MZ-80B Power Supply Repair

gregctaylor
Posts: 27
Joined: Tue May 29, 2018 5:38 pm

MZ-80B Power Supply Repair

Post by gregctaylor »

I need to get into my MZ-80B to fix the power supply (I suspect the tantalum capacitors have blown but have electrolytic caps and ICs ready for repair) but cannot fathom how to get in the case. I thought it would be like the K and A with two screws securing the front then hinge at the back but does not look like it. I have removed around 8 screws around the base but still nothing seems to want to move. There are two things that look like clips but nothing slides to allow them to clear the slot. There must be a trick but I am not seeing it. Tried the exploded diagram in the service manual but it was not helpful.

Greg
gregctaylor
Posts: 27
Joined: Tue May 29, 2018 5:38 pm

Re: MZ-80B Power Supply Repair

Post by gregctaylor »

Well that was easy! To help anyone else, it hinges at the back! There are two large screws round the back on the plastic part, undo those and lift. I am always amazed at the build quality of these computers.

Now to the power supply which looks like a few well built connectors then will have to explore further. Only criticism is that they did not leave as much room to work in as the K and A.

Greg
gregctaylor
Posts: 27
Joined: Tue May 29, 2018 5:38 pm

Re: MZ-80B Power Supply Repair

Post by gregctaylor »

Actually, access is fairly good. First remove the back panel as the power supply is going to slide out that way. At the very back near the power switch there is a small metal cover which has to be removed. Straighten up the two small tags and then they slide through the two slots. Now you can see four large screws around the big transformer cover. Remove these and the power supply comes back and up a bit and slides out. Normally, it is the two mains voltage capacitors which fry (0.047 mfd, 250V) and these are on a small board between the on/off switch and the transformer. I found it easier to slide out a bit by unscrewing the power switch. Now just waiting for the new capacitors to arrive.....
gregctaylor
Posts: 27
Joined: Tue May 29, 2018 5:38 pm

Re: MZ-80B Power Supply Repair

Post by gregctaylor »

All back and working but it does not load from cassette. I choose C, the tape starts loading but nothing happens. The tapes came with my old 80K. Does anyone know if the 80B uses a different tape format?

Greg
hlide
Posts: 681
Joined: Thu Jan 25, 2018 9:31 pm

Re: MZ-80B Power Supply Repair

Post by hlide »

Yes, the speed is 1800 bauds for MZ-80B and 1200 bauds elsewhere. There are small differences in timings in the encoding too (shorter GAP). You can find this information here.
gregctaylor
Posts: 27
Joined: Tue May 29, 2018 5:38 pm

Re: MZ-80B Power Supply Repair

Post by gregctaylor »

Thanks for that. It explains things perfectly. Great to see the 80B working again. I now have the 80A PSU apart to change the same smoothing capacitors. The K continues to be a problem. I have reduced it back to a 16K machine and still no luck. It seems to be writing a pattern to all the RAM then, I assume, failing the checksum on read - normally a sticky bit. I am ordering some spare RAM chips for the first 16K (MB8116E) and then will swap around to see if I can trace the fault. I did try your ram check program but the 80K will not try to load a program from startup. I assume there is some basic POST procedure and it is getting stuck there.
hlide
Posts: 681
Joined: Thu Jan 25, 2018 9:31 pm

Re: MZ-80B Power Supply Repair

Post by hlide »

I made a program RAMTEST to determine which row and which chip has an issue: the program is loaded into VRAM (so you'll see some weird characters at first lines) and will check some patterns (D0 chip: 01 - D1 chip: 02 - D2 chip: 04 - D3 chip: 08 - D4 chip: 10 - D5 chip: 20 - D6 chip: 40 - D7 chip: 80).

Patterns tested in memory are the inverted bits of 01, 02, 04, 08, 10, 20, 40, 80 because when a chip is missing, it gives a bit 1 as a result, so we check only one bit 0 per chip.

In the examples, I suppose you should have 48KB (3 rows of 16KB).

Some examples of faulty chips which halt the program with a message in the center of the display:
- 0400503F: pattern 04 ->D2 testing, read byte 00 says D2 is faulty because its chip is giving 1 while it should give 0. Faulty address row: (503Fh-1000h)/16KB = 1.
- 0111905C: pattern 01 ->D0 testing, read byte 11 says D0 is ok but D4 is faulty because its chip is giving 0 while it should give 1, so D1 is faulty with a bit stuck at 0. Faulty address row: (905Ch-1000h) / 16KB = 2;

The major issue is how to load this program.
viewtopic.php?f=6&t=180
viewtopic.php?f=17&t=171

EDIT: I didn't read your post properly.

Can you elaborate how you get the program and how you load it? give the more details so I could understand the issue. And more important, what is the memory configuration (CS1 and CS2)?
- 16KB - ROW I: 16KB - ROW II: empty - ROW III: empty.
- 20KB - ROW I: 16KB - ROW II: 4KB - ROW III: empty.
- 24KB - ROW I: 16KB - ROW II: 4KB - ROW III: 4KB.
- 32KB - ROW I: 16KB - ROW II: 16KB - ROW III: empty.
- 36KB - ROW I: 16KB - ROW II: 16KB - ROW III: 4KB.
- 48KB - ROW I: 16KB - ROW II: 16KB - ROW III: 16KB.
Last edited by hlide on Tue Jun 12, 2018 9:18 am, edited 1 time in total.
gregctaylor
Posts: 27
Joined: Tue May 29, 2018 5:38 pm

Re: MZ-80B Power Supply Repair

Post by gregctaylor »

I am beginning to wonder if it might be the monitor ROM (EPROM?) at fault. I have reduced it to a 16K version and no difference. The monitor displays a very regular display (two characters ABABABABAB... - not A and B) then nothing, no option to press C (and I have tried pressing it anyway with no result). I have heard of EPROMs needing re-burning after many years, might this be the problem? I don't have an EPROM burner to try.

It looks like VRAM is fine as the display is perfect. I just cannot understand what else might stop the monitor program completing its loading as I assume the memory test is in ROM?

Thanks for your kind help.

Incidentally, I plan to sell on the 80A and 80B when fully fixed and tested. eBay is the obvious route though there are some idiots who expect perfection from a nearly 40 year old computer. I will post on the forum as well.
gregctaylor
Posts: 27
Joined: Tue May 29, 2018 5:38 pm

Re: MZ-80B Power Supply Repair

Post by gregctaylor »

Just noticed the last part. It ws 16K in all 3 rows but now I have 16K in row I then 2 and 3 are empty to try that. No change.
hlide
Posts: 681
Joined: Thu Jan 25, 2018 9:31 pm

Re: MZ-80B Power Supply Repair

Post by hlide »

For the program to be loaded, I suppose you got it recorded on a tape. If not and you have an MZ-700 with a working tape recorder, there is a way:
enter the bytes with M command and save them with S command in a tape. Then you should be able to play it on MZ-80 K.

There is a small complication: if your chips in row I has an issue in the range of 0x1000 and 0x11FF. The monitor ROM uses them as system variables and the program relies on them to set a cursor. Worse, the loading process when issuing LOAD command also needs that range to work.

If the issue is around that range, you won't be able to load a program anyway. RAMTEST is usable if and only if there is no issue in that range.

Unlike MZ-700, monitor ROM has no way to enter bytes in memory.
Post Reply