Changing pinsfile on OEM Parhelia

[tugboat]

I'm new here, but have read all the related threads on making changes to the pinsfile to change certain BIOS settings. I have tried a number of times to make the board a retail by making the changes noted, but there are a few issues.

1. I have no line 380

2. Line 12 has my serial number, but not sure what it should look like as a retail (any prefix?).

3. I made changes and added line 380, but the only ones that changed were the clock settings. All others stayed the same and line 380 wasn't there.

If I set to 220/550, it locks up running the latest 3dMarks. Set back to 200/500 and it works okay, but I only get 871 scoring.

I never really gave much thought to changing things, but I recently upgraded from a P4 2.26 to a P4 3.06 w/HT and the system started randomly rebooting. A couple of times XP was able to make a minidump (nice term) and each time it pointed to the matrox driver. I have not overclocked my system and am running the latest BIOS on the system and Parhelia. I have the current Intel INF files loaded. I figured that if there was something in the BIOS that regulated voltage or something, that might help the problem. Up 'til now I haven't had any problems with my system or the Matrox drivers (currently using 1.06).

Anyway, any help would be apreciated.

[spadnos]

You probably need to get a new power supply. I have a machine that I upgraded to a Parhelia, and it refused to boot (vs the old G450). Replaced the supply, and voila! it workd flawlessly now. The old supply was a 350W unit, which should have been plenty but wasn't.
(got an Antec TruPower 430 as a replacement)
- Steve

[Peter Aragon]

hi Tugboat

try and disable HT through the BIOS, I had encountered an issue once with an obscure Wireless adapter from which the drivers would blue screen with HT enabled.
If the restarts don't occur anymore, it would be a waste to replace the power supply.

Good luck

[UtwigMU]

What's your Parhelia BIOS? I think that at some point they changed voltages of bulks, so they don't overclock as well, but I can't confirm it. When I got Parhelia, it ran fine @ 220/275 overclocked with MTSTU. After a few BIOS flashes it now maxes at 5% where she makes it stable through Aquamark. Although I tested the first overclock on very slow box (PII@467), which perhaps was not able to stress the card. It ran reef demo with 10% overclock for a few hours without problems. Now with 10% overclock, she locks up quickly into Aquamark.


WRT to system random rebooting - graphic card drivers run all the time and even though your crash may be heat or unstable CPU related, minidump shows graphic drivers as the culprit.

Give the box a day of Prime95 torture testing.

[tugboat]

Thanks for all the quick replies!

I am using an Antec true 350w power supply and it ran the 2.26 just fine. Does the 3.06 use much more? Or do you mean the voltage getting to the AGP slot?

I have disabled HT and it still does it. It was one of the first things I disabled after reducing ram speed, etc. Still randomly reboots and usually when closing a program. Last reboot was when closing media player 9 for XP.

I am using v1.4 of the BIOS which I believe to be the most current. Just can't figure out why it won't make the changes for me. Maybe there's something in the version of pbioswin.exe that came with the BIN file that looks at your current settings and won't change anything that would make it retail? Hmmm....

So does anyone know what the serial number format looks like? mine is just a single quote tick, the serial number and another single quote tick.

Does anyone have the original pbioswin.exe from the CD? If they made changes to the current one, the old one should still work.

Thanks again!

[UtwigMU]

My serial number looks like ABC12345

EDIT: Sorry, it's ABC12345 - 3 letters, 5 numbers, the first letter is K.

[tugboat]

Okay, so if I understand correctly, your serial number is a total of 9 places. Is the ABC portion the same as all others? Kind of like a prefix for a phone number? Mine is 7 places if you don't count the ticks, 9 if you do. I wonder if this is what's keeping me from making the changes. The udater might read it to know if it's OEM or retail.

[frost_ii]

You changed what fields? There it is required to change many fields for correct work. Not only values of frequencies.

