Warning: main(php/nav/nav2.php) [function.main]: failed to open stream: No such file or directory in /home/pcx4010/public_html/845ultra-aru2.php on line 46
Warning: main(php/nav/nav2.php) [function.main]: failed to open stream: No such file or directory in /home/pcx4010/public_html/845ultra-aru2.php on line 46
Warning: main() [function.include]: Failed opening 'php/nav/nav2.php' for inclusion (include_path='.:/usr/lib/php:/usr/local/lib/php') in /home/pcx4010/public_html/845ultra-aru2.php on line 46
MSI 845Ultra-ARU Review - Page2
0000
Here are the standard IDE and floppy connectors, as well as the badly placed ATX power connector. It seems ATX power connector placement is different on almost every motherboard. I prefer it to be on the opposite side on the board near the rear USB connector. In my opinion, MSI put this ATX power connecter in the worst possible place.
0000
Promise MB FastTrak ATA133 RAID controller, possibly this motherboard's best feature.
0000
Here we have the D-Bracket diagnostic on top and the 4-port USB2.0 bracket on the bottom. Even though the diagnostic D-Bracket is a nice thought, it doesn't always work like it was meant to work. In fact, I would not put much stock into what it reports, and I will tell you why near the conclusion of this review.
Performance and Overclocking:
Test System:
- MSI 845 Ultra ARU Intel i845 Socket 478
- Pentium4 Northwood 1.6A Retail boxed
- *Board #1: 2X 256MB PC2100 DDR CL2.5 Nanya memory modules
- *Board #2: 1X 512MB PC2700 DDR CL2 Kingston ValueRam
- 2X Maxtor 40GB ATA100 7200RPM hard drives in RAID-0
- Abit Siluro GeForce4 Ti4400
- Fortron/Source SFP400-60PFN (With power factor correction) PSU
- HP 9100i CD-RW
- Cisco Aeronet 340 Wireless LAN card
- Kingston 10/100Mbit NIC
- WindowsXP Professional
- DetonatorXP 29.42 NVIDIA drivers
- MadOnion 3DMark 2001
- SiSoftware Sandra 2002
This motherboard has some nice overclocking features. The BIOS allows adjustment of front side bus clock frequency from 100Mhz to 200Mhz, in 1Mhz increments and also allows CPU core voltage adjustments from 1.5v (stock) to 1.75v. There has been much talk about Pentium4 CPU's dying from too much juice. The general concensus seems to be anything over 1.7 volts will kill the Northwood, but I have been overvolting them to 1.85 volts for about 6 months now with no fatalities - maybe I have just gotten lucky. But, if you need more than the 1.75 volts that this board allows, you can find out how to do it right here.
In the BIOS, you also boost the DDR and AGP voltage by .1 volt if you need to. It's not much of a boost, but it is better than nothing. On to the good stuff, benchmarks...
So, anything over 1.7 volts will kill a Northwood right? Well, behold a 1.6A Northwood running at 2.53Ghz overvolted to 1.85v! This benchmark was made on June 22nd 2002. This CPU is still alive and kicking in this PC that I am writing this article on right now, 2 months later. At these settings (158FSB) this little 1.6A Northwood outperforms a stock 2.53Ghz Pentium4. (533Mhz data rate vs. 632Mhz data rate)
These memory benchmarks are impressive for PC2100 DDR. I used 2 sticks of 256MB Nanya PC2100 rated CL2.5, but they are running CL2 at 316Mhz, with no errors or crashes at all. The system is completely stable.
The numbers are nearly as good as PC800 RDRAM!
Impressive is not the word for this benchmark. Amazing would be more appropriate. The Promise RAID controller comes through with flying colors here! Maxtor drives are known for mediocre performance in RAID-0 configurations, but that is certainly not the case this time. Promise does an outstanding job optimizing the 2 Maxtor drives in this performance array, and you can really tell the difference when opening big applications such as Adobe Photoshop!
I decided to overclock the Abit Siluro Ti4400 to 300/600. Not a major overclock, but very stable. I managed to get 12398 3D Marks on this run, which is the highest score I have ever gotten!
Too bad I didn't record it at MadOnion, it would have gotten me on the Fastest Webmasters list...
Well, "Why don't you just do it again and record it"? you might ask.
Now, I will tell you why! And I will tell you about the major problems I mentioned at the beginning of this review.
In the course of this review, I have gone through not one but TWO motherboards. Let me say this: I have been overclocking for 3 years now, starting back in the day when the Celeron300A was getting kudos for easily O/C'ing to 450Mhz, and in all that time I have never seen a motherboard completely FAIL. As in completely and utterly DEAD...until now. Not only one complete failure, but 2 boards in just 6 weeks.
The failures were identical in nature - memory controller failure. The board wouldn't post, and the D-Bracket said there was no memory in the DDR slots when there was. I said earlier not to depend on the D-Bracket for info, but since I had no other way of diagnosing the problem it's the only indication I had. The first time it happened, I thought maybe I had in fact killed the CPU but after removing it and testing it in another board along with all the other major components, it turned out to be the motherboard that had failed.
