Xtreme Twinmos Memory!

Quote:

---

Originally Posted by longsiew

TQ for the reply. But can you enlighten me on this. Does this mean my mosfet is faulty or my RAM is faulty? So how to remedy the mosfet problem. Last night I could run my comp and benchmarks at 260MHz with ONLY a stick of RAM. Of course, the performance is affected by this. My sisoft is at around 36** and it is half compared to dual channel which was around 71**.

---

Sorry, but ^don.k's^ affirmation is not true, because my new DFI i have got RMAéd (because the first one did the same things as yours) is still doing same strange things, and I havent got changed the vdimm jumper to 4v yet, so this is not the cause of the problem...
I have also changed my CPU (thinking about faulty mem controller) and the problem is still here.
I have tried different PSU´s, because my main PSU is 20pin, and DFI says that the 4v for the vdimm shoudnt work with this type of PSU´s (but it works perfectly... :D ) and have tried with an OCZ 520 24pin, a CoolerMaster Real Power 450 24pin and it does the same things...
I have tried other ram, and the problem hasnt gone.
Resumming, I am thinking about incompatibilties between DFI NF4 and UTT...
When my mobo starts to give me the problem, I swap the modules in the slots, and then it boots fine...
Is not a BIOS problem, because my new DFI RMA´ed is still with the BIOS from factory, and it does the same weird things...
I think that we will have to live with this problem, or change our DFI by a MSI Neo4 + DDR Booster... :nono: :D :D
Hope that all of my tries to find the cause of problem could help some1... ;)
Tell me any idea that you have about this problem... :P
Bye!

MDave, you were right, Im re-testing with bank-interleave ON. Working fine.

Also dropping Vcore as low as I can go, so far down from 1.64Vcore to 1.52Vcore and still 3D/Prime stable. 32M SuperPi is 26m54sec with 2T on 2Gigs of ram, 250Mhz x11 1:1. Searching for the sweet spot of lowest stable Vcore at these tight settings.

The Sandiego is truly amazing, and it's mind blinwing how well these $208 2Gigs of ram are working :banana:

Glad to hear its going well mate, once you've maxed it out lets see some benchies.

Quote:

---

Originally Posted by Lord_RTKK

Sorry, but ^don.k's^ affirmation is not true, because my new DFI i have got RMAéd (because the first one did the same things as yours) is still doing same strange things, and I havent got changed the vdimm jumper to 4v yet, so this is not the cause of the problem...
I have also changed my CPU (thinking about faulty mem controller) and the problem is still here.
I have tried different PSU´s, because my main PSU is 20pin, and DFI says that the 4v for the vdimm shoudnt work with this type of PSU´s (but it works perfectly... :D ) and have tried with an OCZ 520 24pin, a CoolerMaster Real Power 450 24pin and it does the same things...
I have tried other ram, and the problem hasnt gone.
Resumming, I am thinking about incompatibilties between DFI NF4 and UTT...
When my mobo starts to give me the problem, I swap the modules in the slots, and then it boots fine...
Is not a BIOS problem, because my new DFI RMA´ed is still with the BIOS from factory, and it does the same weird things...
I think that we will have to live with this problem, or change our DFI by a MSI Neo4 + DDR Booster... :nono: :D :D
Hope that all of my tries to find the cause of problem could help some1... ;)
Tell me any idea that you have about this problem... :P
Bye!

---

Good to hear from you. Come to think of it, I think you are right. Because I tried a lot of Bios including 310(official), 316, 331-1,331-2,331-3 and 509-2 and I can only stick to ONE stick. Anyway, I can boot and load into windows with two sticks of TMSP without overclocking. (This is the most suffering part). Now I can only overclock my graphic card. BUT, I won't give up and I want to bring back my previous DFI that could do 453MHz(HTT). Hopefully, my DFI get a lot of rest now (because I did not oc) and come back to life after a week. For the time being, I have to live with it.

