Fehlersuche 16kb-SRAM Karte SKS KISS 2116

    CAD, first part:



    This corresponds to logic at the edge of the board, opposite side from the connector. Enable pin number 6 at the decoders is connected to the rest of the logic located near the connector. Not gate (pins 8 and 9) is unused. This part is coming after lunch.


    Notation is weird if the letters are to be considered... First because in the case of gates another letter is appended. Second, "U" must be prefixed, as in CBM documentation because ICs at row "C" would have a reference conflicting with the capacitors.


    I imagine there is a lot of room for improvement. If you see anything wrong, or you have any idea that could improve it don't hesitate to say it. ;)


    Regards,

    Jaume

    When I tried to list all retro systems I have at home, the "The message is too long, must be under 500 characters" error appears! :lol:

    The logic is done. The remaining not gates are there to negate addresses. I hope I got it wrong and those gtes at the read and write signals aren't like I think they are. =O



    I have many obsolete components in my custom libraries. The 2114 memories are present but thee 8226 are missing and I have to create them.


    Regards,

    Jaume

    When I tried to list all retro systems I have at home, the "The message is too long, must be under 500 characters" error appears! :lol:

    Gefällt mir 1

    Hello Jaume,

    basically the columns or rows are with numbers and then the others are labeled with letters.


    Unfortunately I only have the complete switching documents for the "sks displaycontroller" as the axes are called.


    I have the entire circuit diagrams for the DS 2069 (Dr. Ing. HELL) and clearly the construction cards including the fully functioning display unit.

    The basis is a development stage from sks to HELL and TA with the MC80-BUS.

    But some revisions are possible SIGNAL may have changed.


    Would the circuit diagrams per construction card from DS 2069 be helpful for projects like this now?


    Best wishes

    helwie44 †

    Hello helwie44 †


    Yes, I thought the floppy controller unit file would have also the position diagram... but no. U + Alphabetical character + Number + Gate (if required) is a notation found in the schematics for PETs. I don't know if the DS schematics would be helpful in this case... every single manufacturer used its own reference system (TA used just numbers for the P3's ICs). Still, I would be interested in having a look on them, if possible.


    If the reference methodology is not understandable or clear and is found to be confusing, we could always revert to the standard notation. I tried to resurrect the old one just for historical purposes.


    Thank you very much,

    Jaume

    When I tried to list all retro systems I have at home, the "The message is too long, must be under 500 characters" error appears! :lol:

    Gefällt mir 1

    Hi Jaume,


    Good Job!

    Zitat von jlopez

    The logic is done. The remaining not gates are there to negate addresses. I hope I got it wrong and those gtes at the read and write signals aren't like I think they are. =O

    I did not understand, what you mean with "wrong" and the read write signals?


    I took a look on the logic diagram.


    Following points are suspect to me (I did not cross check with the PCB, because I am on holyday and my possibilities are very limited):


    UE6C pin #9 and #10 are connected in your diagram. Please check again.

    Plausibly for me would be a connection from UE6B/#6 with UE6C/#9 only!

    UE6C/#8 will deliver the signal to control the buffers 8226.


    The outputs of UE6B and C are connected to parallel (to get more driving capacity). Are they connected to the RAMs pin #10 (write enable)?


    A word to the reference numbers on the PCB: did you see that some copper letters are on the front side of the PCB: B D F G ?


    Kind regards

    Peter

    Zitat von PAW

    A word to the reference numbers on the PCB: did you see that some copper letters are on the front side of the PCB: B D F G ?

    jes - i have the 16kB s-ramcards (sks , manufactory date 29.01.1980).


    These partial designations on the component part are only available as auxiliary information - so that there is better space for the conductor tracks.


    The full designation of A B C D E F G is available on the back.


    Zitat von PAW

    A word to the reference numbers on the PCB: did you see that some copper letters are on the front side of the PCB: B D F G ?


    Hello!


    I did not miss those letters! But what I have missed completely is the numeration... :wand: Yes, it is present and albeit being present and just in front of me I wasn't aware of it. And got it wrong!:D Now I must change all references in my schematics...



