Can mobo beep for temperature even when alarms are disabled?

[Aussie Allan]

   Cool ..... keyboards are just so cheap now ..... but borrow one for testing by all means ..... it's all about testing through a logical path , piece by piece and continually narrowing down possible causes ..... keyboards next!

  Do you have any-other bits of kit connected up while were on the subject ?

  Aussie Allan

[Dark Mantis]

Stabilisers just stop any peaks and troughs from appearing and putting a strain on your system but to do a proper job need a battery built in which is where the UPS comes into play. It is not just for blackouts but for brownouts and overvoltage and undervoltage situations. I always run everything through a UPS and am confident that electrically it is safe from the mains.

[Rolo42]

I think pretty much everything turns off with no input voltage.  

UPS are good for:
- Automated graceful shutdown/hibernate during a power outage (or just keep running if power is out for only a few mins)
- Boosting power during a brown-out
- Disconnecting your equipment during a surge (switches to battery)
- Line conditioning
- Sacrificing itself to save your equipment when lightning hits your tree (true story--six of my neighbours lost equipment from that; I did not; they all bought APC UPSes after that)
- Peace of mind and not having to worry about unplugging equipment during heavy storms/bad power instances

Without knowing how your stabilisers are built, I'd wonder if they filter sags/surges that would damage a complex PSU (or other sensitive electronics) vs. keeping your household appliances from frying.

Realistically, line input would be the last thing I'd suspect if a motherboard had power problems; I'd look at the PSU itself first, then its load (PC components, fans, etc.).

[Dark Mantis]

Realistically, line input would be the last thing I'd suspect if a motherboard had power problems; I'd look at the PSU itself first, then its load (PC components, fans, etc.).

I would have to agree with that sentiment. Although possible it is rare for the mains supply to be the cause of a problem on the computer. Look for the common causes first.

[www]

Thanks for the UPS info.

It has been 2 days of intensive use and no beeps so far, the only thing that changed is that the weather, now it's not so hot and electric input is normal, that's why I suspect the cause could be the electrical network. Because temperatures of the components never went above 63c, normally max temp was 58c

The temperature I still couldn't get  is the one from the motherboard, all softwares show "25c" for motherboard and never moves, but I touched the northbridge and it's damn hot.

I guess I will have to wait for another really hot day to keep testing because it seems the only common thing when beeps occur.

I also contacted coolermaster to get info about the protections the PSU has but didn't got a response yet, it has temp and voltage protections but I don't know the values, I only know the operative temperature is 0-40c, but don't know if there's a way to verify the current temp on a PSU.

Does someone knows how I could get the motherboard temp?

[Rolo42]

CPUIDs Hardware Monitor has everything and logs min/max.

Another possibility is that heat could indirectly cause power problems if components are broken down enough at higher temps but not enough to be in tolerances at lower temps (inverse temperature coefficient).

You can test this by either lowering the temps (remove side panel, place a floor or desk fan blowing into it or raising temps with a heat gun or hair dryer (be careful on that one).

[Aussie Allan]

 Hay Rolo42

  ....."inverse temperature coefficient" ....? .... do you mean ....     \operatorname{R}(T) = \operatorname{R}(T_0)(1 + \alpha\Delta T)

      or are you referring to thermal breaks/switches when tracks or wires become fatigued due to thermal oscillation ?

  If you are talking about inverse temperature coefficient of materials used in motherboard manufacturers .... they do all they can to match thermal coefficients of mating materials used ..... but an interesting though none the less.

  In the original post .... everything is new except the PS2 keyboard.

  Aussie Allan   

[Rolo42]

As heat increases, resistance decreases.  

If you have a diode, for example, in a rectifier and its broken down just enough to work at room temp but as ambient temp in equipment rises, the diode's resistance decreases (when it shouldn't), letting current--in this case, negative voltage from AC--thorough when it should be blocked.  This will put a sine wave on your DC voltage (it should be flat).  

If this were to happen with TTL semiconductors, signals (your 1s and 0s) would get garbled.

"New" means "untested"; you can't assume something works because it is new.  Standard procedure is to test it before deploying/sparing it.

This is an interesting problem (one that would drive me nuts; it reminds me of the Stormtroopers of Death track, "What's That Noise?"); I can't wait to see how this turns out.

[Dark Mantis]

"New" means "untested"; you can't assume something works because it is new.  Standard procedure is to test it before deploying/sparing it.