Quote:

---

Originally Posted by Lord_RTKK

Simply: because they are untested CH-5, and they only do 2-2-2-5 over 3.1v (apart of magical sticks, of course...)
If you need 2-2-2-5 at a low voltage you should go for BH-5 die or TCCD (always at low FSB´s..., if you need 2-2-2-5 at higher FSB´s, then BH-5 or CH-5 will work right, but with 3.4v+... :D )

---

And i must buy another RAM ? :( This one can reach 250 2-3-2-2@1T with only 3.2V , not too bad ??

Quote:

---

Originally Posted by xgman

I thought I saw someone pot in one of the VX threads that the VX had the dimples also?

---

VX has no dimples...I own some they are blank...

Dragon

Quote:

---

Originally Posted by esdee

they are UTT, but i don't think it's brainpower PCB good luck ocing them ;)

---

PCB's on UTT make no differnce only the chips themselves...

Dragon

I am trying to burn in some twinmos sp 3500 bh-5 pcb ko-6633.
i am using a p4p800 mobo with a ocz booster.

testing so far at 2-2-2-5

200fsb 2.85v

210 fsb 2.85v

215fsb 2.95v

220 fsb 3.0v

225fsb 3.0v

230fsb 3.1v

235fsb 3.15v

240fsb 3.3v but with some errors and higher volts will not remove them.

if i lower the timings to 2.5-2-2-5 it gives no errors.

at 245fsb 2.5-2-2-5 i get some errors

at 245fsb 2.5-4-4-8 about 100 erros each pass

at 250fsb 2.5-4-4-8 i get about 2000 errors each pas of test no 5 at 3.5v and higher.

when testing at 240fsb 2-2-2-5 3.3v if i lower the the ddr volts i get errors like

fffdffff fffdff80 but if i then increase the ddr volts it changes to 00100000 00000000 and if i increase the ddr volts a little higher these erros increase too.

also when idle the ddr volts increase to 3.45. then when i start testing they drop to 3.0v and then by then end of a test #5 run stabalise at 3.3v.

is this normal for the ddr booster ?

is there any reason the errors change at higher fsb when i increase the volts i thought it could be a cooling problem but after adding a 12cm fan and 8cm fan i still get them.

I am soooo confused.

I have 4 sticks TMII Twinmos UTT, and 4 sticks of Twinmos 3200 SP also supposed to be UTT but 44D chips.... AA4T etc.

The SP's wont run for :banana::banana::banana::banana: in the A64 system, but the ordinary TMII's run 260Mhz @tight flawlessly, 2xsticks 1T 4xsticks 2T.

WEIRD!!!!

But these SP's run great in NForce2, fast pair does 250Mhz slow pr does 245Mhz without any problems (though one pr has some high voltage issues over 3.3V, freeze-up/lock-up).

BH5 Kingston acts the same, wondering if I got SP's that are actually BH5 before the chips were marked, so Im actually trying to get Apples to taste like Oranges?

i ordered 2 twinmos 512 Speed premium from newegg. E44B bh-5. I was hoping for ch-5 since that's what is advertised at twinmos. stupid me to think i was getting what was advertised! TwinMos Ad


what can i expect to get out of this?

3500 winchester, DFI nf4 SLI, MSI 6800gt

Quote:

---

Originally Posted by icyroy05

Just got my 512mb stick of speed premium from newegg. It reads AA4T.

---

Quote:

---

Originally Posted by nugzo

i ordered 2 twinmos 512 Speed premium from newegg. E44B bh-5. I was hoping for ch-5 since that's what is advertised at twinmos. stupid me to think i was getting what was advertised!

---

Dang! I just ordered 2 more sticks this morning so I could try a 4x512 set-up. I already have two AA4T (CH) that do 250 at 3.3 rock steady. If the new ones don't match I guess I'm screwed.

I found the twinmos sp 3500 is running to warm from high volts and speed.

If i run at 200fsb 3.4v no problem as i increase the fsb when i get to 235 fsb it starts to get warm and the 01000000 00000000 erros start. if i then drop the volts i can make the errors go away at 3.1v
I had 1 80mm fan on the ddr booster and a 120mm fan over the ram but it made no differance.
Heat spreaders allowed me to run at 240 but for a short time but it didnt take long for the errors to start and i couldnt lower the volts any more.