[tugboat]

I changed the following lines to the values below:

10 0x030A
36 220000
56 550000

140 0x9F810002
156 0x2A2A0000

I added the line below since I didn't have it:

380 0x52383231

The only ones that took were lines 36 and 56. All other lines stayed the same and line 380 wasn't added.

I'm at a loss, that's for sure. Would love to get my hands on the original pbioswin.exe to see if it will make the changes for me.

Thanks

[frost_ii]

You are not mistaken?

At you original PINS?

10 0x030A
36 200000
56 500000

140 0x9F810002
156 0x2A2A0000

And such line is?

376 xx 'PHA-128B'

Remove a line 380. She is not necessary. Replace values

10 0x030A
36 220000
56 550000

140 0x9F810002
156 0x2A2A0000

376 xx 'PHA-128R'

NB! The line 380 is not necessary.

If not will work, we shall spend experiments.

[borat]

Originally posted by UtwigMU
What's your Parhelia BIOS? I think that at some point they changed voltages of bulks, so they don't overclock as well, but I can't confirm it. When I got Parhelia, it ran fine @ 220/275 overclocked with MTSTU. After a few BIOS flashes it now maxes at 5% where she makes it stable through Aquamark. Although I tested the first overclock on very slow box (PII@467), which perhaps was not able to stress the card. It ran reef demo with 10% overclock for a few hours without problems. Now with 10% overclock, she locks up quickly into Aquamark.


WRT to system random rebooting - graphic card drivers run all the time and even though your crash may be heat or unstable CPU related, minidump shows graphic drivers as the culprit.

Give the box a day of Prime95 torture testing.

Im not saying that your conspiracy theory is false, but from personal exprience the 2 cores i have played with decay over time, same bios settings both OEM, my 256 could push a 17% overclock when i got it and loop aquamark at infinitum, not im lucky if i see 14% out of it. The same is true for the 128 although the drop off on that was probably more like 14% at peak to 9% now. Never really noticed any changes when i changed bios, but if there are surely they will show up as descrepencies in the pin files?

[K6-III]

I've noticed the same trend with my Parhelia.

Originally, it could hit 240/320.

Now, it is having trouble maintaining 300mhz memory clocks, even with Tweakmonster BGA ramsinks.

With a 92mm fan on the Parhelia, it can still maintain 300mhz mem clocks.

As for core, I now run it at 230 mhz, but it is stable upto only 234mhz.

[tugboat]

Okay, I've been able to set most of the lines by first changing line 12 and put PBM in place of the A and then changing line 376 by taking out the A in front of 128R.

Now I have all the settings noted to make it a retail and setting the clock speed to retail. Still no go running 3DMark with retail speeds. Can only run it successfully at 200/500.

I tried changing other lines to match the file I found on this site, but it won't let me make those changes either. I guess I have to live with it at this speed. Kind of sad.

Hey, I noticed something in the design of the board (or more like how PCI in tower cases position the cards). I know from experience (don't ask) that you should never mount an audio amp upside down. Since heat rises, it will overheat the circuits and shorten the the life of it. In looking at my card, the fan is mounted on the chip and this side of the card is facing the bottom of the case. Doesn't seem like a good thing as the fan will reduce the heat somewhat, but the heat will rise back towards the board and it won't stay as cool as it could if it were mounted facing up or on its side like a desktop case.

Thanks!

[UtwigMU]

Heat doesn't rise. Hot air rises. That's called convection.

The heat transfers in two way:
- convection: heated gas or liquid has lower density so it rises above cooler gas or liquid
- transfer: when two components are in physical contact and one is hotter, the heat will tranfer untill both are of same temperature. Some materials have good heat tranfer, some don't.

As for upside down - it's AGP/PCI standard. ISA boards were initially upside down as well in AT boxes.

[tugboat]

True enough, but wouldn't the heat removed from the GPU by the heatsink and fan (in the form of hot air) be ineffective as the hot air would be blown down by the fan and then would rise only to create an area of hot air that would then be pulled back by the fan and over the heatsink again. Seems it wouldn't be very efficient.

Not my area of specialty.

Thanks