|
|
Post by Dragget on Feb 25, 2016 9:14:33 GMT
I just replaced a bad chip on my 128DCR motherboard to get my internal 1571 working again, so I thought I'd load a few programs and format a disk or two to make sure my fix was successful. Here is the result: Post on C128.comI'm cross-posting here in case someone has any suggestions on what I need to do to fix this. For some reason, this forum won't allow PhotoBucket embeds, so follow the link for the photos I took to document the issue. |
|
|
|
Post by hydrophilic on Feb 27, 2016 2:10:48 GMT
Wow! That is totally warped!!
It just boggles my mind thinking about how the image would warped in 6502 (native) mode, but be glitch-free in CP/M mode.
But hey, the image of Word Star looks weird... I assume this was from CP/M mode? Well maybe it is suppose to look like that?
Assuming the "Word Star" image is a warped CP/M image, then we know it isn't a problem (directly) with either CPU (8502/Z80). In which case, it could just be a bad VDC... Bil Herd reported tearing of VDC images with flaky chips in the debut CES. If you have another C128, try swapping VDC chips... that is the easy/only thing I can think to try...
Good luck (and keep us posted)!
|
|
|
|
Post by Dragget on Feb 27, 2016 8:32:42 GMT
In the Photobucket album, I commented on why I thought that WordStar screen looked weird in the photo. Maybe I should have included that in the post. Basically, I took that one while the screen was scrolling. The distortion wasn't visible to the naked eye: it only showed up in the photo.
Good idea about swapping out the VDC. The one in my C128D is socketed, hopefully the one in my flat 128 is also.
|
|
|
|
Post by Dragget on Feb 28, 2016 3:55:43 GMT
Well, it seems the VDC chip in my flat 128 doesn't work in the 128DCR. I swapped it in and could get no display on the 80 column screen and garbage on the 40 column screen. The one from the C128DCR has MOS 8568 printed on the top and the one from my C128 has MOS 8563R9 printed on it. They both have another MOS chip with 8564 on it, but apparently, these two are not compatible with each other: en.wikipedia.org/wiki/MOS_Technology_8568So much for that plan. The distortion only seems to occur if there is disk access while text is being drawn on the screen. I input a short BASIC program to print some text on the screen in a loop and let it run for awhile, but there was no visible distortion. As soon as I load a disk and access it, that's when I am seeing this problem crop up. |
|
|
|
Post by Dragget on Mar 2, 2016 6:43:05 GMT
So I dug out my C128/C128D/1571 diagnostic utility disk and manual (provided by Commodore), and ran the 80 column VDC/memory test, and it reports that the memory and the VDC ICs are good. I ran it for 4 passes with no errors reported. My only theory at this point is that while a disk is being accessed, noise or interference is somehow being transmitted from the disk subsystem, causing the distortion on the display. I had to run the diagnostic in 40 column mode because it blanks the 80 column display while it's running and displays the results on the 40 column display. I also ran the logic tests on the internal drive (PASS) and the read/write tests that don't require a special analog test disk (PASS). This is beyond me...
|
|
|
|
Post by hydrophilic on Mar 2, 2016 12:17:57 GMT
Right, the 8563 and 8568 are not pin-compatible (can't swap them). Well that is a lesson for us both!
I really don't understand how disk-access would foul-up the video display!! So I am like you ... confused!
Going into imagination-mode, perhaps something is wrong with disk drive, such that it draws too much power while the disk is spinning (and/or the head is moving)? If it draws more power than can be supplied by AC->DC PSU, then voltage levels may drop... for example the 5V line might drop to 4V. If this did happen, you should expect a variety of "random" failures. Just a wild idea...
Anyway, good luck, and let us know if you find out anything new!
|
|
|
|
Post by Dragget on Mar 2, 2016 18:09:54 GMT
Well, I speculated on that myself, so I hooked up my 1571, swapped the device numbers so that the internal drive was unit 9 and booted GEOS. Same problem popped up: screen distorts when the external drive is accessing now. Since the 1571 has its own PSU, it can't be that. Also, if it was a PSU issue, the distortion should appear on the 40 column screen too and in all modes includeing CP/M, right?
|
|
|
|
Post by designcraft on Mar 2, 2016 21:16:54 GMT
I just replaced a bad chip on my 128DCR motherboard to get my internal 1571 working again, so I thought I'd load a few programs and format a disk or two to make sure my fix was successful. Here is the result: Post on C128.comI'm cross-posting here in case someone has any suggestions on what I need to do to fix this. For some reason, this forum won't allow PhotoBucket embeds, so follow the link for the photos I took to document the issue. How can i test the drive in my 128dcr? I have a problem, but only in 128 mode what i can't understand why: c-128.freeforums.net/thread/296/c128-dcr-strange-drive-problemHow can it work in C64 mode flawlessly if 128 mode all i got is fault? Which ICs do you change, or how can i investigate? thanks... |
|
|
|
Post by hydrophilic on Mar 4, 2016 12:10:42 GMT
Well your symptoms point to one thing... the CIA. In particular, CIA #1 at $DC00 (the schematics don't label which U# this is!) ... it should be the one closest to the keyboard / joysticks.
The reason I say this is because only 128 (and CP/M) modes use fast-serial hardware. C64 mode does not (for obvious compatibility reasons).
Well, it may not be the CIA chip itself which is at fault, it could be some of the supporting circuitry which enables fast-serial. For example, the SP pin (#39) of that CIA is connected to U58 (a 74LS03 Quad-NAND) which itself is connected to U16 (a 74LS14 Hex-Inverter with Hysteresis).... and also connects to U37 (a 7406 Hex-Inverter).... and eventually is buffered by U60 (a 7407 Hex-Buffer). Another example, the CNT pin (#40) of that CIA is connected to U58 (again) and U37 (again)... and the U58 is fed by U16 (again).
Finally the direction (fast-serial input/output) is determined by U8 (a 74LS08 Quad-AND)... which is based on the lines /FSDIR (pin# 44) and 'MS3' (pin# 47) of the MMU (U7).
Sorry if that is confusing! If simply swapping CIA #1 doesn't fix the problem, then I would suspect a problem in the support chips (which are a bit more difficult to replace!). The U60 Hex-Buffer seems the one most closely related to output signaling, and something I would suspect of failure... beyond that, U37 and U58 are used multiple times for fast-serial (so one, or both, may need to be replaced).
DISCLAIMER... this is base on prior knowledge of playing with fast-serial programming, and looking at the schematic in the C128 Programmer's Reference Guide... I did *not* actually open my 'flat' C128 to confirm anything, and I do not own a D(CR) model at all!
Anyway, I hope this may be helpful...
|
|
|
|
Post by Dragget on Mar 4, 2016 21:27:54 GMT
Well your symptoms point to one thing... the CIA. In particular, CIA #1 at $DC00 (the schematics don't label which U# this is!) ... it should be the one closest to the keyboard / joysticks. So in the image below, would that be the one labeled CIA2? (photo is from Ray Carlsen's troubleshooting site. CIA1 is way over on the upper LH side in the photo) Too bad mine aren't socketed like the ones in the photo. I will have to see if I can remove it without damaging the chip, then put in a socket and a replacement chip and see if that solves the problem. In any event, thanks for the information... it's at least a place to start.  I'll post here with results once I have tested it.  |
|
I'll post here with results once I have tested it.