Dear community,

I bought a few 6502 CPUs on eBay. I have two 6502 units by Rockwell and two by Synertek. I have hand wired the CPUs and tried them with 555 timer, but later I found out that the pulse from 555 timer is not fully compatible with 6502. I know, that Synertek CPUs had built-in clock generators, so I have tried these CPUs without any external clock generator. Unfortunately I didn't get any result, any of LEDs connected to 16-bit address bus didn't shine. I have also tried connecting data bus to ground and 5V with 1k resistors to simulate RAM. So then I connected crystal oscillator circuit according to R6502AP datasheet. (see attachment) Synertek CPUs did not show up any artifacts of life, however, both processors have warmed up slightly. One of Rockwell 6502AP units, connected with CMS 1.8432 MHz crystal circuit, showed up some artifacts of life, a few random LEDs were shining. Other LEDs began to light up after the reset, but they never shone all. I expected every LED to shine. What could be wrong? Crystal generation circuit and connections in general should be okay.

Is it possible that all of these processors are defective? Can I ask for refund?

Thank you in advance.
Nogard

You don't have any clock input on the processor, which has to come in on pin 37. Even if you use the onboard oscillator, it needs external passives to control the frequency, even if it's just an RC. See the clock-generation page of the 6502 primer. It covers all this.

_________________
http://WilsonMinesCo.com/ lots of 6502 resources
The "second front page" is http://wilsonminesco.com/links.html .
What's an additional VIA among friends, anyhow?

First of all, your two Rockwell chips are probably counterfeits. The date code on both is Week 01 of (20)19, long after Rockwell ceased to exist as an independent concern, never mind stopped making NMOS 6502s. This is commonly seen from Chinese eBay sources; they really like Rockwell branding but some show up with counterfeit WDC markings too. They *might* still be working 6502s, but at the very least the markings have been changed on them, and you should ask yourself why.

The two Synertek branded chips look more likely to have original markings on them, as they have date codes from roughly August or September 1984, consecutive weeks. In other words, they're more likely to actually be what they say on the box. It's probably worth trying a little harder to get these working.

When wiring up any chip, you normally need to connect all input pins to something, otherwise they'll float around and cause digital havoc; output pins can be left open-circuit if you don't need them. On a generic 6502, that means the /RST, /NMI, /IRQ, /SO, BE, Phi0 (marked Phi2-in by WDC), and all eight of the data bus pins (that's from memory; I might have forgotten one or two). The data bus should be set using resistors, not wire shorts, because they can also be outputs some of the time. Pin 1 should be left open-circuit; it's marked Vss or GND on some 6502 variants, but /VP on others.

The term "built-in clock generator" here means that you can connect a crystal and a couple of other components, or an RC tank, between two pins to generate a clock. It's easier to use a canned crystal oscillator, which is not what you've got, but a plain crystal wired up in a circuit - another potential source of failure.

I'll let other regulars guide you from here…

Please try to look closer, the wire goes directly from clock circuit to the pin 37 on the CPU.

The thick white wire? That's going to the complete opposite corner of the CPU to where the Phi0 pin is. Take note of where the notch in the end of the chip is - that marks which end pins 1 and 40 are.

See the little half-circle on the left end of the 6502? That marks the end with pin 1 and pin 40. You have it going to the wrong end. The pin-1 mark is normally on the left when the printing is right-side-up.

_________________
http://WilsonMinesCo.com/ lots of 6502 resources
The "second front page" is http://wilsonminesco.com/links.html .
What's an additional VIA among friends, anyhow?

You are right, I am sorry. OMG I am so stupid, I messed up the direction of pin 37. I hope, that other pins are connected right. According to the new testing, one of Synertek 6502 CPUs works. Both Rockwells are probably fakes. Thank you so much for this note!

The Rockwells ARE counterfeit. In 1999, Rockwell Semiconductor was spun off into Conexant and MPUs with the Rockwell mark were no longer produced. Ergo any Rockwell 65C02 with a date code later than 1999 IS A FAKE. Conexant itself was bought out in 2017—65C02 production had ceased well before then. I've not seen a Conexant 65C02 in the flesh.

