high AGP freqency can fry video card?

[oops]

hey,

my friend's Geforce4 Ti4600 start to have artifacts all over the monitor. but the video card was NEVER overclocked, the only thing overclocked is the CPU, the FSB is 151MHZ, so the AGP frequency is 100MHZ.

and we tried to reset the BIOS to original and format everything in hard drive, still give those artifacts on both monitors(dual Flatout monitors) even in BIOS.

so we tried to switch the video card with another computer which has a Radeon 8500 in it, then the artifacts are gone for the original computer, and the second computer starts to have artifacts with that Ti4600.

So I think that's the video card is damaged......never heard about video card can be damaged like this without directly overclocking to video card but the high AGP frequency.

so any advise? should just use the warranty to get a new one? or any thoughts think that it will recover?? :)
(my Radeon 8500 was recovered once from -------black and white only with 3dmark shows 4MB video memory, but it recovered by itself 2 days later)

[spazzkid]

i dont think it kud fry a vidcard. but hardlocks are annoying. back down the fsb a bit.

[Top Hat Theater]

Try running everything at stock and see.

~THT

[oops]

I've mentioned we set everything back to normal, still didn't fix the problem.(everything include clean BIOS memory and even format the hard disk)

We also tried the same card on the other computer, which proved that the problem is on the video card. And the video card from another computer runs fine on this computer.

[Maxvla]

Originally posted by oops
We also tried the same card on the other computer, which proved that the problem is on the video card. And the video card from another computer runs fine on this computer.
that right there proves its the card.

it does the same thing on another computer. card is toast.

not really much else to talk about. sorry :(

time to save for a new card.

[Maxvla]

regardless to the fact that you didn't directly overclock the card you did via the fsb. returning the card as defective (warranty) would be wrong. unless you have a complete warranty.. which covers anything you should not send it back.

[oops]

that's not mine, it's my friend's Gainward Golden Sample Geforce4 Ti4600 which was ordered in early summer. Good thing is... there is warranty on the card. So if get a refund, may get the original $350+. That may be a good thing ....after all the same card worth around $270 now :)

[Samoyed]

The Golden Sample cards are sold as OCing cards but even if this was not a GS card I would return it. 150 FSB makes the AGP run at about 78 or so. Since virtually every other GF4 can run at this FSB speed, I think he has a defective card.

[john240sx]

Originally posted by Samoyed
The Golden Sample cards are sold as OCing cards but even if this was not a GS card I would return it. 150 FSB makes the AGP run at about 78 or so. Since virtually every other GF4 can run at this FSB speed, I think he has a defective card.

if the card can run at 66Mhz AGP then it's not defective. when you buy something that everyone else is getting good overclocks with but you can't get it over stock speed then RMAing it is wrong. the card does everything it's supposed do to.

[oops]

definatly, you can see "Overclocking" in the descripition of Golden Sample, and we didn't even overclock the video card yet, just the 100MHZ AGP frequency for only 2 days, the card starts all those artifacts....even in BIOS

[Taylor]

Originally posted by Maxvla
regardless to the fact that you didn't directly overclock the card you did via the fsb. returning the card as defective (warranty) would be wrong. unless you have a complete warranty.. which covers anything you should not send it back.


I'm afraid I have to respectfully disagree. Today's ddr mobos are often billed as "333 FSB", (which is a bit of a white lie) and so it's reasonable to expect a new card (esp. if billed as an overclocking card) to perform adequately with the fsb set to 151 (for 302).

Unless the manual specifically lays out a maximum agp frequency, and assuming the core/mem clocks were left at factory spec, then I just don't see a problem with returning it as a defective card. It is defective.

It should not have had any problems.

In my humble opinion, anyway.

I would most certainly RMA the thing, and I wouldn't lose any sleep over it. And I do consider myself an ethical overclocker.

Who's to say it was the high fsb that made it go buggy? Seems as though no one else's does.

Anyone who's fried their card by high FSB only, feel free to jump right in and correct me...

Is it not reasonable to wonder it may very well have been some defect in the card, which would have manifested itself anyway?

The only thing for certain is, he paid $350 for a card marketed as a good overclocker, and before he could even try, the thing went tits-up. There's no moral issue, given the situation as Oops has described it.

[PhobMX]

Originally posted by Taylor

the thing went tits-up.

ROFL

well, on the topic, its an overclocking card, RMA it for sure.

And hey peeps please help me out, my asus a7v333 mobo has 1/5 divisor, if i run the fsb at 200MHz, whats the AGP Speed gonna be???

[iceman2g]

how do you determine your AGP speed? My FSB is at 165 with only 1/4 divider!

[PhobMX]

165mhz fsb on 1/4 divider = 165/4 = 41.25

It looks kinda high dude, i wouldnt stress pci/agp bus more than 40MHz

[iceman2g]

no I mean what's my AGP bus? Isn't stock for AGP 66mhz?

[DarkArctic]

Originally posted by iceman2g
no I mean what's my AGP bus? Isn't stock for AGP 66mhz?

The AGP speed is twice that of the PCI bus.

So AGP=PCI*2

-DarkArctic

[Taylor]

Originally posted by PhobMX
...1/5 divisor, if i run the fsb at 200MHz, whats the AGP Speed gonna be???

80 I think

[iceman2g]

thanks for the help!!:D

[PhobMX]

Originally posted by Taylor


80 I think

would it be too high??? can it fry the video card???

[ThePunkGeek]

333 changes the pci/agp divider also i think. :\

Originally posted by Taylor



I'm afraid I have to respectfully disagree. Today's ddr mobos are often billed as "333 FSB", (which is a bit of a white lie) and so it's reasonable to expect a new card (esp. if billed as an overclocking card) to perform adequately with the fsb set to 151 (for 302).

