| Tip | ||
|---|---|---|
| ||
This section is being updated continuously. We are up to Z88 Hardware Investigation. Please return regularly, to see the latest information. |
| Table of Contents |
|---|
Latest NEWS
03/04/17 Having a BREAK from this project.
15/03/17 More tests undertaken in Hardware Investigation
11/03/17 Initial results published
09/03/17 Fernando sent in more results
08/03/17 Fernando sent in his results
07/03/17 Benny sent in more results
28/02/17 Benny sent in his results.
Introduction
Overview
The Z88 software team have been working tirelessly, to bring you the latest and greatest upgrade the Z88 has seen to date, 0Z 4.7.
This version, which uses the RAM intensively, has exposed either a hardware design fault on the board, and/or the specification of the chips used are not fast enough or a software oversight . This affects all versions of OZ 4.x but is only noticeable when running applications in RAM from OZ 4.6.1 onwards.
It works perfectly OZ 4.7 works perfectly if run in the following combinations
- internally ROM.0,
- externally ROM.1 in a 256K EPROM
- on the emulator
This issue (which is due to the method of writing to the flash chip) has only affected a handful of user, the users. The majority, who use the normal built-in applications i.e. PipeDream BBCBASICBBC BASIC do not have any issues.
Steps already taken
The fault was initially thought to be due to the decoder on the card. Was it a timing error switching between the RAM and the Flash decode chip? Designing and building Turbocharging the decode chip a faster decoder helped, but did not fully resolve the issue.
We had overlooked that the address line (A19) that selecting them, was not generated by the processor itself, it was generated by a latch. This gave a longer time than expected.
Test Procedure & Results
You are invited to "Send in your results."
This issue is currently being has been investigated, as some cards work well, whilst others fail.
In the past, these errors have been attributed to
- External card connector being dirty - this is not true
There is a way to see if a card is affected or not, by doing this test. There is no solution yet, but when there is - it will be published here.
Take part in the testing. See test section.
Z88 Hardware Investigation
To discover how the Z88 works it was decided to wire up an external card. Using an oscilloscope and see what the signals are really doing, rather than relying on what Cambridge Computer had documented. See what is happened here.
Hardware Modification Options
Both the 1M and 512K cards were modified to see if gating OE with CE would solve the problem. It made no difference.
Use an OTP EPROM
There is a fall-back option which will be tried next. This is using a If there is any demand for users who wish to use RAM applications, a One Time Program EPROM (mounted in a socket as shown) . They are available in 2 sizescould be tested and produced. All options use the 1M address space as there are no issues using the RAM on the same card. The proposed combinations are:-
- 256K - which could have just for OZ 4.7 in it and
768K of RAM - 512K - which could have OZ 4.7 and a 256K - for OZ 4.7
256K file area which could store the zipped version of all the games which could then be run in RAM for example.
- and
512K of RAM - 256K - for OZ 4.7
256K file area of user files (no ability to erase files) and
512K of RAM