Quite correct and something people often fail to understand.

You are also right about the way faults can occur in a circuit even though they can be very hard to track down. Often it requires reverse procedures where you need to understand the effect to find the fault.

[Aussie Allan]

  While it's not too busy ...  .... I agree ! ...........to a point!

  Years ago I was a project manager with Union Carbide (not Bhopal India!)  ... one of my duties was implementing new technology, quality controls for the then new AA/AAA "Energizer Battery lines"

  Each component ..... end-caps .... inner/outer casing .... carbon black mix .... all have a reject tolerance standard ..... 1 in 1000 requires a separate inspection of the batch lot.....in the closer inspection......5 in 1000 requires a rejection of the entire lot.

  This is on top of the final inspection of the assembled finished product..... PCB products will go through an individual testing station (Nested conveyor) ..... yes ! all the boards will be tested prior to boxing for basic connectivity and operation .

  I think most people have little insight into the level of automation and testing that goes into even our humble motherboards.

  Aussie Allan 

[Dark Mantis]

Agreed! Some companies however have a much stricter testing regime than others I am sure you will agree. Some purely batch test whilst others do a multiple test but still in batches and some go much further.

[Rolo42]

Without knowing the QA process to our PC boards, we can assume:

- Assembly line diagnostics are likely basic (might not be any more than POST pass)
- Bench tests can't check every possible configuration or use (especially given the variety of components that will be plugged into it)
- Diags won't always catch intermittent faults
- There is a point where RMAs will be more pragmatic than QA/validation
- There may be issues the manufacturer is willing to accept (why I quit buying Asus)

Finally, no matter how thorough, validation can't account for anything that happens when the product leaves the factory--shipping, handling, end-users' fondling...

[www]

Wow, thanks for all the great info. I went to a technical school but I still only understand 50% of the last posts buts its REALLY interesting, you guys know a lot of stuff.

Anyway just wanted to make an update on this topis, first... I didn't got any beeps since my last posts and I have been using it from 10am to 6am every single day, mostly to watch TV. Weather is MUCH cooler now, so thats interesting,  but also the electrical network is normal at 220v most of the time. I am still waiting for a hot day because I am kinda afraid of shooting the pc with a heat gun

I am planning to place a vga cooler over the northbridge heatsink anyway, would it be better to use those special glues or just screws on the slots? ( I am talking about this heatsink : http://images.bit-tech.net/content_images/2010/11/gigabyte-ga-p67a-ud4-p67a-ud3r-preview/p67a-ud4-sata.jpg btw what are those white things with springs on the two corners? )

Also I got a response from gigabyte support, pretty useless compared with the info you guys gave me but at least they tried jeje, one told me to clear cmos and another to test another ram and update bios.

Thanks again for all the info guys, I enjoyed reading.

[Rolo42]

The white pegs are what keep the heatsink mounted to the motherboard, much like the really annoying pegs Intel CPU coolers use (only these aren't annoying).

That chip is the southbridge ("PCH" now); cooling it will do absolutely nothing.  There is no northbridge (it is now the "uncore" in the CPU).

Seriously, fire up HW Monitor and use a hair dryer to increase ambient temp in the case to see if the problem returns with increased temps.  You could then switch of the heat and cool it back off.

[Aussie Allan]

 WWW.....just another opinion...

  The south-bridge does come with a heat-sink for a reason ....although quoted correctly by Rolo42 that a lot of the overhead has been relieved by the CPU , the south bridge still has a fair bit of work to do handling the traffic from the SATAs it supports ..... if you have just one or two drives.....there should not be a problem ....... if you load up all these ports ...... and say add 4 more drives via a RAID card ..... the south-bridge has oodles of work to do.

  As the designers envisage say up to 4 drives running through the bridge..... the sink is formulated to overcome that load ..... if you exceed that mark/parameters.....and clock as well ..... a little help could be needed.

  If you have access to a non contact or contact temp prob/monitor (or use the on-board monitor) ...... check the temp under idle and load of the sink ...... if it rises over 50C ....it's worth looking at ..... even for extending the life of the chip ...... it could range from a better thermal paste, small after market 40mm fan (or both) right up to water cooling.......I have a lot of drives so went the Enzotech bridge cooler route  and never see temps rise above 36C even when encoding .....

  Aussie Allan