• RAM Disk reliability

    fjkraan03/20/2026 at 20:52 0 comments

    With the CR2032 backup battery installed I did some experiments with the data retention and how the PX-4 handles RAM Disk disconnects and reconnects. The Epson RAM Disk manual states that a disconnected RAM Disk does not retain its information. With some simple experiments it shows the fully static replica is more reliable. Two scenario's are tested:

    - Disconnect and reconnect with the PX-4 switched off.

    Here the PX-4 didn't even notices the disconnect.

    - Disconnect the RAM Disk, start the PX-4 and initialize the Internal RAM Disk. Then switch it off and reconnect the RAM Disk.

    Here the PX-4 re-initializes and asks if the RAM Disk should be formatted. When answered No, the old RAM Disk is accepted and it works as expected. Internally the RAM Disk has check sums for each section/page and if these match the contents, the contents is assumed valid. This is BTW another test for the proper operation of the auto-increment feature in the GAL.

    Allowing disconnects of the RAM Disk is an useful feature as the connection between PX-4 and the RAM Disk in its current form isn't as robust as the original.

  • PX-4 RAM Disk complete

    fjkraan03/14/2026 at 12:05 0 comments

    There are some developments with this project since the last log. With these, the RAM Disk for the PX-4 is more or less complete. The base board works, there is a compact battery solution and a case. 

    For the PX-8 a new board will be produced. The existing board works OK, but when installed with the case, it blocks some other useful ports. The new board will be less wide.

    Backup battery

    The backup battery is based on a CR2032 3V Lithium cell. It uses a custom board and a clip purchased in China. It is very compact, cheap and easy to assemble.

    backupBatteryClipParts.jpg:

    Extension board

    The Extension board is assembled and the plan is to fill it with data mimicking an 1 M Byte ROM, which should be recognized as drive K: on the PX-4. The other memory ranges are not supported by the Operating System, but can be used as extra non-volatile memory. One option is to keep backups for the active RAM Disk. 

    px-4RAMDiskExtensionBoard.jpg:

    Board correction

    When designing the new PX-8 board, it was found /OE line is not connected. That is what you get if you use more than one name for the same signal. The wire is from J2-26 to U11-24 (the /OE pin).

    px-4RD_OEbodgeWire.jpg:

    Show case

    The current case is large enough to contain the base RAM Disk board, Extension board and backup battery. For just the base board a lower case could be designed, but this is left to the user :-).

    PX-4RAMdiskShowCase.jpg:

    New board for the PX-8

    The planned new PX-8 board uses the same circuit ast the original, but the PX-4 connector is omitted. Instead the PX-8 connector is moved to the edge, allowing it to be used next to the PX-8 case, without the need for a cable. Designing a case like the original wedge is possible, but more complicated. 

  • Completed the cases (first version)

    fjkraan02/28/2026 at 11:28 0 comments

    The cases are finally completed, alas not in the typical Epson-non-white, but a kind of beige (Devil's Design, PETG). The PX-4 version is Ok, but the PX-8 case (and actually the board) are less optimal. It is too wide and partially obscures the RS-232c port and the BCR port. Nothing that a board and case redesign can't fix, but one case for both the PX-4 and PX-8 was worth the try. The circuit is the same, only the bus pin-out differs.