FWIW-Originally Posted by Biggles
I ordered a 1gb kit of TMll from Newegg about 2 weeks ago, and both sticks were AA4T, on my kit the dates on the chips were the same also...
They are NOT on BP pcb ( KY6663 or something).
I haven't done any clocking with them yet, beating up on my XMS3500 and GSkill LE nowadays.
Zen2 Has brought AMD back!
If you're looking for the CH-5 and BH-5 sticks that link isn't it. These are...
http://www.komplett.co.uk/k/ki.asp?sku=308281&cks=PLS
http://www.komplett.co.uk/k/ki.asp?sku=308282&cks=PLC
i've a question for Alonso:
i'm experiencing some weird problems running 512mb aa4t stick on Nforce2 mobos...
this is what's is written on the label of the stik i've now:
i'm running it on an NF7-s rev2... but there's no way to get it working with command per clock 1T
i'm using only that stick, no dual channel, nothing!
with cpc off they run great:
but i loss a lot of performance...
i'm not the only one with that problem... also on dfi lanparty ultra seems that some sticks doesn't works...
http://www.dfi-street.com/forum/showthread.php?t=6708
on this mobo i've tested only 1*512 of vdata without any problem and i've run 2*256 twinmos (powerchip) for about 1year a 233mhz 2.5-3-3-3 all the modules with cpc on!
now this stick won't work with command per clock 1t...
the only difference i've noticed is that the spd appear to be programmed in a different way:
it doesn't read the manufacture date
could be that the source of the problems? all the others stick i've used where showing correctly the manufacture dateCode:DIMM #1 General Memory type DDR-SDRAM Manufacturer (ID) (0000000000000000) Size 512 MBytes Max bandwidth PC3200 (200 MHz) Part number M2G9JA8AMK9F081CAD Manufacturing date Week 00/Year 00 Attributes Number of banks 2 Data width 64 bits Correction None Registered no Buffered no Timings table Frequency (MHz) 200 CAS# 2.5 RAS# to CAS# delay 3 RAS# Precharge 3 TRAS# 8
PS: this stick does 260 2225 a 3.4v on dfi nf4 ultra-d
EvolutionCrazy,
I'm in a similar situation to yours (NF7-s) but have the PC3500 1A4T... see my post and some of its replies in this thread and page...
http://www.xtremesystems.org/forums/...page=106&pp=25
...seems the recommendation is to have CPC Off (2T) although I haven't got round to trying it yet since it would mean a flashing of the BIOS (no Abit bios that I know of has a simple CPC toggle setting so a complete flash with specific 1T and 2T versions are required).
thanks for the reply [timko]
uwackme in that post said that he was running 2*512 on nf7 with 1t without any problem...
however i've tried 3different bioses with this stick:
D26 MantaRays XT => memtest freeze on test #5 (all the others OK)
D27_TaiPan_0.3_ED_Bpl_3.19_28.12.04 (CPC ON) => memtest ok... but when i try to load windows i get blue screen :/
D27_TaiPan_0.3_ED_Bpl_3.19C_27.12.04 (CPC OFF) => memtest ok, windows ok...
run prime95 for 8hours at 230*8.5 2225 a 2.9v vdimm without any error...
however i've some friends that run 2*512 ch5 (the olds ch5... the ones produced in 2003) without any problem...
i think the problem is between
spd <-> bioses
i need to know how uwackme's sticks where spd programmed... with/without the manufacture date...
Timko, I've tested 5G's of BH5 and 1G BH6 and none of it will run for shieeet in MY NF7-S....with Vdimm Vtt Vdd mod's!!! Stock NF7 is even worse. But the UTT does 245Mhz 1T 8,2,2,2.0 on the SAME board.
I have to burn the CPCoff bios back into the board when I want to switch back to CH5/BH5 sticks. Using D26 MantaXT cpcon and cpcoff versions.
For you....
First, it NEEDS the CPCoff bios. For some reason BH5 almost never ever runs 1T on the NForce2 chipset. Second its gonna need voltage. CPCoff it will do 200Mhz+ 11,2,2,2.0 2T @ 2.9Vbios. But to shine on NForce2 its gonna need TheVolt and it will never shine 1T on NForce2.
On an A64 board the results will be much better, 1T, lower Vdimm, etc.
These UTT/VX/SP sticks are a different breed and shine 1T high FSB.... but again NOT on a stock board, you gotta get to 3.2+Vdimm to make them work on a NForce2.
