PC builders please help: new build, endless power cycling, no POST.

[Jim Roseberry]

33 minutes ago, Jack Stoner said:

13th gen is out.

 

Yes.  13th gen has been out for several months.

13900k performance is amazing (scores over 40k with Cinebench R23 multi-core).

Requires robust cooling (as in large AIO)

[hockeyjx]

1 hour ago, Jack Stoner said:

13th gen is out.

 

Then +1 to what I said ?

[Starship Krupa]

17 hours ago, hockeyjx said:

Intel is on Gen12 ...yours is a Gen6,  so I wouldn't throw a ton of time/resources/money at it frankly.

You would not.

I'm a frugal dude, I like to tinker, I like to get extended use out of hardware that other people might consider outdated, and I only care about generations of processors insofar as the next one I get should be a performance upgrade over the one it's replacing (I usually like to at least skip a generation when upgrading). Part of the fun of the hobby: how much can I do with how little? I don't have a day job, my spare time is plentiful.

I let other people spend the big money; 5 years later I'll spend a couple hundy for what they spent a thousand to buy. If I even need to, often enough folks are kind enough to give their old hardware to a good home.

My sweet spot seems to be about 6 years from the processor's introduction date, but until the middle of last year, I was happily using a 3rd generation Ivy Bridge processor for DAW, NLE, and indie gaming (with the help of a $15 CL'd GTX550Ti). One of the Cakewalk devs (a true Scotsman) uses an i7-3770 system as his main studio computer.

I recently put together a Core 2 Quad system with 8G of RAM for a friend to use for web browsing and video watching. I installed Cakewalk and it runs just fine. He records a couple of tracks of audio and maybe one or two soft synths per project, so he'll never hit the wall. If he wants to get more sophisticated, he can have my old i7-3770 system.

With the i7-6700, programs start faster. I added a GT 1030 and I can even run any game I want (Obduction, Quern, Pneuma, Firewatch, Lightmatter, Manifold Garden, Turing Test) at ultra settings. It was well worth the outlay of about $200 (for motherboard/CPU/RAM/SSD, PSU and graphics card; I already had a donor case and all the peripherals).

Is there anything other than the latest FPS games I can't run on my i7-6700 system with 32G of RAM? I don't think $200 was a poor investment for a computer that does everything I want it to (and even more).

8 hours ago, Jim Roseberry said:

I'd reseat the CPU...

That's more of a long-term solution (that was my point).

You might also reseat all power connections to the motherboard (worth a try)

I guess it was tl/dr, but I said in my first post I reseated the CPU. While it was out, I checked for bent pins and found none. I also reseated all power connections, multiple times. Including fans.

I'm probably looking at a used LGA 2011-v3 motherboard from CL or eBay. No pressing need, I can wait until they come down in price a bit more (maybe Intel will be up to Gen14 by that time ?). When the ASRock motherboard inevitably mysteriously begins working perfectly the minute I get its replacement home, I also have an i7-6800K that will fit nicely.

[Jim Roseberry]

17 hours ago, Starship Krupa said:

I guess it was tl/dr, but I said in my first post I reseated the CPU. While it was out, I checked for bent pins and found none. I also reseated all power connections, multiple times. Including fans.

Sorry, I missed those details.

If you've already reseated the CPU and all power connections, the only reason loosening the heatsink mount could cause things to "start working" is cold/bad solder joint/s or problem with circuit trace/s.

[Starship Krupa]

1 hour ago, Jim Roseberry said:

the only reason loosening the heatsink mount could cause things to "start working" is cold/bad solder joint/s or problem with circuit trace/s.

Yeah, that's what I fear. Some surface mount component or other probably cracked its solder joint. My guess would be something that causes the power supply to fail to get a "power okay" signal back from the board.

I guess it's a writeoff. I can get a nice used board on eBay for a little over a hundy, so it's all good anyway. Yesterday's technology today!

