ASRock Taichi X399 suddenly dead?

F

Flapjack

2[H]4U
Joined
Apr 29, 2000
Messages
3,207
I woke up this morning to an unresponsive PC. No fans, no response from power or reset buttons, nothing. The only board does is cycle the RGB on the RAM and pulse the blue LEDs around the northbridge. I've pulled the power cable and let it sit for a few minutes, but as soon as I plug it in, it does the same thing. Nothing on the diagnostic LED... nothing. It doesn't even try to boot. I normally leave it running 24/7.

Everything has been working fine, with the exception of XMP. I built the system a few weeks ago and it's been running solid. The G.Skill memory is *supposed* to do PC3200, but if I try to set it to the XMP profile in BIOS, it just beeps four times, then boots up and the normal 1066mhz. I haven't had time to set the RAM manually, so I've just left it for now.

The only other problem I had was a C0 on the diagnostic output when I tried to boot with 4 x 8GB sticks. I updated the BIOS and that was resolved.

Here is a video of it doing absolutely nothing. Disregard the fan noise in the background. That is from my servers, and not from the PC:

 
I tried pulling out half the RAM, but it still won't even recognize the power button being pressed. As soon as power is plugged in, it starts the slowly pulsing LEDs on its own.
 
try a different power supply maybe to take that variable out of it. could also try checking for something grounding out maybe.
 
Pulled a 750W PSU from another computer... no boot. The motherboard must've decided to just die overnight. Still hoping someone can chime in with a fix, but from what I can find online, this is an indicator of a dead mobo.
 
Flapjack said:
Pulled a 750W PSU from another computer... no boot. The motherboard must've decided to just die overnight. Still hoping someone can chime in with a fix, but from what I can find online, this is an indicator of a dead mobo.
Click to expand...

weird, yeah i read through that whole thread.. that poor guy going through hell dealing with all those peoples questions, only for the problem to be a bad 24 pin connector on the motherboard, lol.. wonder if you're having that same issue.
 
sirmonkey1985 said:
weird, yeah i read through that whole thread.. that poor guy going through hell dealing with all those peoples questions, only for the problem to be a bad 24 pin connector on the motherboard, lol.. wonder if you're having that same issue.
Click to expand...
Yeah, they meant well, but they just didn't understand it didn't even TRY to post. Even a bad CPU or memory should spin the fans. Reminds me of an MSI Athlon board that died on me years ago... the CPU fan would spin for a split second, then nothing.

As for the 24-pin connector, I tried wiggling it when plugged in to see if I got anything from the fans, but got nothing. There's really not much to go wrong there, just a plug soldered to the board. I'd like to know what he meant by it being defective... loose solder should make some contact and start something, if even only for a second. Plus, in my experience, soldered connections don't just go bad unless you're disconnecting/reconnecting them.
 
Flapjack said:
loose solder should make some contact and start something, if even only for a second. Plus, in my experience, soldered connections don't just go bad unless you're disconnecting/reconnecting them.
Click to expand...

Solder joints can wear out due to CTE mismatch between the joint/part/PCB. Certain packages are way more susceptible to this than others (large non J leaded resistors/caps are terrible so are parts with alloy 42 lead frames). Time at temp will age solder and accelerate fatigue. BGA have multiple mfg issues that can cause failures in the field like head on pillow, champagne voids, etc. Thermal cycling can cause vias to pull away from traces. Warping of the PCB can accelerate many solder failure modes say if standoffs weren't even.

I'm to lazy to look up public sources and it doesn't help with the issue but feel free to google.
 
Unter Dog said:
Solder joints can wear out due to CTE mismatch between the joint/part/PCB. Certain packages are way more susceptible to this than others (large non J leaded resistors/caps are terrible so are parts with alloy 42 lead frames). Time at temp will age solder and accelerate fatigue. BGA have multiple mfg issues that can cause failures in the field like head on pillow, champagne voids, etc. Thermal cycling can cause vias to pull away from traces. Warping of the PCB can accelerate many solder failure modes say if standoffs weren't even.

I'm to lazy to look up public sources and it doesn't help with the issue but feel free to google.
Click to expand...
Lol, I'm aware of all that... but this is a 24-pin connector. Of course, there could be something else that just died within a few weeks of use, but the case in question regarded a bad 24-pin (from a different manufacturer). Is it possible? Yes. Is it likely? No. At least not on something as simple as a 24-pin connector.
 
Even though the reviews on NewEgg are overall good, digging in a bit deeper reveals many with exactly the same issue. ASRock's reply is always the same.

Dear <customer>,

We sincerely apologize for any inconvenience this may have caused you.
Please trouble shoot the mainboard whether the issue is resolved.
• Unplug the ATX power, remove all components from mainboard.
• Remove the mainboard from chassis, place it on mainboard’s box.
• Clear the CMOS.
• Reseat the CPU and heatsink fan.
• Test with single memory module on A2 slot.
• Reseat video card firmly and 6+6 PCIE power (if applied).
• Connect the monitor.
• Plug the 24 + 8 or 4 pins ATX power.
• Turn on the power (using onboard power button) to check if its post.