So confirm 200Mhz 5,2,2,2.0 with CPCoff 2.9V first, if it still screws up consider RMA.
Yesterday i got 2x SP3500 512's from komplett.nl (not a dual pack, cant find it in holland)
Dual channel started like disaster so i'm only talking about single channel
2.5-3-3-8 217fsb @ 2,9v memtest 5 errors...
2-2-2-5 200fsb t/m 3,3v memtest 5 errors...
2.5-4-4-8 217fsb @ 2,9v works.... but this aint within specs.
This was all @ 1t on a hellfire rev3 (i think) bios.
Just started reading about the cpc off (2t) mode... so the mem could work properly but this just aint righti want 1t
My KHX3000/512 and corsair 256 pc3500 (all bh5) all work fine on my board with the same settings.
My hopes where so high... and now this
Will keep on testing when i have time and hopefully i can test them @ a friends A64 system. First i need some fresh air...
Ps: is this the right thread to post in?
Last edited by Vandread; 04-08-2005 at 11:52 PM.
Dutch Specs
The "budget" rig (had to go eco to buy a ridiculas expensive case so the rest is 2nd hand)
E6600 // ebayd 2x1gb pc4200 D9DQW halfway on the board tricked // Asus P5B dlx // EVGA 8800GT 512MB SC// Silverstone Temjin TJ07S // Tagan TG480 // WD740GD // 2x Samsung 500Gb T166
in my case with the D27_TaiPan_0.3_ED_Bpl_3.19_28.12.04 (1T bios) i can run memtest a 230 2225...
but i can't load windows...
so i hope some bios modder like tic tac or merlin read these post and help us having twinmos working with 1T at least at their default specs
@Vandread does your sticks come with manufacture date or not?
i don't think this is a normal behaviur:
The modules is rated to run at 200MHz 2.5-3-3-8 1Tin my case with the D27_TaiPan_0.3_ED_Bpl_3.19_28.12.04 (1T bios) i can run memtest a 230 2225...
but i can't load windows...
so i hope some bios modder like tic tac or merlin read these post and help us having twinmos working with 1T at least at their default specs
@Vandread does your sticks come with manufacture date or not?
i don't think this is a normal behaviur:
but.. You are pushing at at 2-2-2-5 1T timing at 3.01V Vdimm
You need more Voltage to do so...
Like I said.... TheVolt.
I have my NF7-S mod'd to 3.25Vdimm (using 3.5V [3.3Vrail] to power circuit) and Vtt mod to give 1.625Vtt. Also Vdd is mod'd to the "HAPPY PLACE" for NForce2 which is 1.8-1.85Vdd and this will impact your results.
I use Tic'ys D26MantaXT cpc and non-cpc to "switch" 1T 2T modes. Easy flash with Flashmenu 1.20 so no clearcmos is required, makes it a snap.
BH5 simply dont do 1T in NForce2 (with RARE exceptions), regardless of Vdimm...If been as high as 3.6Vdimm and no go on 1T. The Twinmos AA4T stuff does do 1T no problem....provided it gets TheVolt.
My Infinity has Vdimm mod'd to 3.5Vdimm (and Vtt tracks) and the circuit runs off 5V rail. but I have 3.3Vrail still set to 3.5V because part of the NB chip uses 3.3Vrail as a supply point, as well as Vdd. On Infinity my Vdd is 1.9V as well....its "happy place".
You determine the "happy place" by indepth testing with memtest to find the true PEAK in FSB result stable. IE: once you are at max FSB, diddle Vdd and see what voltage lets you get even further. NF7's seem to like 1.8-1.85V while Infinity likes 1.9V.
But trying to use this stuff on a STOCK NF7-S with no Vmod's is just gonna frustrate the hell out of you. Try the simplest... Vdimm=3.3Vrail mod and see what happens.
Helps to have a PSU with 3.3V rail adjustment.... Fortron530 or OCZ Powerstream.
I have 5V rail at 5.5V too, as the X800XTPE really likes having higher 5V rail. And my 12V rail is 12.50V so at max load it dips no lower than 12.16V.
Its a complicated system, and if you dont get it all right, it aint gonna play well.
I've been through 6 TMII sets, and 2 SP sets all the AA4T stuff. 4 TMII's did 240-245Mhz 1T 8,2,2,2.0 @ 3.4V 1set did 245Mhz 2T and one set was bad. Of the 4 sticks of SP, 3 do 250Mhz 1T and the 4th was bad and is in RMA.
