• FIC's Official PA-2007 Large Size DIMM Report 

SDRAM DIMM 
Vendor IC           Part No.                         Module No.                ECC       Size(MB)    Test result 
Toshiba         TC59S6408FT-10         THMY728071EG-10        Yes          64MB           OK 
Hyundai        HY57V658020TC-10    HYM7V72A801                Yes         64MB           OK 
Micron         MT48LC8M8A2-10      MT9LSDT872AG-662D3  Yes         64MB           OK 
Mitsubishi     M5M4V64S30ATP-8   MH8S72ALD-8                  Yes         64MB           OK 
SEC             KM48S8030AT-G10    640808S09SA31                 No         64MB           OK 
SEC             KM48S8030AT-G10    U8                                       No        128MB           OK 
Hyundai        HY57V658020TC-10   U8                                       No        128MB           OK 
Fujitsu          81164842A-100FN                                                  Yes        128MB           OK 

Note: Below module was not tested because it was not available. 
SEC           KM416S4030AT-G10   KMM366S824AT-G0          No        64MB 
 
 

• The PA-2007, the 113cd133 beta BIOS and the K6-2

SYSTEM OVERVIEW 
MB-FIC PA2007 rev1.2 1meg cache 
Via chipset vt82c586a-586 
ICW--w48c67-01h-0703 1731T 
Bios v1.13cd133 
60ns edo standard mem 32meg 
HD fireball 1.2gig 
S3 virge 3d 4meg video card 
8xcdrom 
linksys nic card pci -10bT 
USR 33.6 modem 
ess sound card 
Enlight case 250 watt 
standard cpu fan + thermal compound 
4"x4" fan in front of case 
waterfall v1.23 
 

Wintune 98 run off the internet 
clock rate 308mhz 
Dhrystone 775mips 
Whetstone 358mflops 
overall mem speed 471mb/s 
avg mem read   515mb/s 
avg mem write   591mb/s 
avg mem copy   306mb/s 

Disk 
cached performace  43.81442 mb/s 
uncached performace 4.178481 mb/s 

Wintune 97 gave a clock rate of 623mhz "yeah right" 

I have not found a bench to test true bus speed.I've been told clock 1  1-2  2   1-2   3   2-3   =68.5.  Take note, more ram 64meg  and fast sdram etc would really boost benchmarks!  Iam also running win95a.  The frame rates in unreal appear to be much faster, I will do a DM later on our network with my son.  I feel I will see a difference there also. I need a voodoo2 "bad". I will try 75mhz 4.5 again later maybe 2.4v.  This chip and system @ 68mhz 2.3v 4.5x after 2 days of intensive running is "rock solid" not just sitting and running, but running hard!  Their is a lot of life left in the PA2007!  Thanks to Jesse Mars my son--testing stuff and thanks to Antonio Cataldo--great PA2007 site! 

Regards 
Frank T. Mars   aka Smoker 

 
• Proof the core voltages I found works

Good on you, 

I'm glad the information is helping prolong the life of your system.  Now if only we can push FIC to release a BIOS revision that supports the new K6-266/300 cpus! 

What are the markings on your cpu both on top and on the underside?  What revision PA-2007 board do you have? - markings on top left hand top side of motherboard near the rear of the last isa bus slot?  I need this info for a faq I'm doing on the PA-2007.  Thanks for your post. 

regards 
Antonio Cataldo 

epdf1 wrote: 

A Cataldo wrote in message <3523B33C.DD881A13@bigpond.com>... 
Hi all, 
 
Yes a PA-2007 can be configured to supply 2.2V as well as other core 
voltage settings.  So read on. 
 
 With regards to the PA-2007 not supporting the K6 266.  I have been using the K6 266 for over a  week now in my PA-2007 that I bought last year. 
 
 PA-2007 version 1.2 flashed with downloaded bios 1.09cd12 
 
 On top of the K6 I fitted a monster heatsink and fan combination, more suitable for a K6 233. 
 
 Initially I tested the chip at 4 * 66. Just to check that the chip actually worked. The ID on boot up was a 486dx16.  Ignoring this I installed win 95 rev 2 on my hard disk. 
 
 When everything was installed and working I upped the clock 4 * 75.  Still no problems although each time I booted up the Chip ID would be any of  the following 486 dx50 or 60 or 66. 
 
 Running Wintune97 showed how well the processor was performing.  The clock reading always read double 602MHz when actually it was 301MHz.  This gave a rating of 566MIPs and 180MFLOPS. 
 
 I tried the 75 * 4.5 but I was getting a silly clock reading of 188MHz and a vast drop in performance.  After investigation I found that one of the jumper plugs was loose so I replaced it with a new one. 
 