In the past, many Rockwell 65C02s were NOS originally intended for use in modems with a Rockwell chip set. These were the 65C02s being sold by JAMECO, who listed them in their catalogs for many years. That supply, as far as I know, has been exhausted. Any genuine Rockwell 65C02 being sold today will be a pull from used equipment. As the operating and handling conditions to which these MPUs were exposed would be unknown to the buyer, it would be quite risky to purchase such product.

It completely baffles me why folks insist on purchasing 65xx MPUs from eBay. The excuse about price doesn't wash, as the genuine article is worth more than the one or two dollars being charged, which means the odds of getting a counterfeit part are very high. The only 65xx MPUs in current production are the WDC products—W65C02S and W65C816S—and are only available through authorized distributors. Chinese eBay sellers are not authorized WDC distributors and are likely selling counterfeit product. Caveat emptor!

_________________
x86?  We ain't got no x86.  We don't NEED no stinking x86!

I have come to the conclusion that modern systems are simply too cheap and in no way represent the value and experience the engineers who designed them.

I have witnessed this more in the Raspberry Pi world than here though. The Pi is $35, or $5 for the Pi Zero.... so people using it to build systems think that their peripherals must also be cheap and when I was maintaining a popular C based GPIO library for them I had countless emails from people buying cheap (mostly) Chinese relays and other IO boards that just did not work on the Pi. Other SBC makers competed with the Pi and I think it's taken them some 7 years to catch up in terms of reliability, yet people still buy them because "cheaper", not caring about support.

So in the past few years people are now used to getting stuff cheap and almost disposable. Why pay $8 plus shipping from Mouser when you can get it for $2 and free shipping from random ebay supplier?

I see this in other worlds too - I run a small home based artisan bakery - yet people in my town will not buy my breads because they see a loaf of mine, hand made with organic flour for £1.50 then they see a mass-produced fluffy white sliced loaf in the local supermarket for 79p so guess what they buy... I can't make/sell croissants for under £1 each, yet in the local supermarket you get a pack of 4 (utterly disgusting looking) ones for £1. And so the story goes on.

People (in-general) now get cheap, expect cheap and ignore quality.

My view, anyway. It's one reason I've stopped supporting my Pi stuff and stopped trying to supply bakery products locally.

-Gordon

_________________
--
Gordon Henderson.
See my Ruby 6502 and 65816 SBC projects here: https://projects.drogon.net/ruby/

Yes; I think the amount of time we put into a project dwarfs the price of the parts, although we're back to the horrendous shipping charges some in European countries have to pay. That's gotta be for insurance! I'm almost (but not quite) inclined to offer the service to buy stuff and ship it to them without insurance, placing my bet on their honesty that they won't say they never received it when they really did. I've sold a few of my modules to people in Europe, and according to my records, it looks like the most expensive shipping was $14.25US, a recent one to the UK. Dajgoro bought some modules seven years ago, and shipping to him in Croatia was $5.27US.

_________________
http://WilsonMinesCo.com/ lots of 6502 resources
The "second front page" is http://wilsonminesco.com/links.html .
What's an additional VIA among friends, anyhow?

So I tested (just with NOPs and an ammeter) the rest of my 6502s (except for the ones in computers), and it was interesting.

• The R65C02P3 and R65C02P2 at the top draw about 1.3 mA, and even at 4 MHz the clock output and A0 are quite clean. (Adding a bypass cap at the Vcc input did help all the CPUs quite a lot. :-P) Date codes are 9529 and 9208; both are clearly well-used chips, no doubt pulls from something.
• Of the two R65C02Ps below that, both look identical (the slight difference in colour in the picture is due to the lighting) and have exactly the same markings, both in the numbers and the font/logo. But the upper one draws 3.7 mA and has clean edges even at 4 MHz; the lower one draws 102 mA and the Φ2 rising edge starts to get slow at 2 MHz, becoming useless at 4 MHz. (I tested both twice, because I at first thought I must have confused a chip.)
• Of the two R65C02P4s at the bottom, side by side, the right-hand one is the one I tested earlier, and the left-hand one tests out the same: they draw around 100 mA and Φ2 again goes downhill at 2 MHz and is terrible at 4 MHz.

