250 FSB DC SYNCH and ABIT NF7-S

[saaya]

my pc crashed at 245dc 2-2-3-9 3.3v vdimm and 1.74vdd with delta on a7n8x nb hsf after 4 days. i was playing a few hours et every night, on sunday i played et for 7 hours and i got a bsod... could have been the 8500 as well... the waterblock is just laying on the pci slot, its not fastened to the gpu at all

[Zeus]

Ok, just some more proof that high Vdd isn't necessary, don't mind the score's though, VC was at stock speeds.
This was all run with just 1.7Vdd, 10,5x230
Haven't tried higher yet, just wanted to see if it was 3dmark stable.

edit: you got a point there, Jason, still i think 1.7Vdd can take this board very high, anything higher than 1.9V requires some real good cooling to the NB i reckon. (pelt)

[SAE]

No. Not necessarily a pelt. Just ensure you lapped the nb fine and pray the mosfets won't blow Water or a decent cpu cooler is enough.

But never heard of an abit nf7 blowing mosfets... Sometimes had 2,7V through my 1.1

[xigfrid]

Originally posted by Zeus


You killed your's?

Last one was the best one i ever had, doing 230Mhz w/ bios 1.6 stock and reporting the voltage exactly the same as the PSu's rails are measured, overvoltage on Vcore and 3.07V Vdimm stock

but when i decide mod it to extrem lapping, watercool, etc this one was running sweeeeet but by error I decided to measured some Vcore at Mosfets ... error, i short circuited two legs, then the mobo wouldn't boot

The one i'm using is completly different although the northbridge is the exact same week :/ it couldn't even do 200 Mhz stock, but atm i 'm benchmarking some 240Mhz with it

[SAE]

Originally posted by xigfrid
Last one was the best one i ever had, doing 230Mhz w/ bios 1.6 stock and reporting the voltage exactly the same as the PSu's rails are measured, overvoltage on Vcore and 3.07V Vdimm stock

but when i decide mod it to extrem lapping, watercool, etc this one was running sweeeeet but by error I decided to measured some Vcore at Mosfets ... error, i short circuited two legs, then the mobo wouldn't boot

The one i'm using is completly different although the northbridge is the exact same week :/ it couldn't even do 200 Mhz stock, but atm i 'm benchmarking some 240Mhz with it

What nb week was that???

Had a similar board I killed with a vdd mod in the dark
Did it and got my 7th rev2 then...
But none that good again... same nb too...

Somebody knows if these with the week23 nb perform better? Board circuitry is always luck of the draw

Thx again.

[xigfrid]

my two last were week 16, hopefully , i've my rma return in 2 weeks, prolly i ll get a new week

[Zeus]

Originally posted by SAE
No. Not necessarily a pelt. Just ensure you lapped the nb fine and pray the mosfets won't blow Water or a decent cpu cooler is enough.

But never heard of an abit nf7 blowing mosfets... Sometimes had 2,7V through my 1.1

2.7V??
You must be the most careless overclocker of Germany..lol!!!

Never heard of blown mosfets on Abit mobo's myself, this seems to be an Epox issue...

Lapped the NB nice'n flat but i think this Crystal Orb isn't up to the job, even though it's one hell of an improvement over the stock cooler.
The stock Abit cooler pushes 3.5cfm against 12.4cfm for the Crystal orb.
If i go over 1.85Vdd i get bluescreens all the time when i hit high FSB which gives me the impression the NB needs some serious cooling at these voltages.
Since you've been trough quite some NF7's and achieved some nice OC's ( ) i assume you know what you're talking about.

Must try some better NB cooling then, for now low Vdd seems to do fine.

[Zeus]

Originally posted by xigfrid
Last one was the best one i ever had, doing 230Mhz w/ bios 1.6 stock and reporting the voltage exactly the same as the PSu's rails are measured, overvoltage on Vcore and 3.07V Vdimm stock

but when i decide mod it to extrem lapping, watercool, etc this one was running sweeeeet but by error I decided to measured some Vcore at Mosfets ... error, i short circuited two legs, then the mobo wouldn't boot

The one i'm using is completly different although the northbridge is the exact same week :/ it couldn't even do 200 Mhz stock, but atm i 'm benchmarking some 240Mhz with it

Damn, how do you quote 2 people in one reply? i'm a n00b...
What a shame of your last board mate, sorry to hear that.
I always have a piece of small tube over the pin of my red probe, just leaving the very tip bare, works great every time.
I killed a NF7 a few weeks ago by having one stick of RAM pushed halfway out of the slot, i think it happend when i inserted my videocard, killed one of the 2 memcontrollers.
The board just got every mod one could come up with, so i know how it feels...

Keep me informed what you're up to with your new one.

[SAE]

Yes. I think I know

Careless

Hmm. I think I got it. But you are wrong. I could not get it right this time cause I read out a standard vdd already around 2,17V when set to 1,7 in bios So I assume it was a readout error. Later after some board swapping I could get it right. Maybe the pot was... crap

NB cooling is important. But not as important as you should go for pelt or vapo on it

I found a nice one in the komplett online shop. It states mcp-t2. Do you also think it's rev a4 sb???

Maybe I'll order one - or get a new one from a friend at k&m??? With wk23 nb! I am clueless.

[Zeus]

Sure, go and get nr8, lol!!

Don't know what the heck mcp-t2 is supposed to mean and also i'm not so sure wheter later NB steppings will necessaraly be better.
Pc ice recently got himself 3 new NF7's but none of them could really satisfy him.
Btw, i thought you already hit 250 with your current mobo?

