Diagnostic Cart

[nbla000]

There is 1K of RAM installed with ability to add an additional 2K.

The installed RAM is wired up to appear at /RAM1 location.

So this is the reason because in vice works with Vic+3K...

So, ROM image could be included on MegaCart and people may build these test pieces as separate little projects. They should not be difficult to build.

Ok finally i know what i must do this evening/night

For a simply and safe work, i will add it in ROM+3K, i think that is not required to select another memory configurations for the user...

[Leeeeee]

I thought I would have a look at the cart ROM, specifically this one and see if I could work out the connections for the IEEE port. I haven't done that yet but I have found some things about the cart.

It outputs a value to the I/O-2 area if it fails, this suggests that there may have been plans for a cart with a POST display of some sort, perhaps a single digit seven segment LED or similar, so you could still get some idea of what was working even if the video wasn't.

It polls the I/O-3 area during one of the IRQ routines and stops if one bit is set, I'm guessing this was for a pause switch or suchlike.

I think this cart image is for the later VIC as there only seems to be reference to two RAM chips apart from the first 1K and colour RAM, there is other code that could be for more but the message pointers reference other code and not messages.

The code is a mess, there are at least three clear screen routines two of which get called in succession. There are at least three ways it outputs "OK" to the screen. There is a basic screen output routine that can handle cursor x,y positioning, scrolling etc. but this is bypassed for a lot of the screen output. There are registers set and then the values never used before they are set again and the NMI routine trashes the X register every time because someone wrote TAX and not TXA in the code to save the registers.

Subroutines are used before the RAM at $0100 to $01FF is tested.

At one point the code tests to see if the byte at $F75D is $4C, none of the VIC ROM images I have have $4C at this address.

There is a days/hours/minutes/seconds/ticks counter that rolls over from 98/23/59/59/59 to 00/00/00/00/01. I seem to remember that the VIC kernal does something similar in that it counts one jiffy too many per day.

When a problem is discovered the code usually dead ends in a loop.

There are a few areas that seem to be messages that aren't referenced. There's a whole table of things from $ACB1 on but the only bit that seems to be used is the word "BAD" at the beginning.

That's it for now, I'm sure there's more. 8^)=

Lee.

[nbla000]

mhh... then i need to investigate on MegaCart with the file image before to change roms...

[nbla000]

I've tested the diagnostic cart and these are results....

The VC-20 diagnostic cart starts its tests and after a while it reports an error on Serial BUS, this is the screen:

Code: Select all

RAM TEST  OK
COLOR RAM TEST: OK
ROMCHECK FF
ROMCHECK E0
ROMCHECK C0

KEYBOARD TEST: OK
SERIAL BUS   BAD

00 DAYS    00:00:02

The test stops here with a White/Blue flashing screen with cyan border.

I've tested it with all my 4 vics and with Eslapion Super Expander with 3K option too, i always got the same test error.

I've tested it with Flexi-Cart too but since the diagnostic cart requires 1K in /RAM1, tests finish on ROMCHECK with an error.

Since my serial bus works very well, talking about Brian, we strongly suspect that some accessory like these are required for tests proper works, maybe a connector to the serial port too, who know ?

As things stand now and considering that the diagnostic cart, like Leeeeee said, writes in IO2/IO3 where MegaCart preferences and music are stored, even if preferences seems OK and music too we cannot add this cart.

Btw we may consider it for a next nice project all inclusive, connectors too.

Sorry for this news btw Brent had a nice idea and with Carlsson collaboration we want to add a last surprise for all of you but in this case not an easter egg, it will be public

[channelmaniac]

Hi everyone,

Carlsson from the vintage-computer.com forums pointed me here and I must say it was very helpful as I was trying to get a diagnostic cart built. Now that my account has been activated I can tell you how I built it.

I had a Kantronics cart board that had a 2732 chip on it but had the extra holes to put a 2764 chip in. I took the 2732 out and put a 6264 SRAM chip in. I cut the trace to pin 31 (it was tied high) and connected that to the RW signal. Next I wired pin 20 to the RAM 1 line. (Pin 22 was grounded by moving a jumper and pin 30 was already tied high)

After that I pent pin 18 out on a programmed 2732 and soldered the chip on top of the 6264 - piggy back style. Pin 18 of the EPROM was soldered to Bank 5.

It's an 8k chip but is only selected when the RAM 1 bank is needed so only 1k is being used.

The cart works well! Now I just need to build the loopback connectors.

Raymond

[nbla000]

Hi everyone

Happy to see you on Denial too....

The cart works well! Now I just need to build the loopback connectors.

What does mean "works" ? do you pass the Serial test ? loopback connectors are thus required ?

[channelmaniac]

Thanks!

What does mean "works" ? do you pass the Serial test ? loopback connectors are thus required ?

Works in that it tests RAM and Color RAM. No, it doesn't pass the serial test. I have to build the loopback plugs for that to work.

I'm guessing that in the pics where the bare wires are showing that those bare wires should be connected to the DIN connector for the serial port. There are a LOT of combinations to work through to find which ones go where...

I wonder if anyone that knows assembly could disassemble it and tell what wires would connect to which pins based on what is being output and input. That sure would help me build the loopback for the serial port.

RJ

[Leeeeee]

I wonder if anyone that knows assembly could disassemble it and tell what wires would connect to which pins based on what is being output and input. That sure would help me build the loopback for the serial port.

I've done just that and will be adding a page to my site about it soon.

I've also made a new version of the cart that doesn't use the I/O blocks, doesn't need the expansion RAM, fixes bugs in some of the tests and tests expansion RAM in RAM1, RAM2, RAM3, BLK1, BLK2 and BLK3.

BTW, the pictures of the test harness don't match the program.

Lee.

[nbla000]

I've also made a new version of the cart that doesn't use the I/O blocks, doesn't need the expansion RAM, fixes bugs in some of the tests and tests expansion RAM in RAM1, RAM2, RAM3, BLK1, BLK2 and BLK3.

Very interesting work, is still required the loopback plugs for serial test or you may use a simple powered on disk drive connected to the serial port for this purpose ?

[hawk]

I've also made a new version of the cart that doesn't use the I/O blocks, doesn't need the expansion RAM, fixes bugs in some of the tests and tests expansion RAM in RAM1, RAM2, RAM3, BLK1, BLK2 and BLK3.

Presumably this makes the assumption that the internal RAM of the VIC is good. The reason that the cart would include RAM is because it's not making any assumptions about the validity of the internal RAM for the correct operation of the tests.

BTW, the pictures of the test harness don't match the program.

Have you come up with what the test harness should look like?

Are the photos of the test harness for a different diagnostic cart?

[Leeeeee]

Presumably this makes the assumption that the internal RAM of the VIC is good.

No it doesn't.

The reason that the cart would include RAM is because it's not making any assumptions about the validity of the internal RAM for the correct operation of the tests.

Actually the original disgnostic cart assumes that the stack page works before it tests it.

Have you come up with what the test harness should look like?

I think so but I haven't tested it yet, I need to find a spare serial port plug.

Are the photos of the test harness for a different diagnostic cart?

I don't know. They seem to be nearly right so they could be something someone made by trial and error.

Lee.

[Leeeeee]

Very interesting work, is still required the loopback plugs for serial test

Yes a serial port plug is still needed.

Lee.

[channelmaniac]

Leeeeee,

Have you finished pinning out the loopback plugs? I checked your website and didn't find anything on it.

Raymond

[gungahlin]

Does this link help?
http://cbm.ko2000.nu/firmware/misc/index.html

[channelmaniac]

I already have that but it doesn't help 100% as it doesn't have the pinout for the silly serial port.