Types of 6502 processors?

[Keatah]

Is there a comprehensive list of which companies made versions of the 6502 and 65C02 and what what different about those versions?

 

Offhand, immediately, I can think of GTE, MOS, ROCKWELL, NEC, SIGNETICS.

[SeaGtGruff]

Is there a comprehensive list of which companies made versions of the 6502 and 65C02 and what what different about those versions?

 

Offhand, immediately, I can think of GTE, MOS, ROCKWELL, NEC, SIGNETICS.

 

Somebody at http://www.6502.org might be able to answer that question.

[+Stephen Moss]

Without going through and comparing every specification in the data sheets from every manufacture it would be difficult to identify any differences and without good electronics knowledge to know if any differences found were significant or not.

 

I general of course they should all be interchangeable thus pin out and functionality should be identical regardless of manufacturer, however in reality as every manufacturer may use a slightly different production process/levels of doping in their silicon mix there will most likely be small difference in voltage range tolerances and speed performance.

For example I once built a circuit using a National 741 and it worked fine, built a second using a Phillips 741 and it did not because it turned out that the Phillips had a 20% lower GBP, quite shocking considering they should be interchangeable. I would hope that example is a more extreme case, if you are enquiring as you need to replace a device then in general I would advise using one from the same manufacturer wherever possible, at least then you should not see any decrease in performance.

 

Another small difference between manufactures may be encasement, some manufactures may only use plastic while other may use ceramic (or both), ceramic is a better heat conductor however that is only useful if they device is being used in a manner that it generates a lot of heat i.e. at maximun performance values. So again of you are looking to purchase a replacement and the original was attached to a metal plate (Heatsink) then choose a ceramic encapsulated replacement, otherwise plastic or ceramic will do.

Edited March 26, 2014 by Stephen Moss

[Keatah]

I was curious from a standpoint of what a diagnostic program might see. IIRC there is an APPLE II program that can tell the GTE 65C02 apart from the ROCKWELL 65C02.

 

I'm not too interested (at the moment) about the electrical characteristics.

[+Stephen Moss]

I don't know the answer to that although I would think it likely that it is reading a memory location within the device that houses a unique manufactures ID.

[Rybags]

I would assume that the pre-CMOS ones are probably the same.

 

It'd be interesting to know how this program can detect between the 2 CMOS ones - I doubt there's any CPUID type function, probably more like different result reading an unassigned memory location or maybe a different cycle count or result from one of the illegal instructions - although they were supposedly all assigned as NOP.

 

From memory though, I think there was some cycle-count difference in 1 or more instructions if in decimal or binary mode - can't recall if it was relevant to the original or CMOS variants though.

[Keatah]

Yeh Mr. Rybags, that's what I'm getting at. I'm all over the place and disjointed and doing a data recovery job on a semi-cross-country bus tour. Ridiculous 2bsure.

 

But yes, I couldn't dare imagine any CPUID function in these chips. A diagnostic program would need to look at a corner case or detect a certain behavior with an illegal instruction. From what I overheard, the Rockwell chip is easily singled out. Also the Rockwell chip is supposed to be the better 65C02. Better, how? Idunno.

[Rybags]

Seems the Wiki page might have some answers at least:

 

http://en.wikipedia.org/wiki/65c02

 

Over various periods of time, the 65C02 has been second-sourced by NCR, GTE, Rockwell, Synertek and Sanyo.

 

Some variants of the 65C02 (including the WDC W65C02S and the Rockwell R65C00 family) feature individual bit manipulation operations (RMB, SMB, BBR and BBS). The 65SC02 was also available, which lacked these operations.

 

Inclusion or lack of the extra instructions an obvious way of narrowing down 65C02 types, although possibly not necessarily how that Apple program works.

 

I'm not sure MOS were prolific at all with 65C02 manufacturing - apparently the 4510 which would have been used in the Commodore 65 was their version, but I don't know of any other usage of it.

Of course, there's the large embedded processor market that takes a back seat to the computers that used 6502, even though total numbers for embedded use might well be 5 times or more that of computers and peripherals.