Of all the BH5/BH6 Ive had, none did 1T in NForce2 at high FSB, none.
GA-MA790GP-D4SH, 965BE, 3.8Ghz
Ultra120E, 4x2G Gskill 1066
2x 150G VRaptor raid, WD640AALS
8800GT 735/940/1685, LG-BluRay
Yeh it mentions Winbond and the BH-5 is not in stock so i thought this might also be bh-5?
there's no BH-5 or CH-5 anymore , it's all UTT.
They call the better UTT 'BH-5' but that's just marketing. You should read that as 'BH-5 alike UTT'
Here my results so far Twinmos Speed Premium 3500 2*512MB 3,3v
(I'm not finished yet) :
i know... that shoot is refered to 2225 2t 3v works..The modules is rated to run at 200MHz 2.5-3-3-8 1T
but.. You are pushing at at 2-2-2-5 1T timing at 3.01V Vdimm
You need more Voltage to do so...
but my problem is that i'm not able to get this stick running at its rated speed!
2.5-3-3-8 (timings by spd) @ 2.6v (minimun for my mobo) @ 200mhz
(multi 10x with a barton mobile + L12mod)
memtest #5 freeze (leterally crashes using mantarays xt, using merlin mod i get just tons of errors)... if i try to load windows i get BSOD...
ram@166mhz 2.5-3-3-8 1t pass memtest but random bsod.... 133@2.5-3-3-8 works like a charm... but 133 isn't like 200mhz!!!
they are rated to 200...
everything in spec!!! nothing out of spec
i don't know if my mobo really need 1.9v of vdd and 3.4v of vdimm to get this stick working at 200 2.5-3-3-8 1t like they are rated... but similar high voltages to run in spec aren't expected
Also trying to get them running at their rated speed?
200 2.5-3-3-8 1t?
i just want to see this stick passing memtest #5 with an official abit bios (cpc on) at stock speed (200 2.5-3-3-8) at stock voltage (2.6v)...
I can't believe it's impossibile
I'll try them with the ddr booster asap i'll "recompose" it (pot breaked...)...
@uwackme: does the sticks that you have tested on the nf7 correctly show the manufacturing date win cpu-z or they have a different spd chip and don't show anything like mine?
Last edited by EvolutionCrazy; 04-09-2005 at 06:44 AM.
i've tried to flash the D27 taipan 0.3 1T over 2T version without clearing the cmos...
this is the result:
there's no more optimal/agressive/turbo presets.. only expert setting...
but now:
220 2-3-2-8 1T @3v memtest #5 ok and windows load
yesterday i've tried that bios but i remeber that i wasn't able to load windows (@200mhz 2.5-3-3-8)... and i remember that i've all the optimal/spd/agressive... settings selectable!!!
now i'm writing here at 220 2328 1T!!! woowoow
i've just flashed this bios:
http://members.lycos.co.uk/bios/Merl...9_28.12.04.rar
over this one:
http://members.lycos.co.uk/bios/Merl...C_27.12.04.rar
using flashmenu without clearing the cmos...
what's happening?
Last edited by EvolutionCrazy; 04-09-2005 at 01:04 PM.
I'm confused.now i'm writing here at 220 2328 1T!!! woowoow
i've just flashed this bios:
http://members.lycos.co.uk/bios/Merl...C_27.12.04.rar
over this one:
http://members.lycos.co.uk/bios/Merl...9_28.12.04.rar
The first link ending in 27.12.04.rar has a file named "Nvidia Bpl 3.19 Cpc Off.bpl"
The second link ending in 28.12.04.rar has a file named "Nvidia Bpl 3.19 Cpc On.bpl"
If you have indeed flashed the first one over the second one then you are actually on 2T at the moment surely?
(CPC On = 1t and CPC Off = 2t)
sorry, i've inverted the 2 links
this is what i've done:
tried the ram with the d26 mantarays xt => memtest #5 freeze at any speed above 166mhz...
flashed (clearing the cmos) D27 taipan 0.3 cpc on => tons of errors in memtest #5, windows wont load at any speed...
flashed (clearing the cmos) D27 taipan 0.3 cpc off and let the system burning for a night... (torture test prime95 -> large fft -> 8hours @ 230 2225 1T @3v)
flashed (clearing the cmos) D27 taipan 0.3 cpc on to see if the night has changed something: nothing... lots of errors in memtest, windows wont load at any speed/timing
then today i've opened flashmenu, deselected the clear cmos then flashed D27 taipan 0.3 cpc on...
went to bios... and see that only expert settings where usable... putted the ram 220 2228 and tried memtest => badwidth higher then before (+50mb/s) so i think it's really cpc on
then i've tried also to flash the lastest D27 cpc on by tic tac (without clearing the cmos)... and also in that case everything works fine...
don't know how a similar thing is possible... but now it seems to works
i hope i can get the booster repaired soon... so i can see if the mobo can take this rams
Last edited by EvolutionCrazy; 04-09-2005 at 01:03 PM.
when can we buy this stuff on this side of the lake? Preferably a dealer in canada for all us canucks up north?
