Jump to content
HWBOT Community Forums

Recommended Posts

Posted (edited)

Ah but these dimms work with C2-2-2-5 1T. I allready tested them. 
I thought so that they probaly didn’t do any c2-2-2-5 but C2-3-2-6. I never saw Promos doing a boot on these speeds. This Kingston runs 213MHz C2-2-2-5. Not a speed monster but still. 

Edited by Sparks.nl
Posted (edited)
52 minutes ago, digitalbath said:

Most ProMos are not good at all. Did you test your Kingston RAM with 2-3-2-6?

No not yet, but I don’t think I will because while pre binning anything that does not boot 235MHz C2-2-2-5 1T or higher on my test socket A bench is not selected for further testing. I will be running these dimms on NF4 as well just to be sure. After that all not properly binned will go. 
All Promos and others I had which actually posted on C2-2-2-5 1T only ran 200MHz. Nothing ever did >205MHz on those timings. So I think it really should be CH5.
I have got a box with all dimms that did post >205MHz C2-3-2-6 1T. Waiting somewhere to something or nothing. I hardly use those. 

Edited by Sparks.nl
  • Like 1
Posted
8 minutes ago, Sparks.nl said:

Nothing ever did >205MHz on those timings. So I think it really should be CH5.

Don't underestimate CH-5. Good CH-5 can do 280+ 2-2-2-5.

Posted
Just now, ZFeSS said:

Don't underestimate CH-5. Good CH-5 can do 280+ 2-2-2-5.

I know, but those C2-3-2-6 dimms were no Winbond dimms. 
For the others. Those are Winbond, but some CH5 do 210-215 with 3.2v while I have CH5 dimms doing 235-250 with 3.2v. 

Posted (edited)

I got 3 sticks of Corsair 3200C2 with CH-5, the best stick reaches 255Mhz at 3.5xV (as usual more doesn't help on CH) on my NF2 setup. My 6A sticks are better, the best pair is able to pass 32M at 257Mhz and 3.45V. Raising the VDimm to 3.6V gives me 261Mhz on the best stick and 259Mhz on the 2nd. However that might be the limit of my chipset and Vdimm regulators, not the ram itself.

That's why i prefer 6A over CH. Sadly i don't own genuine BH sticks to test with.

 

16 hours ago, 4n0nym0u5b3nch3r said:

Will any 512KB chip work?

The PLCC32 512K chips i am using on my A7N8X Boars are these:

PM49FL004T-33JCE

Edited by Tzk
Posted (edited)

I just counted: I own 18 BH chips. The rest is CH. These are relabled on the ic’s sometimes. But never as Infineon for BH. 
2 Kingston BH were able to run PI 1M @3.6v and 270MHz. 

Edited by Sparks.nl
Posted

I forgot to add a bit of interesting information to the Promos/Winbond/Infineon discussion:

Promos was a JointVenture of Infineon and MoselVitelic. Afaik the deal was that Infineon grants ProMos access to its interlectual property and patents, but promos has to sell its chips through infineon (so Promos -> Infineon -> any 3rd party). Instead Promos started selling directly to Mosel Vitelic which then resulted in a termination of the joint venture. Infineon even sued Mosel Vitelic and Promos for contract violations back then.

Here's the source for that (sadly in german, sorry!):

Link 1: https://www.infineon.com/cms/de/about-infineon/press/market-news/2003/128273.html

Link 2: https://www.infineon.com/cms/de/about-infineon/press/press-releases/2002/128489.html

So I think it's no coincidence that some Promos Chips look like Infineon/Winbond. Afaik Infineon and Winbond also hat a licensing and/or manufacturing agreement and the AT-6 chips are the result of that. However i'm not sure if infineon manufactured these for winbond or vice-versa. Still Promos/Infineon and Winbond/Infineon had close ties around the year 2000.

  • Like 1
  • Thanks 1
Posted (edited)
18 minutes ago, Tzk said:

I forgot to add a bit of interesting information to the Promos/Winbond/Infineon discussion:

Promos was a JointVenture of Infineon and MoselVitelic. Afaik the deal was that Infineon grants ProMos access to its interlectual property and patents, but promos has to sell its chips through infineon (so Promos -> Infineon -> any 3rd party). Instead Promos started selling directly to Mosel Vitelic which then resulted in a termination of the joint venture. Infineon even sued Mosel Vitelic and Promos for contract violations back then.