I'll see if I have an old PSU around I can sacrifice to an experiment: pull the CPU and trick the PSU into coming on anyway and see if I can let out some smoke. ?

BTW, for the "latest and greatest" crew, Intel have announced that Gen. 14 is due before the end of the year. Make room in your dustbins.

[Jim Roseberry]

6 minutes ago, Starship Krupa said:

BTW, for the "latest and greatest" crew, Intel have announced that Gen. 14 is due before the end of the year. Make room in your dustbins.

Tight competition between AMD/Intel is really going to drive performance. 

[Starship Krupa]

I ordered a Gigabyte UD (ultra durable) motherboard on eBay. After I did, I tried this one again and it’s still hosed.

(Based on my empirical observation that sometimes the act of buying a replacement will bring the dead one back to life)

[kitekrazy]

6 hours ago, Starship Krupa said:

I ordered a Gigabyte UD (ultra durable) motherboard on eBay. After I did, I tried this one again and it’s still hosed.

(Based on my empirical observation that sometimes the act of buying a replacement will bring the dead one back to life)

So is this the end of the story?   Maybe CPU is hosed?

[Starship Krupa]

6 hours ago, kitekrazy said:

So is this the end of the story?   Maybe CPU is hosed?

Since the motherboard acts no differently when I remove the CPU entirely, and computer issues are so seldom “bad CPU,” I doubt the CPU is shot. It would be my first bad CPU in 34 years of building and servicing PC’s. Not impossible, though. The board doesn’t even beep when I have a little beep speaker plugged into the header.

If it is, the box of components the guy sent also included an i7-6800K that I can put in.

Based on my Googling and what Jim said about them, the takeaway is probably going to be “I will steer clear of ASRock motherboards in the future.”

Pretty much every thread I found where someone was having this issue ended with “got a replacement and it booted right up.” One person loosened their cooler a little and it booted, which implies a motherboard defect (bad solder joint), and it wouldn’t surprise me if they had issues with that board down the road.

[Starship Krupa]

This has turned into one of the weirdest troubleshooting-by-substitution adventures I've ever encountered.

I got my Gigabyte motherboard today, loaded it with the i7-6950X and 16G and installed the cooler. Then removed the ASRock from the case and installed the Gigabyte. Guess what: same damn thing. Only perceptible difference is that the Gigabyte has spiffy orange LED's on it that light up and go dark again in sync with the fans.

Okay, so at that point the only things I hadn't tried swapping out were the CPU and the power supply. In goes the i7-6800X. Same damn thing.

Well, we haven't swapped out the PSU, did it somehow maybe get compromised at some point? Disconnect PSU, remove Corsair from daily driver system and install it....and....same damn thing! Pulled the RAM stix just in case, nope, same damn thing.

This is with two motherboards, two CPU's, two  different video cards, two SSD's and two power supplies!

The only things not swapped out at this point are the cooler and the case, although the original motherboard did the same routine before I put it in the case. I only have two DDR4 RAM sticks, but I've tried it without them; these motherboards should at least POST to a memory error if no RAM is installed.

WTH is going on here?! I'm an experienced clone wrangler; 6 months ago I built a new (to me) system out of various parts, some of which I've tried in this build.

Is it possible for a CPU cooler to cause this kind of trouble? It's a Noctua tower, and for a while, I used the nice and quiet Noctua fan from the cooler as a case fan in my daily driver and it worked a treat. It spins up and back down with the rest of the fans (case and PSU). But it's the one component common to every phase of this fiasco. The cooler mount attaches to the standard 2011-v3 socket, so there's not much chance of error in installing it.

I did notice after removing the cooler that it doesn't seem to be perfectly flat; the paste gets thin-to-nonexistent in a stripe across the center of the CPU and contact surface of the cooler. A short between cooler and CPU case shouldn't be a problem....should it?

