LOL Hipro:D
I don't know if it's the launguage difference or if you trying, but sometimes you really make me laugh :toast:
Printable View
LOL Hipro:D
I don't know if it's the launguage difference or if you trying, but sometimes you really make me laugh :toast:
I'll get 2 x 512 Twinmos UTT so as to test them 24/7 @ 3.5Vdimm+ from a friend and then I could say.....! ;)Quote:
Originally Posted by bachus_anonym
:D :DQuote:
Originally Posted by topboy
This is GOOD.......Laugh is helth..... :D :toast:Quote:
Originally Posted by G H Z
P.S. I'm not a gennie in English though...... :)
i think most of that dead ram issue happen when the system power ON sequence begin ... or maybe power OFF sequenceQuote:
Originally Posted by hipro5
make sure that 24/7 also include
- reboot
- shutdown & power up
each day... :hehe:
Heh, hipro5, what can I say :).
Your RAM has probably been toughened through long-term abuse... ;) http://www.xtremesystems.org/forums/...ies/hehehm.gif
I'm thinking along the same lines as bachus - BH-5 and UTT are two different beasties. UTT is way sketchier due to a complete lack of a standard for operation and integrity right after fabrication.
So is this the final answer from DFI? And what are we all who bought this board for the extra VDIMM supposed to do if we don't have adjustable PSU?
Quote:
Originally Posted by Badong
Also can somebody please direct me to a thread that talks about of danger that Badong is talking about, of using +5V line with less than 3.2V? Because this is the first time I am hearing about it and ever since my UTT sticks died (2 months ago) until about 2 weeks ago, I have been running my TCCD RAM from +5V line at only 2.7V to 2.8V VDIMM.Quote:
Originally Posted by Badong
Thanks.
@hovo73
Try to interpret what Oscar Wu and AG said about "using 4v source is not advised"...
There are two voltage regulators, 1st that supplies voltage to 2nd, which supplies it to memory sticks. If you only need 3.2vdimm then switching to "5v supply" will result in 4.1v flowing thru vdimm mosfets. This makes them run VERY HOT. Why do that and POTENTIALLY shorten it's life (which MAY result in failure after some time) for only 3.2v? Stupid, to say the least.
So, for 3.2v just stick with "3.3v supply" and get adjustable PSU or mod rails if you don't have any means to adjust them. For stable 3.2vdimm you would need 3.4v on your PSU's 3.3v rail.
3.4v vs 4.1v is a LOT of a difference and will probably make you feel better.
I do use "5v supply", because I need 3.7-3.9v to play with my BH-5. Not many PSUs can deliver 4.0-4.1v on 3.3v rail without OverVoltageProtection kickin in and shutting PSU down. So I decided to take the risk (OCing is a risk already) and run 5v. Not that I should not use it, but I can use it because I know, that it MAY kill the board and I 100% realise it.
If you are not willing to take ANY risk related to high-voltage operation, waaaay beyond specs, then just don't use DFI NF4's ability to deliver 4v to your memory.
ill try my best to find it for you, but trust us when we say you should not be using less then 3.3 on +5. i once heard squealing from my dimm caps, because i kept it on load for too long at 3.1-3.2. you can imagine if one were to keep it like that for a few weeks to months.Quote:
Originally Posted by hovo73
ill try my best to find it for you, but trust us when we say you should not be using less then 3.3v while on +5 feed. i once heard squealing from my dimm caps, because i kept it on load for too long at 3.1-3.2 while on +5. you can imagine if one were to keep it like that for a few weeks to months.Quote:
Originally Posted by hovo73
the caps/resistors have to work that much harder to convert 3.2 - 2.5 (lower the worst it is) from +5. and although high life qual. caps were used in that area, there is just too much for it to bear if it is really low.
Darn! :(.
I have switched to 3.3V line recently, but I have been doing just that using only 2.7V with +5V line. I have been using it for about a month and I kept running stress tests (memtest, prime95) almost every day or every other day for 8 + hours (having problems getting my Patriot XBLK to go much higher then 260 MHz with 2.5-3-3-7-1t). Do you think I could have already damaged something?
Where can I find some more info on how to mod my PSU?Quote:
Originally Posted by bachus_anonym
EDIT: Is this it: http://xtremesystems.org/forums/showthread.php?t=37574 ? Are there any other ones?
PS. I would like to add that I still think DFI has responsibility to all its customers to get the 4V thing working.
as long as your on +3 now, while using 2.7vdimm, your ok.
the 4v thing works. you must use it with care though.
@hovo73
There are guides in Xtreme Mods section on how to mod your rails. It requires soldering and being confident in your skills, though ;)
The "4v thing" as you call it is working and doing just great :) If you're not familiar, DFI has already used "5v supply" in the past, with it's NF2 mainboards (Infinity and Lanparty) but vdimm was capped in BIOS at somewhere around 3v (I can't remember exactly, even though I had this board and volt-modded the hell out of it :D ). They run just fine too, even with 3.8vdimm after vmods ;)Quote:
PS. I would like to add that I still think DFI has responsibility to all its customers to get the 4V thing working.
