Talk:WDC 65C02

Title change: "65C02 Microprocessor" --> "WDC 65C02"

Why the change? In short: because <company name> <processor name>, where <company name> belongs to the originating company, is the de facto standard for microprocessor articles in WKP. Note that this also applies to processors which are/have been second-sourced (i.e., manufactured by one or more companies other than the originating one). --Wernher 14:48, 16 July 2005 (UTC)

This is a wrongheaded decision. The 65C02 is a family of related designs: any suggestion they are licensed derivates of WDC's particular design should be supported by citations. The presumption should otherwise be that the designs were independent implementations in CMOS of the 6502 instruction set, with enhancements. This article may mention WDC's specific choices but should not elevate them as definitive. — Preceding unsigned comment added by 86.148.64.214 (talk) 16:14, 21 January 2013 (UTC)

This is an error. There was a 65C02 (and in fact a whole 65Cxx family) from MOS Technology (a.k.a. Commodore Semiconductor Group) in 1982, as shown in the MOS Technology 1982-1983 Data Catalog, a scaned copy of which is available from bitsavers.org: http://bitsavers.trailing-edge.com/pdf/mosTechnology/_dataBooks/1982_MOS_Technology_Data_Catalog.pdf. WDC was founded in 1978, according to their website, so I can't deduce which is the case: that MOS/Commodore licensed the 65C02 from WDC, vice versa, or the two companies developed CMOS implementations of the 6502 independently. Regardless, the article excludes the Commodore product without any cited reference to support the only possible justification for this, that the Commodore 65C02 was a licensed second-source version of the WDC 65C02 (which later developed into the current 65C02S). The article title needs to be changed back, or else a new article needs to be created with the title MOS Technology 65C02. I don't think the separate articles are warranted. 173.59.13.98 (talk) 15:09, 2 June 2017 (UTC)

Update: I have examined the MOS Technology 65C02 (MPS65C02) datasheet from the 1982-1983 Data Catalog and closely compared it with the datasheet for the NMOS 6502 in the same book, and I found the two to be nearly identical, except for power supply current and clock speeds. According to these datasheet The MOS Technology 65C02 has the exact same instruction set, pinout, feature list, and electrical characteristics as the NMOS 6502: it is practically identical except that it is implemented in CMOS instead of NMOS, so it uses much less power but is lmited to 1 MHz, whereas the NMOS version (as of 1982) was available in 1 MHz (no suffix), 2 MHz ("A" suffix), and 3 MHz ("B" suffix) versions. This MPS65C02 is not covered at all by this article, nor am I aware of any other article that covers it. For historical and legacy support purposes, it needs to be described on Wikipedia. 173.59.13.98 (talk) 00:42, 5 June 2017 (UTC)

Dead link

The datasheet's link on 65xx.com seems to be dead. --Thomas, 4:02pm, 30 March 2006 (UTC)

Fixed. --StuartBrady 20:58, 12 April 2006 (UTC)

G65SC02

I've checked the datasheet for the G65SC02[1] and it seems the TSB and TRB instructions were not removed, and that only RMB, SMB, BBR and BBS were removed. I've updated the article accordingly. --StuartBrady 20:58, 12 April 2006 (UTC)

Power consumption

The power consumption is listed at 150uA, which I presume is supposed to mean 150 microamps. However, following the link for uA directs to nanoamps. A nanoamp would be written as nA, not uA. Shall I assume it's supposed to be microamps? What is the source on this anyway? Mofoburrell (talk) 22:29, 4 April 2008 (UTC)

Nanoamps (i.e. less than one microamp) would not be reasonably plausible for any microprocessor, even the most efficient CMOS design implementing the simplest architecture. 173.59.13.98 (talk) 15:11, 2 June 2017 (UTC)

merge

I suggest merging the stubby article W65C02S into the WDC 65C02 article. The "W65C02S" is a kind of "WDC 65C02", and I think the similarities and minor differences between all the kinds of "WDC 65C02" can easily be covered in this one article. —Preceding unsigned comment added by 68.0.124.33 (talk) 15:05, 26 April 2008 (UTC)

Apple II frequencies are wrong

The given frequencies for the Apple II models are slightly incorrect. They don't run at 1.023 MHz but at 1.020484... MHz, i.e. 65 machine cycles per 228 cycles of the NTSC color carrier frequency. I'd correct it myself if the article wasn't locked. -- 77.187.133.27 (talk) 09:11, 23 March 2011 (UTC)

Is this microprocessor still in use???

