Unhappy AIM-65

[Pamphonica]

Does anyone here (preferably local) run a working AIM-65?

I posted a wanted ad recently for chips to help diagnose this unhappy AIM-65. https://www.vintage-radio.net/forum/...d.php?t=200027

This machine was a gift, complete with plastic case and power supply, and a heap of manuals. It was stored in a barn but seems not to have suffered much at all. It's so sweet that I would definitely like to get it up and running and doing something useful. Maybe running some of my collection of HP X-Y plotters.

I started by printing out large copies of the circuit diagram, and downloading a tech manual. Many decades ago I was a microsystems designer (6800, 6802, 6502, X80), so this is quite a nice return to old haunts.

On power up the display shows nothing, or sometimes the odd spurious character on start up. I have one spare 6502 (marked "?" but seems to run) and some spare 2114, and a couple of 6821 (should work instead of 6520). So I started by removing as many LSI chips as I could to see if I could get a simple config going. Just 6502, 1K RAM, Monitor ROMs and keyboard PIA. I also checked that the board is getting a good supply (5V almost exactly).

Despite some swapping around, I still got no display. I even bought some spare displays as I found that at least one was faulty (using a breadboard). I then broke out the big guns. A 16-bit logic analyser (built in to my Keysight scope). It has a good sized memory so can trigger and then scroll through what happens afterwards.

I could see bus activity so triggered the logic analyser on the display side of the display PIA. No sign of any write pulses at all, so not promising. I also armed myself at this point with the disassembled Monitor listing so I knew what to expect on startup (eg a "Rockwell AIM-65" message on the display). Then triggered on the (not)CSAC strobe which controls access to the display PIA, and monitored the data bus and various other bits like A0, A1, R/W, RST, clock. You can see 3 pairs of two writes as it sets up the PIA, then a series of 20 sets of 5 writes (that looks like the 20-characters going out, per the assembly listing).

However, the data being written does not have the correct ORA bit set and reset to strobe the write pulse to the displays. Also, the data register just has 00 in it so just null characters would be written.
So - I am waiting for an adapter card from China (with other stuff) to allow me to read the monitor ROMs. I am also waiting for a custom designed pcb to make my Polar "uP Bus Tester" speak 6502 instead of 8085. (This device, which cost me £5 at Audiojumble, will do proper tests of RAM, ROM etc and allow writes to specific locations, and does dis-assembly, so I have high hopes!).

Anyway, after all this rambling, if anyone has any ideas, or could even be brave anough to let me put my chips in a working AIM-65 to see what does work, I would be most grateful. It is probably too much to hope that I find any more LSI chips to try, but who knows?

All comments and help very gratefully received!

-Jeremy

[Pamphonica]

And by the way, does anyone know if the data from the two 2332 4Kx8 Monitor ROMs is dumped anywhere online? Or does anyone have binary or hex dumps of good ROMs?
thanks
-Jeremy

[SiriusHardware]

Hi Jeremy - this is a better place to roam far and wide on this subject.

On the subject of ROMs, I believe that a couple of the regulars here have AIM65s and one of them might be kind enough to dump the contents of the ROMs from a working system.

Although you've already ordered your 27xx to 2332 adaptors, it may eventually be possible to replace the 2332 ROMs directly with 2532 (not 2732) EPROMs which were made for a while by Texas Instruments and Hitachi - they are often usable as drop-in replacements for the original 24 pin ROMs used in older systems -if your programmer supports 2532 EPROM devices you should be able to read your ROMs by selecting 2532 and reading your 2332s as though they are 2532s.

In your 'wanted' thread you mentioned that you managed to get some more life out of the system by replacing the original 6502 with another which was marked '?' which obviously means that either you or someone else must have had doubts about it in the past - you may in fact have replaced a completely duff 6502 with a semi-duff 6502, so I think you are going to have to get another one from somewhere else as a 'casting vote'.

Can you point to where you found the circuit diagrams?

[Phil__G]

Mouser have brand new DIP 65C02's in stock, a tenner a throw. Worth it to know that you're starting out with a good chip
https://www.mouser.co.uk/ProductDeta...o9AEKH5w%3D%3D

[SiriusHardware]

I realised that Jeremy has the original paper manuals, so I went looking for digital ones and found them here on this very interesting looking site dedicated to SBCs:-

http://retro.hansotten.nl/6502-sbc/a...-65-manuals-2/

In the short time I spent looking I still didn't find the AIM-65 diagram but I am sure it will be there somewhere.

[SiriusHardware]

I still didn't find a one-sheet circuit diagram but section 7 of the User Manual (available via the link above) has a walk-through description of the hardware - each sub-part has its own technical description including the circuit diagram of that section.

