New vs Old NF7-S v2's

[Sentential]

lmao.... probably is.. the light is poor in here

 

[james.miller]

it sounds like i do have a hybrid v2.15 then....unless you change the revisions again lol.

ive just checked and i actually have two different types of vcore mosfets. both have writing on them, but neither have the markings they are acustomed to v2.1/2.2 boards......

 

[Sentential]

Im just about to give up and enjoy what I have. I cannot keep track of an infinate amount of revisions with an infinate amount of OCing potential....this is just ridiclous. Congrats on the good mobo.

Any chance you could read off its NB stepping?

<raises the shotgun>

I give up....

My old mobo's NB and SB dont even have the same stepping

NB : G41638.1 0344A1
SB : FB8844.1 0347A4

 

[james.miller]

i honestly think it's just a case of use whatever components they have at the time. IT ssems to be pot luck what board you end up with, though im convinced there is something that ties all the boards together with regards to overclocking....

as for the northbridge stepping: not at this moment i cant, no.

SB: FB2226.1 0306A3

 

[Sentential]

<shrug> What can I say... Abit isnt very organized

 

[wildfrogman]

Has anybody "modded" the nf7-s so lets say they removed the mosfet~surface Ic things to make it like the later boards layout? Maybe the chips below the southbridge push more voltage into the southbridge? My southbridge before installing a heatsink had a 120mm 90cfm panaflo blowing directly on it and still the SB was hot enough that I could hardly keep my finger on it. Chances are the board's trace layout is the same just with the removed surface IC's so in theory atleast removing the chips could produce large gains. Unfortunatly I am not quite the pioneer type yet, rather have somebody else with more experience test this "Ultimate Nf7-S v2.0 mod"

 

[Rabbi]

I think allocating revision numbers isn't going to work....

A better way to see what's limiting some boards and not others would be to have a "checklist" for every board...

e.g.
does it have SATA MOSFETS
does it have CPU Socket MOSFET
does it have Metal Socket Lever
does is have the SB MOSFET
L12 done yes/no
RAM used
CPU used
Max FSB achieved


If we could start putting together results with all these variables for each board we may start seeing a common denominator...

 

[Suma]

Another thing I noticed.
On the NB after the stepping code it says: "A Taiwan"
On the SB After the stepping code it says: "C Korea"

What does yours say there?
Suma.

 

[Suma]

Well, I had a theory, which just broke into 1000 pieces.
I thought maybe the 2.x,1.0 is a NF7-S thing only, so I asked a friend of mine who owns a S so send me a pic.

Here it is:

He bought his in MAY 03, and he also said that his lever is made of plastic, not metal.

Suma.

 

[Sentential]

Finally starting to see a trend. All the "high-clocking" NF7s / NF7 have this string of codeing where it sez model ## one CPUID.

"2.X,1.0"

Low clocking NF7 have just plain ole "2.X"

Thats what I am begining to see....someone prove me right / wrong

 

[Suma]

Man mine is metal and it say only 2.X
Sorry..

Suma.

 

[Sentential]

Its starting to look like:

CPU-Z
________

2.X = 230s max FSB

2.X,1.0 = 240s stock.... max ??
________

Plz test ur mobos pplz, if this is correct we are starting to make major headway.

Lets start seeing some tabulation of max FSB based off of model numbers.

 

[Panzerknacker]

Man this is really getting too complicated right now. I'm thinking of selling my NF7-S. I'm running a 2800 @ 3200 right now, but when I up the FSB 5 more mhz, it's already Prime unstable, even with higher voltages and with my SLK-947-U @33c. I'm using kingston HyperX pc3200, dunno if it is BH5, timings are default. Bios 18.

This means my mobo sucks, right?

 

[Sentential]

What FSB? Odds are you have it at an unstable setting

 

[Clevor]

I have three boards in front of me, bought at different times, and all have plastic levers. I don't think metal or plastic has anything to do with revisions, it's just whatever batch of sockets was available at the factory at the time.

Also keep in mind that your ram has a LOT to do with an overclock. Everybody's saying my board can do this, mine that, but we are all using different ram. Unless someone has tested the exact same components with maybe 4 boards from each revision, you can't say for sure one revision is better than another. Then there's board-to-board variation.

My first board has all the mosfets and was bought a year ago. It has the plastic lever and does 228, 2-2-2-5 out of box, box stock. But this is 1GB XMS3500, and with the max 2.96 VDIMM of the board, you will not get much higher. Same ram has gone DDR500, 2-2-2-5 on P4 boards with 3.2 VDIMM. On the 8RDA+, where max VDIMM is 2.86, sure enough, I max out at 220-223 or so.

I just got what I thought was an old board (since I thought they were better), but nope. It does not match any description 'cept maybe Rev. 2.4. Got it in Japan. It has plastic lever, missing mosfets EXCEPT the ones at the bottom of the board, teal green IDE slots and black floppy, with blue/purple DIMMS with small latches. Haven't checked it out yet, but I will use exact same case, components, CPU, and ram as my first board.

You'd think the later the board, the more stable, but hard to believe missing mosfets increase stability. I thought my first board was Taiwan, but nope, both boards are made in China.

 

[Cyrix_2k]

Well, I have a 2.2 that I ordered in August of '03 from GoogleGear (now ZipZoomFly), haven't pushed it very hard but I do know it will do 225mhz FSB no problem. I guess that's good.

I'm willing to bet we could get some GOOD NF7-s's if we ordered refurbs

 

[Sentential]

I dont know.... Ive seen alot of differences between the 3 NF7 I owned. Each OCed differentally with the same parts, so I dont know.

One thing is for damn sure..... I love mine

 

[Clevor]

No need to be apologetic, as I'm sure we all appreciate your efforts here on this post!

I sure didn't know there were these minor differences myself. Heck, if this was Epox, there would definitely have been 4-5 distinct revisions of this board.

Just puzzling why Abit has not bothered to formally issue them. While not as prolific as Epox, on something like the IS7 P4 board, Abit has issued Rev. 1.0, 1.1, and 1.2 boards. All it takes is changing the label, as Abit does not silkscreen PCBS like Epox and Asus.

 

[Rabbi]

Finally starting to see a trend. All the "high-clocking" NF7s / NF7 have this string of codeing where it sez model ## one CPUID.

"2.X,1.0"

Low clocking NF7 have just plain ole "2.X"

Thats what I am begining to see....someone prove me right / wrong

Sorry dude, i dont think this fits.
My "OLD" one with black everything, metal lever, all FETs etc did 229FSB 2-2-2-11 stock with 1GB BH5 and could post and benchmark at 250FSB 3-4-4-8 with 256MB Hynix BT-D43 like yours.

But i killed her

Oh yeh, my point was that CPU-Z showed it as jus "2.X"

 

[Sentential]

Sorry to hear ... well atleast your odds are greater with the "2.X,1.0" mobos I guess...