DFI nF3 Lanparty UT isn't ALL that...

So after testing it for 2 days, there are a few issues that really annoy me:

1) Vcore fluctuation. Watchin CPU-Z jump from 1.728v to 1.792v while IDLING is rather uncomfortable. My Gigabyte K8NS-Pro almost NEVER twitches. I realize that the DFI increases vcore during load, but even @ load, it fluctuates 0.02v, and very frequently..

2) vDimm fluctuation. Jumps from 3.30v to 3.34v while in BIOS... Don't know how much it dips while doing memtest.....

3) High pitched squeeking noise when booting WinXP during resolution changes. Very annoying and rather loud. I assume they're coming from the coils, and I think I've read about a fix somewhere by taping down the coils? :confused: I've tried 2 diff CPUs and 2 diff sets of ram, happens everytime I bootup whether overclocking or not, never happened on any other mobo I've used. PSU is an Antec Tru550. Probably an issue related to this specific board, but it gives you an idea about DFI's quality control :stick:

4) Build quality: the caps used are "decent" I guess, they come from United Chemi-Con which should be made in Japan, but most of the MoSFETs are small and no name, which is probably what's causing the vcore and vdimm fluxing. MoSFETs also get HOT while underload, but I guess it's quite normal for all boards, DFI should of included heatsinks with it though since it's aimed @ the enthusiast market....
All the caps on the K8NS are made by Nichican which are high quality made in Japan, and the MoSFETs are big, made by NEC. I believe the Epox board uses Sanyo caps, and the MSI uses....??

5) Northbridge heatsink has thermal paste underneath it, which is GOOD, but it desperately needs a fan. If you turn the vChipset up and run the board outside of a case without air circulation, the thing gets HOT!

6) There is no difference in overall performance against other boards. I benched both the DFI and my K8NS-Pro in 3dmark05 and CS:S stress test using the same equipment, settings, everything. DFI was 7points faster in 3dmark, and the K8NS was 1FPS faster in CS:S.

Don't get me wrong, the DFI is a good board, but it isn't the HOLY GRAIL that most people think it is, and it probably WON'T get you a higher CPU overclock or +1000 3dmarks. I see all these people who were waiting for the DFI to switch to A64, or thinking that the DFI can hold them off from S939, but I just don't see this being the case. I've also seen quite a few people having issues with the DFI board..

Personally, I despise the vcore fluxing and the relatively "cheap" components used... I actually needed 1.728v to run my 3200+ mobile stable @ 2.52Ghz whereas my K8NS only required 1.696v. Maybe the board I tested is a dud?

On the plus side, the endless BIOS options are nice, and so are the voltage options. But frankly, almost all these tweaks can be done via A64Tweaker + a soldering iron... Of course, people will say how unstoppable the DFI is when paired up with TCCD ram, but I won't get into that right now since A64 and memory is such a hit and miss issue.

Again, these are my own observations and opinions. I know all you DFI lovers will probably start flaming this thread, and I don't care. Just want to inform potential buyers of certain issues. NO motherboard is perfect.

I just ordered one in hopes it will be better than this an50r I have.

Quote:

---

Originally Posted by wolfgang

I just ordered one in hopes it will be better than this an50r I have.

---

I thought the same thing and honestly even with the additional tweaking options I haven't gained a thing. To be honest I am a little dissapointed but my main reason for trying this board was just an attempt to run a gig of ram with my original 3000+ CH (512 l2 cache) CO stepped chip. No luck so far so I will have to wait until I can get some Ballistix/EB/TCCD sticks to try. The extra memory voltage is nice but my BH5 sticks are crap and won't run stable at 2-2-2-x at 230+ so it really didn't help me any. Oh well, live and learn.

Yeah, the price ($130-150) and reports of limited cpu overclocks (like all the DFI boards have had) kept me away from it. I'm trying the epox board. I'm hoping dfi gets 939 right. Though don't mention it on dfi street, apparently they nominate their admins based on who is the biggest :banana::banana::banana::banana::banana:.

I love the stripped down cooling, it allows you to design your own cooling solution to best suit your needs. Remember the Abit OTES? Many people loved it, I despised it, it was terrible for my needs.

HK - if your 3.3 line is not at least .15v higher than you set your ram in the bios you will have fluctuations.

Its a good idea to test with a multimeter too, as the monitor chips could be inacurate. .02v fluctuation is nothing.. i've had several nf2 boards that were .15v or more ;)

LOL, another person who complains about voltage fluctuations and all they go by is software in Windows and the motherboard's own hardware monitor.

