Why cant i use a higher vcore than 1.9?!

[DisposableHero]

i have to disagree.. if the system is on... the computer will be running

[Bondo]

No.

I see it all the time. NB fan still spins and current is going thru it.
If the siren is on, then electricity is powering the speakers so you hear the siren.

[Bondo]

Wait a min.

I just read the pdf file on this VRM. It's only rated to 1.85v !!!

No wonder the OVP is tripping. They used the wrong Vcore voltage regulator.

http://www.irf.com/product-info/data...ta/iru3055.pdf

[Bondo]

We're going to have problems any way we cut it pun intended.

Even if you cut the OVP, the VRM will fail and cause instability. A 1.85v VRM is only rated for perhaps 1.90v before OVP kicks in. You think it will function right at 2 to 2.2v? Never...

[DisposableHero]

well mine wont even work at 1.85.. pretty sad... actually let me test 1.85 now that ive cut the pin 16

[Photon]

Originally posted by Bondo
Wait a min.

I just read the pdf file on this VRM. It's only rated to 1.85v !!!

No wonder the OVP is tripping. They used the wrong Vcore voltage regulator.

http://www.irf.com/product-info/data...ta/iru3055.pdf

Yeah I saw that too, makes you wonder what tricks they have to do to get 2.2v out of it.

DH,

I don't know what to say. I only cut the trace leading to pin16 (OCSet). I'm still running fine at 1.975v and no more shutdowns. If you wanted to retain the overcurrent protection, you'd have to change the value of the resistor connected to pin16. You could use a pot (based on the value of the stock resistor) and keep turning it up a little every time you get a shutdown until the shutdowns stop. I'm running with the overcurrent protection disabled though, I'm nutty like that.

You might want to hold off on any more changes until somebody can post a close up pic from the board and I can draw in where the cut on the trace goes.

The shutdowns you guys are describing is exactly what I had before the mod. System would power down and sometimes would get the siren, sometimes not. Seems like the whole system went dead though, I don't think the fans were running even though the siren was going. Whatever is driving the siren may be getting power from the 5v standby circuit, I think that is active even when the PS is "off".

[reverend]

For how long have you been running 1.975V Photon? And have you tried 2.2V? Mine runs at 1.9 for 5 - 6 hours then shuts down. 2.0 and 2.2V doesnt work at all, might run the system for 30 seconds or something then it shuts down. Havent tried the mod yet.

[Bondo]

These are defective boards no doubt about it but I know we'll hear differently from some people here. I expect it. I mean it's user error right??

Anyhoo, I'm willing to cut pin 16 3rd up on from the botton on the left side with the notch being on top of the VRM but it didn't seem to work for Hero. So far it's 1 for 2.

Perhaps he didn't make a clean enough break and we're getting
intermitent connections?

[reverend]

OK. So now I've noticed a few things. I tried running the CPU at stock speeds, XP 1800+ (1533 MHz) with vcore set to 2.2V and additional cooling. Entered windows, ran TOAST for a few minutes without the system shuting down. Before I ran my cpu at 2.2V and I barely made it into windows without the system rebooting. Might have been the CPU all the time or am I wrong on this one?
Since I overclocked it quite bad last time with 2.2V and now it ran 2.2V and stock speeds without rebooting.

[DisposableHero]

well... i tried to cut the trace... cut everywhere before the pin... then that didnt work so i cut right at the bottom of the pin.. which cut the pin...

no difference



and about pin 15.. what voltage is that monitoring... cuz maybe since the voltage chip is for 1.85... it might be trippin something

[Bondo]

Read the description in the pdf file I posted. It also has to do with
monitoring the vcore for faults but I don't know what would happen if it got cut.

[DisposableHero]

Originally posted by Photon


Yeah I saw that too, makes you wonder what tricks they have to do to get 2.2v out of it.

DH,

I don't know what to say. I only cut the trace leading to pin16 (OCSet). I'm still running fine at 1.975v and no more shutdowns. If you wanted to retain the overcurrent protection, you'd have to change the value of the resistor connected to pin16. You could use a pot (based on the value of the stock resistor) and keep turning it up a little every time you get a shutdown until the shutdowns stop. I'm running with the overcurrent protection disabled though, I'm nutty like that.

You might want to hold off on any more changes until somebody can post a close up pic from the board and I can draw in where the cut on the trace goes.

where did u cut the trace? right where the pin touches the mobo?

[Bondo]

Photon,

can you try 2v vcore or more and see it it shuts down under Prime?

[Photon]

Originally posted by reverend
For how long have you been running 1.975V Photon? And have you tried 2.2V? Mine runs at 1.9 for 5 - 6 hours then shuts down. 2.0 and 2.2V doesnt work at all, might run the system for 30 seconds or something then it shuts down. Havent tried the mod yet.

I did the mod and cranked it up to 2v right away so I wouldn't have to wait for results. See the screen capture I posted in the thread. At that point I was running Prime for about 20-30 minutes. Before the mod, I couldn't run at 2v for 30 seconds. I have not run at 2.2v, I have no need to, that's way too high for air cooling. I have been running stable at 1.975v since doing the mod yesterday (over 24hrs). No shutdowns at all since cutting the trace.

BTW, no other changes were made. No new cooling or anything.

[Bondo]

You know what I don't get?

I'm running Prime right now at 2.025v vcore and BUT the CPU at default and it's OK. I don't get it now. I know this 2400+ did 2500mhz Prime stable in the 8K3a+ Now it can't do 2300 with out the shutdown and over 1.90v.

[Bondo]

BTW here's bench off my rda+

water cooled North Bridge and vdd mod at 1.87v

[Bondo]

BTW,

I always test a board at Default before I mod or Overclock it. I had these shutdowns then too.

[Photon]

Originally posted by Bondo
These are defective boards no doubt about it but I know we'll hear differently from some people here. I expect it. I mean it's user error right??

Anyhoo, I'm willing to cut pin 16 3rd up on from the botton on the left side with the notch being on top of the VRM but it didn't seem to work for Hero. So far it's 1 for 2.

Perhaps he didn't make a clean enough break and we're getting
intermitent connections?

As I said when I wrote this up and recommended to DH, I cut the trace leading to pin 16, NOT the pin. You'll see the trace (green raised line) coming from the pin. It passes between two "vias" or holes. I cut the trace right where it passes between the holes. I never touched the chip itself.

I used an x-acto knife and made several passes at light pressure until I was through the trace. Don't get crazy with the knife, these are multi-layer boards.

[Bondo]

Right after the holes, or in the middle of them?

[Photon]

Originally posted by Bondo
You know what I don't get?

I'm running Prime right now at 2.025v vcore and BUT the CPU at default and it's OK. I don't get it now. I know this 2400+ did 2500mhz Prime stable in the 8K3a+ Now it can't do 2300 with out the shutdown and over 1.90v.

One thing to keep in mind...

There can be other reasons for the shutdowns. What kind of temps are you getting at 2500Mhz on the other board? The 8RDA will watch the AXP's internal diode and shutdown based on that too (who knows what temp). The reported temp is from a diode under the CPU socket.

The real bummer is that I don't think any of these shutdown circuits can be adjusted through the BIOS (assuming Epox would adjust them for us). They are designed as simple, brute force circuits, they have no logic that the BIOS can talk to.

DH,

Don't get discouraged yet. You've eliminated the overcurrent circuit as the cause of the shutdowns, maybe try looking at other causes. Run a log in MBM and see what Vcore looks like at shutdown. Is there a big drop (or rise)?