[Pamphonica]

Retro computing has copies of circuit diagrams for all the variants.
http://retro.hansotten.nl/6502-sbc/a...m-65-hardware/

[Pamphonica]

Originally Posted by Phil__G

Mouser have brand new DIP 65C02's in stock, a tenner a throw. Worth it to know that you're starting out with a good chip
https://www.mouser.co.uk/ProductDeta...o9AEKH5w%3D%3D

Is a 65C02 fine in place of a non-C 6502? I have found one on an old board.

[SiriusHardware]

I just found that Hansotten site myself, fantastic site although sometimes quite hard to find exactly what you are looking for there.

I tried and failed (so far) to find ROM images there so I have dropped an email to the site owner to ask if he does have them there somewhere.

[SiriusHardware]

Now that Jeremy has pointed me in the right direction, a direct link to the complete circuit diagram:-

When the diagram is open, click on it to zoom into it.

http://retro.hansotten.nl/wp-content...it-diagram.jpg

[SiriusHardware]

And I've just had a very quick response from Hans pointing me to this page for the ROMs for the AIM65:-

http://retro.hansotten.nl/6502-sbc/a...im-65-software

I would assume that the following files listed there:-

Monitor ROM R3222
Monitor ROM R3222-11

and possibly

Assembler ROM R3224

-are the ones Jeremy needs.

[SiriusHardware]

To attempt to answer the question about 65C02 vs. 6502, have a look here:-

http://forum.6502.org/viewtopic.php?t=124

[Michael Haardt]

In case you are wondering if something seriously weird is going on, you can remove the 6502 and put an adapter wired to an Arduino Mega in the socket. That allows you to try arbitrary bus cycles to use RAM/ROM or IO right from the Arduino IDE.

It is also worth checking if the logic analyzer sees the expected code being read during the first few executed cycles.

[Pamphonica]

Small update. I lashed up an adapter that let me dump the Monitor ROMs (looking like a 2764 with the top address bit as zero) and compare them to the binaries from Hans' site. The ROMS are identical to Hans' binaries so for now I'll assume I have a RAM problem or the fault is elsewhere.
The program was clearly running, as I could see the display PIA write sequence quite clearly.
Off to buy or make a 2114 RAM tester. Onwards and upwards!

[SiriusHardware]

I'm glad the ROM images allowed you to rule out problems with the ROMs - even though you seem to be able to see the 6502 running code on your analyser it only takes one or two dud bits, therefore broken instructions, in the firmware to make everything go haywire beyond a certain point so a byte for byte comparison, as you've done, was what you really needed.

A 2114 tester, you say?

https://github.com/Danjovic/2114-Tester

Based on an Arduino Uno - it directly addresses the ports of the underlying micro on the Uno so that it can do parallel writes to and reads from the ports in order to run the tests at a realistic speed. I have a version of that code which I have modified to test 2111 SRAMs, it works quite well.

Also if you have a high-end device programmer, does it have a chip testing function? While all the ones which do can test 7400 and 4000 series logic at a minimum, some of them also have test routines for common RAM ICs as well.

[Pamphonica]

My TL866II Plus EPROM programmer does test SRAM but not this old, only 6xxx series.
I worked out that a ready made £19 SRAM tester from EBay based on a UNO was a good bet, as I can reprogram it to do other devices if required. If it’s not the RAM then I need to start checking the glue logic. The decodes that drive devices on the bus look OK, but there may be timing issues. I still don’t know why the display writes are wrong, and it looks as if the data sent is all “00”.

[SiriusHardware]

How are you with writing 6502 code yourself? You could replace the boot ROM with an EPROM containing test code specifically to exercise the areas you are interested in, such as the display section.

[Pamphonica]

No need to write code. The Polar uP Bus Tester does all that clever stuff, and allows memory mapping and modification. Also disassembly of ROM contents. I’m just waiting for the new personality plugboards from China due in a few days. As it stands, it “speaks” 8085 not 6502!

[River25]

Hi,

I forget the actual chip number but go check the logic outputs on the ROM and RAM select chips. I have repaired a few AIM65 and these chips have failed more than once, resulting in a dead AIM65.

rgds
River

PS: As for 65C02 vs 6502, I would replace the 6502 with another older NMOS 6502 or R6502 (Rockwell NMOS version) rather than the 65C02. I am not sure if I have a 65C02 but will check. If I do will swap it into one of my AIM65 units to see how it goes, as I think (IIRC) there were some slight signal variations which may cause issues.

[Pamphonica]

The chip selects look OK, but they could have timing issues. I am waiting till I can get the bus exerciser working later this week. I can then work out if RAM, ROM and peripherals are doing what they should. If I need to substitute some of the glue chips, I will put in quality turned pin sockets in each location, starting with the select decode chips, of which I have a plentiful spares stock. I have tried the "?" marked 6502 and the 65C02 as well as the original. all give exactly the same result and all show the characteristic write cycles to the display PIA, so are following the monitor correctly. So the instructions seem to be correct but the data is wrong. That smacks of a RAM fault, as there is a lot of use of the stack in RAM.