EP-8rda6+ Pro Problem

[franky]

Hi all,
I'm stuck with a problem on my Epox 8rda6+ Pro. I've got an XP-M 2500 and there's no way to set multiplier lower than 13x. It can be freely set higher, while setting it lower overclock fails and boots in safe mode. I even tried with modded Merlin bios with no luck. Can anyone please help me?
Thanx

[Wolf2000me]

That's odd. I've never heard of such a problem. It's usually the other way around. What does you chip default to? Did you do a pinmod like the L12 mod? Maybe your used arctic silver 5 and spilled some on a bridge or two resulting in this kind of odd behaviour?

[franky]

Yes, very odd. It defaults to 1400 (100 x 14) as 14 is default cpu multiplier. I did no pinmods. I just put some thermal compound on the core, anyway I'm going to clean cpu and retry.

[Wolf2000me]

Normally the 8RDA6+ Pro should not recognise that default multiplier It should post 6x100 unless you use a bios with L12 mod. Let us know how you did. Of course if you didn't use a conductive thermal compound it's no use reapplying and reseating the heatsink.

Which board revision do you have? It's in the corner near the pci slots on the long side of the board. Although I'd be greatly surprised if it is anything else than 1.2 or 1.4. And even then, an older bios should do the trick which you already tried.
Really, I'm rather puzzled by this.

[franky]

I'm going mad...I tried again with same results. Then tried with an athlon xp 2700 and works fine: multiplier is free. So I think it should be something in the cpu. motherboard revision is 1.4 and now i'm using the merlin modded bios.
cpu label says: AXMH2500FQQ4C IQXEA0327TPMW...

[Wolf2000me]

I do find it a very strange CPU. The first code is a completely normal mobile XP code but the other one is an "I" stepping I had never seen before. If you've got a digital camera, could you take a picture of the CPU? Make sure the label on the cpu is clearly visible.

Where did you buy the CPU? I'm intriged because a mobile should be multiplier unlocked. A mobile also has a different multiplier FID signal than regular XP processors hence the 6x100 default. What is its default voltage?

[jnansley]

The IQXEA is a very early stepping - notice the week 27 of 2003. With a bit of googling I don't see that anyone else has the same problem. It sounds exactly like you've done the AJ27-AH28-bridge trick. Try taking the CPU out and making sure there's not any crud in the socket? I'm guessing you've obtained this second-hand - check for conductive ink on the bottom?

[franky]

I tried with merlin modded bios, original bios and latest bios from epox and still get the same result. CPU defaults to 100 x 14. I gave a look at the socket and under the CPU but there is nothing. CPU is second hand while motherboard is brand new. I'm sending a pic of the CPU. Don't know what else try.

[ajopz]

sounds like you have a dodgy cpu. <it happens!
Ive got a 3000+ that just will not run at the stock speed of 13.5 x 166mhz. If i want it to run stable it needs to be set at 161mhz.
Just a case of the bad chip in the bunch.
Is the xp2700 you tried yours or did you borrow it?

[franky]

Originally Posted by ajopz

sounds like you have a dodgy cpu. <it happens!
Ive got a 3000+ that just will not run at the stock speed of 13.5 x 166mhz. If i want it to run stable it needs to be set at 161mhz.
Just a case of the bad chip in the bunch.
Is the xp2700 you tried yours or did you borrow it?

I don't know as former owner remembers he set multiplier lower than 13. It's not a matter of speed as it doesn't accept lower multipliers even at 100 mhz bus. It seems like motherboard can't handle this cpu very well. I think I'm going to try this on an Abit AN7 and see what happens. Anyway the 2700 it's mine, why?

[Wolf2000me]

A mobile should be a mobile for an nforce2 board. I'd be surprised if the CPU would react differently with the multiplier on another nforce2 board. If both CPU sockets are clean.

[ajopz]

Originally Posted by franky

Anyway the 2700 it's mine, why?

How does the 2700 oc?

would it be that bad to stick with the 2700?

[franky]

Originally Posted by ajopz

How does the 2700 oc?

would it be that bad to stick with the 2700?