Well, I have 2 sets of brand new BH-5/UTT sticks (2x512 OCZ EL Gold and 2x512 TwinMOS TMSP) it would be a shame if I won't be able to use them with this board to their full potential. :)
Well, the board can make it happen but you need to realise, agree and take full responsibility in case your memory dies from over-volting beyond what it's warranted at. I'm not saying that it will, but it MIGHT.Quote:
Originally Posted by hovo73
I signed that deal myself 2 1/2 yrs ago when I first put 3.5v to my 2x512MB Mushkin LvL II PC3500. But it was true BH-5. As far as I know, it's new owner is still enjoying it until now ;)
I think that UTT-CH might be a little bit more fragile, so regardles if I'm right or wrong, be careful :toast:
EDIT: I just looked at mems you got... Well, they seem to be Winbond UTT-BH5, so they might be stronger than UTT-CH5 ;)
as long as it is BH die (UTT BH/classic BH), it should tolerate voltage and last as classic BH has proven. classic is based on .17 fab process, so the UTT BH die should also follow the same process.
i cant comment on UTT CH (.15), but from the looks of things, this is the ram that is not tolerating high voltages for long periods, in certain situations (possibly; not enough data right now to determine that).
edit: when you go over the warrantied voltage of the stick, you automatically sign a 'waiver' so to speak, hehe. be careful nonetheless, and take some precautions, and youll be fine:)
If like what we'd speculated that these UTT/CH base memory chips are weak and can't handle higher voltage load like the fabs BH5, then what difference does it make to use the jumper trick with super high 3.3v rail or with booster? if you're putting the same amount of vdimm through these UTT then technically and eventually they'll still get fried regardless what method you use to supply the load(assuming you're using a good PSU). :stick: UTT reminds me of those scratching lotto ticket that you get from 7-11 :D :D I suppose we will find out in couple months and see what is the memory failure rate with them running under the 3.3v rail instead of 5v, but for now...anyone care to be a test mule and feed 3.5~3.6v through your UTT 24/7 from the 3.3v rail and see how long your sticks gonna last? :D :D :D
is mushkin xp4000 utt-bh or utt-ch?
@ben805
I'm sure, that If you're feeding your sticks 3.6v then if you do it from "3.3v supply" or "5v supply" makes zero difference to them.
What the whole talk of "3.3v trick" is about is, that it forces DFI users to lower strain on vdimm mosfets. I bet than the percentage of folks raising their 3.3v rail to 4.1v as opposed to 4v jumper (plus, the only PSU I know that does not shut down beyond 3.9v is my Fortron 530W :D) will be EXTREMELY marginal, if NONE. For those (like myself) that really need it, 4v jumper was supplied. But If you only need 3.2-3.4v then stay on "3.3v supply" and get mod your PSU. It's probably not absolutely required for the mobo to survive long operation, but everyone will tell you that it's a wise move to make.
from all signs on the earth and sky :D, it's UTT-CH5.Quote:
Originally Posted by stealth17
:toast:
okay thanks...so does that mean it will die when i feed it 3.5v 24/7?Quote:
Originally Posted by bachus_anonym
its warrantied to 3.5v so ill just rma it if it does...
i havent seen any xp4000 die yet :confused:
if they warranty it to 3.5v, then u are insured up until that point. will they die? all depends on the specific IC's being used.Quote:
Originally Posted by stealth17
We do not know :) Not many folks have been playing with XP4000 so far, it's very fresh product.Quote:
Originally Posted by stealth17
On the side note... I've been scratching my head on something... Why Mushkin waited sooo long before releasing their Winbond UTT sticks and let OCZ and others test the waters first??? Mushkin has always had excellent relationship with Winbond :fact:
OK I want to pose a question. If the boards are really fine, if there is really nothig wrong with them and if there really are no voltage spikes, why does DFI suggest such elaborate procedures for things like exchanging RAM (unplugging PSU, draining power etc) and flashing BIOS (removing CMOS battery)? I have used a good number of different motherboards, and NONE ever required anything even remotely close to these.
Quote:
Originally Posted by hovo73
because supposely NF4 chipset is really sensitive to that kind of stuff. However personally I've never down any of those and I am still fine (knock on wood).
check out Charlie's sneak peek thread, atleast 3 people had bad or dead sticks....including mine and I was on 5v rail when it died. Only had mine for 3 days but Mushkin took care of the RMA though, and so far my replacement have been working great, and luckily I can run 250 1.5-2-2-2 @3.1v now...still hesitating to use 5v on these UTT again.Quote:
Originally Posted by stealth17