I tested some corsair xms3500 bh5 (2x256) and they ran perfect upto 260fsb no problem.

One thing i did find was the time it took to run a single memtest #5
at 225 fsb 2.5-2-2-5 with the xms 3500 35 seconds but with the speed premium 86 seconds with no errors.

Quote:

---

Originally Posted by leejsmith

I am trying to burn in some twinmos sp 3500 bh-5 pcb ko-6633.
i am using a p4p800 mobo with a ocz booster.

testing so far at 2-2-2-5

200fsb 2.85v

210 fsb 2.85v

215fsb 2.95v

220 fsb 3.0v

225fsb 3.0v

230fsb 3.1v

235fsb 3.15v

240fsb 3.3v but with some errors and higher volts will not remove them.

if i lower the timings to 2.5-2-2-5 it gives no errors.

at 245fsb 2.5-2-2-5 i get some errors

at 245fsb 2.5-4-4-8 about 100 erros each pass

at 250fsb 2.5-4-4-8 i get about 2000 errors each pas of test no 5 at 3.5v and higher.

when testing at 240fsb 2-2-2-5 3.3v if i lower the the ddr volts i get errors like

fffdffff fffdff80 but if i then increase the ddr volts it changes to 00100000 00000000 and if i increase the ddr volts a little higher these erros increase too.

also when idle the ddr volts increase to 3.45. then when i start testing they drop to 3.0v and then by then end of a test #5 run stabalise at 3.3v.

is this normal for the ddr booster ?

is there any reason the errors change at higher fsb when i increase the volts i thought it could be a cooling problem but after adding a 12cm fan and 8cm fan i still get them.

---

i had my 2x sp3500 replaced today with 2x512 sp3200 (utt's) and im no longer getting the troubles i had with my 3500's

i have to say though, on a nf7-s running 225 with 3.25 volts they are gettings amazingly warm, i have a 92mm fan attached to the top via that band method and i did attach some heatspreaders but took them off for testing, am tempted to put them back on again but have heard they dont help...

still, fairly stable at the moment 225 2-2-2-5 was hoping for a little less heat and tiny bit more speed but still.. cant complain :)

This is the email i sent and the answer i recieved from twinmos about the bh-5 on the speed premium 3200's.

Hi i ordered 2 x 512 TwinMos Speed Premium pc3200 from newegg. i ordered these because at your website (click here) this ram is listed as having the winbond ch-5 chips. these do not have the winbond ch-5 chips, they are bh-5 on the modules i recieved. The only reason i ordered these was because TwinMos advertised as CH-5. Are there any CH-5 available? and if so may i RMA them for exchange? thank you!
Lynn H.

Dear Lynn,

Usually the BH-5 chips is more expensive than the CH-5 chips because of
better performance.

As we don't have the CH-5 chips available now, so my suggestion is to
keep the BH-5 chips.

Regards,
Alice


Do u agree with alice that bh-5 is better than ch-5?

Quote:

---

Originally Posted by nugzo

This is the email i sent and the answer i recieved from twinmos about the bh-5 on the speed premium 3200's.

Hi i ordered 2 x 512 TwinMos Speed Premium pc3200 from newegg. i ordered these because at your website (click here) this ram is listed as having the winbond ch-5 chips. these do not have the winbond ch-5 chips, they are bh-5 on the modules i recieved. The only reason i ordered these was because TwinMos advertised as CH-5. Are there any CH-5 available? and if so may i RMA them for exchange? thank you!
Lynn H.

Dear Lynn,

Usually the BH-5 chips is more expensive than the CH-5 chips because of
better performance.

As we don't have the CH-5 chips available now, so my suggestion is to
keep the BH-5 chips.

Regards,
Alice


Do u agree with alice that bh-5 is better than ch-5?

---

... if every man of this forum tells u to keep them, u wont listen, but u need a women to tell it to you ;). Keep them man.

