Well guess what this does?

Printable View

Show 100 post(s) from this thread on one page

06-17-2004, 09:56 AM
novagamer

Quote:

Originally posted by h0ok3r
Such a shame that vtt doest go up with the vdimm, i think that will be the biggest turn off. Cos people who should buy this thing dont wanna do a mod, still have to do a mod to up their vtt. But many people dont even know what vtt is so beginning OC'ers will prob buy it anyway.

For me it wont be usefull cos i have a Max3, and this already has a problem with vtt, which requires a mod. I dont know how high you can get with a vtt thats way off. Anyone here know that?

The way I figure it (I also have a Max3), since I already have to mod the vtt this module will be useful because I don't want to mod my psu to give higher- 3.x voltages to my bh-5. Plus I can use it on other boards (if A64 compat. is worked out).
06-17-2004, 10:14 PM
n00by

All mighty Hipro would be proud of this contraption :D
06-17-2004, 11:26 PM
Soulburner

Quote:

Originally posted by squeakygeek
I don't think anyone knows for sure :)

A few people here have run 4+ volts...
06-18-2004, 12:23 AM
The Virus

Quote:

Originally posted by h0ok3r
Such a shame that vtt doest go up with the vdimm, i think that will be the biggest turn off. Cos people who should buy this thing dont wanna do a mod, still have to do a mod to up their vtt. But many people dont even know what vtt is so beginning OC'ers will prob buy it anyway.

For me it wont be usefull cos i have a Max3, and this already has a problem with vtt, which requires a mod. I dont know how high you can get with a vtt thats way off. Anyone here know that?

Hook3r , thats why im saying. I dont see the point of buying this product when they want you to do a vtt mod then i might as well do a vdimm myself and save 50 bucks. I read about this product on ocz web site and it talks how it eliminates the risks of vmoding the mobo's and stuff but dont mention anything about vtt being off center..........I think this product is gonna be useless in my opinion.....
06-18-2004, 01:59 AM
Spec3

Quote:

Originally posted by The Virus
Hook3r , thats why im saying. I dont see the point of buying this product when they want you to do a vtt mod then i might as well do a vdimm myself and save 50 bucks. I read about this product on ocz web site and it talks how it eliminates the risks of vmoding the mobo's and stuff but dont mention anything about vtt being off center..........I think this product is gonna be useless in my opinion.....

For everyone that doesn't have a crappy Abit board this thing will work just fine. Sell your Abit and get a real board like the Asus P4C800 which has working VTT :p:
06-18-2004, 02:01 AM
Rabbi_NZ

Quote:

Originally posted by The Virus
Hook3r , thats why im saying. I dont see the point of buying this product when they want you to do a vtt mod then i might as well do a vdimm myself and save 50 bucks. I read about this product on ocz web site and it talks how it eliminates the risks of vmoding the mobo's and stuff but dont mention anything about vtt being off center..........I think this product is gonna be useless in my opinion.....

not all boards will need a vtt... as a general rule (correct me if im wrong here please) any board that tracks Vtt automatically after a VDimm mod will be able to use this module without the need for a Vtt.
Eg, P4C800E, SocketA DFIs, etc..
________
buying vaporizer
06-18-2004, 02:22 AM
cold_ice

Quote:

Originally posted by Spec3
For everyone that doesn't have a crappy Abit board this thing will work just fine. Sell your Abit and get a real board like the Asus P4C800 which has working VTT :p:

Then you have to do the droop mod, for a great oc board :p:
06-18-2004, 02:43 AM
h0ok3r

Quote:

Originally posted by Rabbi_NZ
not all boards will need a vtt... as a general rule (correct me if im wrong here please) any board that tracks Vtt automatically after a VDimm mod will be able to use this module without the need for a Vtt.
Eg, P4C800E, SocketA DFIs, etc..

yeah well my board doest do that after 2.8 volts so it requires a mod(you have to up it yurself)

i didnt know about other boards but i hoped that it would track automatically, cos then it would be usefull. Can anyone conform if this is true? (i dont think i can remember a vTT mod for the asus)

Edit: Yup just checked it its true, asus' vTT is dependent of the Vdimm.

Maybe i should get an asus, true........ Damnit damn reviewers so so damn postive about the abit, but its has a giant bug in it. But im just a poor student so ill mod my ic7 max3, a fully moded ic7 max3 rocks 2 you know. (still havent done it)
06-18-2004, 12:20 PM
texuspete00

Still no insight onto whether this will keep A64's alive, or help to? Starting to assume no because people would be knocking down the doors to buy them if it would help but not a bleep about it so far, except of course for a lot of people asking ;) Hook us up Bigtoe!
06-18-2004, 12:29 PM
squeakygeek

Quote:

Originally posted by texuspete2k2
Still no insight onto whether this will keep A64's alive, or help to? Starting to assume no because people would be knocking down the doors to buy them if it would help but not a bleep about it so far, except of course for a lot of people asking ;) Hook us up Bigtoe!

