Beiträge von Hans Otten

    The MOS image loads fine in the KIM-1 Simulator.

    Procedure used:

    GIMP

    1. Open image file

    2. Image - Scale Image to 320x200 (unlink couling horizontal - vertical)

    2. Image - Mode- Indexed (if color choose dithering)

    3. Image - Mode - RGB

    4. Export - c header file

    KIM-1 Simulator

    1. Settings - Enable K-1008

    2. File - Load C header file

    3. Load C header file

    4, Load into memory

    Note that feeding other dimensions than 320x200 will fail.

    Note 2. When I am back from vacation I need to update the help file in the simulator program, the online help is up to date. And have a look at the geometry of the K-1008 display.

    Eduardo, enjoy the beach!

    Congrats!

    This is awesome, a new KIM-1 with a new K-1008! And both working and looking so good.

    Hans

    The MM5280 ICs working, a miracle with the reputation of Chinese IC sellers nowadays.

    I updated the KIM-1 Simulator to make using the K-1008 more fun.

    Images formatted to the K-1008 video display format can be loaded into the display.

    Any suitable image can be used now. Use GIMP to prepare it, export as C header file and load in the Simulator.

    Bitte melde dich an, um diesen Link zu sehen.

    Examples of what the K-1008 can display.

    Bitte melde dich an, um diesen Anhang zu sehen.

    Bitte melde dich an, um diesen Anhang zu sehen.

    Bitte melde dich an, um diesen Anhang zu sehen.

    Bitte melde dich an, um diesen Anhang zu sehen.

    Images from photos, loaded in GIMP, resized to 320x200, threshold to blackwhite, exported as C header file, converted by commandline program to K-1008 binary, loaded in KIM-1 Simulator K-1008 screen.
    (Eduardo gave me the tip to use GIMP for this)

    Your test of autorepeat Enter is not a good one. Enter will respond with sending the next memory location. Sending that takes time as you see. No line delay will compensate for that.

    A papertape text file upload is a good test.

    With appropriate line and character delays it works fine for me.

    A conservative setting of 1200 baud 200 ms line day, 20 ms character delay.

    2400 baud also works.

    Prolific USB chipset, FTDI chipset. USB is not the troublemaker here, it is the slow KIM-1 serial bitbanging and the processing of especially end of line that requires line delay.

    A Basic text file upload requires even more slowing down.

    Clear the decimal flag always! Place 0 at 00F1.

    Are the Funkschau magazines 1979 1980 scanned? THey contain more KIM-1 articles.

    I found the KIM_Baudot.pdf in the

    telexforum.de

    There are more scanned articles about connecting teletypewriters to apple 2, VC20 and others.

    Funkschau 1979 has a lot of interesting 6502, KIM-1 and AIM 65 articles, as I read in the Contents I did find online. I did not find scanned magazines alas, but with the help of Matthias (SBC) on the VzEkC e. V. SBC forum I may get scanned articles. Thanks for getting me to look for those!

    The discussion of the KIM-1 ROM dumps filler bytes here let me publish the hopefully real KIM-1 ROM dumps with $00 as filler bytes.

    Bitte melde dich an, um diesen Link zu sehen.

    The versions of the KIM-1 ROMs I know from the internet all seem to originate from recent assembled source, typed in from the listing as published in the KIM-1 User manual. I think all this originated from Ruud Baltissen.
    The binaries are correct KIM-1 ROMs and function good in emulators and reproductions (from MICROKIM to PAL-1 to Corsham KIM Clone to KIM UNO and many more, including the designs presented in this forum).

    But as we know now, these are not the original KIM-1 ROMs. The locations not in use are filled by the assemblers/linkers with $FF. That makes sense, an UV EPROM can be programmed later per byte to change a 1 to 0.

    The KIM-1 ROMs contain $00 as filler bytes. Quite a lot in the tape ROM, 3 in the main ROM.

    I knew this already but forgot about it until now. I have real KIM-1 ROM dumps made by Dwight Elvey as part of his 6530 replacement project (see VCFED.ORG).
    I checked these ROM dumps and found dump errors in , several locations had a bit error and the NMI vector was wrong. But the filler byte was $00!

    So I did not publish those ROM dumps then. And did not correct the ROMs as published, since it is not required for the correct functionality. But for checksum calculations it is!

    Hans

    My Simulator was not intended to be a 100% exact emulator. My main goals were emulation, study and debugging: looking inside the processor and memory and I/O ICs, and have TTY programs run. Single step, run to breakpoints and watches, traces, debug, profiling, that is what the architecture and my skills allow.

    It runs fine as is on Windows 10/11 (note the installation hints for high DPI screens and the console), Linux 64 bit , Raspberry OS (a bit too slow) and MacOS.

    With LEDs and keyboard and TTY console it gives me the feeling of working with my old KIM-1. Especially the TTY console part is important for me, since that is what I used my KIM-1 for in the old days. Having Microsoft Basic and MICRO ADE running is fun.

    My simulator does not :

    • run at 1 MHz. Though CPU cycles are counted, the processor runs as fast as the host allows. About 1000 CPU instructions at full speed, then a check for UI events is done. And some KIM ROM reads or write trigger UI events also. So that prevents the need for threads and mulititasking besides what the system/Lazarus does.
    • do bitbanging. Since I/O in the KIM-1 is all bitbanging I intercept those routines in the ROM. Any program that uses the KIM-1 routines will work. So a program that does call LED SCANDS or GETKEY or the TTY I/O routines or the tape routines at 1800 and 1873 do run exact.
    • Programs that do the bitbanging themselves do not run.
    • 6532 timers. The 6532s are not completely implemented yet, on the Todo list.

    An architecture that would allow cycle correct and bitbanging exact behaviour requires multithreading. A main thread running at 1 MHz and a second thread looking at the I/O memory locations, a third thread for the user interface. Beyond my skills, motivation and available time.

    So it is understandable your experiments gave these results:

    • CRC calculation fails since I patched at runtime the ROM. And accessing the tape routines wil activate the tape dialog.
    • LEDs will not change by doing the bitbanging . The diagnostic ROM uses bitbanging (which is understandable, since the KIM-1 ROMS are switched off).

    Since my original, and last time I checked working, KIM-1 is now in a wall display and I do not want to turn it on, it is too valuable for me, I cannot check the ROM contents against what is available now.

    Filler bytes have to be checked by someone with a working original KIM-1.

    Hans

    Another gem to be added tot he 6530/6532/KIM repair pages! Thank you!

    This board will also be very useful to original KIM-1 users with defective 6530 IC's.

    Hi, I've made everything public at my Bitte melde dich an, um diesen Link zu sehen., under the Bitte melde dich an, um diesen Link zu sehen.. I hope you enjoy it :)

    Awesome!

    Link and description to your github repository will be added to my 'KIM-1 clone' page!

    Added the description and link two days ago and did some PR on twitter, mastodon and facebook groups. So many nice reactions and compliments on the design!