I didn't really try extreme oc on the 2700, but I don't expect much as it is the higher t-bred cpu. Wouldn't be bad to stay with it but I expected better result from the mobile and this mistery is really teasing me
I made some tests:
1 - I used the interactive pin mod here and tryed the "Unlock multi up to 12,5X" option with no results, it was just like no pin mod;
2 - this time I selected a 11x multiplier and did the pin mod and system didn't start at all;
3 - set default multiplier in bios and it shows Athlon XP 2500+ (133 x 14).
This is very strange as I supposed default 6x. It seems that motherboard correctly detect a mobile cpu and sets default multiplier to the maximum multiplier sets by L6 pins and probably can't handle the last pin so it's impossible to go under 13x. Could it be possible?

[Wolf2000me]

Originally Posted by franky

I didn't really try extreme oc on the 2700, but I don't expect much as it is the higher t-bred cpu. Wouldn't be bad to stay with it but I expected better result from the mobile and this mistery is really teasing me
I made some tests:
1 - I used the interactive pin mod here and tryed the "Unlock multi up to 12,5X" option with no results, it was just like no pin mod;
2 - this time I selected a 11x multiplier and did the pin mod and system didn't start at all;
3 - set default multiplier in bios and it shows Athlon XP 2500+ (133 x 14).
This is very strange as I supposed default 6x. It seems that motherboard correctly detect a mobile cpu and sets default multiplier to the maximum multiplier sets by L6 pins and probably can't handle the last pin so it's impossible to go under 13x. Could it be possible?

Maybe it is a "malfunction" of the board. Normally the board should misinterprete the FID signal and post at 6x100. It seems it only does this partially as you post at 14x100. You did mention to test it on an AN7. Did you try that?

What you could try for pinmod is the reverse of the mod that is done for accessing the higher multipliers.
Take socket view. Select multiplier 12.5 and look at the leftmost wire connection. Then switch to 13x. You will notice that the left pin changes. You will notice that that pin remains the same for all multipliers over x13 and the same thing for all multipliers below 13x. So that's the one you might want to try. To obtain the higher multipliers like this on a VIA board with a mobile worked for me. I have however not seen or done it on an nforce2 board.
I am sure however that your board should post at 6x100 as anyone else who has used a mobile with the 8RDA6+Pro. As I'm sure you already know since you snooped around the romsip mod thread

[jnansley]

To disable 5th multiplier bit you would need to be breaking a connection instead of making one.

[franky]

In your opinion, could it be a faulty cpu? I suppose I can exclude the mobo as it works with all multiplier with another cpu. Anyway I'm going to try with another mobo (not the an7). Thanks for help

[Wolf2000me]

It could be, yes. A normal mobile in a laptop would need to be able to lower its multiplier, not raising it.

Did you check the bottom of the CPU and more importantly the pins of the CPU for any anomalies?
At any rate. The only difference in bridge connections between 6x and 14x is located at the L6 bridges. The last bit is supposed to be connected instead of cut. Check the ocinside page, select workshop and choose the third socket A guide for tbreds and bartons. There choose a mobile cpu and leave anything else unchanged. Switch between 6x and 14x.
If you compare your L6 bridge setting from your pic with the 14x setting on that guide you will notice that it is the same. Therefore connecting that last connection would make your cpu behave normally. Your L5 bridge config seems to be correct.

[franky]

Originally Posted by Wolf2000me

It could be, yes. A normal mobile in a laptop would need to be able to lower its multiplier, not raising it.

Did you check the bottom of the CPU and more importantly the pins of the CPU for any anomalies?
At any rate. The only difference in bridge connections between 6x and 14x is located at the L6 bridges. The last bit is supposed to be connected instead of cut. Check the ocinside page, select workshop and choose the third socket A guide for tbreds and bartons. There choose a mobile cpu and leave anything else unchanged. Switch between 6x and 14x.
If you compare your L6 bridge setting from your pic with the 14x setting on that guide you will notice that it is the same. Therefore connecting that last connection would make your cpu behave normally. Your L5 bridge config seems to be correct.

Here I am, I made some interesting tests. I tried the cpu with an abit an7 and it behaved the same way, so I can safely say that the mobo is ok. I then tried with an asus a7v8x which has via kt400 chipset and, as expected, it started at 100x14 and I wasn't able to set a multiplier lower than 13 in bios. Then, I started windows and used right mark clock utility and, surprisingly, I was able to lower multiplier to 11x (tested with cpu-z). So, I'm thinking to paint the last L6 bridge as you suggested and try again.

[Wolf2000me]

Great

Good luck, and do post your OC results