I haven't looked into the whole A64 dying issue and what causes it because I have no plans to shell out the cash at this moment, but this isn't a miracle device. It would have the same result as soldering to the dimm pins on the back of the mobo. This device just makes the process a bit more user friendly.
06-18-2004, 12:31 PM
scrible88

I want one BADLY!!!!!!!!!

It'll maybe push my Mushy 2-2-2 Special over 250mhz @ 2-2-2-11
06-18-2004, 02:59 PM
SlackeR

i guess it won't help a tad.. the way i see it, the line from Vdimm-mosfet to dimm is split up and fed into the CPU Memctrl.. and this device provides voltage that is cut at the mosfet. The voltage in the memctrl will always be the same as vdimm.. UNLESS:
1) the mobomakers start making a cirquit dedicated to this Memctrl-voltage. I actually think Soltek did this on at least one board.. their K8T800 board. my friend has one, and we could not adjust vdimm. we found a setting called 2.5v select in bios. when set at 2.7 we could use vdimm up to, and maybe beyond 2.7. as we see this option has something to do with vdimm, but it is NOT the same. so with a vdimm-mod it may stay below 2.9.
2) we could do something ourselves. add diodes, cut traces and feed 2.9v or whatever.
I do not see this as a real problem, it is something that CAN be solved.

EDIT: a64 vdimm i mean ;)
06-18-2004, 03:14 PM
Tony

Guys

VDD/VDDQ and Vref are altered on the dimm...i always thought VBT killed A64's...is that not correct?

VTT is supplied by the board and does not hit the dimm, so boards like the P4c800E do still track correctly, the IC7 does not.

My Alpha dimm is a little big so i just ordered all parts to bring it to final spec..it should fit A64 boards then. I can say we HAVE tested on A64 already with a VERY early sample and it was fine, i will however test again and post my findings here.

It will be like monday though maybe tuesday till i get some results, im decorating all weekend.

finally, we do hope to have some units available soon, we have had a small component supply issue that is now resloved so supply may be a day later than we planned.By the time modules are available we will have the board list up listing the boards we have tested.

Upto now I have tested.

P4P800E
P4C800E
IC7
IC7max3
DFI 875B (OVP on the board kicked in at 3.2V)

I will test more but I know you guys are looking for A64 results..so please hang in there.
06-18-2004, 03:42 PM
Charles Wirth

How soon till DDR2 version?

I think you know why I am asking :D
06-18-2004, 04:38 PM
Tony

OK..soldering iron out and dimm is modded (of a sort) so it fits A64 754.

Aopen AK89max..i ran upto 3.6vdimm but VTT does NOT track on this board so a VTT mod will be needed.It uses a similar control chip to the IC7 so a similar mod is needed to that..i think Macci already posted this.

It seems i was confused with VBt earlier, It seems many term VBt as VTT.

Many new boards actually have the chipset voltage adjustable in bios, the chaintech has it to 1.9V if im not mistaken and the Epox does something similar.This leads me to beleive it actually runs 1.5v default and is controlled independantly.

I will however look more into this.If you know of any board that actually links the memory controller voltage to the vdimm please post where i may check it out.
06-18-2004, 09:06 PM
Raybo

If this is anything like the motherboards I buy, I will wait until 1.1 or version 2.0 comes out.

Raybo
06-19-2004, 05:23 AM
SlackeR

http://www.xtremesystems.org/forums/...threadid=26650
seems there have been some deaths on AN50R..
a guy mentions somewhere in the thread he has found a trace to the CPU reading the same as his vdimm..
06-19-2004, 11:20 AM
ns_ripper

Quote:

Originally posted by bigtoe

I will however look more into this.If you know of any board that actually links the memory controller voltage to the vdimm please post where i may check it out.

bigtoe, I strongly advise you to do some homework about computerarchitectures first before you make such a product.

you guy's just know that more vdd gains stability when operating memory above it's specifications. That's it. you need to know a h*ll of a lot more actually.

Please read

I'm not going to explain a lot about that part. You want to earn money by the device you just made, so it's your task to know what you are doing.

As far as my input goes on this:

A pc is actually a analog device. Yes, there's no dirt in your eyes ;) Analog! :D
Why is that so? Well a pc uses a voltage for it's bits. but how to differ a logical 1 from a logical 0? Well you already mentioned it. Vref kicks in.
Vref is a reference voltage for a logical 1. and it had a little tolerance.
So, the voltages (bits)wich are being read will be compared to the Vref. when it's within the limits (not to low), it's being reconised as a logical 1.
when it's gnd or below the tolerance of Vref it's a logical 0.

