NF4 - Some Help Getting Up & Running (Or Faster?)

[sephir0th]

sephir0th, how is that TCCD working for u on that board? exRoadie , ty for clearing that w/ me i was unaware .

 

Im still testing to find whats stable, will get back to you on that one.

 

278 x 9 with 2.5-4-3-8 and most of the other settings in this guide is 3d and super pi 32m stable, havnt tested prime much (though prime blend doesnt like me very much)

[graham_h]

Amazing to think that we need to go to this much bother to get the board working with two sticks of ram...........

We're using OCZ platinum ram on our board and it refuses to work with two sticks.

That, to me , seems like a huge oversight on DFI's behalf.

Seriously are there any high end users that don't use a gig of ram ?

And yes it'd be nice to use dual channel.

Yet to see a bios upgrade fix the issue either.

We're RAing this board and probably giving up the fight.

Shame really, we had high hopes for the DFI board.

Surely DFI can't be happy with so many users havingthese issues ?

[RGone]

Graham_h, I think you should rma the board you have. I think it has a defect. Now as to the moving on to something else> well that will have to be your call. You obviously have never closely attended some of the other motherboard sites and seen that all mobo's have problems of one persuasion or another. Usually cyclic stuff common to a brand it seems due to a design problem. I know one motherboard that will remain nameless that users that site wear a " I have RMA'd three times" as a badge of honor. Now that one caused LOL for me.

 

'pinnochio' in another thread in these forums said that he was sure there were WAY TOO MANY problems and rattled off here, AnandTech, Exteme and one other I cannot remember.

 

So I went and did a check of the tips, and fixes thread at Extreme, started by TranAm. 344 posts as of 2:15AM CST in the UsA. I went from page one/post one to the fifth (5th) page and post 107 and then to page 11 post 260 to the end at 2:15 AM CST at post 344. That leaves 5 pages in the middle I did not check in that thread at Extreme.

 

So here is the result of 9 pages and 191 posts.

 

1. Fiive (5) real non-working TCCD memory problems and you are there in that group. Won't boot or cannot use 2 sticks.

2. Now of the total 11 problems I labeled as in the TCCD bracket here some examples of the other six (6)

A. Post says TCCD will not boot. User unplugs MOMO Steering wheel and computer boots.

B. One users TCCD symptoms change when he swap out the CPU.

C. One claims it takes him too much voltage to run his TCCD @289FSB.

D. One user hooked up his floppy power connector backwards and wonders why his LED's and other stuff does not work now.

E. One user is complaining he is only able to do 260FSB with his TCCD.

 

You see a pattern there? Well I most certainly can and is why I do not really care for statements saying folks are swearing off DFI when they should RMA the board if it is bad and statements like there are WAY TOO MANY Problems. Well any problem in my computer is one too many and I understand that for sure but to go further than that and broadly assume that all is in the shetter on a board is not right nor factual likely.

 

Ok of the five (5) real problems one is cannot boot with anything but Cas 2.0 with Patriot memory. Seen that for myself. Thought it was the NF4. Put the sticks in my NF3 and low and behold the sticks were dorked in it TOO. Another of the 'real' problems I have given credit to the board as a TCCD problem is one user with Mushkin V2 in slot #2 BUT he is trying to run ADATA DDR500 in slot #4 with the Mushkin. Do I need to go on about the TCCD problems?

 

Ok 4.0V problems which truthfully I am not surprised about when people try to put 3.7Volts on memory but I guess that is there 'mindset'. Two (2) of the problems assigned to the 4.0V category and I list 5 of them are because of the ZIPPY 700 Watt power supply and they know its' rail is bad. Do I need to go on about the 4.0V problem?

 

OK one problem I put in the BH-5 category of problems and the poster said his problem was cured with flash to the 02/01 bios. So much for that one.

 

Post #325 was a user who hooked the P4 connector up to the wrong socket on the board and blew up the Ocz Power supply and probably took the board out as well and maybe his memory too.

 

OH yes I did keep a list of the problems by the post number in that thread but I do not intend to embarass anyone.

 

Two posters make reflection that Hynix based memories seem to work very well.

 

Now if I went and took the time and effort to do a thorough check of the other forums mentioned by 'pinochio' I likely would find similar results. Some problems with TCCD memory that this thread and what I posted has helped quite a few get going and move out of the problem category. I did not expect to cure all the problems but some is better than none. But 100's of bad boards? I doubt it. A pain for the one in the problem zone? Yes and amen.

 

So before you jump ship and tick DFI off of your list, think about what you see above and act accordingly. That is all I can ask. Thanks.