I imagine that the start-up sequence for a PC probably goes something like:

1. power button momentarily shorts pins on the PSU
2. PSU sends juice to the mobo
3. mobo sends signal back to PSU saying that the power is all okey dokey, no shorts, noise floor low enough, I'll go ahead and boot now, you can stay on
4. If the PSU doesn't get that signal, it kills power to the mobo and tries again

People with greater knowledge please correct and/or fill in the blanks.

If this is true, then my motherboards must be somehow telling the PSU that the voltage is too low, hence likely a short somewhere.

The wattage of my PSU's is more than adequate; one is an EVGA 500W and the other a Corsair 650W.

[Xoo]

5 hours ago, Starship Krupa said:

WTH is going on here?! I'm an experienced clone wrangler; 6 months ago I built a new (to me) system out of various parts, some of which I've tried in this build.

Partial shorting to the case?

[Cookie Jarvis]

My first build many moons ago wouldn't power up....forgot to put the standoffs in so the MB was contacting the case Put the standoffs in and voila!

Bill

[Starship Krupa]

7 hours ago, Kevin Perry said:

Partial shorting to the case?

Before the case was here I tried to power it (meaning the ASRock) up with no case at all, sitting on a box. Haven't tried that with the Gigabyte yet, but I will at some point.

[Steev]

On 2/4/2023 at 5:18 PM, Starship Krupa said:

I ordered a Gigabyte UD (ultra durable) motherboard on eBay. After I did, I tried this one again and it’s still hosed.

(Based on my empirical observation that sometimes the act of buying a replacement will bring the dead one back to life)

Hmm, I'm pretty sure the Gigabyte UD with PCIe 3 is designed for 8th- 9th gen i7, and don't think this mobo supports older Gen 6 i7, or vice versa.. ⁉⁉⁉.  I'm not sure though, I've never had much luck with quality or longevity with Gigabyte. However, once you get past the setup quirks and stuff that come with any new and or Nuused/repurposed computer builds, I can't remember EVER having an ASRock build fail, and shortly after building my 1st AMD FX 8300 running Window 7 in x64 machine, I did some pretty insane utterly STUPID overclocking "crash test dummy" stunts with an AMD Phenom II 965 4 core Black Edition, which did not crack. But Windows XP did. XP became very unstable at around 4.6 g/hz? I still have that old beast, with no overclocking of course, it's native 3.5 g/Hz really is FAST enough with Win 7 x64 installed, but that old, old, OLD ASRock mobo only supports a whopping 4 gigs of RAM (which was monumental at the time) is bare minimum today.  Although I haven't used it in at least a decade I'd be willing to bet it'll fire right up and run if the old Western Digital VelociRaptor 10k RPM hard drive spins up.

However, after carefully reading thru this thread, and having similar various difficulties with sparking up ASRock builds, particularly with older UEFI BIOS Secure Boot v1 chips made before 2018, and the VERY unfortunate power hit targeting hyperthreading side effects from Intel's patch forced by the introduction of the "Meltdown Exploit" which led to another level of Intel Hell temporary slowed down by a deeper level of security of the introduction of the "SecureBoot" chip v 1, which now, in case you didn't notice yet, Microsoft deepened security a little deeper to drop support of v1, and will only allow upgrades from Windows v10 to v11 if we have the newer and yet to be determined better "SecureBoot" chip v2 installed on our mobos. 

Before starting make sure you have all the motherboards systems drivers handly on a CD or USB 2 stick. OR simply plug an internet cable into it,, I have no idea if the following applies to Gigabyte silicon, but I have had success getting into BIOS with building/rebuilding ASRock motherboards without them quitting before getting started..

HOPEFULLY, but not always, just by connecting a CAT 5 cable before hitting the start button, it will automatically connect and report to the ASRock "Mothership" shortly after hitting the power button and check & install the latest BIOS version if need be and system drivers, all to basic most likely to succeed default settings. Just follow the on-screen instructions. (This auto method will not work behind a Netgear of Linksys or any hardware "Internet Firewall".)

1.) You "must" have a mouse and keyboard plugged into their dedicated USB ports. The  Fatal1ty has is own patented "Key Master" technology that controls and finetunes the mouse and keyboard. and they are the 1st thing the BIOS looks for before starting.

