pro, knock on wood, even though I'm posting from my Asus Maximus II Gene.

[shadow]

Very interesting find! I find it impressive that someone would come up with such a solution.

By desoldering the transistor and wiring a 3.3V source straight to the drain pin. It seems like you are bypassing the whole Chassis Intrusion circuitry and making the motherboard 'happy' so to speak by saying that there has not been any Chassis Intrusion at any time (tying the relevant IO pin to high).

I assume the Chassis Intrusion 'feature' will not work anymore with the transistor gone. And I also assume that the real cause of the issue will be either that transistor which you desoldered or some part of the circuitry downstream of that.

[kc8adu]

and nobody i know uses the feature anyway.good fix.

[Th3_uN1Qu3]

Originally Posted by shadow

I assume the Chassis Intrusion 'feature' will not work anymore with the transistor gone. And I also assume that the real cause of the issue will be either that transistor which you desoldered or some part of the circuitry downstream of that.

Yes, it does not work anymore, but who cares about that again? A feature that was useless to begin with. At least the way Asus implemented it. I promised i'll get into more detail so here goes:

First of all, there is no other circuitry downstream. The gate of that transistor goes straight into pin 3 of the chassis intrusion connector and that's it. And the drain connects directly to pin 104 of the super IO controller, which would be the 3rd pin counting from the right side, where you see 102 printed.

And that transistor was just doing its job. It was fine, i checked. The way it was implemented was the problem.

Now with that settled, let's see just how many wrongs in a row could Asus get...

1. On most other boards that support a chassis intrusion sensor (switch), there is an option in the BIOS to change the chassis intrusion sensor so that it merely reports that your case has been opened, rather than preventing you to boot up the damn computer. There is no such option with ASUS boards. The mere inclusion of this feature would have prevented all this trouble.
   
   
2. The Super IO controller works with ridiculously low pull-up currents on those pins, i measured 3.3uA (microamps). Merely measuring with your multimeter, or even just TOUCHING them WILL pull the pins down, and make them report false information!
   
   
3. The transistor doing the pull-down when the chassis intrusion switch is OPEN (or the jumper removed for whatever reason), is a MOSFET that conducts at very low gate voltages, and has no gate-source resistor. At 1.5v the transistor already conducts and pulls the pin down on the super IO. You can charge it up to that 1.5v by merely touching the pins of the connector.
   
   
4. Given the very low currents the pins on the super IO controller, just the drain-to-source leakage current as that MOSFET gets old, or simply dust and dirt accumulating on the board, can create enough leakage current to make the IO pin appear as "low" and generate a chassis intrusion message which you can then never get rid of.
   
   And the icing on the cake...
5. The chassis intrusion detection is a flag in NVRAM, and it can get set even if power is removed from the board. As long as the CMOS battery is there, the chassis intrusion "feature" will work. Or shall i say, prevent YOU from working?

Asus deserves a big and for this one.

Extra bit of fun info: The transistor right next to it will make the board report "New CPU installed!" every time until you clear CMOS, if you touch it or probe it. But at least it lets you boot the system.

Originally Posted by kc8adu

and nobody i know uses the feature anyway.good fix.

Thanks.

[kaboom]

Originally Posted by Th3_uN1Qu3

and has no gate-source resistor

Yep. That's ASUS for ya.


Were you expecting them to get it right even after their "USB ESD" problems?

[cheapie]

Why don't you just turn the intrusion detection off? Every board with it that I've seen has that option.

[Th3_uN1Qu3]

Originally Posted by cheapie

Why don't you just turn the intrusion detection off? Every board with it that I've seen has that option.

If you would have actually read my explanation of the issue, you would have noticed that...

Originally Posted by Th3_uN1Qu3

On most other boards that support a chassis intrusion sensor (switch), there is an option in the BIOS to change the chassis intrusion sensor so that it merely reports that your case has been opened, rather than preventing you to boot up the damn computer. There is no such option with ASUS boards. The mere inclusion of this feature would have prevented all this trouble.

That's right - i'm not "just turning the intrusion detection off" because it cannot be turned off. Trust me, i've looked everywhere in the BIOS setup (and manual too) of that board, the option to turn it off or ignore it is missing.

