980625    5560-W2B Modified BIOS and Clock up

     I am totally blind about HEX code, assembler,  programing and such matters. I really admire those people who can read or write computer language. Here is one of such a guy and his name is Mr. S.  He  joined the Room #4  in 1997 ( almost same time when I joined the Room) with his 5560-W0/W1.   5560-W looks like PS/2 8590 but the planar and complex are a bit different from that of 8590.   He tried clock lodding of these machines.  5560-W0 can take OSCs rated between 30MHz to 32MHz  without 102 error but the complex is a older type like PS/2 Type-1 25MHz.  The complex used in 5560-W1 is derived from PS/2 Type-2 H and is faster than W0 complex but it can not accept any OSCs rated higher than 54MHz( 27MHz system clock ).   Once a higher rated OSC installed  61-W1 gives a charming 102 error code  on the screen and stops there.  I had tried my self and succeeded on W0 complex to use it with 31MHz OSC but completely failed with W1 complex.

   Mr. S's initial attempt was adding a sub-board with 60MHz OSC and shifting  OSC from 50MHz on the complex to the added OSC on the sub-board manually right after the system finished POST.  He succeeded but it ws not so easy to shift OSC from A to B on a working system manually.  Then he examined  BIOS if there were any chance to use a higher rated OSC on the complex.  He analyzed the BIOS and it's "algorithm" (  please don't ask me what it is ) then reprogramed to skip 102 ERROR.    His 61-W1B with his reprogramed BIOS EPROM  run successfully in 32MHz base clock without 102 error.  I was given his handmade ROM on 1998 Jun 25.   His  magic really worked on my 5561 too.   Type-2 H'ish  Japanese complex run in 32MHz system clock and drove am5x86 around 130MHz speed with 64MHz OSC. 

980700  Type-A, Type-B and other models

     The story is a bit long so let's make the long story short.
I got  reprogramed BIOS ROMs for Type-A and Type-B complex.   Type-A, the fastest complex among all of  PS/55 complexes, got boosted further with 80MHz OSC and  a BIOS ROM. free from 102 ERROR 
With an am5x86-133 ADZ on an interposer,  5580-"N" with  BooDoo magic'ed Type-A run flawlessly  in 160MHz inner speed.    "Nancy  the  second"  also had a  modified ROM in it and run in 160MHz speed too. 

      Note;
        5551-N uses Flat Packaging Write-One-Time-PROM, a tiny SOJ chip.  Chip No. is HN 27C301AFP-12. 
        It's belongs to P571000 series and  it's hard to find 27C301 ( and other P57C1000 equivalents ) these days . 
        P571001 and it's equivalents are common and easy to get.  In order to use 1001- type ROM in stead of 
        1000-type ,  soldering rework  is needed  to exchange connection of copper traces to pin #2 and pin #24.
 
 
 

5580 with DX2-66 UpGrade Type-A complex
			WindSock	HDBENCH		ETDIAG
CPU	Clock	Inner clock	CPU	F/P	Integer	Dhrystone MAX
DX4-100    x3	33MHz	99MHz	814			37878
am5X86     x4	33MHz	133MHz	957	3743	5645	41666
am5x86       x4	40MHz	160MHz	1162	4497	6762	51020
5551-NAB
Cyrix 5x86  x3	33MHz	99MHz	857	3130	4693
Cyrix5x86   x3	40MHz	120MHz	1017	3752	5643
am5x86      x4	40MHz	160MHz	1160	4155	6695	51020
am5x86      x4	42MHz	168MHz	1205	4490	6762

 CPU & Memory Test Under PFM486 ( DOS program) 
 1) 5551-N 40MHz
  Performance of 486-PC Ver1.77 (C)1992-1994 DysanKeihin
 H----------------------------------------------------H
 | i486Dx Frequency  = 160.26 MHz   Machine:AT   v1.77|
 |    CPUID Results  = AuthenticAMD Fam 4 Mdl14 Stp 4 |
 | Ecache Read Time  =  47 ns/dword ( 6.3wait, 85mb/s)|
 | Ecache Write Time =  80 ns/dword (10.9wait, 50mb/s)|
 | Memory Read Time  =  47 ns/dword ( 6.3wait, 85mb/s)|
 | Memory Write Time =  81 ns/dword (11.1wait, 49mb/s)|
 H----------------------------------------------------H

