Retail Abit IN9 32X-MAX - Pictures, observations and testing...

Well, I removed two RAM modules and was able to run memtest about 25 minutes just fine atleast.. which is progress..

In windows now running the StressCPU test in Orthos on all 4 cores.. Been running for 10 minutes (at STOCK cpu speed mind you) and my PWM temps are as high as 142 degrees.. This with an extra 90mm fan hanging a couple of inches above my CPU waterblock.. (Without this fan they are 100-110 degrees idle.. with the fan they are usually 60-70 degrees idle.) CPu temp is 38-39 at this point and system temp is 48.. All according to the uguru software from Abit.

This is poor engineering** imo!



**Disclaimer: I know nothing about engineering.

Ok i was putting off installing my board till the rest of my w/c parts turned up but, seeing what is happening here i`m gonna put it in tommrow. I hope i haven`t wasted £235 on an unstable pile of crap!

Would someone be so kind to post the NB voltage is using?

I managed to (miracle) boot into Windows @533 5-5-5-15 2T 2.2v, but as usual my NB voltage goes to 1.45v in this setting, and it won't move down from there.

Is not something I do, is something the BIOS decides to do, regardless of my input.

Quote:

---

Originally Posted by uberowo

Well, I removed two RAM modules and was able to run memtest about 25 minutes just fine atleast.. which is progress..

In windows now running the StressCPU test in Orthos on all 4 cores.. Been running for 10 minutes (at STOCK cpu speed mind you) and my PWM temps are as high as 142 degrees.. This with an extra 90mm fan hanging a couple of inches above my CPU waterblock.. (Without this fan they are 100-110 degrees idle.. with the fan they are usually 60-70 degrees idle.) CPu temp is 38-39 at this point and system temp is 48.. All according to the uguru software from Abit.

This is poor engineering** imo!



**Disclaimer: I know nothing about engineering.

---

HOW HOT:eek:
Surely that amount of heat for any length of time is gonna kill componants real quick. Do the digital PWM on servers run this hot or is it just these 680 boards?

My NB voltage is 1.36-1.38..

So yeah, unless someone makes a SB/NB/PWM all in one waterblock for the 680i I'm :banana::banana::banana::banana:ed it seems.

Quote:

---

Originally Posted by \/ENDE][A

HOW HOT:eek:
Surely that amount of heat for any length of time is gonna kill componants real quick. Do the digital PWM on servers run this hot or is it just these 680 boards?

---

What Irks me about all of this is the "Otes" crap is not even solid copper, just coated that color! For a board that cost over $300 US that is awful! I hope someone comes out with a aftermarket solution for it soon. or a solution is found before my board arrives next week.

Ok i`m gone for now, post back tommorow when and if i`m up and running.

Ok, so now is not only the NB which goes to 1.45 because it feels like doing so...

It's also the CPU VTT, which is set @1.20v and uGuru is reading 2.20V... WTF?

Is this some sort of misreading?

This whole business is starting to piss me off... big time...

I'm going to bed... tomorrow I'll decide wtf I do.

mm yes i believe you could fry a very small egg on the nb without a fan ...teehee

re below; well yes you seem to have answered your own question...i'm teetering myself toward a revised 965...

580 then 590
680 then 690?
i think 680 is incompatible with sb..but i know nothing about engineering :) maybe nb and vrm just too damn hot?

i've had my windows totally :banana::banana::banana::banana: up with corrupted .hal and other system files after an unsuccessful oc on my old x850xt..had to repair with windows xp cd....i thought i had fried the gpu..but found that various windows files had just somehow become corrupted all of a sudden; no viruses; did windows repair evrthing hunky dory..still using it now.

i wonder if corruption problems to hdds could be related to chipset running out of spec. / too hot? 680 is very hot and some peopledont do anything extra to cool their chippies down...i'm just curious why some people seem to have problems with 680i and some dont???????????? some dudes still having trubble with latest bios's..do they just wait in hope of a miraculous bios update???

i am really considering this mobo however after my EVGA 680 nightmare, i'm a bit reluctant to go with another one, i do not need SLI my system is mainly use for photoshop, video editing and VMWARE running UBUNTU, i have the BAD AXE 2 right now, maybe the COMMANDO is a better choice

It's too early to tell if this thing is a total dud or not. The same thing people said about the Evga 680i and I had no problems with it. The only way is find out for myself, and I guess I'll be overclocking with some faith when this board arrives.