If you bothered to get a multimeter and read the voltages off the mosfets, like I did, you'll see that the voltages don't fluctuate anything like what the hardware monitor says.

Quote:

---

Originally Posted by STEvil

HK - if your 3.3 line is not at least .15v higher than you set your ram in the bios you will have fluctuations.

Its a good idea to test with a multimeter too, as the monitor chips could be inacurate. .02v fluctuation is nothing.. i've had several nf2 boards that were .15v or more ;)

---

I had it set to 3.6v and had vdimm set to 3.3v in bios. Yes, I have an NF2 board that goes crazy too, and guess what? It can't overclock for crap :(

Quote:

---

Originally Posted by Walrusbonzo

LOL, another person who complains about voltage fluctuations and all they go by is software in Windows and the motherboard's own hardware monitor.

If you bothered to get a multimeter and read the voltages off the mosfets, like I did, you'll see that the voltages don't fluctuate anything like what the hardware monitor says.

---

As far as I know, most A64 mobos use the same hardware monitor chip, I see the same ITE IT8712F-A Super I/O chip on almost all of them. (DFI, K8NS, AN50R) Now, if this chip's HW monitor functions are faulty or inaccurate, wouldn't it show the same kind of voltage fluctuation on all mobos that use it? :confused:

depends on signal routing.

I need an osciliscope.. multimeters arent fast enough to read fluctuations really unless they are pretty big, but they do read better than some monitoring chips.

Quote:

---

Originally Posted by Walrusbonzo

LOL, another person who complains about voltage fluctuations and all they go by is software in Windows and the motherboard's own hardware monitor.

If you bothered to get a multimeter and read the voltages off the mosfets, like I did, you'll see that the voltages don't fluctuate anything like what the hardware monitor says.

---

yeah that's what alot of other users of this board have confirmed... the fluctuations are software errors as the voltages don't fluctuate

i can live with that, only annoying problem for me is using OCZ DDR Booster >3.6v on DFI

yeh this board rules!
even oskar admits he used a cheap winbond chip to cut down costs. the volts are really rock steady
and it never claimed to be a good cpu ocer, its hailed as the BEST MEMORY ocer which it is. 330HT for the chipset without mosfet cooling is damn good too
also the bios is so much easier to use than other boards, plus its a really stable board (more so than abit) and it looks awesome

This time i agree with HK. I have noticed some problems too.
First jumping Voltage on Vcore wich might be reason of instablity i have.
Whine sound, but its probably cpu. I have problem with nic , it does work only with 10 FD . Layout could be lil better like, ATX power connectors and 4 pin rail in same place like EpoX 8KDA3J have. Post could be better like on Epox 8KDA3j you see what speed your cpu have ie (10x240 for cpu and 200 MHZ in 166ratio ).

I have not been super impressed with my DFI Lanparty UT either. I do like the layout and excellent bios options including a wide selections of voltages and working cpu multiplier. I do not like the weak northbridge cooling and its close proximity to the AGP slot so I cannot fit a decent aftermarket cooler on there. My bios for some reason cycles through twice before going into Windows and ever since moving to this board from my VNF3-250, I get a clicking noise on startup from my hard drives. My Chaintech VNF3-250 had the same exact overclock and I no longer need massive vdimm voltage since picking up 1GB of PQI TCCD's. Good news is that I lucked out and picked it up for $107 when it first showed up at Newegg. Overall I am fairly pleased, but not overly impressed.

I don't know if I'd say it is "all that" myself, but overall I am quite happy with it. I picked it up to replace my 8KDA3J, which was just a boner with the quirky VCore. From boot to boot I experienced the common VCore fluctuations that a lot of people see on that board, regardless of bios rev. Not a huge problem with a moderate o/c, but when trying to find the top, too little voltage would hose stability, as would too much, and I never knew what it would give me. I tried to compensate with a VCore VR, but found the fight with the random volts just too much to deal with effectively. With the much more reliable VCore on the DFI, I have managed higher clocks. Whether I'm limited in clocks compared to other NF3 boards, I don't know, but I got more then I could with Epox. I agree that the mem tweaks on this board are nice, but nothing that can't be done with CodeRed's tweaker, but there is some benefit to being able to shake out timing adjustments in memtest86+ prior to getting into Windows. This can't be done with the Tweaker. It's also kind of nice that this is the first mobo I've run in quite some time that doesn't have any VR's soldered to it, I could actually RMA this thing if I ever run into a problem.

This board has been over-hyped in the biggest way, but still, not bad at all.

My :2cents:

edit to add: sobol has an excellent point regarding the post screen too. I did love that with the Epox, at least I knew quickly which direction I was going to need to turn my VR to compensate for the random VCore.