    I'll check them after renumbering... But the single connection between two '04 and the '00 surprisingly are as stated, the not gates at B also provide the negation coming from the '86 xor.


    Regards,

    Jaume

    When I tried to list all retro systems I have at home, the "The message is too long, must be under 500 characters" error appears! :lol:

    DS 2069 part diagram


    I only added the circuit diagram of the DS2069 CPU_3 and for the 128kB memory below - as an example.


    A "circuit diagram" with the component positions on the board is easier to recognize.


    I hope - there won't be much more to a labeling system?

    Decide on a system that suits you best.

    Good luck and until then

    @ helwie44

    Sorry for the confusion. I found the original and will stick to it.

    When I tried to list all retro systems I have at home, the "The message is too long, must be under 500 characters" error appears! :lol:

    Hi Jaume,

    Zitat von PAW

    UE6C pin #9 and #10 are connected in your diagram. Please check again.

    Plausibly for me would be a connection from UE6B/#6 with UE6C/#9 only!

    I checked the connection against the PCB (there is no connection). See the following PDF:


    Correction of Diagram.pdf


    This is now plausible!


    The reference numbers of the chips in my document should match the new numbering. (I hope so.)


    Kind regards


    Peter

    Zitat von Toshi

    jlopez : Nice work!

    What is your intention here? Just to understand and to reverse engineer the circuit or do you plan to create new PCBs ?


    Regards

    Stephan

    My goals are to understand the circuit, generate schematics and obtain any information that could speed up diagnostics and repairs. Maybe it's late for you this time, but maybe next time someone (including yourself) comes around with a KISS with faulty memory this process may save his/her time and reduce the risk of damaging the boards during the process.::solder::


    About recreation, I don't use to center new designs around components whose life cycle has ended. I make some exceptions, but the 2114 is not in my plans. Still, if some KISS came incomplete, this would also help if its owner wanted to recreate a board looking similar to the original unit. Talking of memory boards... have you been able to test the prototypes on the KISS?


    Regards,

    Jaume

    When I tried to list all retro systems I have at home, the "The message is too long, must be under 500 characters" error appears! :lol:

    Gefällt mir 1

    Hello,


    I'll revise for the next update tomorrow. Yes, this matches the original numbering. Being focused at both edges of the board and especially on the component side I missed completely numbers 2-5. Didn't saw 1 and while I saw 6 I thought it was some defect because seemed some sort of via but present only at one side... Thinking there was no numbering I used the same decreasing pattern I saw in the video card, albeit with numbers instead of letters.


    You are right, I revised pins 9 and 10 of UE1 and they were connected when they hadn't to. I suppose this error was made while moving the gate. All references are updated and the remaining issues you reported are to be checked tomorrow before updating again.


    Regards,

    Jaume

    When I tried to list all retro systems I have at home, the "The message is too long, must be under 500 characters" error appears! :lol:

    Gefällt mir 3

    Hello Jaume,


    I also did not see the numbering on the back side. It is very smal. It was good that helwie gave us a hint.


    Zitat von jlopez

    My goals are to understand the circuit, generate schematics and obtain any information that could speed up diagnostics and repairs.

    If you need some help for some special diagnostic software for Z80 systems, please tell me.


    This year I helped Aquarius to find out the errors in the RAM of his Olympia BOSS.

    A similar job I did last year for my own Philips P2500.


    Finding Error on Olympia BOSS Z80


    I hope, you can translate the thread from German to English.



    The specialty on those Z80 programs, which are have to be burned into the boot EPROM, is that they must be run without any RAM! So you can test all the RAMs and other components. Because each system is different, the program has to be adapted from case to case. Simply routines may also be written for the 8080, but there are not so much registers available and the therefore the functions will be reduced strongly.



    If "nothing" is working in the Z80 system, I have written a small routine for the boot ROM. Documented in this thread:

    First test for Z80


    For the test you need a scope and an EPROM programmer.

    This small test program is running in an endless loop, jumping between two address ranges each second (from 180 hex to 200 hex and back).

    So it is possible to check the address lines A7 and A8 with the scope, while triggering to M1 pin of the Z80. On the scope you should see a blinking of the two address lines in a rhythm of approx. 2 seconds. In this case the chance for a working CPU is given.


