Talk:Fairchild F8

Source added
This was prompted by the former second sentence (now the third), which I found confusing as the Intel 4004, introduced in 1971, is often attributed as the first single-chip microprocessor and the F14A Central Air Data Computer multi-chip microprocessor claims to deserve the title of "first microprocessor" on the basis that a microprocessor need not be one chip. The mention of Noyce being involved with the F8 and also leaving Fairchild before 1971 seemed to imply that F8 development was at least well underway before 1971, which seemed implausible given Intel history. Therefore, a date for the F8 (conspicuously absent from the WP article) was needed, and I added one. The only source I could find in a casual, 10-minute search was the one cited, which while perhaps not the most authoritative, at least does seem to have enough in-depth information and an extensive catalog of images of chip packages (and a correct date for the 4004, as a spot check) to indicate that the author has done serious research and bears significant depth of knowledge. I submit that the article as I have augmented it (in a minor way) is superior to the previous version, despite the non-bulletproof nature of the source I have cited. If a more senior editor than I (I have been around as an anonymous reader/editor for about two years, making, I estimate, probably a few hundred edits to roughly 50-100 pages) contests this source as inadequately authoritative in context, then they should feel free to remove it, and the facts it supports, as necessary. (However, of course, it would be better to find a better source as a replacement.)

Also, the tag that says "this article does not cite any sources" is no longer literally correct, though I don't think it's intent has really been fulfilled, so I haven't removed it.

71.242.27.236 (talk) 01:01, 24 September 2008 (UTC)

Coprocessors
I think we need more about the coprocessors seeing as it says the processor had no address bus and instead had each coprocessor keep its own address counter controlled by 5 control lines. Does anyone know what the 5 controls lines were? RJFJR (talk) 02:15, 31 August 2011 (UTC)
 * OK, I just saw we had the datasheet as an external link. A summary would be nice.  RJFJR (talk) 02:20, 31 August 2011 (UTC)

ABOUT ROMC
Rather than referring to the set of three integrated circuits in the Fairchild F8 microprocessor as co-processors, they can be viewed as a partitions of the microcomputer. In 1975 the F8 designers were trying to put an eight-bit computer onto the fewest number of ICs that was possible given density constraints in the NMOS IsoPlanar Technology. The Fairchild designers settled on dividing the F8 logic between four ICs: F3850, F3851, F3853, and F3854. The logical duties given to these were CPU, PSU, SMI, and DMA (Central Processing Unit, Program Storage Unit, Static Memory Interface, and Direct Memory Access). These ICs were connected using two buses: the time-multiplexed 8-bit data bus which handles all addressing and data functions in the system, and the 7-bit control bus which unifies the system by coordinating and synchronizing the activities of all the system's major components, relative to the data bus. In the F8 system, the control bus regulates the use of the data bus by providing two kind of information to all components of the F8 system: timing signals and state control. The timing signals are derived from the master clock generator on the 3850 CPU. Timing uses 2-bits of the 7-bit control bus. The remaining 5-bits are used for state control signals. These five lines are referred to as ROMC, that is, Read-Only Memory control. The five bits allow 32 discrete states that permit other devices in the F8 systems to match their own responses to those of the CPU. The phi clock line, the WRITE clock line, and the ROMC signals direct the flow of data throughout the system. This is essentially all about managing the program counter register and data counter register. The program counter points to the next operator (machine instruction). The data counter points to the location of the next operand (data). What is neat here, is that the control bus and data bus are available as signals in pins on the IC. Michael Edward Zeidler (talk) 20:26, 21 June 2015 (UTC)

Date of introduction?
The topic as written implies that the F8 was introduced in 1977. My copy of the spec-sheets detailing the for this is copyright 1975. The package includes a paper (Fairchild Systems Technical Bulletin 5) "Microprocessor Testing": A Testing Turnaround&mdash;Smart Dut Runs the Tester by Robert Huston, dated February 1975. Based on that, I wrote a simulator in late 1975. "1977" needs some tuning. TEDickey (talk) 10:50, 18 August 2013 (UTC)
 * MK 3850 central processing unit
 * MK 3851 program storage unit
 * 3852/3853 memory interface
 * 3854 direct memory access