The interesting thing to me was that the two middle ones I suspect came from the same source and they look like they were remarked by the same vendor, yet, at least by power usage, one is CMOS and the other isn't. I guess they just dump massive batches of "whatever 6502-type things" into their remarking line and mark them all the same, even if they're possibly better than what they're remarking them as.


Or at least re-marked, I'll believe. But there's no way an NMOS device pulls only 3.7 mA, as far as I know, so that third processor down, with a 1040 date code, is some sort of CMOS 6502. The lottery is subtle!


Well, no, it's not. (Or, at least, is only sometimes worth more.) An article is worth what someone's willing to pay for it given full knowledge of what it is. And for my playing around a 6502 of any kind is worth about $2, whether it's an old 1 MHz NMOS CPU or a brand-spanking-new 14 MHz W65C02. Either one is fine for the little projects I build (or at least am currently building or anticipate building this year).

That said, it is sad that one does have to understand that parts get remarked or have a bad experience like Nogard did above. But that's really just a part of building hardware, it seems to me. There's a reason that system vendors carefully qualify their vendors and the parts they receive, even when not buying from cheap and dodgy distributors.

_________________
Curt J. Sampson - github.com/0cjs

Imports into the UK (where I am) from the US are taxed highly. Import duty from the east isn't taxed much (if any).

So I was looking at the Neon816 board - shipping was $79, then on-top of that is the UK VAT at 20% then another tax (and I've forgotten exactly which one - might be another import duty) at I think 2.5% then on-top of that is the local delivery charge which is usually £18. That took a $350 product completely out of my toy budget range. So even if you didn't have any sort of insurance you can't get away from the import duties/taxes without breaking the law )-:

What will happen now post Brexit is anyones guess...

-Gordon

_________________
--
Gordon Henderson.
See my Ruby 6502 and 65816 SBC projects here: https://projects.drogon.net/ruby/

...which by definition make the part a fake.


When I say "worth" I refer to the price at which the real McCoy is being sold, not what it's worth to you or anyone else. For example, Mouser's current price for a W65C02S in DIP-40 is $6.95. That's what it's worth, as Mouser isn't going to sell you one if you aren't willing to pay $6.95. If you are unwilling to pay $6.95 for the real McCoy you can take your chances with an eBay seller and likely wind up with counterfeit product.

_________________
x86?  We ain't got no x86.  We don't NEED no stinking x86!

I think you're missing that different people have different valuations of money, time, and risk, BDD. It's mildly interesting perhaps to know what your perspective is, but it makes no sense to expect everyone to share your perspective.

I'm guessing that Ben Eater's projects has brought a number of people into the world of building 6502 systems, and it's good that some of those people have found these forums. I hope we present an overall positive experience for them, which helps them succeed in their first project and continue with others.

Online marketplaces have been around for a while now, and I'd imagine most people are aware of the risks and the ways of assessing them. What people might not be aware of is that old chips get re-used and sometimes get re-marked, and that the sellers may be unaware of the needs of the buyer, or unaware that buyers might be unwary and unaware of how markets work.

I think most people would be careful if buying food from a street stall in a tourist trap, and would know that the sales effort is aimed mostly at making a sale, not primarily at making a happy repeat customer. I think it's the same kind of thing: it's the mechanism of reputation which keeps transactions honest, and the mechanism of refunds which reduces the risk.

It's a fair point that most people don't have a feel for the true cost of low volume electronics.

But it still doesn't follow that the One True Way is to buy the way someone else chooses to buy.

I think the micro-economics are actually worth diving into a little deeper. The market value is not necessarily the selling price, rather the price at which a transaction actually occurs.

Mouser prices a W65C02S at $7 - plus shipping & handling. The latter could raise the perceived price rather a lot over the nominal one, even if the order also contains some glue logic chips and a breadboarding kit. It really is enough to make a novice think "I don't need a 14MHz 6502, maybe a second-hand one that runs at 1MHz or so would be cheaper." And they're not yet experienced enough to know the finer differences that come with it, nor how to distinguish a counterfeit from the genuine article, nor where to reliably obtain the latter at a reasonable total cost.

For a $50 starting price that turns into a $100 transaction through S&H charges, many people would give up and buy a train set instead; it's simply less hassle. I think that's why Ben Eater sells starter kits himself - they contain all the basic needs for $85, and are known-good parts.