    HWTEST 002 for IPL Z80.zip


    The program needs an EPROM or NVRAM with at least 1KB size (e.g. 2716). The rest of the BIN file may be omitted. The EPROM has to be in a memory range starting at 0000 and should be started after reset.


    For special tests to find out memory errors etc., you need a possibility to communicate to the user. Because you cannot use the screen (is not initialize at this time, or the video card has been removed) you have to look for an other way. May be a serial interface or a parallel port. Because they are different from system to system, they must be programmed individually. Also for a complete memory test it is necessary to find out the RAM bank switching mechanism.


    So Aquarius and I avoided a lot of soldering jobs and reduced the risk of destruction of the hardware.

    After finding out the defective chips, we could replace them exclusively.


    Also for re-engineering it will be helpful. So we found out the I/O addresses of diverse components, by polling the complete I/O range (while using a scope).


    Have a nice day!

    Peter

    Greetings!


    I have revised the whole thing.



    Zitat von PAW

    The outputs of UE6B and C are connected to parallel (to get more driving capacity). Are they connected to the RAMs pin #10 (write enable)?

    Do you remember that part of "I hope I got it wrong"? The write pin signal is wired to UE1#4. The outputs of those not gates doesn't seem to be connected anywhere... I can't understand why. Unless I got it wrong even after multiple revisions (like that thing with the numbers ;)).


    Both 8226 units have been added and the remaining components to add are UD1 (74LS04, not), the memories (that will be a fast copy-paste job), the connector and the capacitors.


    I just found an error with the switch's pinout and it is to be changed.


    helwie44 † wanted a diagram with the positions of the components, which is an understandable request. I have bad calligraphy and I'm not that good with image manipulation so I resorted to Kicad's own PBC design utility to generate it. Only two drawbacks: it only shows components present in the document and the grid needs to be adjusted. The first is to be solved when the rest of the parts are documented and the latter is not needed at this stage.



    I hope this is useful and if revised and found correct the document could be finished this same afternoon.


    Regards,

    Jaume

    When I tried to list all retro systems I have at home, the "The message is too long, must be under 500 characters" error appears! :lol:

    Gefällt mir 1

    The circuitry you have drawn cannot be true...

    I agree with the adress deconding until output of UE1A. This output will be low only if /A14 and /A15 match SW1.

    This will also enable the RAM Selection via E2 of the 74xx138.

    This Signal is further inverted by UE1D and does an additional enable via E3 of the 74xx138. This seems somewhat odd, but is no problem.


    a) Now the /MemR and /MemW should be both inverted twice to reach UE1B to cause its output (6) to go high on either Read or Write and enable UE1C to cause enabling of the data buffers 8226.


    b) The RAMs /2114_W should have the same polarity as /MemW and I suspect it comes from the pins 8 and 10 of UC1D.


    c) The signal /DIEN of the 8226 should come from /MemW and not both.

    I have just retraced and corrected it. Thank you.

    As said, I was wrong.;)

    When I tried to list all retro systems I have at home, the "The message is too long, must be under 500 characters" error appears! :lol:

    Greetings,


    I think this time I got it. Some footprints may need to be replaced. I would like to know what do you think about this revision.



    The full schematics for the board: KISS - MC80 1010E.pdf


    Thank you very much,

    Jaume

    When I tried to list all retro systems I have at home, the "The message is too long, must be under 500 characters" error appears! :lol:

    Gefällt mir 1

    I am still not happy ;)

    UE1B should get /MemW effectively non-inverted (aka inverted twice), otherwise /CS to the 8226s won't go low on writes. Please check the connectivity there.

    Zitat von helwie44

    Hey, PAW did you test the circuit diagram with simulation software?


    No, I checked some parts of it (those are not plausible for me) on "paper" against the printed circuit board. I copied the scans of the PCBs into a word document, also parts of the schematics of Jaume and checked the obscure connection against the board.

    Then I modified the schematics like a picture, making a white overlay at wrong connections and drawed new lines with word. At last I printed the word document into a PDF document.