Printable View
i did some more research on the manufacturing process of the diferent chips, and this is what i found:
AH-6 0.180µ
BH-6 0.166µ
CH-6 (doesnt exist?)
AH-5 (doesnt exist?)
BH-5 0.150µ
CH-5 0.130µ
this explains why bh-5 and bh-6 perform almost the same, they are made in almost the same manufacturing process.
CH6 does exists.... found in some xms3200 :)
yeah, ive also found some pics of it, but its almost impossible to find it somehwere :(
i wish i had some to add it to my test.
I know that Corsair XMS pc3200 v. 1.2 uses CH6 :)
yeah but its expensive... and i have like 20€ left for this month and still need to buy some food ^^
hey, u guys need to separate DC from SC results
I think i´ve got access to some.. I´ll test it for ya.. What do you need tested?Quote:
Originally posted by saaya
yeah, ive also found some pics of it, but its almost impossible to find it somehwere :(
i wish i had some to add it to my test.
Well, we all know that BH5 and BH6 are kiss ass chips, but... Which manufacturer make the best PCB?
My vote goes for Kingston HyperX...
HI
here is mine
2432D W942508BH-5
stable speed, DC
"2432D" 2-2-2-5 3.35v 255mhz (Corsair 3500 256@875)
3D benchmark speed
SC 270 2-2-2-6 3.60v I865
DC 260 2-2-2-6 3.60v I865
SuperPI 1M benchmark speed
SC 276 2-2-2-5 3.78v I875
DC 270 2-2-2-5 3.70v I875
I've got some (XMS3200 v1.2), but haven't been able to see its full potential yet, still need to vmod my nf7-s.Quote:
Originally posted by saaya
yeah, ive also found some pics of it, but its almost impossible to find it somehwere :(
i wish i had some to add it to my test.
It will do 2-3-2(rtc) 231@2.9v
Kingston HyperX 3200 stepping 316we 256mb
2-2-2-5 250MHz SuperPi stable w/ 3.2V, 257MHz SuperPi stable w/3.31V
TwinMos 3200 stepping 249we 256mb
2,5-2-2-5 256MHz SuperPi stable w/ 3,2V
TwinMos 3200 stepping 249wd 256mb
2,5-2-2-5 256MHz SuperPi stable w/ 3,2V
Tested both TwinMos sticks together, so one of both might do better than that.
All tested in my Shuttle AN50R.
Interesting thread Saaya. ;)
corsair xms pc3500 2x256 sticks
steppings are 249we and 250we both do 250mhz 2-2-2-5 @3.45 volts
I think I've got a new goal here. It says: Find some AH-6 RAM and give it 4V VDimm. If BH-5 can take upto 3,6V without damage, then the AH-6 should easily stand 4V.Quote:
Originally posted by saaya
i did some more research on the manufacturing process of the diferent chips, and this is what i found:
AH-6 0.180µ
BH-6 0.166µ
CH-6 (doesnt exist?)
AH-5 (doesnt exist?)
BH-5 0.150µ
CH-5 0.130µ
this explains why bh-5 and bh-6 perform almost the same, they are made in almost the same manufacturing process.
It would really be interesting to see what they would do with such a high voltage.:rolleyes:
BTW, I think I finally found me some nice dual kit of 2x 512MB Twinmos PC3200 RAM with rebadged BH-5 Chips.
Too sad that you only want the scores for original labeled modules. :D
No... It's possible to post the results too IMO. saaya ist kein Unmensch ;)Quote:
Originally posted by celemine1Gig
BTW, I think I finally found me some nice dual kit of 2x 512MB Twinmos PC3200 RAM with rebadged BH-5 Chips.
Too sad that you only want the scores for original labeled modules. :D
If you are sure those are winbond chips... just show us some results :D
I have a unique stick of mushkin pc3500 level one
It has bh5 chips ;)
Its a 512 stick and on the pcb it says:
18-15180A-NL
REV A
and the bh5 stepping is 320WF on one side, and 319WF on the other
Is this weird to have two different steppings?
Forgive me if this is normal, but I didnt see any in this thread like this.
Well right now I'm limited SOMEHOW at 225fsb with my setup (no matter the timings I cant get higher so I'm pretty sure it aint the ram) with 2-2-2-11 vdimm=3.3rail
I did post about my level one being bh5 at the mushkin forum and got this reply:
Yes, what else could he say, but I duno if its true. I couldnt run 200fsb 2-2-2 with 2.6v before, but I couldnt test anything under 200fsb with 2.6v and I didnt have any higher vdimm to try. I did get into windows though, but it would crash after heavy load for awhile. Mushkin alows like 2.75v anyways.Quote:
when we do large batch tests some modules do not pass at the timings common to that modules, due to variable chip potential. instead of tossing these modules out we test them at the next step down and see if they'll work as that kind of module. it sounds like you got a modules meant for a bh-5 level 2 but that it could not pass as such and was chosen a level 1 instead. this module might be able to run past level 1 spec but it probably won't reach level 2. it does tell you that you should always test your modules out to see what they are capable of. who knows, maybe you've some chips that can scream a little faster than normal under those heatsinks
My thoughts where that this was just an old stick prior to using ch5 in production.
Yea, well anything fishy about the different steppings on each side? :stick:
hi i got this stepping:
340WF
W942508BH-5
it does 210 2-2-2-11 dc, tested on a vdimm-modded nf7-s 2.0 and a lanparty revb.
clocks the same at 2.9v and at 3.3v, fan cooled.
this is the the worst pair of bh5 i've ever seen, none of these sticks alone will do no more than a few mhz above 210.
it's adata (says 'selectra' on a sticker too) btw.
Hi I have corsair xms3500 512x2 the number are
24120 W942508BH-5 the max fsb are 244 2-8-4-2 @ 3.20volts on IC7-G , the number 24120 it mean this are good chips ??? :confused:
PD: sorry for my spelling
thats ch5 thenQuote:
Originally Posted by CTKP
thx :toast: i think i will move it to the xtreme overclocking section so it gets some more attention. i added your results as 260mhz stable as all the other ones a 3dmark stable.Quote:
Originally Posted by Zeus
can you try with less vdimm then to see how much they need for 225mhz?Quote:
Originally Posted by bleeder
and having different chips on one stick is nothing special :) ive seen up to 8 different stepping chips on one single sided stick with 8 memory chips ;) hehe
2-3-4-2 timings :confused: what about 2-2-2-x?Quote:
Originally Posted by David302
and make sure its not 2412D instead of 24120
no, if the dx oc is lower than the sc oc its the chipset/pmw/psu limiting, not the memory.Quote:
Originally Posted by Blaster
BH5:
Stick 1: "315we" SingleSided KHX3000
Stick 2: "0251PQ1" d328d(v?)-50 BothSided KHX3000 (remarked)
DC async: 2.0-2-2-5 3.64v 264MHz (Kingston KHX3000 2*256mb@NF2) Memtest 1h stable.
Same sticks 2.0-2-2-5 3.7V 274MHz SP1M stable.
DC:
Adata 256 stick:
338WF
227fsb 2-2-2-5 2.9 @ nf7-s poor ocer :(
is on your sticker A.00 ?Quote:
Originally Posted by JenkaRiPS