Fairchild F8 Design Evaluation Kit Number One
By 1975 advances in NMOS IsoPlanar Technology allowed small chips with a large number of circuit elements. In early 1976 the Microsystems Division of Fairchild Camera & Instrument Corporation was shipping an 8-bit microprocessor development system named the Fairchild F8. The Fairchild F8 Design Evaluation Kit Number One, which included a ROM containing the Fair-Bug development software, was sold and used as an 8-bit microprocessor development system during the years 1975-1977.

The F8 Kit 1 had the microprocessor functions divided over three ICs to be followed by the 3870 which had the same logic compressed to one chip. The central processing unit contained circuits for interrupt, status register, accumulator, clock, scratchpad, and input/output. The program storage unit contained circuits for ROM, program counter, data counter,, memory page select, and control logic. The memory interface unit contained circuits for dynamic RAM control. That names of circuits that passed signals between the three ICs were data bus, control bus, and control and interrupt lines.

The physical package of the F8 microcomputer was a circuit board. An edge connector with 44 contacts provided circuit access to the F8 microprocessor. The board was powered by +5v and +12V with a common ground. The edge contacts allowed connection to a Teletype. Two parallel ports allowed circuit expansion to include circuits such as a floppy disk controller. The board included an interrupt switch for restarting the computer. Large pads at the opposite side of the board allowed addition of RAM or ROM to 64K.

The F8 module ROM contained the Fairchild debugging software called Fair-Bug. Fairbug was essentially random access management routine for RAM. Fairbug gave the developer the ability to specify the contents of every memory cell. The 70 machine instructions allowed both byte and bit manipulations of memory cells and registers. The hardware stack was limited, but a stack software routine could easily be written to use RAM memory. After setting memory cells, the developer could use the Fairbug command G to branch to any location in memory which was treated as a start address. Basically the Fairbug provided an environment for writing routine in machine code. The programs could do what small computers do.

The F8 board was considered a modulo of the Formulator system even though it was a self-contained computer. In fall 1976 the microporcessor successor to the F8, which was the second sourced MOSTEK 3870, came to market. The Motorola 6800 appeared in late 1976 with Fairchild being the second source. The rapid evolution of 8-bit microprocessors coincided with a diminished interest in the Fairchild F8 in 1977 even though its design is suitable for applications today.

F8 documentation from Fairchild ranges from 1975 through 1977. Fairchild Semiconductor advertising for the F8 was strongest in middle to late 1976. The Fairchild F8 was innovative and well engineered product. The Fairchild F8 Design Evaluation Kit 1 set standards that moved the microprocessor manufacturers toward greater complexity and high quality.

SOME FAIRCHILD F8 PUBLICATIONS IN CHRONOLOGICAL ORDER

1. June 25, 1975. "Preliminary F8 Microprocessor Data Book", Revision A. Fairchild Semiconductor. 2. June 27, 1975. "A Guide to Programming the Fairchild F-8 Microcomputer", Revision A. Fairchild Semiconductor. 3. 1975. "F8 Microprocessor". Fairchild Semiconductor. 4. August 7, 1975. "A Timeshare Users Guide for Fairchild F8 Microprocessor", Revision A. Fairchild Semiconductor. 5. September 1975. "User's Manual: F8 Design Evaluation Kit Number One", Revision B. Fairchild Semiconductor. 6. September 1975. "Appendix H: F8 Design Evaluation Kit, Assembly and Installation of PC Board" and "Fairbug Errata". Fairchild Semiconductor. 7. February 13, 1976. "Instruction Sheet for The One-Card F8 Microprocessor Evaluation System". Micro Systems Division, Fairchild Camera & Instrument Corporation. 8. April 12, 1976. "Fairchild Releases Microprocessor Design Kit". Computerworld. 9. May 15, 1976, "Fairchild F8 Kit". EDN. 10. June 7, 1976. "Microprocessor System" (Patent Application), US 4086626 A. U. S. Patent Office. 11. September 15, 1976. "F8 Fairbug User's Guide". Fairchild Micro Systems. 12. November 20, 1976. "Fairchild F8". EDN 13. Date not determined. "F8 Formulator", EDN. 14. February 5, 1977. "MOSTEK F8 Microcomputer Kit". EDN. Michael Edward Zeidler (talk) 00:56, 18 March 2014 (UTC)