Unless the manual specifically lays out a maximum agp frequency, and assuming the core/mem clocks were left at factory spec, then I just don't see a problem with returning it as a defective card. It is defective.

It should not have had any problems.

In my humble opinion, anyway.

I would most certainly RMA the thing, and I wouldn't lose any sleep over it. And I do consider myself an ethical overclocker.

Who's to say it was the high fsb that made it go buggy? Seems as though no one else's does.

Anyone who's fried their card by high FSB only, feel free to jump right in and correct me...

Is it not reasonable to wonder it may very well have been some defect in the card, which would have manifested itself anyway?

The only thing for certain is, he paid $350 for a card marketed as a good overclocker, and before he could even try, the thing went tits-up. There's no moral issue, given the situation as Oops has described it.

[iceman2g]

I've had it runing this high for only three days, but i've always had it of of spec.

[john240sx]

Originally posted by oops
definatly, you can see "Overclocking" in the descripition of Golden Sample, and we didn't even overclock the video card yet, just the 100MHZ AGP frequency for only 2 days, the card starts all those artifacts.

that 'overclocking' in the name refers to the Core and Ram ability to overclock. not the FSB.

[Taylor]

Originally posted by ThePunkGeek
333 changes the pci/agp divider also i think. :\



Yes that's my point- er I think it is...well my point is any modern day manufacturer selling a card as an overclocking card has got to realize the agp freq is like to be elevated to some degree, and if that's going to be a problem they should publish recommended agp freq max.

But again, are we even quite sure it was an elevaed agp freq that made it go buggy? Has anyone else had this problem?

I'm pretty sure different 333 boards handle the divisor differently, some better than other , help here?

[oops]

That was a new Abit BD7II borad.
The Gainward Geforce4 Ti4600 was bought in the summer.

Guess my friend just isn't so lucky.

My Gainward Ti4200 64MB 3.5ns can handle the same AGP frequency no problem at all. (100MHZ AGP frequency)

And that Ti4600 was starting show artifacts for only 2 days under 100MHZ AGP frequency. Don't see any other reasons, we cleaned the BIOS memory and reformat the hard disk, so everything is back to normal, and the same card shows artifacts on another system too... not just in windows....but also in BIOS, and a Radeon 8500 from the other system works fine on this system...so it's just the card

I thought that's one good card, because I saw it has really nice RAM sink on it......

guess video cards are just like processors, it all depend on luck, every card is different....

[Maxvla]

100mhz agp?

the numbers don't add up.

at 100mhz its 66/33
at 133mhz divider kicks in and its once again 66/33
at 166mhz divider kicks in and its once again 66/33

how exactly do you get 100mhz agp?

[oops]

FSB/AGP/PCI = 150/100/50

something like that

[[OC]This]

Aey oops, I know what the problem is.


If you really ran 100mhz AGP frequency(I'm not saying you didn't, just making sure..), that is critical, high-end AGP OC'ing.. which is for advanced OCers.. as long as they know what they are doing.

Here's the rule:

GF4 families and Radeon 9700pro cards do not handle super high AGP speeds very well. The GF4s are usually max out at around 86+mhz. R9700s are even worse, they are at lower AGP speeds.

GF1/GF2/ GF3 series handle high AGP speeds very well, as high as 100+mhz. Some go 105mhz+.

Radeon64mb ddr/R7000/R7500/R8500s also handle high AGP speeds very well.. i.e. 90- 95mhz AGP + or -.

You are damn lucky with your Gainward 4600 doing fine@ 100mhz AGP. Few of those cards can co-op like that, hit or miss. But i assume your card isn't OC'ed either correct? if not, maybe a little.. that could get you by okay. You'll probably see your card start to puke when you OC that card with that AGP speed.


Your friend's Ti4600 is damaged due to high AGP speed interrigation. yes,.. they "can" go bad. If you or your friend is planning to stick with the late models like the GF4 cards, i suggest you seek for 5:2:1 ratio boards i.e. KT333/ KT400/.. or, NF2. Or, you could get a card like R8500 or GF3 ti200/500. That will solve the problem as well.

If you have a board like a i440bx chipset platform, and if you want that nice tasty sweety 145- 150+mhz FSB, you would definite need those older cards like GF2 or GF3s.. :)

Good luck.

[oops]

Originally posted by [OC]This
R9700s are even worse



lol,

You really should recall that one earlier....

I just messed up my computer again......by replaced my Ti4200 with a Radeon 9700 Pro....

That's the system I was talking about 150/100/50(mine 2.4B, not my firend's 2.26B), and it wouldn't even boot :D
I had to shut down the power supply to turn it off, then clean the BIOS memory.
And it finally booted at 66MHZ for AGP.

Then I got into windows and installed the driver for Radeon 9700 Pro, and it asked me to reboot.....

then it could never get into windows again......it says "cannot load profile" something like that.....guess shutting down power supply damaged the windows boot files.

......

[breez]

I tried my 8500LE card at 100MHz AGP-bus and it locked up. Just did this to see if I could run 150FSB with another processor (1.1GHz tB1 stepping). I guess not :/

[PhobMX]

this makes things clear to me, my 1/5 divisor mobo at 200MHz FSB would make agp 80mhz and pci to 40mhz. My r8500LE and my WD hd must handle those speeds respectively. So just get a mobo with good dividers and good hardware, agp/pci lock isnt that much necesary...

[toyotanova]

Gainward has seriously gone downhill quality-wise IMHO.Sell your replacement on eBay and get a MSI card.