 

EDIT: and I see now from your post elsewhere that you are running OCZ platinum 3200- rev 1.0 and that is not TCCD anyway. I don't think so. I had put your memory in the TCCD fails above and you really are not in that group. Not sure the memory modules on that stuff. You know what memory module is on the stick?

 

RGone...

[pinnocchio]

Rgone

 

I think you need to read my post again.

 

I didn't mention anything to do with TCCD, I specifically mentioned OCZ Platinum but there do seem to be a high instance of problems.

 

I said specifically I believed the problem would be sorted out.

 

I specifically drew attention to the fact I'm not familiar with the platform.

 

I worked the problem through very thoroughly yesterday and spent a lot of time testing on two mobo's, two processors, two sets of memory etc...

 

I agree that problems like this are of a cyclical nature and not every processor/memory set combination get's tested. But having taken the time and trouble to check what was supposed to work together before buying it and then spending nearly $3000 in total last week buying bits I think I have a right to be a bit miffed.

 

Many users you get here would just shout and scream and not try to help by doing extensive testing and then giving feedback so the problem could be solved. If you read it any other way then I'm not sure why.

 

Oh and if you read some of my other posts I've been trying to draw users attention to the stickies and existing solutions to try to ensure you don't get too many repeat questions.

 

P

[RGone]

'pinochio' man, I have read your posts and now the question still arises that I asked Graham_h above and that is what is what have you got for memory modules on the stick. See I was under the impression you were talking about Ocz platinum Rev 2 the current stuff as I know it and it has been TCCD for some time. Are you talking some lower performance level of Ocz currently or are you talking about Rev 1 which I think is in Ocz EOL grouping.

 

So in the long run the exact nature of the memory will need to be known. I know Ocz is currently working diligently to unearth any problems memory related as they usually always are. And if the memory you have is less than 2 months old and is PC3200 Rev2 it is supposed to be TCCD module populated unless there has been an unknown change. That is why I am presuming your memory is of the same TCCD category as a couple of users here in reporting problems with Ocz PC3200 Rev2.

 

RGone...

[pinnocchio]

Mate

 

I can't tell you without removing the heatspreaders which invalidates the warranty (and if I end up having to return them I'm scre*ed), the memory I have was bought from OCUK Tuesday last week specifically for the mobo as it was on here as being suitable.

 

http://www.overclockers.co.uk/acatalog/Enh...dth_Series.html

 

I have one each of the top two sets as they didn't have two of the Rev 2 in stock.

 

I have got a message into Tony from OCZ saying I'll drive to him tomorrow with one of my Mobo's, Proc's, GPU's and both memory sets for him to test (as long as he's not more then 200 miles from me), if he's happy to remove the heatspreaders I'm OK.

 

I know you're probably fairly frustrated at the moment (I used to do your type of job), you get a lot of . and little credit. But I (and the majority of people on here) want to work to finding a solution to the problems because we're excited about the product and it's potential (I want to use one of mine to try to drive 5 TFT screens all using PCIe). The thing to keep in mind is for each person on here trying to find a solution is you've probably got 10 more out there that will probably just RMA a board and give up.

 

So please allow us a little frustration, keep in mind that we've paid for the kit and are even keener then you to make it work so we can start to get the advantages of it!!

 

P

[RGone]

I have one each of the top two sets as they didn't have two of the Rev 2 in stock.

'pinocchio', not sure I follow that statement correctly. When I follow the link, I see the top memory set is Rev 2 (TCCD) and the second set of memory is not listed as REV2 and is a 1 gig kit and the third is not a 2 stick kit but a single stick of the kit listed as number 2 down on that page. You have confused me. But I presume you do not have Ocz's better Rev 2 memory.

 

And I doubt if you want a fix more than I do. I run the boards myself and as soon as I know they are useable, (mine anyway) they are used daily. And when you guys get "fixed" or handled> well my days go way better when I don't have to troubleshoot or try to some user's cannot get going configuration but can tweak my own board in a configuration that " I " prefer.

 

Good luck in your possible journey, if it is not north of Brum. Hehehehe. I do not have a clue where that is but Tony is a sharp dude and sure he is in touch with Taiwan.

 

RGone...

[pinnocchio]

That's it...I have one of each.

 

P

[RGone]

One of each of which? Number 2 on the list is a 1gb kit with two sticks in it and not TCCD I don't imagine. And the Number3 on the list is a single stick of 256Mb memory and also not likely TCCD but who knows? You buy two sticks of Number 3 on the list? Number one on that page link is what I would have had had I bought and it is the Rev2 Pc3200 with TCCD but you do NOT have it right?