I got BH5 too but I am keeping them if they don't error out at 200 in memtest! You will have to pry the bh5 out of my cold dead hands.

argh u guys are killing me with this ch5 bh5.This mem worst i ever bought,but for the money hey ok.I ate this stuff and its bh5.My freind proved it to me as we put them together.IN my opinion, wich is whatever.spend like 10 more euros or 15$ more and get some mem that runs.This is the second and last time i buy this mem.Sry to disapoint u guys getting low timingns and high fsb from this stuff.I have had no luk with this stuff

1AA4t <---------::{ GOD ihate u ;-)

Quote:

---

Originally Posted by gundamit

Dang! I just ordered 2 more sticks this morning so I could try a 4x512 set-up. I already have two AA4T (CH) that do 250 at 3.3 rock steady. If the new ones don't match I guess I'm screwed.

---

I'm screwed. Opened my package yesterday and saw I had received 1A4T BH-5. So much for trying 4x512 this weekend. So I figure I just head back to the NewEgg site and order up another pair ... OH FOR F**Ks sake! They're sold out. No joy. This 2gig memory project is turning into a pain in my butt, as well as my wallet.

I'll trade u my 2 bh-5 i just recieved for your 2 ch-5 :)

Quote:

---

Originally Posted by chester

argh u guys are killing me with this ch5 bh5.This mem worst i ever bought,but for the money hey ok.I ate this stuff and its bh5.My freind proved it to me as we put them together.IN my opinion, wich is whatever.spend like 10 more euros or 15$ more and get some mem that runs.This is the second and last time i buy this mem.Sry to disapoint u guys getting low timingns and high fsb from this stuff.I have had no luk with this stuff
--

---

2,958+ replies, 129,300 views - all for nothing!
Oh well. :rolleyes:

Quote:

---

1AA4t <---------::{ GOD ihate u ;-)

---

1A4T is the BH-UTT. (...or was that a misstype?)

given the choice, and ive had both now (CH5/BH5 UTT) i would stick with the ch5 i have now every time :)

Well I tried putting in both the TMSP CH and the BH UTT and it actually seems to to be showing a bit of promise. Set everything to auto except
CAS = 2
DRAM Bank Interleave = Enabled
DRAM Drive Strength = Level 8
DRAM Data Drive Strength = Level 4
Max Async Latency = 8ns
Read Preamble Time = 6ns
R/W Queue Bypass = 16x
Bypass Max = 7x
32 byte Granularity = Disable(8burst)

I guess I will be playing with it over the weekend. I might have to go with this instead of a 2x1024 set-up. I'm liking the 3v but just like the Geil 2x1024 pair I tried I can't use a divider lower than 3:4.
http://img18.echo.cx/img18/9574/1sttest2bp.th.jpg

I must be lucky, SP BH-5 3200 2 x 512 @ 245 3.25v. They will go higher, just cpu memory controller won't.

Quote:

---

Originally Posted by slunge

i had my 2x sp3500 replaced today with 2x512 sp3200 (utt's) and im no longer getting the troubles i had with my 3500's

i have to say though, on a nf7-s running 225 with 3.25 volts they are gettings amazingly warm, i have a 92mm fan attached to the top via that band method and i did attach some heatspreaders but took them off for testing, am tempted to put them back on again but have heard they dont help...

still, fairly stable at the moment 225 2-2-2-5 was hoping for a little less heat and tiny bit more speed but still.. cant complain :)

---

With alittle more testing, im able to do 223 2-2-2-5 1T or 243 2-2-2-5 2T on my nf7-s with 3.25 volts (seems to drop to about 3.19 under complete load with Vio=Vdimm mod).

Quote:

---

Originally Posted by mitgsx

I must be lucky, SP BH-5 3200 2 x 512 @ 245 3.25v. They will go higher, just cpu memory controller won't.

---

Mine top out @ 265 @ 3.45v :)
Running @ 260 24/7 is no problem, a 12cm Enermax fan pointed at them @ 2 inch distance, 1500rpm... very quiet but still good airflow :)

Best bang for the buck :D