[Topcat]

good work! while I've never had this problem from an Asus (one of the few quirks I havent had from 'ass backward asus'), thanks for sharing this fix!

[cheapie]

Originally Posted by Th3_uN1Qu3

If you would have actually read my explanation of the issue, you would have noticed that...



That's right - i'm not "just turning the intrusion detection off" because it cannot be turned off. Trust me, i've looked everywhere in the BIOS setup (and manual too) of that board, the option to turn it off or ignore it is missing.

Oh.

[momaka]

Neat fix!
If only I knew of this fix a few years ago. Someone gave my friend this exact same motherboard since it wasn't working. And because no one could fix it, it just got torn apart for random parts.

IMO, ASUS motherboard can be real stupid sometimes. I have an older AsRock S939 motherboard designed by ASUS (at least that's what it looks like), and one of the regulators for the RAM Vcc gets so hot that the PCB has severely darkened around it. And this is not limited to just my motherboard model. I've seen other ASUS boards with this problem as well.

[neuron]

Hi,

This problem is very common in Asus site and in many forums in Internet.

Since I crack many software - Mac and W$n ... I would like your advice and expertise to solve this in the BIOS.

Disassemble BIOS - the ROM code and change the jump ... in the code. Not in Menu


Anyone test this procedure - Transistor in P5K E

Best Regards,

Neuron

[Th3_uN1Qu3]

Yes this could also be solved thru BIOS modification, but disassembly of a BIOS ROM is beyond my software skills. If you show me a clear, high resolution photo of your board, i could point out where you can check to find that transistor.

[neuron]

Originally Posted by Th3_uN1Qu3

Yes this could also be solved thru BIOS modification, but disassembly of a BIOS ROM is beyond my software skills. If you show me a clear, high resolution photo of your board, i could point out where you can check to find that transistor.

Thank you for your quick answer.

If we solve in BIOS, many P5K owners ... appreciate this.

http://www.asus.com/Motherboards/P5KEWIFIAP

Regards,
Neuron

[Th3_uN1Qu3]

Found a bigger one on the 'net. Check all SOT-23 transistors in the circled area, one of them must be the chassis intrusion detector. There's only 6 of them so it should be easy to find.

I could probably learn the knowledge required to patch the BIOS in a few weeks (i've done this before), but my time is very limited nowadays, and also, i don't have such a board to test.

[neuron]

Originally Posted by Th3_uN1Qu3

Found a bigger one on the 'net. Check all SOT-23 transistors in the circled area, one of them must be the chassis intrusion detector. There's only 6 of them so it should be easy to find.

I could probably learn the knowledge required to patch the BIOS in a few weeks (i've done this before), but my time is very limited nowadays, and also, i don't have such a board to test.

Thanksssssssssssssssssss

If you could get me some guidelines - BIOS... since you studied this in past.

Regards,

Neuron

[anomaly7]

Thanks for this post. I received my first Chassis Intruded ! alert this morning. I think I'll cheeck back for the BIOS fix mentioned...

[tmiha71]

Nice thread, but I see lot of Panasonic's FJ caps...
If there are at the rear USB connectors, I would replace them with Rubycon XYG 8x11 1000uF, even used ones are good for that place...

Reason for replacement - too good caps (??), on wrong place ...

To add, through time, inductive, capacitive, or even resistive bond is created, and is main cause of that behavior...
Most likely some caps died, and can not swallow current... And because they're FJ's - most likely PSU caps took big hit...

[3d0]

My friend has an ASUS P5-N32 Sli Deluxe board with the same problem. He can't make any changes in BIOS and save them, because on the next restart the chassis intrusion message pop's up. So no overclocking could be done. We tried everything from flashing the BIOS, replacing all the jumpers etc, nothing helped.
Wish I could find this 7-8 years ago, when it was brand new. It cost around 250-300 euros.

Thanks for the explanation.

[Th3_uN1Qu3]

Originally Posted by anomaly7

I think I'll cheeck back for the BIOS fix mentioned...

If you're waiting for *me* to fix the BIOS, i won't.