 

RGone...

 

If I count it right you would have 2x512 of Non-TCCD memory bought in a 2 stick kit.

 

And you have 2x256 of the same Non-TCCD stuff bought as 2 sticks separately.

 

I got the count correct now?

 

RGone... Again...

[pinnocchio]

I have one of each of the top two i.e. the 1gb sets.

 

1 set (2 x 512mb) of MY-048-OC

1 set (2 x 512mb) of MY-005-OC

 

P

[RGone]

Ok I got the memories straight now. The 048 is supposed to be TCCD that will start out around 2, 2, 2, 5 at 200 but quickly has to loosen at elevated FSB.

 

I have totally lost site of your original problem as I have made 4 different setups across 2 motherboards in the last 5 hours and with 4 different kinds of memory and a quick test with each one. Lot of hassle.

 

Was your original problem that you could not run 2 sticks of 048 in Orange slots 2 and 4? If so that should have been doable if you ran the exact timings I have at the top of this thread. MOST assuredly turning off Bank Interleave and Max Asynch Latency to 7x and the Read Preamble to 5x. I had to do that for every stick of TCCD memory that I had and is what 048 is supposed to be. Bear in mind I have no clue what so ever as to what the 005 memory is my memory chipset on those 005 sticks. No clue at all man.

 

If you followed exactly the settings I posted and are sure of the bank off and the 7x and 5x > heck I am just about inclined to believe you have a defective motherboard because I have tried about 6 different sticks of TCCD thru my SLI and my Ultra-D and both had to have bank off to work and to get to 270 or so FSB it required 3.0 volts and on two sticks it took 3.1Volts on the memory. Again bear in mind that I set timings and volts and all to start with and headed up the ladder. I do not fiddle around in the lower regions and so I turn up volts and loosen timings and begin ascent.

 

I think it sounds as if you have a bad board in the nutshell. That is based on the precise premise that you exactly followed the settings I gave in this thread. Good luck with Tony if you meet up with him. He is as I said earlier one sharp dude for sure.

 

RGone....

[Aldersix]

Just want to add my experiences over the last week. Initially had problems booting and loading windows but beta 201 bios helped that. With Optimized Defaults. Then had issue with SW IDE drivers and my plextors. Solved that by un-installing those and using the default MS drivers. Did some read write test with the plextor and all fine there.

 

 

At optimized defaults had trouble passing memtest at tests 2,4&5 with both sticks and single stick not all errors on each run. Ran Prime95 and had good and bad results after 6 and 12 hours. Default timings on ram were 2-3-2-8 @ 2.6v.

 

But could get boot fine and get into windows multi time throughout the week and load programs etc.

 

 

I changed settings in bios to set timing at 2-2-2-6 @2.8v per the spec of the ram manual set DRAM freq to 200 1:1,. Again no boot issues could get into windows fine etc. But again could not pass memtest and mix results with Prime95. This time I documented a little better.

 

 

Got Prime 95 error 8 hours into test.

So ran memtest on both sticks got errors like before, sorry for blurry image. 35 error on first pass with in multi tests

 

 

So took one stick out and ran memtest again. Got errors on stick 1, 2 errors Test 5 first pass. Stopped tested and I tried the following:

 

 

1) Changes went to 2.5-2-2-6 @ 2.9v Error at test 3 stopped test. Then

 

2) Left timing the same and lowered to 2.8v, Error test 4, stopped test then

 

3) Set timing to 2.5-2-2-7 2.8v Error test 4 stopped test.

 

 

Took out stick 1 put in stick 2. with the mentioned 2-2-2-6, 2.8v

 

 

Memtest ran with no errors on Stick 2 for 12 cycles, 2 hour 41 minutes and got three errors on Test 5.

Also could not set my multiplier manually when I did I could not get into windows would get that endless loop and re-boot over and over, so left in auto. Did try Rgone setting in his post titled NF-4 SLI-DR getup on Micron Ballistix and could not boot with those.

 

 

As a side note: this is my first build and I bought this board as I wanted a better gaming rig. My old computer is an HP w/ xp2400+ and a PNYFX5200. Not sure if I will OC until I learn and read more. But after reading some of the above comments just wanted to say got board because of the rep. of the other lanparty 754 board and that it was designed well in my opinion and I will stick through this as if I ever do decide to OC the rep. of DFI and Wu are what swayed me to this board. Hope above may help some as I have gotten a lot of info from the street and the assistance of Rgone, Thanks man.

 

 

 

Test Used:

 

Memtest86+ v1.51 from: http://www.memtest.org/

 

Prime95 v23.8.1