 * "This circuit board performed as a 64K 8-bit personal computer " reads as if this board had 64K or RAM/ROM, but in fact it had just 1K ROM, and 1K (8 pieces of 2102 RAM chips) of RAM! Mahjongg (talk) 18:16, 11 November 2016 (UTC)
 * I agree. I was shocked when I read that because I had never heard of an F8 being paired with 64K of RAM in anything since RAM was ridiculously expensive back then. Just because it could use 64K of RAM doesn't mean it functioned like a 64K computer, when no one would spend that much on RAM in the first place for this CPU. It's like saying the Mac SE/30 was a 128 MB machine. In reality, no one ever installed that much RAM in one. — Preceding unsigned comment added by 107.77.194.65 (talk) 07:43, 13 January 2020 (UTC)

Which F8 are we talking about?
This encyclopedia entry is really about the second-sourced F8 from MOSTEK; it is not so much about the Fairchild F8 from Fairchild. The semiconductor manufacturer MOSTEK introduced a single integrated circuit version of the F8 microcomputer in 1977. This version (MK3850) had significant differences in comparison with the set of four integrated circuit chips (F3850, F3851, F3853, F3854) that comprised the F8 microcomputer that was put on the market in 1977 by Fairchild Semiconductor. One major difference was that the pins for two buses on the original set were not included on the 40-pin version from MOSTEK. The second-source version of the F8 did not have pins for either the address bus or the data bus. The MK3850 appears to have been designed for large volume buyers who would present to MOSTEK the machine code ROM instructions that they wanted inside the MK3850. MOSTEK would create a mask for etching the custom ROM and RAM circuits inside the MK3850. This strategy meant that only the well-capitalized users of integrated circuits could apply the MK3850 from MOSTEK. In contrast, the 1975 version of the F8 allowed the developer to intermix ROM and RAM memory in any useful configuration for all 65536 addressable byte locations. The 1975 F8 put out by Fairchild Semiconductor was a real 8-bit microcomputer. It was programmable using 70+ machine code instructions. The computer was easily connected to a Teletype which served as the terminal for both programming and executing programs. It could also be interfaced to diskette systems using available port space. The Fairchild F8 appeared in the market place just as ICs were beginning to get density that was high enough to include the many circuits needed for a computer. The Fairchild F8 came after the Intel 4004 and before the 8008 and the 6800. — Preceding unsigned comment added by Michael Edward Zeidler (talk • contribs) 12:03, 20 June 2015 (UTC)

Difference between Fairchild F8 and F3850
The title of this Wikipedia entry is not strictly correct. The Fairchild F8 refers to the Fairchild F8 Microcomputer which is constructed from a minimum of two integrated circuits identified as F3850 and F3851. The F3850 is the Central Processing Unit. The F3851 is the Program Storage Unit. Both of these are required to be in a circuit that makes up the core computer. Other integrated circuits can be added to the core circuit. Fairchild manufactured the F3852 (Dynamic Memory Interface), the F 3853 (Static Memory Interface), and the F3854 (Direct Memory Access) to expand the F8 Microcomputer system beyond the two IC core. Knowing this makes it obvious that the picture of the F8 at the beginning of this article is not the whole microprocessor, but only the PSU. It would be technically correct to show both the F3850 along side the F3851. Or alternatively to picture the F8 Microcomputer Design Evaluation Kit Number One which is a complete 8-bit microcomputer system. Also, it would be instructive to show an image of the Mostek MK3870 which puts the logic of the F3850 and F3851 on one IC, but does not let designers access the two internal buses. — Preceding unsigned comment added by Michael Edward Zeidler (talk • contribs) 20:56, 22 June 2015 (UTC)