I was at a loss for words because I had never killed a motherboard before, much less in such a short period of time (five days). So, I RMA'd the damned thing and patiently waited on the replacement board which took 3 weeks to get back.
Ok, so now we are into mid-July and I finally get the replacement board. I promptly put the system back together, and it ran at the same overclock as before with identical performance. Three weeks later, the 2nd board dies...
Well now I am convinced that it is a design flaw of some sort, or perhaps a weak component on the board. I talked to Iggy over at Overclockedpc.co.uk and he reported to have a similar problem and an AMD based MSI motherboard. He said he thinks the problem may be bad or cheaply made mosfets on the board, which is a strong possibility.
I also want to add that I had problems with cold boot system hangs which I attributed to the radical CPU overclock, but in retrospect it may have been caused by the motherboard itself. The problem was not consistant. Sometimes it would hang, and sometimes it wouldn't.
So, if you have had similar failure with any MSI 845 based motherboards please let me know!
Conclusion:
All the performance this board is capable of use totally useless if you can't depend on it to stay running. Granted, I was pushing it beyond its advertised specs, but as I said I have been doing this a while now and have never seen failures like this. This is a serious problem for the performance crowd that this product is aimed at. This is like a car salesman selling you a Corvette, and then you blow your engine after hitting 70MPH! And then, you take it back and it happens again!
This would tend to leave a bad taste in anyone's mouth. I think this indicates a lack of quality parts and/or quality control. But then again, it could just be really really bad luck on my part. But I'm kinda doubting that.
Let me also add that this motherboard is completely stable at advertised specs. Overclocking killed both motherboards, there is no doubt of that. My main rant here is that they were manufactured and marketed for overclocking, but overclocking kills them... rapidly. In light of all I have had to go through to complete this review, I am giving this board a 1X rating unless someone can shed some light on the problem, or I can get more information on why these boards failed so quickly.
Update: Sept. 4, 2002
So far, I have received 22 reports from people that have had or still have this motherboard. Some report that they have had failures like mine, some say they have the warm boot problem but no failure, and others report no problems at all. Here is the tally so far:
- Total failures - 9 (one was not overclocked at all)
- System hangs, but no failure - 8
- No problems at all - 5
I also asked one of my suppliers about the RMA rate on this particular motherboard. He said that based on the units he sold, the RMA rate was a little high but not extremely high.
And lastly, I have written MSI personally to ask for help and/or information in this matter. Here is a copy of what I sent to them:
"Dear sirs,
I recently did a review on the 845 Ultra-ARU. While this motherboard is an excellent performing board and scored some outstanding benchmarks, I had 2 of them fail completely during the course of the review (about 6 weeks). I would like to know why these motherboards failed so quickly. Before I go any further, I will tell you that both were overclocked running at 158FSB. I have been overclocking for about 4 years now and have never seen failures like this. It is not my intent to defame MSI or any other company, so any help or information in this matter would be greatly appreciated, and if possible I would like to obtain a few more boards for testing because I realize it could have been a bad batch, or just really bad luck on my part, and if so I will amend the review to that effect.
Thank you for your time.
Best regards,
Dwayne Sessom PCExtreme.net"
I have other information, but I am not going to publish any more on this until I hear back from MSI. If they do not respond within a week, I will post the rest of my findings.
Update: Sept. 17, 2002
Well, it has been nearly 2 weeks, and still no word back from MSI. So, here is all the info I have found on the problem:
This mosfet that is marked "C5001" with an "R" underneath is located between the 1st DDR slot and the CPU socket, and is the cause of the premature failures that I experienced, and has been confirmed by multiple readers who had similar problems.
Here is a closer look at the mosfet. You can clearly see the "bubble" on top, which was caused from overheating. There is also a hairline crack down the center of the mosfet as well.
You can see the crack in this picture. I removed the mosfet from the motherboard, and showed it to several electricians and an electrical engineer. I wanted to find out the specs, and manufacturer to see why it had failed so easily. As it turns out, the "C5001" marking on the mosfet is apparently remarked by MSI to desguise its true identity. This is not an uncommon practice among manufacturers, as it aids in protecting the way in which the boards are made (patents, etc.).
At any rate, it was impossible to tell for sure what the rating was on the mosfet, or who made it. One thing IS certain though. These mosfets are not able to handle the job they were intended to do. They fail prematurely even on non-overclocked systems, but also consider that not all 845-Ultra boards are made using this mosfet. Some are, and some aren't.
So, now that you know all the gory details of the failure, and I am 100% sure that it wasn't just a bad string of luck on my part, I am reinstating the original "Piece of Crap" award that I gave this motherboard!
000000
Warning: main(php/nav/footer.php) [function.main]: failed to open stream: No such file or directory in /home/pcx4010/public_html/845ultra-aru2.php on line 254
Warning: main(php/nav/footer.php) [function.main]: failed to open stream: No such file or directory in /home/pcx4010/public_html/845ultra-aru2.php on line 254
Warning: main() [function.include]: Failed opening 'php/nav/footer.php' for inclusion (include_path='.:/usr/lib/php:/usr/local/lib/php') in /home/pcx4010/public_html/845ultra-aru2.php on line 254