R65C02P4 fake chips

[Chromatix]

With NMOS, bus loading would be an important factor in how well the CPU would run in an actual system, TTL logic imposing significantly more load than CMOS. The test harness they used probably had relatively light loading, and they probably tested at room temperature to boot, so calculating a generous margin makes a lot of sense to cover more complex systems over at least a commercial temperature range.

[JuanGg]

JuanGg wrote:

I purchased five of this ICs off eBay. This is what arrived: Manufacturing date 2019. On the top, Mexico, on the bottom, Taiwan and a handful of scratches.
I wouldn't mind much if they worked. I hooked them up on a breadboard, providing power (5V) and tying reset, IRQ, NMI and SO high by 1K resistors. 1 Mhz clock courtesy of my function generator and later by a crystal. Data bus set to EA by 1k resistors. Power consumption is about 30 mA...

It's been some time, and the discussion has gone elsewhere, but I thought this may be interesting.
Cleaning up yesterday I came across these chips and thought I'd have a go at decapping one. No acids or heatgun around, so I just sanded the epoxy off to reveal the (obviously damaged) die.

The best photo I could take with a regular camera is the first one attached. The second one is an "artistic interpretation" of the die, product of a toy microscope, a camera looking down the eyepiece, and my expertise in MS Paint. Of course, I was pushing the ic around with a mechanical pencil.

I couldn't find any markings or numbers, most likely scratched away (by me) if there were any, but I'd say it looks like a microcontroller, with built in ROM, as Martin A pointed out:

Martin A wrote:

I check some of the 40 pin chips I've acquired over the years, those reverse markings look a lot like those on an Atmel AT89C51 micro controller.

In fact, it looks similar to this 87C51 die: https://upload.wikimedia.org/wikipedia/ ... FC_die.JPG
Couldn't find a die photo of an 89C51.

Only thing I know for sure is that there's no way this is (was) a 6502. Also 6502's die should be square... The other 4 I got could very well be completely random parts as well.

Juan

[BigEd]

Nice work! As you say, not a 6502, surely a microcontroller, and quite plausibly a relative of the 8751 or 8951.

One of the tactics for decapsulating epoxy is to heat and twist - there are videos out there. For example
https://www.youtube.com/watch?v=ZQeHHYJYWXo

[JuanGg]

BigEd wrote:

Nice work! As you say, not a 6502, surely a microcontroller, and quite plausibly a relative of the 8751 or 8951.

One of the tactics for decapsulating epoxy is to heat and twist - there are videos out there. For example
https://www.youtube.com/watch?v=ZQeHHYJYWXo

Seen that. The hair dryer I have may not do the trick. A hot air gun is on the list though. But without more sophisticated imaging means, it's not really worth it...
Juan

[lagomorph]

cjs wrote:

Tor wrote:

@cjs: "I have a couple of parts with 2010 date codes on them"
Those are fakes (or at least the date is fake). Rockwell stopped existing in 2001.

Perhaps I'm reading the date codes wrong?

I guess it's possible that the R65C02P4 chips in my photo above were relabeled to a faster part number, so as to attract customers (they were about the same price as other vendors), but that hardly seems worth the effort.

On the other hand, the CPU in my Apple I replica is labeled, "R6502P 13500-13 MEXICO 1349 B50323-8 [Rockwell logo]." And it definitely works at least as far as the monitor, BASIC and my own minor hacking so far goes. What on earth that works reasonably well as a 6502 would you re-label as a 1 MHz NMOS part?

I know this is an old thread but I also received a R6502P as part of a Apple I replica kit and it's thoroughly confusing. It should be NMOS right? It works fine but treats undocumented opcodes as NOPs (such as the various JAMs, LAX, etc.) and it seems to implement the opcodes I'm familiar with from the 65C02 such as BRA, STZ, PHX, PLY, etc. Is this chip actually CMOS and mislabeled?

The markings are:
R6502P
13500-13
MEXICO
1643B50323-8

[jds]

A Rockwell 65C02 has the part number 11450, so the 13500 is suspect. The NMOS Rockwell parts normally didn't have this internal part number and had R6502-xx on the second line. The xx, and the 13 on your chip, I believe, is a revision number. Most R6502's I have seen have a revision number of 13, some older ones have 11.

There is a seller on eBay currently that has a picture of a chip with the same markings as yours. https://www.ebay.com/itm/1PCS-R6502AP-R ... Sw8BNdCly6. To me the logo looks too large for that time period, so I'd suspect that someone is remarking chips and yours is one of those. I don't have any idea why someone would remark a R65C02 as a R6502, but maybe it is another brand of 65C02 that is harder to sell? The price is still very cheap.

There are some instruction differences that could help identify the chip even more. Rockwell added bit manipulation instructions to their NMOS processors, but not to the 6502, just the microcontrollers. Rockwell 65C02 added BBR, BBS, BRA, PHX, PHY, PLX, PLY, RMB, SMB, STZ, TRB, and TSB, and added some addressing modes to other instructions. I think the WDC 65C02 was similar but not identical?

[Dr Jefyll]

lagomorph wrote:

It works fine but treats [NMOS] undocumented opcodes as NOPs (such as the various JAMs, LAX, etc.) and it seems to implement the opcodes I'm familiar with from the 65C02 such as BRA, STZ, PHX, PLY, etc. Is this chip actually CMOS and mislabeled?

Seems to me you just answered your own question.

(I personally would be delighted to have the CMOS version, but I can understand you'd be disappointed if you were hoping to use the various JAMs, LAX, etc.)

Edit: some 65C02's omit the bit manipulation instructions, so until you test those instructions your chip is to some extent still unknown. But we're sure it isn't NMOS.

-- Jeff

[lagomorph]

Interesting. I think it must be remarked since it seems to act like a 65C02 but yeah, I'm not sure why someone would do that. Maybe the old NMOS chips are getting scarce/more expensive? I've used WDC 65C02s which do have those new instructions. I didn't expect this one to have them.

[lagomorph]

Dr Jefyll wrote:

lagomorph wrote:

It works fine but treats [NMOS] undocumented opcodes as NOPs (such as the various JAMs, LAX, etc.) and it seems to implement the opcodes I'm familiar with from the 65C02 such as BRA, STZ, PHX, PLY, etc. Is this chip actually CMOS and mislabeled?

Seems to me you just answered your own question.

(I personally would be delighted to have the CMOS version, but I can understand you'd be disappointed if you were hoping to use the various JAMs, LAX, etc.)

Edit: some 65C02's omit the bit manipulation instructions, so until you test those instructions your chip is to some extent still unknown. But we're sure it isn't NMOS.

-- Jeff

Looks like BBS works.

[BillO]

How much current does it draw? That is a dead giveaway. A CMOS chip will draw less than 20ma with about 13ma being typical and an NMOS chip will draw over 70ma with about 100ma being typical.

[lagomorph]

BillO wrote:

How much current does it draw? That is a dead giveaway. A CMOS chip will draw less than 20ma with about 13ma being typical and an NMOS chip will draw over 70ma with about 100ma being typical.

Looks like it only draws 5.2 mA when running at 1MHz. I just have the data bus tied to a value of EA and reset it watching the address lines toggle at their appropriate rates. It also seems to allow the clock to stop and run super slow (1Hz and below) with no issues.

It also runs fine at 2MHz but for some reason only draws 3.8 mA running in the same configuration other than a faster crystal oscillator.

Looks like a CMOS 6502. ¯\_(ツ)_/¯