  2) 5551-N Original 33MHz
  Performance of 486-PC Ver1.77h(C)1992-1995 DysanKeihin
 H----------------------------------------------------H
 | i486DX Frequency  = 133.13 MHz   Machine:AT  v1.77h|
 |    CPUID Results  = AuthenticAMD Fam 4 Mdl14 Stp 4 |
 | Ecache Read Time  =  57 ns/dword ( 6.4wait, 70mb/s)|
 | Ecache Write Time =  96 ns/dword (10.8wait, 42mb/s)|
 | Memory Read Time  =  57 ns/dword ( 6.4wait, 70mb/s)|
 | Memory Write Time =  96 ns/dword (10.8wait, 42mb/s)|
 H----------------------------------------------------H

  3) 5580-Yx With Souped up Type-A 
    Performance of 486-PC Ver1.77h(C)1992-1995 DysanKeihin
 H----------------------------------------------------H
   |   i486DX Frequency  = 160.26MHz    Machine:AT  v1.77h| 
 |   CPUID Result    = AuthenticAMD Fam 4 Mdl14 STP 4 |
 | Ecache Read Time  =  47 ns/dword ( 6.3wait, 85mb/s)|
 | Ecache Write Time =  80 ns/dword (10.9wait, 50mb/s)|
 | Memory Read Time  =  47 ns/dword ( 6.3wait, 85mb/s)|
 | Memory Write Time =  80 ns/dword (10.9wait, 50mb/s)|
  H----------------------------------------------------H

  4) 8595-J Type-3
   Performance of 486-PC Ver1.77 (C)1992-1994 DysanKeihin
 H----------------------------------------------------H
 | i486Dx Frequency  = 150.00 MHz   Machine:AT   v1.77|
 |    CPUID Results  = AuthenticAMD Fam 4 Mdl 8 Stp 4 |
 | Ecache Read Time  =  26 ns/dword ( 2.6wait,154mb/s)|
 | Ecache Write Time = 179 ns/dword (24.7wait, 22mb/s)|
 | Memory Read Time  = 122 ns/dword (17.0wait, 33mb/s)|
 | Memory Write Time = 179 ns/dword (24.7wait, 22mb/s)|
 H----------------------------------------------------H
       Wait! Oh yeah wait a minute Mr.Postman. Wait! Oh No..

  5)FYI
   9595 Type-4 N complex 
     33.333MHz
   Performance of 486-PC Ver1.77 (C)1992-1994 DysanKeihin
 H----------------------------------------------------H
 | i486Dx Frequency  =  66.61 MHz   Machine:AT   v1.77|
 | Ecache Read Time  =  38 ns/dword ( 1.3wait,105mb/s)|
 | Ecache Write Time =  60 ns/dword ( 2.0wait, 67mb/s)|
 | Memory Read Time  = 134 ns/dword ( 7.7wait, 30mb/s)|
 | Memory Write Time =  60 ns/dword ( 2.0wait, 67mb/s)|
 H----------------------------------------------------H

  40.0MHz
  Performance of 486-PC Ver1.77 (C)1992-1994 DysanKeihin
 H----------------------------------------------------H
 | i486Dx Frequency  =  79.96 MHz   Machine:AT   v1.77|
 | Ecache Read Time  =  31 ns/dword ( 1.2wait,129mb/s)|
 | Ecache Write Time =  50 ns/dword ( 2.0wait, 80mb/s)|
 | Memory Read Time  = 126 ns/dword ( 8.8wait, 32mb/s)|
 | Memory Write Time =  50 ns/dword ( 2.0wait, 80mb/s)|
 H----------------------------------------------------H

    CPUID results were erroneousely deleted.  But it wasa stock DX2-66 with a gold heat sink 
    loaded on my stock N complex.