[SAE]

Originally posted by Zeus
Damn, how do you quote 2 people in one reply? i'm a n00b...

I personally prefer opening multiple windows per right klick on the quote button of those ones I wanna quote

Then copy the other quotes to one reply window and add ya answers

[Zeus]

Thanks mate!

[SAE]

Originally posted by Zeus
Sure, go and get nr8, lol!!

Don't know what the heck mcp-t2 is supposed to mean and also i'm not so sure wheter later NB steppings will necessaraly be better.
Pc ice recently got himself 3 new NF7's but none of them could really satisfy him.

Who really knows

Originally posted by Zeus
Btw, i thought you already hit 250 with your current mobo?

Yeah. I did. But I found out lately my ram can take 250 single channel at 3,23V and that with memtest for windows. But I never get that dc stable with any vdd up to 2,26V by now

So it must be stability probs of the nb

[Shroomalistic]

Highest i can get dual channel is 496, the highest i can single channel is 512. I cant figure out for the life of me how to 500 dual.

Pc Ice is going to let me play with his 2 new nf7-s's before he takes them back. Well see what i can do with them. He has them fully modded 3.3, vmem, and vdd and still cant get over 230 dual. Ill post some screenies of my 256 if you guys want. I have two sticks that will do 500(single channel) but one of them will do 512.

[Zeus]

Originally posted by Shroomalistic
Highest i can get dual channel is 496, the highest i can single channel is 512. I cant figure out for the life of me how to 500 dual.

Pc Ice is going to let me play with his 2 new nf7-s's before he takes them back. Well see what i can do with them. He has them fully modded 3.3, vmem, and vdd and still cant get over 230 dual. Ill post some screenies of my 256 if you guys want. I have two sticks that will do 500(single channel) but one of them will do 512.

Screenie's are nice!
What timings and multi? CPU interface enabled?
256FSB is cool, shame it won't do DC.

[SAE]

Sigh...

I am at memtest for win for 2hours now and no error at 252MHz 2,0-11-2-2 3,23V on one of my twin/wb in slot3 If there are no errors here I can easily do prime in my exp.

Single I can do that easily but never dual Watch the timings
Have two other twinmos laying around waiting to be tested the same way. For that I'll use slot2. When putting them in together... I wanna cry when... erm, nevertheless I wonder if the newer boards like sold on komplett.de (seems to have new sb a4 - mcp-t2 stated) would do fine? Such fine that I can up my dual channel mhz a bit

[SAE]

Originally posted by MickeyMouse
I can do memtest86 at 2.9v in dc at 236mhz dc 10-2-2-2 but it wont post to windows.. and this i left on overnight with not one error...

I don't mean memtest86, I am speaking about memtest for windows.

Try it out.

[SAE]

lol

Never mind

[kup]

I take it that Mem Test is a good burn in Proggy too?

I usually use Sandra and loop it 1000 times, but is this better?

[STEvil]

I dunno.

Over the last couple nights i've looped my ram in sisoft for over 5000 times at 3.23v... burning in for when my nF2 board gets to visit my case 8-)

[SAE]

Originally posted by kup
I take it that Mem Test is a good burn in Proggy too?

I usually use Sandra and loop it 1000 times, but is this better?

Yes. I assume it's good. It seems way more sensitive to mem errors though ...

If you don't use a burn in speed 100% stable proggie's goin' to drop you out of loop - Error. Should replace Ram -

But I also have read somewhere here in the forums about it being more intense and way more effective in burn ins

[Zeus]

Originally posted by SAE
Sigh...

I am at memtest for win for 2hours now and no error at 252MHz 2,0-11-2-2 3,23V on one of my twin/wb in slot3 If there are no errors here I can easily do prime in my exp.

Single I can do that easily but never dual Watch the timings
Have two other twinmos laying around waiting to be tested the same way. For that I'll use slot2. When putting them in together... I wanna cry when... erm, nevertheless I wonder if the newer boards like sold on komplett.de (seems to have new sb a4 - mcp-t2 stated) would do fine? Such fine that I can up my dual channel mhz a bit

To me this is a proof that it's the NB that can't handle DC at such speeds.
A single stick only transfers half the amount of data to and fro the NB that 2 sticks in DC would do.
The NB just can't handle the stress from 2 memcontrollers givin and askin data.
Maybe some very good cooling on the NB could make a difference but honestly i have my doubts on that.

It looks like 240-245 DC is the absolute max for these boards.

If i'm talkin out of my ass right now, it must be the few drinks i had tonite...sorry

[SAE]

Originally posted by Zeus
To me this is a proof that it's the NB that can't handle DC at such speeds.
A single stick only transfers half the amount of data to and fro the NB that 2 sticks in DC would do.
The NB just can't handle the stress from 2 memcontrollers givin and askin data.
Maybe some very good cooling on the NB could make a difference but honestly i have my doubts on that.

It looks like 240-245 DC is the absolute max for these boards.

If i'm talkin out of my ass right now, it must be the few drinks i had tonite...sorry

I couldn't say it better

No matter about the drinks... I did not even notice it

To your health

[Zeus]

Originally posted by MickeyMouse

also anyone find that 1+3 slots give higher bandwith that 2+3??

I've tried that once, didn't notice any difference.
Same with the 3/3, 4/4,5/5 and 6/6 settings, never found one to be more stable than the other or yieling better performance.

[STEvil]

whats the difference between 3/3 4/4 5/5 6/6????

They all look like 1:1 to me.. lol.