You should have been able to exchange it under 30 days - you're located in the states?
--fibo
Printable View
You should have been able to exchange it under 30 days - you're located in the states?
--fibo
Nope, Sweden...
We've got 14 days to return the merchandise after we paid for it. 25 days has passed (yes, dumb, but I thought I'd be able to bump the speed more). The only thing I can do is to "break" the card in some way and return it for warranty, but that's not really something I'd do unless it really breaks.
I guess I'm stuck with this card until it's time for another upgrade. And it's not really fair to complain, I can run it past 4400 speeds, and I doubt I'd notice enormous difference (unless I could bump the core to 330+ and the memory to 700+). Stock 4200 speed is 250/500 (250/444 for 128mb), and I run mine at 300/575 without any mods (except the loud OTES system). Perhaps there's a need for more voltage or something, I don't know.
The reason why many of us can't get good oc results is not psu or agpvolt related. It's simply due to the fact that Albatron later used different components to manufacture the 4200p turbo card, specifically the ram.
The response written by Roger Chien to Dissolved isn't exactly true:
Thank you for choosing Albatron product.
In regards to your issue, we manufacture all the Ti4200P Turbo video cards in the same standard and configuration, there is no special or different batch among these cards...
Sincerely,
Roger Chien
---
I, like many of you, bought a 4200p turbo from newegg.com, hoping that it'd perform as well as the ti4600. But when I received it, I noticed that it wasn't the same card being reviewed on all of the websites. The main difference on mine is the absence of the hole and different design of the front ram heatspreaders. On all the review sites, http://www.neoseeker.com/Hardware/Pr...P/reviews.html, all of the cards have front heatspreaders with the hole.
I decided to investigate a little further by prying one of the heatspreaders and found what may be the key reason why many of us can't get good overclocking results. The ram used on my card without the holed heatspreader is 128 mb Samsung 238, K4D26323RA-GC33 300MHz 600Mbps/pin SSTL_2 (VDD/VDDQ=2.8V) 144-Ball FBGA
I don't exactly know much about this ram, except that it's also used on the Geforce 4 MX440 8X. Go to the bottom of this page http://www.overclockers.co.nz/ocnz/v...x4408x/1.shtml.
Now, the ram used on the cards with the hole is Etrontech, http://www.guru3d.com/review/albatro...g/DSC00842.JPG. Notice the difference.
So my guess is the Etrontech ram has a huge influence on the overclockability of the card. Note that the revision on my card is A3, same as the card that was reviewed at Guru3D and perhaps all the other sites.
grapeape,
Welcome to Xtreme :)
Great first post...used to watch that cartoon religously back in the day BTW ;)
hi n8. thanks for the kind welcome.
ya grape thanks for that info.
Quote:
Originally posted by grapeape
The reason why many of us can't get good oc results is not psu or agpvolt related. It's simply due to the fact that Albatron later used different components to manufacture the 4200p turbo card, specifically the ram.
The response written by Roger Chien to Dissolved isn't exactly true:
Thank you for choosing Albatron product.
In regards to your issue, we manufacture all the Ti4200P Turbo video cards in the same standard and configuration, there is no special or different batch among these cards...
Sincerely,
Roger Chien
---
I, like many of you, bought a 4200p turbo from newegg.com, hoping that it'd perform as well as the ti4600. But when I received it, I noticed that it wasn't the same card being reviewed on all of the websites. The main difference on mine is the absence of the hole and different design of the front ram heatspreaders. On all the review sites, http://www.neoseeker.com/Hardware/Pr...P/reviews.html, all of the cards have front heatspreaders with the hole.
I decided to investigate a little further by prying one of the heatspreaders and found what may be the key reason why many of us can't get good overclocking results. The ram used on my card without the holed heatspreader is 128 mb Samsung 238, K4D26323RA-GC33 300MHz 600Mbps/pin SSTL_2 (VDD/VDDQ=2.8V) 144-Ball FBGA
I don't exactly know much about this ram, except that it's also used on the Geforce 4 MX440 8X. Go to the bottom of this page http://www.overclockers.co.nz/ocnz/v...x4408x/1.shtml.