Is this microprocessor still in use? I am not technically minded enough to know, but I ran into this circuit from this page:

https://en.wikipedia.org/wiki/Transistor_count

It shows that it has had an updated manufacturing process in 2009 using the 0.22 µm process. Thus it appears that it is still being manufactured - but the "Notable uses" section only lists ancient products such as the Apple II. Does anybody know what this is still being used for? And is it still used in large quantities? Thanks in advance to anybody who knows. — Preceding unsigned comment added by Betathetapi545 (talk • contribs) 09:30, 8 February 2014 (UTC)

Section on New addressing modes

This section is very inaccurate and confusing.

The first paragraph starts with "variation on indirect addr. mode" but describes a direct addressing mode. Furthermore, the (zp,y) mode did not exist. I suppose the author meant "LDA $12,Y". But then the statement "Unfortunately, this was the only way to load data from the ZP" would be wrong, since most ALU instructions worked with ZP and Absolute addressing modes.

The 3rd paragraph is also incorrect, as the original 6502 also included a (zp),y postindexed mode. In fact it was the only indirect addressing mode that was used regularly on the C64.

The 4th paragraph is also incorrect. The JMP instruction was updated with (absolute,X). The JSR instruction was not affected. There was no JMP or JSR (zp,x) on the original 6502.

All these comments can easily be verified with any copy of the original 6502 and 65C02 datasheets (widely available on the internet).

I suggest a complete rewrite.

DiederikH (talk) 18:28, 7 September 2019 (UTC)

There are a few more issues regarding 6502 instructions:

• The 6502 has load and store instructions using zero page for the X ($a6/$86) and Y ($a4/$84) registers, so the claim of using an index register to get at a zero page address is simply wrong.
• The terms "preindexed" and "postindexed" sound alien to me, "indexed indirect": `LDA($10,X)` and "indirect indexed": `LDA ($10),Y` are what most datasheets use, as well as the C64 PRG.
• The example "A program can LDY #$10;JSR ($12),Y to call "routine number 10", is very confusing. Unless one manages to construct a table of words where the MSB of the routine called is the LSB of the next routine and so forth, this is a bad example. (Either do LDA routine_num; ASL; TAY, or come up with a better example, or delete it (my preference). Also, the "This new mode not only allowed for a larger table, but also meant it could be placed in ROM." doesn't make a lick of sense, ROM vectors have been fine since the PET and CBM-II.

Hopefully someone who's a native English speaker can fix this. I can help with the details.

Compyx (talk) 19:55, 10 October 2019 (UTC)

I did a complete rewrite, and it at least passes Compyx muster. Somebody else can figure out the proper closed/resolved/archived/whatever wikipedia process.

70.166.159.104 (talk) 19:11, 17 October 2019 (UTC)

References

Inaccurate historical reference in intro

Development began in 1981[a] when it was known as the 65802.

That is untrue.

Evidently, whomever wrote that into the article completely misread the reference source. That source did not say the 65C02 was named the 65802. The 65C802 designation was given to the 65C816 when packaged for use in 65(C)02 circuits. The 65C02 was never referred to as "65802" at any point in its development.

The 65C02 design commenced shortly after Bill Mensch founded the Western Design Center in 1978—the 'C02 was WDC's first product. Mensch is the sole patent-holder on the 65C02 design, with Rockwell and others acting as second-sources. Functional product was available through distribution in late 1982—I bought one at the time to try out in a Commodore 8024 computer—and Apple started purchasing 65C02s in 1984 to power the Apple II_C and later the II_E. As soon as the 65C02 had gone into production, Mensch (and his sister) commenced design work on the 65C816.

Although the commonly-repeated story is one where Apple wanted the 65C816 and that is why it was developed, the fact of the matter is Mensch already had the 65C816's design worked out and realized he needed a (relatively) high-volume application for the processor in order to proceed. It so happened that Apple was looking to develop a "poor-man's MacIntosh" and when Mensch learned about their plan he pitched the 65C816. Steve Wozniak revealed in January 1985 that they were looking at the 65816. The result was the Apple II_GS, which went into production in 1986.

The 65C802 came later on, after WDC had amortized the costs of developing the 65C816. The 65C802 is essentially a 65C816 die in a package that is pin-compatible with the 65C02. No bank address is emitted by the 65C802 and it lacks the ABORT input, as well as the VDA and VPA outputs of the '816—the latter two are replaced by SYNC. Also, the Rockwell RMB, SMB, BBR and BBS instructions are missing in the 65C802 (and in the 65C816).

