65xx parts sources, genuine and fake

[jds]

I've bought quite a few things at AliExpress, normally with good results. I've had a few problems, once on a higher price item where the seller said he'd send another one and then eventually shut his store down. The only dispute I've managed to submit was resolved in just a few minutes and refunded. So you have to be careful, but it's mostly worth it.

I've had good results with UTSource, bought some Rockwell chips including 65C02P4 and 6552 and they all worked as they should. Bought some WDC65816's too, but haven't had a chance to test them. Prices are probably not as low as eBay or AliExpress, but so far I've got a lot more confidence in what I get. I don't really want to spend time on faulty parts when I'm trying to debug my own design with faults of it's own. I built up a breadboard based 65C02 system with UTSource parts and I can only think that I was incredibly lucky as it worked relatively easily.

[BigDumbDinosaur]

gbm wrote:

UMC W65C02S

There is no such an animal. Only WDC has produced a part with that part number.

[gbm]

Thanks for clarifying. So all these W65C02 offered on ebay are counterfeit?
Checked mine thoroughly - it has no VPB signal, nor it does support WAI/STP. It does support BBS, BBR, RMB, SMB and INC A. So it's just some remarked 65C02.

[BillO]

Will those UMC ones run at their rated 10mHz? If they do so reliably then they are probably worth the $2.70

[gbm]

I believe they are SOME 65C02, rather not UMC. I run them at 1 MHz and 3.3 V.

[BigDumbDinosaur]

gbm wrote:

Thanks for clarifying. So all these W65C02 offered on ebay are counterfeit?

Genuine WDC parts are sold only through authorized distributors (Mouser in the USA and elsewhere, as well as Coltek, DEMA, TRG and others in the EU). If you get something on eBay with WDC markings you should be very cautious, as you will have no clue about its provenance.

[unclouded]

I bought a W65C02S8P-10 on eBay for about $2.90 because I was fascinated to see an earlier incarnation. It might be a pull or NOS. I've only tested to 6 MHz but it seems OK so far. I've got some proper parts from Mouser as well.

[BitWise]

On eBay UK there is a seller called 'toucano76' who usually sells sets of 5 devices. They are always new WDC stock. I've had several lots from them over the years.

[gbm]

I have nothing against pulls from old equipment, as long as they are not remarked. Unfortunately it looks like 80% of old ICs sold on Ali or ebay are remarked. I don't really understand the idea behind remarking Z84C00 as Z84C00 (yes, new marking of the same type as original), 65C02 as W65C02S or V20H as V20HL, but it seems this is a common practice nowadays.
It's quite easy to distinguish NMOS from CMOS CPU. If you touch it 30 seconds after applying power, NMOS will be warm and CMOS - cold.

[BillO]

Here is an R65C02 I bought as part of a batch from polida2008 about 6 years ago.

It is either genuine, despite the 2008 manufacture date, or it is a real fine copy. I know this because:

A) It only draws 10mA while running at 4mHz
B) It works perfectly in an enhanced Apple IIe with a microdrive card installed. An NMOS 6502 or one missing the R65C02 instructions will not even boot under these conditions.

If it is genuine then we can assume that someone was still cranking these out as late as May of 2008 or that the date code is in error.

If no one with the rights to the name Rockwell was making these back in 2008 and it is a 'copy', then it is a damn good copy and worth the ~$2 I paid for it. However, I cannot see someone starting up the process to make these for $2 or less a pop considering the demand for them might be in the 100's per year.

So I'm going with - if it looks like a duck, quacks like a duck and walks like a duck, its a duck. I think someone with the rights to the name Rockwell and the logo were manufacturing these in Mexico as late as May of 2008.

Maybe we are all getting too cynical.

