6502.org

Hi All, have been busy for a while so not been doing much until recently.

However I've just spent 3 evenings scratching my head over random issues which looked like software but I was convinced was hardware. I checked the EEPROM and RAM over and again, but finally this evening decided to swap out the 6502 with a fresh one. It worked! So then I decide to solder a bypass capacitor to this one, plug it back in and find exactly the same weird problems! So I swap that one out for yet another fresh 6502 and my system is working again.

I'm not going to try and put a bypass cap on this one - I think I might have heat damaged the previous one. I have no idea what happened to the original, but I am not the most careful when it comes to anti-static and have even shorted the board before! However, I must have been lucky not to have any damage to the '02?

I always thought the damage would be obvious - e.g. random issues. However I could count the number of cursor flashes was always the same before the system froze - hence why I was convinced that this was software initially.

Has anyone else blown their '02 - not catastrophically, enough so it seems to work but starts giving odd behaviour making it look like a software or peripheral chip issue?

Anyway, lesson learned for me!

That could be it.

Part of my job in about 1985 was to do infrared thermal scanning of RF power transistors in operation under maximum-stress loads and to check for voids under the die where heat was not getting transferred into the substrate as it should. The die was uncovered, just exposed to the air while operating. In the worst case, I saw a transistor actually operating (not just in storage or getting soldered or brazed) with junction temperatures of over 350°C, or 660°F. Granted, it probably would not have lasted even a half hour doing that, but it did not destroy it immediately. (The whole thing was about the size of one of my fingernails, and it was dissipating about 350 watts. (The heat sink had water running through it for cooling.)

When you solder an IC by hand, one pin at a time will get that hot, but most of the leadframe and the body of the IC will be absorbing most of the heat, and the die won't get anywhere near that hot, let alone while operating like what I observed with the IR microscope. You don't have anything to worry about there. As for hours, I've been using the same WDC processor on my workbench computer for probably 20 years. BDD has told how many thousands of hours he has run his POC computer without re-booting.

Just take care with the static handling and absolute maximums given in the data sheet regarding voltages and so on, and you'll find these things are completely reliable, not delicate at all.

My experience with defective ICs has been extremely limited, despite have worked with them for close to 50 years. Almost always, defects are actually damage caused by ESD (most common) or operation at excessive voltage, not soldering. As Garth pointed out, devices can tolerate some amazingly high temperatures for short periods of time, even under bias. 450 degrees F is a typical reflow temperature in many production environments, so you know the device has to be able to withstand that.

When I built POC V1.1, I used an SOIC-24 version of the MAX238, as well as an SOJ-32 SRAM, both with 50 mil pin centers. I dragged-soldered these parts, which subjected them to considerable heat for about 12-15 seconds per side. Adding insult to injury, I then used desoldering braid to clean up my handiwork, adding more heat to an already hot part. Yet, I have experienced absolutely no problems with the computer, which has exhibited excellent stability. In fact, and as Garth pointed out, POC V1.1 ran 24/7 for over 300 straight days without being rebooted. That's nearly 8000 continuous operating hours.
There was a time when chip yields were poor and most manufacturers cheerfully admitted that some of what they sold would turn up defective. It is possible your 6502 was defective from the day it was manufactured and only a certain combination of operations uncovered the problems. However, the bad old days of low yields had pretty much subsided by the latter 1970s. So I think it is unlikely you got a device that was defective as produced.

As you admitted that your handling techniques aren't the most cautious, I'm inclined to say you accidentally zapped the 6502 at some point and the "wound" is now "festering." It isn't going to get better, no matter how many bypass capacitors you connect to it. Try to be more careful with the next one. I'm not obsessive about ESD handling precautions, but I am careful to ground myself before touching anything. I have never zapped anything.

Well, don't solder the cap on the 6502. Put it on your board instead.

Thanks all. I too have had excellent reliability, same CPU for the past 2 years. Strange that I had a similar problem with 2 in a row, but perhaps have been unlucky.

Arlet, not sure I blew the CPU due to soldering but I will avoid as it seems to be working fine without the cap. Or should I put in some kind of cap else suffer other odd behaviour later on?

Not sure, but why take the risk ?

If it's easy to put it on the board, I'd do it right now. If it's tricky, and you risk breaking something, then I'd wait until you see something odd.

Am I reading that right, your soldering the bypass cap to the actual 6502 itsself?

I've seen examples where some IC's (not the 65C02) have the VCC and GND next to each other in the middle. Soldering a capacitor to the pins is about as good as you can get in that scenario.

Although, I can't bring myself to solder anything to a '02. Mainly because the VCC is in the worse possible place (IMHO).

It may be as good as it gets, but it's certainly better than it needs to be. You'll almost always see commercial 6502 boards made with the bypass caps next to the chip, at one end or the other.

Indeed, look at old 6502 boards like C64, Beeb, PET, or anything else from that era. You'll see plenty of double sided boards, cramped with components, no ground planes, and caps near the ICs, but not in "ideal" locations.

Don't worry too much.

I was referring to this statement from the original post. I was going to suggest soldering the bypass cap to the board rather than the chip to see if that helps lesson any damage he's doing to the ship.

Just out of curiosity:
- if you use one of the flakey one again - does the problems reappear?
- if you then (flakey cpu) cut off the cap (don't desolder, simply snip), does that change anything (more errors/less/different)?
- at what clock is your board running?

But be careful: discharge yourself and your tools and use condictive foam for your good cpu while it is replaced

Also, check that your iron's AC outlet is wired correctly, otherwise your soldering iron tip may be a very high voltage probe.

You could check with a multimeter between earth and probe tip (check both AC and DC settings).

Well that would be a ... wait for it... "rude shock", wouldn't it?