If you have any technical support inquiry, please contact us at http://www.ASRock.com/Support/tsd.asp
We will assist you shortly.

Thank you

ASRock Support
Tech Support Email: http://www.ASRock.com/Support/tsd.asp
Click to expand...

I got them on the phone fairly quickly (they didn't respond to my online request yet) and they sound like they'll RMA it. Problem is, they don't cross-ship, so this fiasco is going to cost me weeks.

I'll probably run up to Microcenter in Denver, buy a new board (with an extended warranty, as it seems I'll need it with any of the TR boards out there... which all seem to die in a similar manner), and just sell the replacement when it comes back.[/quote][/quote]
 
So... weirdly enough, I unplugged the 8-pin and 4-pin power connectors and left just the 24-pin connector hooked up. Immediately after plugging the AC cord in, the power supply and radiator fans kicked on. I thought maybe this was a safeguard for when you're doing a BIOS flashback, but holding the flashback button did squat. Curious, I unplugged the power and plugged the 8-pin connector back in. As before, no fans when I connected the AC... just LEDs. This time, however, the flashback button started blinking and it read the flash drive. It ran for a few minutes and just like the flashback FAQ said, it stopped when it was done. Still, no post on version 1.30 of the BIOS. I tried everything, including removing the 8-pin connector so the fans would come on. Nothing ever came up on the diagnostic LED, so I'm pretty sure it wasn't even trying to boot (even though the fans were running. I also tried reflashing the 2.00 BIOS, hoping that would help. No luck.

Just for shits, I unplugged the 8-pin power connector and tried the 4-pin one (in addition to the 24-pin, of course). No fans with that either. So if either the 4 or 8-pin connectors are hooked up, the fans don't come on. If I plug in power with just the 24-pin connector hooked up, the LEDs AND fans will come on, but it still doesn't post.

Since it does the same thing even with different power supplies, I wonder if something is shorted on something shared with the 4 and 8-pin connectors? I wonder if a working board would boot without the 4-pin connector.
 
I finally gave up on the ASRock and ordered an Asus STRIX X399-E Gaming board. It's EATX, but it fit perfectly in my Thermaltake P90. Everything booted right up and it seems like a great board. Still have the memory stuck at PC2133, though.

Interestingly, it had a presets for Stilt's "Safe" and "Fast" PC3200 settings. I tried safe, but CPU-Z still show 1066 mhz instead of 1600. Oh well... I'm just happy to have my TR back up and running.
 
Flapjack said:
I finally gave up on the ASRock and ordered an Asus STRIX X399-E Gaming board. It's EATX, but it fit perfectly in my Thermaltake P90. Everything booted right up and it seems like a great board. Still have the memory stuck at PC2133, though.

Interestingly, it had a presets for Stilt's "Safe" and "Fast" PC3200 settings. I tried safe, but CPU-Z still show 1066 mhz instead of 1600. Oh well... I'm just happy to have my TR back up and running.
Click to expand...

which ram do you have and is it hynix or samsung B?
 
sirmonkey1985 said:
yeah they should be if they're cas 14.. i'd just step up until you get back to 3200.. might just take some voltage tweaking to get it back to 3200 but see no reason why they shouldn't run at that speed.
Click to expand...
I actually thought they were supposed to do 3200 at their stock voltage (1.35v). Of course, things seem to get cagey with TR boards, the more RAM you have. Only four sticks doesntd seem like too big a deal, though. Many who run 8 sticks seem to have to lower them a bit to reduce stress on the memory controller.

Check out the NewEgg reviews. Looks like someone got them very stable on a very similar Asus board at 3466.
 
Flapjack said:
I actually thought they were supposed to do 3200 at their stock voltage (1.35v). Of course, things seem to get cagey with TR boards, the more RAM you have. Only four sticks doesntd seem like too big a deal, though. Many who run 8 sticks seem to have to lower them a bit to reduce stress on the memory controller.

Check out the NewEgg reviews. Looks like someone got them very stable on a very similar Asus board at 3466.
Click to expand...

the imc's are really sensitive.. some need more ram voltage, some need more SOC voltage and some need less on both.. just have to play with it to find what works best with your setup. for example my cpu requires the ram to be at 1.37v and requires the SOC to be at 1.21v changing the SOC voltage up or down means my systems completely unstable.
 
sirmonkey1985 said:
the imc's are really sensitive.. some need more ram voltage, some need more SOC voltage and some need less on both.. just have to play with it to find what works best with your setup. for example my cpu requires the ram to be at 1.37v and requires the SOC to be at 1.21v changing the SOC voltage up or down means my systems completely unstable.
Click to expand...
So weird... so even if XMP did work, you'd still have to set it manually.
 
You must log in or register to reply here.
Back
Top