 The system booted a few times at 75 * 4.5 (338) giving the CPU ID 486dx66.  Running wintune97 I got a rating of 628MIPS and 204MFLOPS, it became unstable later the next day when I tried running some programs the whole system kept locking up within 1 minute in win95.  This I believe is down to the supply voltage being on the low side of 2.2V. 
 
 I was wondering about the 2.1V rating of the board before I bought the K6, but if they did error on the low side, they obviously did us a favour if we are able to add resistors in series using the voltage select jumpers to increase the voltage to the K6 266 to within safe operating margins. 
 
 Having read the other posts and what is on anandtech's hardware page.  I have checked the voltage on my board at R267 and it reads 2.13V. 
 
 I will probably be trying to up the voltage within the next few days, and taking temperature measurements within the case to determine if I need more ventilation.  As for the Bios I'm not really that bothered what it says as long as it works at full steam. 
 
 I have just tried the voltage adjustment above my results are as follows. 
 
 short connector 2.13V 
 47K connector   2.198V 
 100K connector  2.553V 
 
 I have stuck with the 47K connector for now as the system worked for long periods with the short connector, but hung every other day.  It didn't always boot properly without pressing the reset switch on switch on.  If with the 47K it still hangs occasionally, I will try other resistors 56K, 68K, 82K, 
100K.  The type of resistor I have soldered to the connector is a Surface Mount 1206 package, to reduce component lead lengths. 
 
 Many thanks to Anand Shimpi for recommending the board 6 months ago, I waited a long time for K6 266. 
 
 And also A. Cataldo for the voltage reference chart he kindly posted on the Newsgroup. 
 
 A Happy Chappy @ 338MHz. 
 
 Epdf1. 
 
 

After reading that there was an unofficial x5 setting I tried it 5 x 66 
which worked but was slower than 4.5 x 75. I have tried 5 x 75 a few 
times whilst upping the voltage a bit to (2.4V) to get 375MHz but just 
get windows IO protection errors. I think I might have stumbled upon 
this 5 x setting before but would not post without upping the voltage, 
which I did not want do above 2.3V when I first got the chip. 

Nice to see that the chip is now a 1/3 of the price I paid in April. 
Intel must be creaming the PC market with their prices. 

The K6-2 3dnow sounds good, but I shall wait for a K6-3 to really 
overshadow Intel parts. It would be nice to know how the PA2007 fair's 
with windows98. 

I hope your site expands, as a good web site originates from a good idea, 
one which I have not yet got. 

Regards, 

Epdf1. 
 

• Overclocking K6-266 using 2.1V

 I would just like to let you people know about my experiences with 
the above combination. 

    I have a rev. 1.2 PA-2007 w/ a K6-166 right now. RAM is 64 MB EDO 
(4 x 16 MB 72-pin). The system has been running very stable @ 240 MHz 
(3.5 x 68) with the voltage set at 3.3v. OS is Win95 OSR2. I'm using 
the beta 110cd119 BIOS from ftp.fic.tw. 

    My neighbour is upgrading his CPU & MB to K6-266 & FIC VA-503+, so 
I decided to use his CPU to check if my PA-2007 would accept this new 
chip. 

    I set the voltage to 2.1v and turned the system on. It worked fine 
at 240 MHz. Next step I changed the bus to 75 MHz and it also worked 
fine (262, bios reports 266). I then tried both 4x and 4.5x settings 
and they worked too (bios reports 4.5 x 75 as 333). At 5 x 75 (375) 
it would not even turn my monitor on, so I decided to try 5 x 68 (340, 
bios reports 333) and it worked fine at that speed too. 

    But the biggest surprise came when I went to remove the CPU from 
my system: the chip was no hotter than the same as touching someone 
else's hands! Compared to my 166 (which will burn your fingers if you 
try to hold it after turning the system off) this was really amazing! 

    I did not try those voltage jumper combinations and/or extra 
resistors, because to me it was enough the way it worked very well 
without any kind of "hacking" on the motherboard. :-) 
 

|_ |\/| 
 
 

• PA-2012 resistor valve for 2.2V

Hi, 

I'd like to try and make my FIC PA2012 deliver 2.2V also. Can you help out a bit and let me know where I should be measuring the said voltage? 

FIC tech support always seems to lose something in the translation when I ask. 

Thanks 
Phil Hoover 

ref 
 Dear Sir, 

     Thank you for contacting FIC Technical Support. 

     Hereunder is the rework instruction for PA-2012 to achieved 2.2V. 

     Change the resistor on R129 from 2.89K Ohms to 2.55K Ohms. 

     Warning! We will never support you if something goes wrong after you have implemented the 
      rework. It was assumed that you are technically capable of doing such rework. 

     Sincerely yours, 

     Tech Support 
     (REV)