Scor­pio News

Cautiously the board was initially installed set to 4MHz. On power-up or reset the Monitor starts by displaying the text ‘ABC’ on the screen. If auto-boot is selected, then both the first floppy and then the Winchester drive are polled in turn, It won’t boot my winchester since I modified the SASI interface on the GM829 about a year ago in order to facilitate DMA transfers. However, booting a floppy ‘sysgenned’ with the winnie on drive A works a treat.

If the monitor setting is selected then the Monitor announces its presence with the text ‘Newburn Monitor’, followed by a rather untidy menu of options on a wrapped round line. Perhaps this is attended to in the new version (1.1) of the monitor PROM, which I am told is already in the post. I gather that Newburn operate a free update service, for the first year after purchase. No problems were encountered in this first period of use, and the clock was soon re-set to 8MHz. The improvement in performance was immediately noticeable. Everything happens much more quickly and display of a full screen of text almost gives the impression of a memory mapped display!

One minor problem soon became apparent – with the floppy drive motors stopped disk access produces a ‘Drive: not ready’ message. Immediately selecting the re-try option causes a resumption of service. The cause of this is the one second delay which the GM829 inserts after starting the drive motors, before allowing the drive ready bit to be asserted. Most of the existing 80-BUS BIOSs incorporate a software delay loop before forcing the time-out. The improved performance of the NE898 shortens this delay to the extent that it is shorter than the hardware delay. The author’s BIOS source code was soon modified, assembled and the whole MOVCPM image re-linked. Users of proprietary BIOSs may not have the facilities to carry out this modification, so I have investigated an alternative solution in hardware. This appears to effect a cure for all the Gemini BIOSs. Newburn-claim that no problems are encountered with the GM849, so presumably the same delay is not implemented on that board. Users of other FDC boards should beware of this possible problem.

Another problem was discovered when using the public domain utility D.COM. If the files were held in the directory in alphabetic sequence, then all was OK. However if a sort was required, then the disk access proceeded as usual but no output was produced before the next system prompt. After much investigation it was discovered that D.COM employs two undocumented Z80 opcodes in the sort. The Z180 detects these and merely sets an internal flag before trapping to address zero, hence causing a harmless warm boot. I intend to modify my OS to report detection of illegal opcodes, which will save much time in tracking down other similar problems. In this particular case the four bytes of offending code were replaced by three bytes of legitimate Z80 code, plus a NOP. Other cases may not be so obliging!

Apart from these two incidents no other incompatibilities have been encountered, although, not surprisingly, Z80A peripherals do not respond kindly to being driven