Wierd problem with 8RDA @200FSB

Hey guys I've just bought a Epox 8RDA which has the new A1/C1 stepping. It ran 200FSB straight out of the box without any mod, together with some 512MB Samsung PC3200 CL3. Anyway, I'm having some wierd problem at this magical 200FSB, read on.

System Spec:
JIUHB 1700+ DLT3C 0308RPMW
Alpha PAL8045 + Elanvital 43CFM fan + Coolingflow
Epox 8RDA (A1/C1 stepping) BIOS 3305
Samsung Original PC3200 CL3 512MB x 1
Chieftec DX-01WD + 3 case fans
Enermax 330W PSU

The 1700+ is now doing 2.3GHz (12.5*184) @ 1.7V, tested with 10hrs of Prime95.
Now the problem is, the system runs rock solid with CPU at 2GHz (10*200) @ 1.5V, even tested with loops of 3DMark + Prime95 overnight. However, it fails at 2.1GHz (10.5*200) even @ 1.85V!

In summary:
2GHz 10*200 1.5V <= OK
2.1GHz 10.5*200 1.85V <= Failed
2.1GHz 12.0*175 1.6V <= OK
2.1GHz 12.5*168 1.6V <= OK
2.2GHz 12.5*177 1.65V <= OK
2.3GHz 12.5*184 1.7V <= OK

What's up? 200FSB failing when CPU is faster than 2.1GHz! A quick look at MBM5 reveals that the voltage readings are actually flunctuating by a large amount, except for VDD and VDIMM. For example, Vcore is 1.65V~1.75V (1.7V set in BIOS), +5.00 is 4.76V~4.92V, +12.00V is 11.98V~12.40V.
Could it be the cause of the unstable voltage supply of PSU? By the way, I just had the PSU serviced/repaired from the warranty, so maybe that's the problem? Later, I will try my system on a SPI 300W and see what happnes.

Anyone have a clue? Anyone seen this before?

Thanks in advance.

Could be the chip not liking 1.85v or your enermax, i hear they have a problem handling overclocking even tho they are rock stable at stock. Maybe since it's a 330watt it also doesn't have as much juice.

livingelf said:

Could be the chip not liking 1.85v or your enermax, i hear they have a problem handling overclocking even tho they are rock stable at stock. Maybe since it's a 330watt it also doesn't have as much juice.

Click to expand...

But why it can do 2.3GHz @ 1.7V while failing 2.1GHz even @ 1.85V (or lower)? I'm sure this isn't the chips problem.

Regarding the Enermax, are the readings just normal? It wasn't like this before I have it replaced and before I switched to Epox from the Abit KT7A.

Will try with a SPI 300W later when I have time.

It will be because of the over current resistor R678 causing it to cut-out. If you remove it then you can get past 1.85V
http://www.xtremesystems.org/forums/showthread.php?s=&threadid=8344

OC Detective said:

It will be because of the over current resistor R678 causing it to cut-out. If you remove it then you can get past 1.85V

Click to expand...

I don't mean it won't get over 1.85V, my problem is that 200FSB can only be attained when CPU clock is below 2.1GHz. It did 2.3GHz (184FSB) just fine at 1.7V, but fails if I have it at 200FSB (11.5*200). Memory in sync and timings by SPD.

But you just said you couldnt run at 1.85V whereas all your successes were below 1.85V - plus you said you did 200 x 10 with no problem.

OC Detective said:

But you just said you couldnt run at 1.85V whereas all your successes were below 1.85V

Click to expand...

OK I mean it won't run is that it fails Prime95, not that it won't goes to 1.85V.

Read the thread I have linked previously - there appears to be similarities.

Hows the cooling on the NB??? If it's the stock cooler, I wouldnt be surprised if thats the prob

yeah and try running low mhz with 1.85 volts and see if it runs smoothly just to rule that out, you have to rule everything out man, you cant just change stuff, wut i would do is overclock the CPU, then put ram async at like 50% and get the FSB up their using moderate cpu speeds so the cpu craping out isnt a factor(although yu may want to give the CPU a moderate voltage increase) once you achieve your max fsb, then you can sync your ram and see wut your ram can handle, then increase timings....just don't go haywire and assume it's this or that, you should know wut it is most likely if you rule one thing out at a time.

OK, I will reiterate my problem again. Maybe I'm did not make myself clear enough, sorry for that.

Problem is, I'm 100% sure that the system/motherboard is capable of 200FSB, as tested at 2GHz 10*200 (default Vcore @ 1.5V) with loops of 3DMark and Prime95 overnight. However, 200FSB will fail if I have the CPU at 2.1GHz, but the CPU is stable at either 12.0*175 or 12.5*168, both at 1.6V only. It just won't do 10.5*200 no matter how high the voltage is at (that's why I said 1.85V, didn't try higher).

Anyone have a clue?

I have the same problem with my NF7-S. It will do either a High CPU speed or high FSB, but not both. I don't know what the deal is. Probably something with the NB.

I understand what your saying(i still think it's the NB). Can you give us a rundown on the type of cooling you're using for everything(NB/SB/mosfets)???

Just because a board can handle 200mhz at one cpu mulitplier, does not mean it can feed a CPU at 200 mhz at higher/different multiplier. At least not without better cooling i think

One thing...the new 1700's hate voltage over 1.8v's. They are at 1.5v default, and run great betwean 1.5 - 1.75 or so. Anything over 1.8 usally wont let them boot...just like you having now...

Just read the NF7-S 2.0 review at VR-Zone, in page 4 of the article (Link), they're having the same problem with me, as stated under "Issue with FSB over 238 Mhz".

here's what happened to me, the first time i did 200 fsb, i tried 10x200, but no matter what the vcore i would either crash upon entering windows, or sandra would read the proc at 2ghz, but 3dmark01 and 03 would read it much lower ~1850 or something. So switched to 10.5x200 and it booted at 1.675 vcore and read correctly and benched correctly and stably in all programs. So then I switched it back to 10x200 at 1.65 vcore and it worked. I don't know what happened as i didn't ouch my memory (512 mb hyperx pc3000 at ddr400 2-2-2-5 at 2.63 volts) only touched the cpu multiplier and it fixed itself. btw i have an epox 8rda+ and a tbred-b 2100 0301.

another a weird thing for me is that i can't get any higher then 2.2 ghz no matter what i try. it won't go into windows and my temps go crazy, even with an slk800/tornado (generally run a 45 cfm delta instead of the tornado)

Whats your VDD voltage at those settings.

I still believe your chip doesnt like 1.85v, anyways im going to read the link you sent.

Theres a big difference between running at 200 and 238FSB! As you know these board dont officially support such speeds but some are capable - especially the Abit. At 238 it is the chipset limitations holding him back not his cpu. However yours at 200FSB may not be the chipset - have you eliminated vcore as an issue (say by running at default or lower but with vcore of 1.9 or so)? If it runs OK then it may well be the chipset.

OK guys, sorry for this late reply.

Now the system is running 2.3GHz (11.5*200) at 1.7V, just passed 15hrs+ of Prime95, full loading temperature around 53C with a case temperature of 28C.
However, it still fails 2.1GHz at 200FSB, even if I have the voltage up to 1.85V, won't even get into Win2k SP3.

Wierd......