Where can I buy one of these? :)

Removed half my RAM as suggested above. (Now using 2x1GB in the blue sockets only on 2.2 voltage and default timings.) I kept all my other components in for now.

I now have two 90mm fans blowing directly onto the PWMs. Allthough they both about 2 inches above them.

I just finished running 4 instances of Super PI at the same time (one on each core) successfully. I've also been running 4 instances of SP2004's Blend test the same way now for just over 45 minutes without any problems. Unless ofcourse you consider PWM temps of 150 a problem..... The CPU temp was as high as 47 at one point, but generally around 42-43, and the system temp is steady at 48. Again this is stock speed on a QX6700. I literally think my board will catch fire if I overclock. :|

So aside from the HORRIBLE pwm temps and noise my system actually seems a whole lot more stable now that I removed 2 of my 4 RAM modules. Though the RAM modules were all fine when I had the EVGA board.. So I duno..

Is this enough to consider my computer "stable"? I'd leave it on overnight but I'm not really comfortable with the ridiculous pwm temps.. Kinda worried I might damage the board permanently..

Quote:

---

Originally Posted by syne_24

It's too early to tell if this thing is a total dud or not. The same thing people said about the Evga 680i and I had no problems with it. The only way is find out for myself, and I guess I'll be overclocking with some faith when this board arrives.

---

mm faith ... start praying then to the 680i godz...

Quote:

---

So aside from the HORRIBLE pwm temps and noise my system actually seems a whole lot more stable now that I removed 2 of my 4 RAM modules. Though the RAM modules were all fine when I had the EVGA board.. So I duno..

Is this enough to consider my computer "stable"? I'd leave it on overnight but I'm not really comfortable with the ridiculous pwm temps.. Kinda worried I might damage the board permanently..

---

i dunno either :rofl:

Uberuwo, do you have the stock thermal pads in place on the PWM portion of the heat pipe, or did you replace it with thermal compound?

Mr. Toad, based on what you are reporting, "something" is very odd about your board. Your VTT reporting etc is making me wonder.

I haven't altered the board or anything on it in any way.

You guys have any thoughts on the 2GB vs 4GB RAM situation?

I cant run memtest for 20 minutes without getting errors if I have all four modules.. Any two of the four modules will work, just not all four together..

Wonder how Bioses engeeners are working ? Did they try several stuff ?

The molex location problem, the EPP issue, Temperature issue....

Quote:

---

Originally Posted by Johnny Bravo

As for the Molex connector for the SLi power stiffening, yeah that was a bad move. Either the could use a 90 degree connector or include a similar cable like that shown below

http://www.chieftec.de/images/_2sata.jpg

For those in the UK, if it's a real problem then why not try this:

http://www.kustom.co.uk/acatalog/info_1951.html

---

yeah, that's going to be the ticket for fitting the molex with the second card in.. that seems to be my only complaint about this board's layout.

Quote:

---

Originally Posted by \/ENDE][A

Ok i was putting off installing my board till the rest of my w/c parts turned up but, seeing what is happening here i`m gonna put it in tommrow. I hope i haven`t wasted £235 on an unstable pile of crap!

---

464$???

Ouch.

I am getting this board in 4-5 days (waiting for it) and I am really teerrified about all those horror stories that appear mainly on SLI setups.

I really wish I had not bought the second 8800GTX and this 30" Monitor with the ridiculously high native resolution of 2560x1600 that forces me to use SLI.

I thought the market and the current products where ready for such a setup but it seems I am wrong. God damn me I wasted money it seems since Nvidia deserves a lawsuit.

How the hell can they release a freaking platfomr that does not even work with their own products and have the adacity to charge those outrageous prices ffs! And what I hate most is they will come out with a revision soon too.