Here's the source for that (sadly in german, sorry!):

Link 1: https://www.infineon.com/cms/de/about-infineon/press/market-news/2003/128273.html

Link 2: https://www.infineon.com/cms/de/about-infineon/press/press-releases/2002/128489.html

So I think it's no coincidence that some Promos Chips look like Infineon/Winbond. Afaik Infineon and Winbond also hat a licensing and/or manufacturing agreement and the AT-6 chips are the result of that. However i'm not sure if infineon manufactured these for winbond or vice-versa. Still Promos/Infineon and Winbond/Infineon had close ties around the year 2000.

Link to english versions:

https://www.infineon.com/cms/en/about-infineon/press/market-news/2003/128273.html
https://www.infineon.com/cms/en/about-infineon/press/press-releases/2002/128489.html

luckily Infineon organizes the way I would do it. ?

Edited by Sparks.nl
  • Thanks 1
  • Crew
Posted

After I stumbled over over a Toshiba DDR PC2100 stick on german ebay, which had exact the same looking of the typical Winbond IC, I think Winbond are actually Toshiba. The ICs were labeled with production year 2002 in Japan. 

In December 1995 Toshiba and Winbond started their first technology transfer agreement and over the next years 3 other generations followed. They mentioned in April 2000 that they were the first manufacturer in Taiwan who successfully offered 0.175µm 256Mbit DRAMs. And me guesses that this was the start of the BH chip, even though the birth of the holy BH-5 should be somewhat later. Anyway they also agreed to develop 0.130µm 512Mbit DRAMs for the first time together with Toshiba engineers and hope to start producing by the end of 2001, later mid 2002. However in the meantime Toshiba decided to exit dram market and Winbond lost its technology partner. This article states that they also ended the production at Winbonds fabs for Toshiba. But apparently they were still allowed to use Toshiba technologies to produce chips as claimed in February 2003, that the share of 0.130µm DRAM chips of the whole output should reach 33% in the first quartal. I think this makes also clear, that BH-5 and CH-5 were produced parallel in different fabs. Me guesses that they developed the 0.130µm process but didnt reached the goal of 512Mbit chips, as CH-5 were still 256Mbit. This forced Winbond once again into action and as result they decided to sign a contract in May 2002 with Infineon to use their 0.11µm process. In 2004 they signed another contract for 0.09µm process. In both contracts, Infineon gained access to the produced DRAM chips. It is very well possible that Winbond gave them some chips from their older fabs as well. I have an AT-6 as low as week 33 year 2002, so I think this matches. I also own a BT-5 "Winbond" week 36 year 2003, so this is probably the CH-5 equivalent while AT-6 should be BH-6.
Anyway Winbond decided to focus on special ICs and virtually exited DRAM market, becoming just another manufacturer for Infineon. Which is probably also the reason there was no successor to their BH / CH chips.

Taiwanese manufacturers simply had not the resources to develop their own chip designs by that time. The plan was to earn money by manufacturing chips for partners but they never had enough market share and revenue to start R&D by themselves. Also DRAM manufacturing was not the healthiest business. In the nineties were a lot of companies but most vanished.

The ProMos Ram definitely got no Winbond IC on it. When Infineon left the joint venture, they made a new arrangement with Elpida. Also if you desolder any other IC, you probably will see those two flat dots on the backside. So its all about which machines were used for production.

  • Like 4
  • Thanks 6
Posted (edited)

@Mr.Scott

Could you give credit to @digitalbath when posting Vmods for the K7NF2? Also that "crap board" does 260Mhz+ 32M... :)

https://warp9-systems.proboards.com/thread/1327/asrock-k7nf2-raid-mods?page=1&scrollTo=39936

https://www.hardwareluxx.de/community/threads/amd-sockel-a.584473/page-99#post-27560667

Maybe the K7NF2 does also work with the 12V rail mod... Actually it's not that crappy as one might think.

 

Edited by Tzk
  • Thanks 2
Posted
42 minutes ago, 4n0nym0u5b3nch3r said:

I wish someone would buy my A7N8X-X so I have more room for fun stuff :^) a better AXP board.

 

Make it validate 300mhz it it will go quickly:)

Posted (edited)
11 hours ago, Tzk said:

@Mr.Scott

Could you give credit to @digitalbath when posting Vmods for the K7NF2? Also that "crap board" does 260Mhz+ 32M... :)

https://warp9-systems.proboards.com/thread/1327/asrock-k7nf2-raid-mods?page=1&scrollTo=39936

https://www.hardwareluxx.de/community/threads/amd-sockel-a.584473/page-99#post-27560667

Maybe the K7NF2 does also work with the 12V rail mod... Actually it's not that crappy as one might think.

 

Certainly. 

Please don't think I was trying to nic somebody elses work. That's not the case. Just wanted to add the pics to the public volt mod section of my site, in case anybody wanted them.