2.) You "must" have a SSD or hard drive installed and plugged into a SATA or SATA supported M2 slot. That is where the UEFI BIOS is or will be stored on it's own patrician before proceeding.

3.) You "must" have a Graphics card installed in a slot that doesn't share a PCI buss with anything else..

And this is VERY IMPORTANT. M2 slots (#) and corresponding PCIe slot # share the same port to the PCI buss, so you CANNOT PLUG & run an SSD in M2 slot 1 and your graphics card plugged into PCIe slot 1 at the same time. I recommend leaving the M2 SSD in the 1st slot, and plugging the graphics card in the last PCIe slot.

4.) At least 1 stick of memory in slot 2 to run in single chan. mode, or slots 2 and 4 for duel chan mode of "Matched Pairs" of EXACTLY  the same speed & size of high quality HIGH PERFORMACE RAM.

I highly recommend going to the ASRock website and only using "approved" memory for the Fatal1ty .

It is the BIOS's job to make find and make sure all components and peripherals are connected to even be a computer, If it can't find ANY of the above components hooked up properly, it will not run the BIOS and simply shut down.

GOOD LUCK! ?

 

 

Edited February 6, 2023 by Steev

[Jack Stoner]

This time of year in the cold climates static electricity can be an issue.

Even in warm Florida, where I live, Murphy's Law applies and I wear a grounded wrist strap when working with computer "innards".

 

[Jim Roseberry]

On 2/6/2023 at 12:52 PM, Jack Stoner said:

This time of year in the cold climates static electricity can be an issue.

Even in warm Florida, where I live, Murphy's Law applies and I wear a grounded wrist strap when working with computer "innards".

On the first build I ever did, I used a grounded wrist-stap.

For the hundreds of builds since then, I've never used one.

Unless you're walking on wool carpet in sock-feet and touching ICs, you'll be OK.  ?

You can discharge yourself as you're working.

 

[Xoo]

I'm with Jim  although I may have used a strap for my first two builds...

[Starship Krupa]

Yeah, motherboards and CPU's pretty much contain only bipolar transistors. It's MOSFETS that can get zapped, but I must say that in my many decades as an electronics hobbyist and professional, never using a wrist strap, I have never killed a MOSFET with static.

This situation has got me baffled. I literally replaced everything except for the CPU cooler. I've just set it all aside. I have other things to attend to and my current daily driver is fully capable of handling anything I throw at it. It was kind of a "for fun" project that has turned into NOTfun.

There is surprisingly little information on the web about the power-up sequence of a PC. All I get are highly detailed descriptions of what happens at POST and after. I want to know about what kind of "all okay" signal the PSU needs to get from the mobo before it will throw the juice to it. Because I think that's what's happening: the PSU switches on, then gets some kind of signal back saying that all is not okay in moboville. Like dirty power, or drawing too much current, or something. I used the tower cooler's fan as a case fan in my daily driver for a while, so I know it's okay.

Baffling.

Edited February 8, 2023 by Starship Krupa

[kitekrazy]

On 2/8/2023 at 10:27 AM, Jim Roseberry said:

On the first build I ever did, I used a grounded wrist-stap.

For the hundreds of builds since then, I've never used one.

Unless you're walking on wool carpet in sock-feet and touching ICs, you'll be OK.  ?

You can discharge yourself as you're working.

 

Don't most of the boards these days protect against static shock?

I live in the desert and always worry since humidity is all or none.

[Jim Roseberry]

On 2/12/2023 at 8:17 PM, kitekrazy said:

Don't most of the boards these days protect against static shock?

I live in the desert and always worry since humidity is all or none.

Motherboards have protective coatings.

This coating technically doesn't prevent ESD damage.

Modern components are a bit more resilient... but not immune to ESD.

 

I've worked on hundreds of machines over the past 30 years.

Never had an issue with ESD.