And the problem with this board is that its not easy to return it in my country after I have bought it.

A few more observations on this board from another user

1. I am using an X6800 chip, and niether the shipping bios nor the 11 beta bios allow upward adjustment of the multiplier (by the way, Johnny Bravo - thanks a lot for posting that bios)

2. I couldn't get a stable FSB above 425 no matter what I did with the voltages (Johnny, your 495 is impressive, can you let us know what settings you used to get there).

3. Using a single 8800 GTX card, the video performance was horrible, no matter what I did with Linkboost or the Nvidia GPU ex option in the bios.

4. I started getting disk corruption errors after an hour or so of playing with the board, and without pushing the memory too far.

5. I didn't get any of the digital PWM temp problems others have reported, and it ran my Mushkin XP8000 memory fine straight off the bat (though given other difficulties I didn't tspend too much time testing it all the way up to 1280 I can get out of it on the EVGA board).

I also have the EVGA 680i board, and my experience with this Abit board feels just like the EVGA board when it first came out - at every turn I feel like I need a better bios. Given that Abit have been later to market with this, I expected it to be much more stable and a better FSB overclocker straight off the bat, but it isn't. After a month or so of bios revisions, I can run rock solid stable at 465 FSB on the EVGA board, and have gone back to that for now. This Abit board feels like it has potential, but for me the bios is letting it down.

One final observation. Abit's failure to include the 40X40X10 fan you need to stick on the Northbridge when overclocking is, in my view, unforgiveably cheap given the price point of this board. Abit should be embarassed about that!!

Neonflx - you say are considering this Abit board for a system that doesn't need SLI, and when you already have BadAxe2. Having played with both the EVGA and Abit versions of the 680i board, I honestly can't see why you would even be considering 680i if you don't need SLI support. If it wasn't for trying to keep the SLI door open, I would be steering clear of them. Just my view....

Update:

-The odd voltage reading was a "glitch" of uGuru. Used a different version of the app to the one included in the installation disk. Once installed the version included on the installation disk the readings are correct.
In BIOS is set to 1.2v and it reads 1.20~1.22v.

-If from the very beginning after a CMOS reset I never enable EPP it doesn't play around with the NB voltage. It allows me to change it freely without reverting to that 1.45v setting.

Now, Yellowbeard, before I start RMA-ing things, could you please tell me exactly what parameters you used to make the 8500C5D work for you?
And I mean all the relevant parameters. Basically, can you include a SShot of:

- uGuru screen.
- Memory Timing Setting screen.

And also tell me which devices have you enabled, which disabled, which settings of the BIOS you've changed from defaults, and finally which version of the nVidia drivers are your using, the ones included on the CD or the newer ones available on the nVidia website.

I know is a lot of data to gather just for the needs of some idiot that simply can't put four pieces of hardware together, but at this point in time I've run out of ideas.

By the way, I've been in Windows for 10 minutes w/o crashes with 533 5-5-5-2T set manually, and w/o oddities in the voltages.

Silly as it may look, as normal behaviour is abnormal in my experience with this setup, I'm in the borderline of an adrenaline rush :woot:

OK a few things...

Who exactly is having 8800GTX problems??? It seems there are complaints about this but the majority of folks running SLi on these cards can do it easily enough.

As for the molex connector, I have some news. This is only needed if your card DOES NOT have an on board power connector. Basically it "stiffens" the power supply via the PCI-E slot. So everyone with big ass 8800 cards there's no need for you to be putting that connector in, that's why it won't fit!

For those of you with SLi compatible/EPP RAM who find they cannot change their voltages....that's because your RAM is changing the voltages for you, that's what EPP does! If you set it to manual or disable the sli compatible RAM feature then you'll get your control back.

For those of you who are running 4 sticks of RAM don't be surprised if your FSB clocks are low, 4 sticks of RAM is a pretty big strain on the NB chip. THE same goes for users of the QX6700 CPU, it's well known that this cpu does not have a great FSB headroom.