when you increase your ramspeed ... ow yeah, I forgot to mention this. bits are stored with capacitance. a capacitor looses his capacity slowly, thats why sdram is being refreshed constantly.
when you increase your ramspeed, it can happen that when you fetch data from your memory, the voltage of the memorycell isn't yet within the tollerance of the vref. Resulting in a misinterpretation of the bit. So we raise the ddr voltage, you shorten the timespan needed to get the voltage in the memorycell within the vref tollerance decreases. (like a battery speedcharger, when you put more voltage on the batteries, they reload faster (lifespan will decrease also)). The memory controller is feed by the same Vref, wich is 2.5V for ddr. When the memory controller fetches bits (wich are voltages --> vdd) from the memory, it compares them to Vref and passes them trough to the cpu.

so these bits (voltages) also enter the memory controller. will it harm the memory controller? I can't answer that, as I don't know. Some devices can stand some overvolting, some can't.

this is a pretty rough explanation. I haven't got time to realy go in-depth. But when you make such a device, make it commercially avaible and make profit on it, ...

imho, it's your duty to know that yourself. These questions should already been tought of before you designed it.

it's quite easy, to find the spots where the ddr voltage is located on the dimm's pins and feed it an external voltage.

I would realy like to mention it again, this is not meant as an insult of whatsoever. But it upsets me a little bit that people are making profit out of things. And they don't even know what they are messing with.

When someone messes his stuff to death ... that's his personal problem. But when you sell things like this. Everyone can just do it. every noob can overvolt his stuff easely with this device. Without even knowing he can kill his stuff. And when he does realise ... I hope that you have got a very big disclaimer on your product and that you are legaly covered.

cheers :toast:

I hope that my input was helpfull.

some graphs:

perfect digital blockcurve:
http://www.xtreme-oc.be/articles/volt-oc/images/1.gif

more real (analog curve):
http://www.xtreme-oc.be/articles/volt-oc/images/2.gif

The device wich needs the bits will always compare it to Vref, to determine wether it's a logic high (1) or a logic low (0).

When the speed is increased, and the capacitors can't reach a value within the tolerance of Vref in time:
http://www.xtreme-oc.be/articles/volt-oc/images/3.gif

When you increase the voltage:
http://www.xtreme-oc.be/articles/volt-oc/images/4.gif

When this is compared with Vref:
http://www.xtreme-oc.be/articles/volt-oc/images/5.gif
06-19-2004, 11:22 AM
trueplaya4ever8

good idea :D, dont want a bunch of ppl complaining to you guys that your product killed my stuff.
06-19-2004, 03:25 PM
Tony

ns_ripper

First off no insult taken but im not an A64 tester, i have been intel for 2 yrs.This doesn't mean i can't test it, it just means i am a little behind to what i know about 875 and 865 boards etc.

I know exactly how the voltages effect ram and i know exactly how it works..the ONLY thing im personnally unsure about is VDD or VDDQ passing back to the memory controller on A64.

Some graphics cards work in a VERY similar way to the booster and have seperate plugable PSU on them supplying the Vdimm and vcore to the card.

With regards to Intel boards with seperate vreg for vdimm etc I see no issue.With the booster set at 2.5v and the board set at 2.8V the booster will report the bios set vdimm and apply nothing. If the booster is set at 2.8 and the board 2.5 the board will have 2.8Vdimm at the mosfets etc as if the bios has set 2.8V.

I see this no different to guys altering the 3.3vline to 3.7V with onboard vdimm mods and pushing 3.6vdimm thru the onboard mosfets etc, infact using a booster will actually create less stress on the board as the mosfets will not be delivering the current ...the booster will.

Regarding Intel boards we have actually pushed EB to the point it fried the IC's with absolutely NO DAMAGE to the motherboard what so ever....so we do test quite thoroughly ;)
Now...i do recognise A64 may be an issue.The modules are not on sale yet so we have time to tweak if needed.

So to conclude...have you posted this same info to Abit, MSI, Epox etc and all the others that allow vdimm of 3.2v or higher ???
06-19-2004, 04:11 PM
zabomb4163

Quote:

Originally posted by bigtoe

Regarding Intel boards we have actually pushed EB to the point it fried the IC's with absolutely NO DAMAGE to the motherboard what so ever....so we do test quite thoroughly ;)

at what point do the IC's die?
06-19-2004, 04:33 PM
Tony

You don't want to know....High thats all I say.
06-19-2004, 04:56 PM
squeakygeek

Quote:

Originally posted by bigtoe
You don't want to know....High thats all I say.

If he didn't want to know, he wouldn't have asked.
06-19-2004, 04:58 PM
Tedinde

Sign me up for the preorder
06-19-2004, 04:58 PM
charlie

The A8V certainly seems like an excellent candidate for the thingamajigger....
Seems that the Vddr circuit is incapable of 3.15V+
Strange too, when set in BIOS to 2.6V, the VR won't change the voltage at the Vddr readpoint...but when set to 2.8v, the read point will respond up to 3.15v before windows locks up... something odd here.

C

Show 100 post(s) from this thread on one page