[gbm]

These falsely marked CPUs are not copies. They were pulled out from old PCBs, their pins were straightened and pulled at an angle to look like new chips, then the chips were remarked to look like new. Many remarked chips are false (like NMOS marked as CMOS) but some of them are real - see my case of remarked Z84C00 and V20. It's just that their marking is new and looks like new. I don't know how many companies back in 1990s were laser-engraving the chips (some did). I didn't see any original WDC, Zilog or ST chip made in 1990s with laser markings. The counterfeit ones are laser-engraved. Well, not all of them - the mysterious Rockwell R65C02s from 21st century are nicely painted.
Most remarked chips have the same date codes. My set of 5 V20s has different date codes, although the rest of marking is identical. And they ARE REAL, working V20, just remarked as HCZ-16 version.

[BigDumbDinosaur]

unclouded wrote:

I bought a W65C02S8P-10 on eBay for about $2.90 because I was fascinated to see an earlier incarnation. It might be a pull or NOS. I've only tested to 6 MHz but it seems OK so far. I've got some proper parts from Mouser as well.

The part number says it's supposedly a WDC product with a static core, in PDIP40 and fabricated in 0.8µ geometry. The switch to the 0.8µ geometry was coincident with the part being uprated to 14 MHz. Therefore, your alleged 10 MHz part is suspect, in my opinion.

Incidentally, currently produced versions of the W65C02S and W65C816S are fabricated in 0.6µ geometry and are all rated at 14 MHz. Due to the smaller geometry, the Fmax curve extends beyond 20 MHz when operated at 5 volts. However, 20 MHz was possible with the earlier 0.8µ geometry parts, as was demonstrated in the old CMD accelerator card for the Commodore 64 and Commodore 128.

[BillO]

gbm wrote:

These falsely marked CPUs are not copies. They were pulled out from old PCBs, their pins were straightened and pulled at an angle to look like new chips, then the chips were remarked to look like new. Many remarked chips are false (like NMOS marked as CMOS) but some of them are real - see my case of remarked Z84C00 and V20. It's just that their marking is new and looks like new. I don't know how many companies back in 1990s were laser-engraving the chips (some did). I didn't see any original WDC, Zilog or ST chip made in 1990s with laser markings. The counterfeit ones are laser-engraved. Well, not all of them - the mysterious Rockwell R65C02s from 21st century are nicely painted.working V20, just remarked as HCZ-16 version.

Hmm...possibly.

I just can't see why someone would go to all the trouble to remove the marking from R65C02P4 chips, re-dress the tops to look like virgin plastic, straighten the pins and place them at an inserter angle then re-mark them as R65C02P4 chips with dates like '86, '93, '06 and '08 then sell them as 'new' for $1.30 shipping included. It seems like a bad way to make money. I don't know about you, but I would pay $1.30 each for them if they were used.

But nevertheless, it is all just speculation unless you have irrefutable evidence that there were no R65C02s made in the 21st century.

I'm sure there are some brain dead folks out there re-marking $8 Synertek, Rockwell or MOS NMOS chips as $1.30 CMOS chips, or undertaking the laborious routine mentioned above, but I'll bet it's not everybody on eBay doing it and I'll bet they won't be doing if for very long.

Again, if it quacks like a duck .... I'll pay $1.30 for a working R65C02P4 regardless of how pretty the top of the chip is.

Sure, I can see some scurrilous folks marking a cheap and cheerful R65C02 as a WDC W65C02S to make a few illegal bucks, but I find the other stuff is, at least with 6502s, to be a bit of a stretch.

[BillO]

Those 10 R65C02P4s I ordered from polida2008 recently have arrived and have been tested.

All are CMOS.
All will run at 5mHz or better.
All have the date code 9326.
Some have had the printing removed from the bottom (or never had any).
All are obviously used and seem to be pulls from soldered-in applications. However, they were sold as 'New'.

All in all, for $1.30 ea. delivered, I'm quite happy with the deal.

[LIV2]

They're not 6502s but I ordered a bunch of Altera EPM7128's off ebay and none of them work at all.
Top-left is one I bought from a local reseller before they went obsolete (this one works), all the other ones are from ebay & are not detected by the programming cable. they could be pulls from some product where JTAG was disabled I guess?

Hopefully I eventually find some that work.