FRAM71B and overclocked HP-71B

[Hans Brueggemann]

As there is currently arising some interest in using FRAM71B in conjunction with an overclocked HP-71B:
i have on my bench an HP-71B 2CDCC, ser# 2852A00xxx w/ HPIL 1B and FRAM71B + 512KB Extender.
the calculator runs on panasonic eneloop AAA cells. clock speed of the HP-71B was modified from stock 625 kHz to 961 kHz by replacing inductor L1 on the mainboard (180 µH) with an inductor of 75 µH which (ironically) comes from a broken HP-41CX half-nut board. the HPIL consists of 2 HP3468A (HPIL# 3,4) and a HP82162A. the MATH ROM is run from FRAM71B's 512 Extender.

i used the following code to do a quick test of the above configuration:

Code:


10  ! FOURTST
20  ! TEST MATH MODULE ON
30  ! FRAM71B IN
40  ! 900 KHZ HP-71B
50  !
60  OPTION BASE 1
70  COMPLEX SHORT A(8),B(1,2)
80  FOR Z=1 TO 8
90  ENTER :3; U
100 ENTER :4; I
110 A(Z)=(U,I)
120 NEXT Z
130 MAT B=FOUR(A)
140 MAT PRINT B
150 END

the code was run successfully mutiple times, without the calculator throwing any errors, HPIL lock-ups, or even MEMORY LOST.
keep in mind though that, despite the HP-71B/FRAM71B combo seeming to possess quite some potential for overclocking, this is just a singular test case and the result shouldn't be taken for granted for other HP-71B/FRAM71B combos. to fully qualify FRAM71B for higher clock speeds would require in-depth test runs with various peripherals and operating conditions, which i did for the nominal clock speed of the HP-71B. Hence, the official and guaranteed max speed of FRAM71B is no higher than the nominal clock speed of an unmodified HP-71B, which is 650 kHz.

EDIT: did another test in same configuration, but this time with an HP-71B 1BBBB, ser#2452A00xxx, 1LF2 0301 (84061) and HPIL 1A. changing the oscillator inductor from 180 µH to 75 µH yields a clock (STROBE) increase from 634 kHz to 943 kHz without encountering any problems while running above code.

[J-F Garnier]

Note that late HP-71B used the 1LK7 Saturn processor that provided more timing margins for memory access and could work up to 1MHz. It is probably the case for your HP-71B 2CDCC, ser# 2852A00xxx (1988).

I'm not sure if an original 1LF2 CPU accelerated at ~800-900kHz would work as well with your FRAM71.

I once wrote a small utility to identify the processor type by software without opening the HP-71B.
See here: thread-1224-post-133336
A binary ready for download with the PIL-Box can be found in the following posts.

J-F

[Hans Brueggemann]

thank you for digging into this, J-F.
indeed, the HP-71B mentioned in my post above has a 1LK7-0001 (88331) processor.
still wondering though as to why its factory-set STROBE was only 625 kHz (fosc = 2.5 MHz), rather than 1 MHz (fosc = 4.0 MHz). will be doing more testing and report back here into my top post.

[Hans Brueggemann]

(02-25-2024 10:32 AM)J-F Garnier Wrote:  I'm not sure if an original 1LF2 CPU accelerated at ~800-900kHz would work as well with your FRAM71.

well, doesn't seem to be a problem for my particular specimen. see EDIT in my first post.