Date Codes on Fairchild F8 ICs
The integrated circuit identification appears on the Fairchild F8 ICs. Thus, inspection shows codes like F38501DC, F3853DC, F3851ADC, etc. A second code on the IC is either a Fairchild Semiconductor Order Number or else a Date Code. For example, the codes 7534, 7541, 7649, 8621, and 8622 were manufactured in years 1975, 1975, 1976, 1986, and 1986. The original F8 ICs have the 75 and 76 as the lead two digits of the Order Code. When Mostek was manufacturing F8 ICs in 1978 their order codes had the leading digits 78, such as 7839. Inspection of the photo at the top of this article shows an order code of 8621. The IC shown was manufactured eleven years after the Fairchild F8 was first available in the marketplace. The image is of a replica, not the original thing! Michael Edward Zeidler (talk) 19:18, 24 June 2015 (UTC)

Patent identification that covers Fairchild F8 Design Evaluation Kit 1
Up to this point in researching the Fairchild F8 it has not be clear which US patent applied to the Fairchild F8 Design Evaluation Kit One that was sold to engineers for less than $200.00 in 1975-1976. The problem has been cleared up: US Patent 3984813 A describes the Fairchild F8 Microcomputer in the same way as the documentation that was provided by Fairchild in 1974 describes the computer. Much of the documentation appears to have been derived from the patent document or else papers written internally at Fairchild. One of the most interesting aspects of the patent document appears to be explanation of the CROM which had output connecting to registers that released information onto the two buses, and also to control system logic. A preliminary examination of the patent document suggests that unanswered questions about the nature of the F8 can now be answered by reasoning about the circuits as described. Over a few days there will be new insights.Michael Edward Zeidler (talk) 20:50, 26 June 2015 (UTC)

About criticisms offered June 6, 2016
The changes to this article recommended on June 6, 2016 by the anonymous person going by the pseudo-name AirCombat are inconsistent with the content found in the Fairchild publications identified in the footnotes of the article. The changes would diminish the value of this article to most serious readers. And it would push aside interesting facts taking place in the early months after the birth of eight-bit microcomputers.Michael Edward Zeidler (talk) 23:22, 7 June 2016 (UTC)

Errors introduced by editors
The date in the lead photograph caption and the date stamped on the integrated circuit shown in the photograph do not agree: The IC shown was manufactured in 1986 while the caption is declaring the IC as having appeared in 1975. The person who provided the picture did not have the original set of three IC's that comprised the F8 microcomputer. In the interconnections of the three IC's the internal data bus was accessible. The single IC version that was from MOSTEK essentially was less design friendly. One can suspect that MOSTEK wanted to limit access to the data bus as a way to have a stronger economic position against potential microcomputer application designers. Those other recent editing changes have diminished the accuracy of the Fairchild F8 article. — Preceding unsigned comment added by Michael Edward Zeidler (talk • contribs) 08:24, 19 June 2016 (UTC)

General Instrument court case and the Inventor section
There is nothing in the article about the court case over the alleged IP theft that resulted in the F8. As stated at CPUSHACK: in February of 1976 Fairchild signed a agreement with Olympia Werke A.G. Why? According to the article, it's because Dr. Chung possibly stole IP from Olympia that resulted in the F8. I would quote the whole section but copyright, so here is the link: http://www.cpushack.com/2013/06/08/cpu-of-the-day-fairchild-f8-microprocessor/