OPB's perspecive about The "real truth" of 5BF IC

[Onepagebook]

hello everyone; I have been working on how to set the Micron 5bf IC;
basically, I just figure out someone said is absolutely NOT ture!
The reason I say this is because I have 100 sets of team group 2gb cronus
micron 5bf ic here. I have been testing about 30 sets, 19 of them all passed not only 250mhz 3-3-3-8 but DDR540 for quite nice setting; if you are talking about 250+, then most of them passed 265 32mb spi on 2.7v.
Maybe you guys doubt why previously some of friends bought TG's 5bf and got bad results;but forgive me say this.----> You guys have no IDEA or just think5bf is kinda like 5Bd.

Like I said, and I don't wanna argue if all my rams are all cherry pick. From now on, I will do the serious check before all this rams go to retail. Let's go check the truth and my results on these two days (over 25 sets)
here are the points you guys gotta be aware of:

1.If you are Opteron user, please DO NOT use divider on running 5bf IC. Why, because OPty is well-known
lack of divider of good results.

2. Keep the settings I post here, you don't have to try any loosing timings, 5bf is a bit weird IC;
it needs a special treaments, I tried many different settings, nothing help on 5bf, you will see.
for ex: the MAL/rp on the auto setting under bios would be 8/6 or even 9/7 after you get into windows
but I tell you what, 8/6 or 9/7 won't let me pass instead of 7/5, weird huh?:D

3.Don't try any weird TREF, only target on 3:
3120: 200mhz, 15.6us----->set this under bios for mainly using purpose.
3684: 100, 1.95----> if you have problem for 3120, then try this one
200, 3.9---> if you really no luck on both above, try this one, most likely you won't until you try over 275 fsb.

4.Watch your PSU rails, especially on 3.3v; don't tell me you don't believe this,
PSU is the only concern when I bench.
** Be careful for MBM5's 5v reading, which is not correct for reading some big power unit 5v rail but 5vsb;
as I know, my zippy 950w and PC P&C 1kw are like that.
let's see some examples: these result are NOT special pick, they are randomly select from the stock I promise.

ddr530 16m testing:



how about ddr540 32m pass



well just read this and be patient for playing this, you would be fine.
get used into it. you will understand its patent such as some fixed settings that I figured out.
I hope this will help those who used to be not happy with it.

update, Running lots of Opteron bench, all forgot to grab screen shots,
here you go, nothing different settings from the info above.



3d05 single 7900gtx also no problem on exactly settings:



And I seriously apologize about the bad experience you guys had before.

OPB

[Cpt Twitchy]

Nice little guide here OPB and thanks for putting in soo much time in order to give the knowledge to everyone whom has been struggling.

Tim

[ozzimark]

i just got a random kit from TankGuys to play with for a few days. 270mhz is as easy as baking cookies, even with 4x1gb (actually, 4x1gb clocks a touch better, but i'll get into that when it's time )


some things to watch out for though, sometimes tighter timings are better. it seems to me that as the voltage goes up, trrd and twtr MUST go down to maintain stability

also, if you choose to use yellow slots.. for me at least, drive strength must be raised to maintain stability

[njkid32]

Very nice Kevin this will help alot of people here. Good Job!!

[tsuehpsyde]

OPB, since you have over 100 sets, feel free to send me a set, preferrably a good clocking one.

[Onepagebook]

Thank you guys, one thing I 'd like to clearify is that:
Yes, the truth is that 5bf might be tough for tuning; however, it doesn't mean
ALL the 5bf will need to be sentenced to death.:D Especially if you place it under the cpu with good memory controll and decent PSU

@tsuehpsyde
basically, those pairs can do over 265+, all will be sent to tankguy for sure.
I will NOT allow any sinigle pair fail on that. BUT, just don't blame on the ram if you can't make it,
narrow the prob, find the reason failed could be occured on your components as well.