Now, the ram used on the cards with the hole is Etrontech, http://www.guru3d.com/review/albatro...g/DSC00842.JPG. Notice the difference.
So my guess is the Etrontech ram has a huge influence on the overclockability of the card. Note that the revision on my card is A3, same as the card that was reviewed at Guru3D and perhaps all the other sites.
gawd damn, thats EXACTLY what i wanted to hear! Now i gotta find one of these cards with that memory..
I'll be back in 1 hour, I have some ideas that I think we need to try, but I wanna research them a bit.
hmmm, Sounds interesting.. your first post.. I cant WAIT! lol..Quote:
Originally posted by Radelon
I'll be back in 1 hour, I have some ideas that I think we need to try, but I wanna research them a bit.
ive opened up the bios, and nothing is disabled, So at Least we know the Bios is fine.
Radelon,
Welcome to Xtreme :)
Ok first off, we have identified that there is a difference. Now, on that note, we need to see if we can solve the problem.
Here's the theory:
1. Newer Cards (Samsung Ram): These cards use a different HS on the ram than the original version. These cards also are only able to hit 640mhz max before having problems. This also seems to effect the core on this version, as the cores cannot get the numbers from the original version. This can all be related to the fact that maybe the Samsung ram requires more power to run than the original ram. Because if the ram being power hungry it could be trying to rob power from the core therefore resulting in an all together lower speed. Now, with that in mind, I did a little research on the Samsung Ram. It seems that on any card the Samsung ram is power hungry. But it also seems that if you give the ram more juice, it will perform better, it just needs more power.
2. Older Cards (Etrontech Ram): The cards use a hs with 2 straps in between the HS. These cards are the cards that everyone wants but just can't seem to snatch up. Now, investigating the matter I found that this Ram may still be robbing the core of voltage when they start to hit around 715mhz.
Ideas/Solutions:
1. Bios extraction: We need someone with each card to backup their current VGA bios so that we can compare them for differences. This could lead to a resolution (most likely won't, but it's a valid shot).
2. VMOD Memory: Now, this is tricky but I'm sure that some people could handle it. I really want to see what happens if you up the Memory Voltage on the Samsung version up about .2v and see if you get an overall increase in both core and memory speeds. If so, we would be able to conclude that the Memory does need more juice and the reason that as a whole, the Samsung version of the card clocks way less is due to a low voltage such that the ram needs more voltage to run and starts to rob the core, therefore causing everything to be slower. If in fact, this does allow for higher clocking on both the core and ram then we also might try the same on the Etrontech version. The cores on these should easily hit 340mhz (my MX420 core hits 350mhz). Maybe the reason the cores can't reach the higher speeds is that in both versions the ram is robbing the core of voltage therefore making the max oc lower.
I hope you all understand my theory, and for the sake of many, I really hope that my theory is correct. If so, we will all be happier.
It's not the disabled part I'm worried about, I'm worried that maybe there's different ram timings, or something of that nature.Quote:
Originally posted by Dissolved
hmmm, Sounds interesting.. your first post.. I cant WAIT! lol..
ive opened up the bios, and nothing is disabled, So at Least we know the Bios is fine.
well Albatron has the bios on there website..Quote:
Originally posted by Radelon
It's not the disabled part I'm worried about, I'm worried that maybe there's different ram timings, or something of that nature.
Do you have aim/icq/msn?
i looked through the bios, and i cant find ram timmings?
i can confirm what rade is saying. More messing with my card has resulted in 250/640 stable on the memory side while on the core side 285/552 seems to be the max.