HI,Evolutioncrazy;
The ram you pictured, is it the one listed in the block of specs?
The one in the picture is AADT, but unless I am mistaken[probably so] the specs;
M2G9JA8AMK9F081CAD
indicate a CADT ram, yours should appear as a AA4 on the last letters.
Also, what are the numbers etc on the chips? I have CADT that are 50I
If I am incorrect, could someone explain, I am very willing to learn.
mike
Last edited by mike047; 04-10-2005 at 06:21 AM.
i've also noticed that thing... but dunno... (the label show AA4T
yes... this photo:
and this shoot:
are related to the same stick...
that's another thing that make me suspicious about "bad spd programming"...
i'll take out the stick and post a full quality scan in the next hour
take a look at this pic:
i've never seen CADT doing 230 2228
Last edited by EvolutionCrazy; 04-10-2005 at 06:22 AM.
HI, I just got the CADT chips yesterday and have not used them yet.
I was told on another forum that they were not good chips[50I]. I am going to vmod an Albatron KX18 PRO board this afternoon and then I'll see what they do.
There is a timing difference in the two CPU-Z shots???
mike
ROADTRIP For my DAD 09-11-07 and MA 10-28-11
comhar a neartú agus a leathnú
200mhz 2.5-3-3-8 by spd...
however mine are 50D
this is the chip (date MUST end with 4!):
front:
back:
i cannot explaint the wrong spd detection under windows (wrong type (CAD instead of AA4) and no manufacturing date...)
does the problem is the spd?
i hope someone from twinmos can tell us
Last edited by EvolutionCrazy; 04-10-2005 at 06:46 AM.
HI,
I have 3 sets;
TMSP400;AA4T 44D TWINMOS chip
TMII400;CADT, 50I TWINMOS chip
These two have the same spd chip as yours.
The other set is;
TM333; AADT, 60D MTEC chip
This one has the spd chip[it is also smaller] in the lower right, below where yours is.
These are only 333 but they both will do in excess of 220 with no additional voltage.
I have another TM333 set[1DDT] that run 200, I think they are 60B MTEC. They run 24/7 in a folder.
mike
ROADTRIP For my DAD 09-11-07 and MA 10-28-11
comhar a neartú agus a leathnú
Did you like not read the first post?
Using the labels on the new BH-5 and CH-5 is mostly refering to the manufacturing process, so it's more than just marketing as the different size processes clearly have different characteristics. 0.175u vs 0.13u, which is different than just speed binning from the same size process.
Cosmos S | Corsair HX1000 | Core i7 920 | DFI X58 | 12GB Super Talent DDR3 1333 Samsung HCH9 | 3x HIS Radeon HD7970 IceQ 2 | VelociRaptor 300GB | BD-RW
TFC360 | HeatKiller 3.0 CPU w/ Koolance Backplate | 2xPMP450S | Koolance 450x2 dual bay res | Swiftech MCR220 + MCR220 Stack | 3x Heatkiller HD7970 Waterblocks | Enzotech NB
@Adamantine: Adding the "-5" after "BH" is purely 126% marketing. TwinMos rides here with the name of a legend.
Yes. The difference is marginal - but it exists.
You were not supposed to see this.
ok , so there's BH5 UTT and CH5 UTT ?
But technically there both UTT , right ?
Confusing stuff
Point is i'm very happy with my Twinmos SP 3500 (BH5 like)
I'm still burnin' in .
So far I'm upto 251MHz @3,3v 2-2-2-5-7-16
Twinmos ROCKS
before winbond used to test there moduals and thats how bh-"5", bh-"6" moduals came about. Now there not testing them so we could be getting bh-5, bh-6, or even bh-7 who knows. that why they are labeled bh- UnTesTed. Same goes with the CH moduals.
Posting Permissions
Reply With Quote
Bookmarks