[ozzimark]

Thank you guys, one thing I 'd like to clearifyis that:
Yes, the truth is that 5bf might be tough for tuning; however, it doesn't mean
ALL the 5bf will need to be sentenced to death.:D Especially if you place it under the cpu with good memory controll and decent PSU

it is indeed exceptionally picky, though i think you just made it a lot easier to tweak for many people
to me, the key with 5b F is to try everything

i'm gonna go off on a bit of a side note here, and point out that i've seen nothing good from asus with F dies. just something to keep in mind

[njkid32]

ozzimark are you running 2T with 4 sticks?

OPB I dont think you will have any probs with that psu you have

[ozzimark]

yeah.

i'm having a problem getting a screenshot though, i think 270mhz is a bit over my memory controller's limit with 4 sticks in... made it to loop 23 in 32m, and it locked. was priming for about 20min, started opening cpu-z screens.. and it locked. oh well. guess i'll do lower mhz with less voltage and tighter timings to make it equally cool
a weird thing though.. i can easily get prime to not fail in blend mode at 280mhz. i'll have to sorta demonstrate later

[[cTx]Philosophy]

As always opb comes thru where he should, nice work and effort put forth to bring some light back on the Cronus.

expert is supposedly able to do 1t with 4 sticks in correct?
Have you tried any of this OZ, id like to see some results at any speed just to know that it actually does work and its not all some smoke up da but

[ozzimark]

it can, though at really low mhz.. like 201mhz
also, what is it that you want to see, some results with 4x1gb?

[[cTx]Philosophy]

yah 4X1g 1T is all I wanna see
I heard expert can do it but I thought this was due to a mem controller issue with amd, now I hear its mobo based, and expert addresses it.
Makes me sad I sold all my bh-5

[don_vercetti]

drive strength is the thing that controls 1t timing, i believe. With high drive strength one can do 1t.

[Supertim0r]

nice guide
I used the exact same tweak with my new ballistix and it's stable windows memtest @ 295fsb 3-3-3-8 2.7v

[el rolio]

it can, though at really low mhz.. like 201mhz
also, what is it that you want to see, some results with 4x1gb?

i do! me sir!

[kovboi]

hey OPB, i appreciate the time & effort you've shown here. but i think there are two varieties of concern with these modules. this guide addresses the first variety of concerns, which was that a lot of people were unable to get them stable in the 500-530 range. hopefully this will help, but i notice the modules you're testing in the screenies are week 10, whereas a lot of people were having problems with week 06. did your results hold for that batch, as well?

the second concern is that 540 isn't nearly as spectacular as the review modules sent out at the end of last year which were hitting 560-600. my modules will go up to 550 (maybe), but now i'm running them on a divider at 520 because they won't go 1:1 with my cpu (and for the record, they are very stable despite the use of a divider). now obviously 520 is not very exceptional, and i could have bought much cheaper memory for comparable results.

also, ozzimark is right about asus boards and this memory. you need to make clear to people (and ask your distributors to do the same) that they will need to get a different motherboard if they don't want to run 2T.

[Onepagebook]

hey OPB, i appreciate the time & effort you've shown here. but i think there are two varieties of concern with these modules. this guide addresses the first variety of concerns, which was that a lot of people were unable to get them stable in the 500-530 range. hopefully this will help, but i notice the modules you're testing in the screenies are week 10, whereas a lot of people were having problems with week 06. did your results hold for that batch, as well?

the second concern is that 540 isn't nearly as spectacular as the review modules sent out at the end of last year which were hitting 560-600. my modules will go up to 550 (maybe), but now i'm running them on a divider at 520 because they won't go 1:1 with my cpu (and for the record, they are very stable despite the use of a divider). now obviously 520 is not very exceptional, and i could have bought much cheaper memory for comparable results.

also, ozzimark is right about asus boards and this memory. you need to make clear to people (and ask your distributors to do the same) that they will need to get a different motherboard if they don't want to run 2T.

Good points:

Answer are:
1.that date code you see merely stamped as Team Group package dated not Micron manufactured date. I can just tell in public all Team Group's IC date code is 0538 no more.