Yep never really understood all the hub-bub around it, sure it does some nice rates but when every entusiast site hailed it as the king of both S754&S939 just for getting over ~300MHz HTT I was like lol.. I mean even my Soltek K8AN2E-GR will do 345MHz HTT without breaking a sweat!

Quote:

---

Originally Posted by mikead_99

I don't know if I'd say it is "all that" myself, but overall I am quite happy with it. I picked it up to replace my 8KDA3J, which was just a boner with the quirky VCore. From boot to boot I experienced the common VCore fluctuations that a lot of people see on that board, regardless of bios rev. Not a huge problem with a moderate o/c, but when trying to find the top, too little voltage would hose stability, as would too much, and I never knew what it would give me. I tried to compensate with a VCore VR, but found the fight with the random volts just too much to deal with effectively. With the much more reliable VCore on the DFI, I have managed higher clocks. Whether I'm limited in clocks compared to other NF3 boards, I don't know, but I got more then I could with Epox. I agree that the mem tweaks on this board are nice, but nothing that can't be done with CodeRed's tweaker, but there is some benefit to being able to shake out timing adjustments in memtest86+ prior to getting into Windows. This can't be done with the Tweaker. It's also kind of nice that this is the first mobo I've run in quite some time that doesn't have any VR's soldered to it, I could actually RMA this thing if I ever run into a problem.

This board has been over-hyped in the biggest way, but still, not bad at all.

My :2cents:

edit to add: sobol has an excellent point regarding the post screen too. I did love that with the Epox, at least I knew quickly which direction I was going to need to turn my VR to compensate for the random VCore.

---

There is a memtest prog for windows, and from my experiences, it's actually MORE accurate than memtest86. I used to pass memtest86 without a hitch but this prog found errors.

http://hcidesign.com/memtest/

Quote:

---

Originally Posted by HKPolice

There is a memtest prog for windows, and from my experiences, it's actually MORE accurate than memtest86. I used to pass memtest86 without a hitch but this prog found errors.

http://hcidesign.com/memtest/

---

Nice! I'll give that a shot. I've been wanting to find something like that. I agree that memtest86+ isn't the be all/end all mem stability test. I definately find the need to continue testing in Windows after passing, but if I can't pass memtest, I know I shouldn't even bother loading up an OS yet, unless I want to risk hd corruption and try one of those marginally stable bench runs.

Quote:

---

Originally Posted by HKPolice

5) Northbridge heatsink has thermal paste underneath it, which is GOOD, but it desperately needs a fan. If you turn the vChipset up and run the board outside of a case without air circulation, the thing gets HOT!

---

Just wondering .... Doesn't this board only have a southbridge?
:stick:

you guys are complaining about a nv bridge heatsink and thermal paste??

have you ever thought maybe dfi went a lil cheap on them to keep the overall costs Down?

This thread makes me sad..

Quote:

---

Originally Posted by HaLDoL

Just wondering .... Doesn't this board only have a southbridge?
:stick:

---

Actually its neither, its an MCP :stick:

Well I just set mine up yesterday. I admit to being a lil nervous after reading some of the poasts about this board, add in the fact I went with a 3400+dtr(even more horror stories with this coupling). I was down right skeerd.

I did the build, used the Diamond Flash Image floppy flashed to the 10/5 beta bios from the site. went in to bios and set it up stock 200x11 1.45vcore 2.7vdimm installed windows haven't had a crash yet only up for 24 hours.

I'm leaving any power flux up to my sucky Aspire 500w PSU. knew it had weak rails going in.

I not counting this board out just yet it only been two beta bios released, lets see what the next offical bios comeout. I allready seen some nice modified sets floating about too.

It's a nice mobo I find,it seems to require a good ~500Watts PSU to get the most out of it tho.

Made it to 310HTTx2.5 at 1T with two dimms of Corsair Twinx PC4400 9/10 ratio(memory top speed 279MHz)3-4-4-8.

My EB3500 are the last kit that OCZ had (in Canada ?),they are far from doing what OC I've seen from other EB3500,end of line memory I guess :( .

Two thing that pissed me off:

-There's no CMOS Reloaded like the other Lanparty mobos,there's soo many BIOS options it would have been nice to have it ,if not just to save time finding the best OC/stability settings,cost me 1 extra day tweaking not having it. :mad:

-They cheaped out and did not include a manual,only the quick setup guide,made it a PITA to setup the front audio in my case,had to DL the manual off their site.

The voltages are ROCK stable. samurai jack and jess1313 tested it with an oscilloscope and its definatley the crappy ITE sensor. so dont say it has fluctations