The weird thing is the video lag everyone is describing, I have it to and i dont know why. I mean i have overclocked my friends msi 4200 64mb card and my other friends visiontek 4400 128mb card and when your upping the core the only thing you will see is video lockup. Now i get the freeze and go type lag of video, it doesnt make sense.
Quote:
Originally posted by kap
i can confirm what rade is saying. More messing with my card has resulted in 250/640 stable on the memory side while on the core side 285/552 seems to be the max.
The weird thing is the video lag everyone is describing, I have it to and i dont know why. I mean i have overclocked my friends msi 4200 64mb card and my other friends visiontek 4400 128mb card and when your upping the core the only thing you will see is video lockup. Now i get the freezing of video, it doesnt make sense.
i was gammimg at 300/650 and i noticed lag too.. like the frames just craped...
dissolved:
I think by increasing AGP voltage through mobo's BIOS has nothing to do with higher MHz for GPU and Memory. All GF4 cards are designed with built-in power supply & distribution IC (Semtech SC1102 or SC1175). So, when you increase the AGP through mobo bios, the IC will get HOTTER. You need a resistor to increase the Vsense.
I think you should OC your Albatron with voltmod before the whole computer break-down due to crashes, re-booting, on-off, etc.
i dont have a Way to do so on my KX7-333.. the AGP ic chip is right in front of the AGP slot, so the card's hsf blocks the voltages chip. I dont solder and wont ever on voltage mods, so i cannot do that enless i buy a mobo with agp in the bios. And the only better board then mine is the Epox nf2, and that only goes to like 1.8. Which is AGP voltage, From what i know that only powers the SLot, not the actual vgpu/vmem on the card. So id have to do a Card mod.Quote:
Originally posted by asw7576
dissolved:
I think by increasing AGP voltage through mobo's BIOS has nothing to do with higher MHz for GPU and Memory. All GF4 cards are designed with built-in power supply & distribution IC (Semtech SC1102 or SC1175). So, when you increase the AGP through mobo bios, the IC will get HOTTER. You need a resistor to increase the Vsense.
I think you should OC your Albatron with voltmod before the whole computer break-down due to crashes, re-booting, on-off, etc.
ive bought the better Albatron today, and when i get it i plan to modd it. at least the memory. If i keep my pelt block, ill vmodd the gpu.
But for now this card i have with the samsung memory is going back for a full refund.
i volunteer to vmod the memory if someone tells me how to do it.
i have a removable vmod kit i made that is like the one in this article: http://www.liquidninjas.com/reviews....owcontent&id=9
what pins on what chip do i hook it up to? i have 2 of them i made
You will need pomona grabbers for this, i think those clips are too big.
What size VR/Pots do you have?
GPU Mod, Connect to that Pin, and the other wire to Ground.
(this is on the back side of the card)
vmem..
this is where to check the voltages..
dissolved, i definetly can EASILY mod my memory w/ the grabbers i have (the one in the liquidninja article was just an example of somethin like what i have).
however, i still am not sure how to check to voltage of the ram. what do i hook up to the card to see? a multimeter? if so how do i do it w/ the multimeter? (i have one but dont know how to use)
using stock cooling what voltage do you think i should set the memory at?
well, you just hook the RED prob to the positive on the card and black to the negitive.
The multi meter should have a setting for DC voltage set it so it can read up to 5/10/20 volts.
ill talk to my friend tonight, and see if we can make a fully Guide tonight if hes online. i have a multi meter i can show you..
when i get the other albatron ill take some pics for you and clearly point this all out.
thanks for your help dissolved yah i still dont really get about how to check for the voltage.
well i tried uppin the agp voltage from 1.5 to 1.7 and that did NOTHING to help increase stability at high speeds
mobo vagp wont do much i think.. it helped alot on my gf3.. but i think its all on the agp bus, i doubt the vagp voltage will do anything for the card.. Probly only good if you had a high agp bus speed.
well if you can figure it out, when i get my new card, ill make a write up for everyone. my bro might buy this card too, and i might make him vmodd it also..