2.I just got Asus An832SLI DX, I think this must be the one some other peeps wanna try or already tried; I can honest tell you, I can get it work @265~270 1t (the same pair I passed spi32m on DFI both ultra-D and Venus)BUT only doing nothing ; all you need to do is change the memory freq to limit as well as first part of memory parameters, that is, the primary timings.
others are all set as auto.
After you boot into windows, use a64tweaker to change the timings and test them one by one. Well, I don't sound Asus is not the board to run anything , but since Asus's first K7 board until know, the only goodness of AMD was Socket 940 Via board--->k8v delux? Asus is still only good for intel.

so ya, here are the picI can show you, all the micron 5bf teamgroup 2gb cronus in US are here
I test them all by from 1m, 8m, 16m, 32m; if it pass, then I sign for that on the sticker. Simply I just don't want to disappoint all the friends who support me.

[ozzimark]

2.I just got Asus An832SLI DX, I think this must be the one some other peeps wanna try or already tried; I can honest tell you, I can get it work @265~270 1t (the same pair I passed spi32m on DFI both ultra-D and Venus)BUT only doing nothing ; all you need to do is change the memory freq to limit as well as first part of memory parameters, that is, the primary timings.
others are all set as auto.
After you boot into windows, use a64tweaker to change the timings and test them one by one. Well, I don't sound Asus is not the board to run anything , but since Asus's first K7 board until know, the only goodness of AMD was Socket 940 Via board--->k8v delux? Asus is still only good for intel.

so ya, here are the picI can show you, all the micron 5bf teamgroup 2gb cronus in US are here
I test them all by from 1m, 8m, 16m, 32m; if it pass, then I sign for that on the sticker. Simply I just don't want to disappoint all the friends who support me.

then i do wonder why almost everyone i've seen with an asus board and 5b F sticks was having serious problems.

however, that must be exceptionally time consuming... testing all the sets in sPI.. i'm assuming at 250mhz? that shows some serious dedication to quality :toast

ps- if you don't mind, i'm gonna start up a thread with my own results with the 0609M dated sticks i'm testing now.. should have some useful info by tonight

[Onepagebook]

then i do wonder why almost everyone i've seen with an asus board and 5b F sticks was having serious problems.

however, that must be exceptionally time consuming... testing all the sets in sPI.. i'm assuming at 250mhz? that shows some serious dedication to quality :toast

ps- if you don't mind, i'm gonna start up a thread with my own results with the 0609M dated sticks i'm testing now.. should have some useful info by tonight

ya, from 250mhz then 260mhz if both all pass, then directly jump up to 270mhz; if failed go down to 265mhz---> that's my first 5 pairs test when I struggle consecutively about 24 hours non-sleep. After I found out the patern, I just simply plug pair in, boot up and test it. So if someone got those passed pair and told me that is not even boot at 250 on dfi, I will invite him bring that pair to Columbus, Ohio, and I can test in front of him.
sure I don't mind you start that.At least you are the user who really get in study of this; shame on the one who just BS and never tried any effort.

[ozzimark]

thanks kev
though these sticks seem weird, even by 5b F standards. i'm gonna do a bit of work at the lower mhz and tighter timings for those guys who don't wanna run high mhz for whatever reason

and see something wrong with this?

[Onepagebook]

example on Opteron, 1:1, ddr540 rocks stable.
Every single setting is just like what I said. Nothing new
the vdimm showing on the mbm5 actually is what I set 2.7 under Ultra-D BIOS,
Now in the middle of process on super pi 32m running!






************************************************** **********************************

Updated: see anything different from the setting I provided? Nope!!

[Supertim0r]

are they similar to Ballistix ?

[bigval]

Very nice work man, I wish I had 100 sets laying around lol.

[Onepagebook]

are they similar to Ballistix ?

basically same thing. but not 5B-D

[OmegaMerc]

Although, this is nice to see, on my book its a dollar late and a nickle short. Why? Because after 3 emails to Team rma, I never got a tracking number, because my IDE 0/1 is dead, and because I spent 270+ 15$ shipping (first time around) + 15$ shipping to send dead sticks back to tank guys, plus another 15$ shipping for some new sticks. So in effect, 350$ (plus 160$ for a newmobo) for 270$ pair of sticks.


If anything, im happy I got to purchase through TG, they have excellent customer service and responsability, and gave me an RMA # right away. As far as im concerned I wouldn't do Team ever again.