There are a few vendors on AliExpress offering them. I only have a few there that I would call "trusted", this one isn't one of those. I know it's a gamble, but I've become good at betting. A part like that isn't going to be a clone, the market is too small. Some parts from Ali are "grey market" or "second shift" but rare parts are far more likely to be NOS.
These came from the store with the imaginative name, "IC electronic component chip Store". 
I'm in Canada, and the big NA distributors don't service us well. Some have branch offices here but I believe they don't offer full service, full catalog. So I refuse to pay 10x the value of the parts for shipping, instead I swallow my politics and order from China.
If those companies gave a *** about us, they would ship USPS and everything would be reasonable. But they are too lazy.
I might give them a try. I currently am designing around a 68B09E that I know is good but Iād consider not having to use the external clock generator.
Itās a pity companies like Mouser and Digikey donāt serve Canada well. I use Mouser almost exclusively because they are literally up the road but I feel your pain. Sparkfun have an almost prohibitive shipping cost policy in my opinion.
I think I saw HD63B09E somewhere there too...
MC68B03 and MC68B50 arrived. As soon as I get some memory, I can start on the first running proto.
I'm conflicted about the VIA. Will the real VIA 6522 please stand up? Seems there were buggy versions and it's not clear which manufacturers fixed it, and how to be sure when ordering NOS parts, how to make sure I get a good one, and also an "A" suffix to run 2MHz.
So my I/O strategy is in disarray. 
I can see why the 6522 is a desirable part. Hopefully youāll be able to find a good one.
I wrote a Python program to decompose S-records and am almost done with an EEPROM programmer. My biggest problem is I havenāt found a 6809 cross assembler that I can figure out how to install and that takes standard Motorola macro assembler source. If you or anyone have any suggestions, Iām all ears.
Hmmm. I think it was xasm I used before, but that probably did not have macros. I agree with using S-records, if you have a bootloader that will read them, you are set to go with installing firmware as well as user code. Easy upload using any terminal emulator. Except, your firmware is in ROM, so the different method needed...
A macro assembler isnāt crucial but would be nice. Iāll take a look at xasm. Iāll have to modify my monitorās Load command to wait for the EEPROM write to complete but I donāt expect any other issues. Most application software will load to RAM anyway.
I found this cross assembler:
asxxxx Cross Assembler
and successfully built it under Linux. It includes all the 8-bit processors and seems to accept Motorola syntax. Still experimenting with various source files and output formats.
I haven't found a version of xasm for the 6809 yet but it's readily available for the 6502.
Good find. xasm is decades old, this one was updated in 2021. I would not be surprised if asxxxx is inspired by xasm. Found some Rockwell R6522A so we'll try those. I guess since I use IC sockets, if I have chip problems I can keep looking. Also I will try to make the 6522 not mandatory for the device to boot and run. Another PIO option now, is an I2C port expander on the I2C bus. But then, no timer. Possibly, I could connect a divided down baud clock to an IRQ then have a periodic interrupt. IDK.
I was hasty in believing that asxxxx obeyed Motorola syntax. There are differences. However, the distribution does contain the source for assist09 the Motorola monitor that they used for all their dev kits that has been converted to the correct syntax and I can work with that for the time being.
Did I miss that you put an EEPROM burner on board from post #147? That's interesting. Another mushroom for the rabbit hole, now I find out (from the assembler documentation) there was a Hitachi analogue of the 6309 in the HD63C03 which, as the 6309 adds secret instructions to the 6809 set, adds secret instructions to the MC6803 set! So off I go again. 
The AT28C64B-15 I intend to use writes just like RAM but with a 10ms delay per byte. After I get the monitor programmed in I don't expect to write to it much but I can if I want to.
My breadboard EEPROM burner just has two 230xx I2C port expanders on it so it wouldn't be hard to incorporate on the main board.
Feed your head!
You could add one of these for RTC, power on reset, watchdog and battery backed RAM.
That is an interesting IC. However, I'm wary of adopting any devices that might become a supply problem. Some semiconductors gain a few big production orders and then "vanish into the night". Somehow, that one feels like it. It's limiting, but I want to stick with "mainstream" components, whether or not they are vintage parts.
The other issue, more subtle, as @EmilyJane pointed out, CMOS parts won't interface directly with the NMOS processors and I/O chips. I'm trying to avoid bus buffering unless I absolutely have to do it.
Some updates. Asia is closed for two weeks. So only about half my stuff reached me. I started putting together a battery backed SRAM module, the LY62256SL TSSOP on a DIP adapter, *CE pin bent up so it can be driven by the DS1210. Unfortunately the DS1210s haven't arrived yet.
I realized the NMOS to CMOS level problem applies to the SRAM, so a minimal NMOS system may need bus buffers. A long time ago, I saw a kludge to fix that, pull up resistors on all the NMOS lines. I spent some time investigating the wisdom of that, but didn't find any conclusive answer. But there is the chance of interfacing the CMOS SRAM with the CMOS processors (such as HD6309), only those also haven't arrived yet.
Another thing, I had a notion to make a dead bug construction, one off just for fun. Maybe paint the components with acrylics!
Also ordered some HD63C03YP, that has on board peripherals including a UART... shows some promise for a limited system, it will run MC6800 machine code AFAIK. It's a 64 pin chip.
The LY62256SL Seems to have TTL compatible I/O. Am I looking at the data sheet wrong?
Sounds like that might have have a rise/fall time issue, youād have to check it on a scope.
Now that sounds like fun!
The 63C03 certainly has a full set of peripherals. The multiplexed address/data buses would add a little to program storage interface and 128 bytes of ram wouldnāt go far but nonetheless a good 6800 project.
Iām just about done breadboarding my EEPROM programmer. Just have the address bus to punch down. Not pretty but it doesnāt have to operate at a very high speed so should work okay for programming.
I have to admit, I didn't know it was so easy to blow an EEPROM. I might have to think that one over. But it would be good to learn how to do that. Only thing I can gain from the NVRAM approach, is one memory IC to rule them all. Thanks for the heads up about the RAM I/O, I will look into that.
Iāve yet to program any bytes but I donāt anticipate any problems. Iāll report back on my results. It works just like ROM on reads and writes just like RAM except for the 10ms write time.
NVRAM certainly cuts the chip count.
(For the pedants,no decoupling on schematic but on breadboard.
)
Runs off 3V3 and 5V so I can program it easily with an RPi Pico.
Edit: I rewired the PIOs to take advantage of the 16 bit writes and added the level shifters to program from the Pico. I couldn't find any wide supply range EEPROMs like the RAM I have.
2 Likes
I just noticed, PIO7A & PIO7B are outputs only on the MCP23017 which is fine for my application but is weird and should be noted for other applications of the chip.
Now I'm wondering about keying in a minimal bootloader using switches. Basic idea is a PDP-like front panel, with a bus access circuit. I'd enter a basic program that would just:
- initialize the ACIA
- accept a fixed number of bytes that come in
- execute them
That would allow you to load a monitor, or at least an S-record loader. After that, enable write protection on high memory, the monitor will remain there via NVRAM backup, ready to boot every time after that...