Didn't really think that was a big deal.

 

Edited by Mr.Scott
  • Thanks 3
Posted (edited)

It isn't a big deal, i just noticed it and thought i'd point it out. That's why i posted it as a question and not as accusation :) 

@4n0nym0u5b3nch3r

Well, the -X is singlechannel. As long as we can't make DC work that board isn't interesting at all. At least for benching. SC will give you a disadvantage in 32M and for valids the NF7 is wayyy more interesting, because 2.25V Vcore. And the -X doesn't even have SATA... I own two of these myself and still got to test if we can mod dualchannel into it or if nvidia somehow blocked it inside the chipset on silicon level.

Edited by Tzk
  • Thanks 2
Posted

Sorry to ask this but i skimmed through the thread looking for a A7N8X Deluxe bios for pcb 1.04/1.06 and didn't find any, am i just blind or is there none? :)

Posted
18 hours ago, TAGG said:

A7N8X Deluxe bios for pcb 1.04/1.06

There are mods for v1.x: http://bierbude.spdns.org:2302/Mod-BIOS und mehr/Asus/

Just be aware that the 1.x boards use the old (non Ultra 400) NF2 chipset. I'd expect about 220Mhz fsb at max. If you want 230Mhz++ then you have to go for v2.x. It's the same as with Abit NF7 v1.x and v2.x. Personally i'd use the Trats bios as it got DFI romsips and a newer sata rom included.

For A7N8X-E i'd go for the digitalbath bios with ED, ED55 or EBED romsips anytime: http://bierbude.spdns.org:2302/USER UPLOADS/digitalbath/ASUS A7N8X-E/

And for the regular v2.x Deluxe i'd use: http://bierbude.spdns.org:2302/USER UPLOADS/Tzk/Asus/A7N8X Deluxe v2.0/

I haven't uploaded my latest bios for the regular Deluxe yet. I added a few other options which digitalbath already included in his -E bios. Still all bios versios are able to hit 260Mhz+ if the boards chipset can handle it.

 

  • Like 1
Posted (edited)

By accident I found 2 GEIL 1 GB dimms which do C2-2-2-5 1T @420MHz and 3v (never tried lowering the voltage). 
 

I never knew this was possible. Let alone on Socket A with it’s limited memory controller when looking at NF4. 
They have a Brainpower pcb as well. 
 

A6D524D6-6B50-4D4C-8936-E1832E99B92A.jpeg

453C0974-6F81-4063-819F-5D52440DDF85.jpeg
When looking at the bottom they are not Winbond. Is this Elixer? 

Edited by Sparks.nl
Posted
43 minutes ago, Mr.Scott said:

Not for nuthin but that's 210 MHz and DDR 420. 

Socket A max's out at 300 MHz or a little better. That's DDR 600.

I think that was what I was saying. But the point was it is running c2 tight. A 1GB dimm is running c2 tight...

  • Thanks 1
Posted

It can't be winbond as those haven't got enough density. Winbond is 512mb doublesided only. That leaves us with the usual suspects: Samsung, Hynix, Micron, Nanya, Mosel, Infineon and Promos.

UCCC got 6 dots and usually doesn't run at 2-2-2 -> no UCCC

Micron hasn't got a straight edge -> no micron

That leaves us with Infineon, Mosel, Nanya, Promos and Hynix.

Maybe you can compare with the chips listed here: https://www.hardwareluxx.de/community/threads/der-ddr1-ram-chip-thread.1244468 Just make sure to compare with similar chips (density!) as for example Samsungs 256mbit and 512mbit chips (TCCD and UCCC) look way different.

My guess is that it's either extremely well binned Infineon (unlikely) or Nanya/Promos/Mosel. I don't think Hynix D5 can handle Trcd = 2 at 200Mhz.

 

Posted (edited)

Didn’t Elixer make memory as well?
Samsung is using dots more close to each other so I ruled them out. 
 

I think I have to take the heathspreader off to really take a good look. I will look at that later (“at work” now). 

Edited by Sparks.nl
Posted
20 hours ago, Sparks.nl said:

By accident I found 2 GEIL 1 GB dimms which do C2-2-2-5 1T @420MHz and 3v (never tried lowering the voltage). 
 

I never knew this was possible. Let alone on Socket A with it’s limited memory controller when looking at NF4. 
They have a Brainpower pcb as well. 
 

A6D524D6-6B50-4D4C-8936-E1832E99B92A.jpeg

453C0974-6F81-4063-819F-5D52440DDF85.jpeg
When looking at the bottom they are not Winbond. Is this Elixer? 

Hynix D43

 

  • Like 1

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

×
×
  • Create New...