New Toy: Oscilloscope
On Friday, some new toys arrived, one of which I was super excited about: an oscilloscope! We had an oscilloscope when I was a kid, an old 20MHz CRO unit that I had a tenuous understanding of how to use it. When I worked in TV repair in California there was lots of things we could have used one for, but we didn’t have one. I’ve been working on fixing a couple of my broken video game consoles, and one would have come in super handy… so when the Coronavirus stimulus money came in and the government said “spend it, for the good of the economy”, I decided to grab a budget DSO for home tinkering purposes.
I’ve watched a few folks diagnose various 8-bit computer maladies with one, and I felt like I had some idea of what I was supposed to do but I decided to watch some tutorial videos anyway, and it’s probably just as well… one of the first videos mentioned being careful where you ground the probes to as if there’s a difference in potential you can blow them up and potentially damage the channel of the scope, or the entire thing… that was news to me! I don’t really recall being aware of this from my younger days, though perhaps I was and I’d just forgotten about it. I’d certainly never managed to destroy one, so who knows?
Anyway after some tinkering, I managed to capture what I’m pretty sure is an NTSC composite video line from my wife’s Super NES. I couldn’t get the machine to work over ethernet or USB to my Windows desktop, but putting a USB thumb drive into the front of it and hitting “print” does give me the lovely screencapture you see linked above.
After some more time tinkering around, I turned my attention to my broken Atari 2600. I bought this “as-is” (read: probably the owner knew it was broken), without a PSU, but with a pile of games some time last year. It’s an old, woodgrain “light sixer” 2600. I think I paid $80 for the lot, and considering that some of the games and peripherals that came with it are worth more than that I’m not super upset the unit itself doesn’t work… if I’m being honest I would never really play it as I didn’t think much of the Atari when I was a kid.
Doing some prediagnosis of it shortly after I got it, I verified that the PSU checked good and that power was where it was supposed to be on all three chips, and basically there wasn’t a lot I could do to diagnose it per the field service guide without known-good samples of all three chips (the TIA, the RIOT, and the 6507 CPU). I put it up and left it alone until the day I got a scope, so yesterday I drug it out again for another look.
First thing to check: do we have a clock signal to the CPU? Yes, and with some clever use of my fancy scope I can actually measure it, to some precision. Do we have a reset signal to the CPU, and does it do what it’s supposed to? Yep, it curves up and down with a short delay as expected. After verifying the ready signal from the TIA (which I think is supposed to coincide with one of the blanking periods?) I was pretty satisfied that all the inputs for the CPU were correct, so I started working out ways to probe the outputs.
So following some advice I found elsewhere on the internet (and since lost), and I pulled the CPU out and stuck it in a breadboard. I ran the VCC, VSS, CLK, RDY, and the RST lines over from the socket with jumper wires. I then used some resistors (IIRC 5x 1kohm for the 1s and 3x 47ohm for the 0s) to create a NOP generator, forcing the data lines to 0xEA to create a NOP generator. This should make the program counter steadily march upwards until it overflows and returns to zero, which will continue forever. In theory, I should be able to see each of the address line pins strobe, each one half as often as the previous one, though I must confess I’m not sure which is the most and least significant bits.
Unfortunately I didn’t see anything, which assuming there’s no error in my testing methodology suggests a dead 6507. Unlike the MOS 6502, which can be found quite cheap at times, the 6507 is basically a $40AUD part, and there’s no guarantee it’s the only thing dead, which basically makes fixing this machine a non-starter as if I’m patient I’ll buy another, functioning one for about that.
However I did learn a bit, and the main thing I’m upset at now is that I don’t have more non-functioning 8 bit micros to work on. :(