Many Apple II_E owners upgraded their machines with the 65C02 or 65C802. II_C owners often removed the 65C02 and replaced it with a 65C816.

The 65C02 core is currently used in embedded applications, including heart pacers and defibrillators. Estimated usage of the 65C02 in discrete and ASIC forms is over 100 million units per year.

216.152.18.132 (talk) 07:39, 9 December 2019 (UTC)

The wording above is unclear - are you claiming that design on the chip started in 1978, that samples were available in 1981, or both? Do you have documentary evidence for these statements?

We do have cited documentary evidence that the chip became available in 1983, and that is also the date it appears in the IIc and various other products. I cannot find any documentary evidence of it existing prior to that date. Maury Markowitz (talk) 23:27, 30 September 2020 (UTC)

Further to the 1983 date, this is the first known advertizement for the 65C02: here. Maury Markowitz (talk) 13:31, 8 October 2020 (UTC)

Yep! My wording was a little ambiguous. Sorry!

In conversations with David Gray at WDC, he once mentioned that Bill Mensch started work on the 65C02 design c. 1980. The point in time when the masks were released to produce wafers is unclear, but supposedly happened sometime in 1982. My above date in obtaining one to retrofit the Commodore machine was a typo that I didn't notice when I made my post.

216.152.18.132 (talk) 06:45, 26 February 2021 (UTC)

Computers -- Atari 400/800/600XL/800XL/1200/XEGS missing -- how did THAT happen?

Atari was the first, for many years the ONLY and by far the LARGEST (by number of units sold) user of 65C02 processors, dwarfing the combined sales of all the other systems mentioned here COMBINED.

Abkedefghi (talk) 13:47, 24 June 2020 (UTC)

@Abkedefghi Can you post some sales figures to substantiate that statement, without the SHOUTING please. ToaneeM (talk) 17:36, 24 June 2020 (UTC)

@Abkedefghi: The Atari machines did not use the 65C02, they used the 6502C. That is a slightly modified version of the original MOS 6502, and has nothing to do with the topic of this article. Maury Markowitz (talk) 23:21, 30 September 2020 (UTC)

Low power modes

This section needs some attention. There is at least one incorrect statement in it concerning STP, and the description of how the WAI instruction behaves following an SEI instruction is not particularly understandable. It might be worth looking at the 6502 interrupts article (can't recall the exact title), which has some discussion about STP and WAI.

216.152.18.132 (talk) 06:59, 26 February 2021 (UTC)

Sales "in the 100's of millions" out of date

The 6502 article talks about the 65C02 in its opening text. It claims that the 65C02/65xx family continues to be widely used, with "with estimated production volumes in the hundreds of millions".

The linked reference is to the Western Design Centre website. The actual paragraph there says: "...Annual volumes in the hundreds (100's) of millions of units keep adding in a significant way to the estimated shipped volumes of five (5) to ten (10) billion units. With 200MHz+ 8-bit W65C02S and 100MHz+ 8/16-bit W65C816S processors coming on line in ASIC and FPGA forms, we see these annual volumes continuing for a long, long time. The 65xx brand is probably the only processor family that has remained loyal to its ISA over the last 33 years."

The last 33 years since 1975, start of the '65xx brand', means this was written in 2008, making it 14 years out of date.

That website redirects you to a newer WDC website, which contains a revised statement: "Through the last 30+ years as one of the most popular microprocessor architectures of all time the 65xx brand is estimated to have over six billion embedded 65xx processors shipped and is growing by hundreds of millions of units per year, provided by WDC and its licensees."

The relevant text on that page looks like a rehash of the original site's text. The '30+ years' looks like a rehash of the '33 years', aimed to make it timeless so they don't have to worry about changing the website. So it's also not proven to be current and is likely out of date.

All in all, there needs to be better evidence that the 65C02/65xx family sells these quantities per annum, either in parts or as licenced IP. Or the statement's wrong and should be removed - not just removing the quantities but also the claim of popularity and any volumes. ToaneeM (talk) 11:23, 25 January 2022 (UTC)

Indexing bug

The article mentions a very specific bug that was introduced into the 65C02 that wasn't present in the original NMOS version. Does that bug remain in samples manufactured today or was it fixed at some point? I've added a Citation Needed tag because if the bug persists it will be documented somewhere. 87.75.117.183 (talk) 05:31, 16 November 2022 (UTC)