github.com/mahjongg2.ZX81plus38
1
0
Fork 0
mirror of https://github.com/mahjongg2/ZX81plus38.git synced 2026-02-13 19:35:48 +00:00
Code Issues Packages Projects Releases Wiki Activity

Update README.md

Browse Source
This commit is contained in:
mahjongg2
2022-07-07 16:45:07 +02:00
committed by GitHub
parent 965d86d266
commit 836f4bc791
1 changed files with 6 additions and 4 deletions
Download Patch File Download Diff File Expand all files Collapse all files

10
README.md
View File

@@ -20,16 +20,18 @@ The file ZX81+38rev1,5.pdf is the schematic of the corrected revision 1.5 (in bl
We found one more error, that I have fixed in revision 1,6:
the keyboard matrix is connected to eight Address lines, A8 to A15, but the error is that for A8 a version is used that comes from the signal line passing through a 10K resistor (R1) so it is signal A8' , that is not correct, the keyboard needs unattenuated signals, so A8, not A8'. otherwise the resulting dataline does not go low enough to reliably register as "low". (actualy this was not essential, as long as the RP1 pullups are not too strong, so I changed them to 10K instad of 4K7.
Revision 1.7 fixes some small errors, and revision 1.8 has the fault fixed that lead to a mix of logic families, now only HC logic is used (in theory you can also use all LS logic, or all HCT logic, all LS logic is confirmed to work.
Because of covid (closing my Makerspace) I have not been able to build my own copy of the ZX81plus38, but I'm several people have built it and reported that revision 1.6 is working well. You can read much more about the development of my designs in my makerspace at www.revspace.nl look up my projects in the projects gallery. For example this project can be found here: https://revspace.nl/ZX81plus38_simple_to_build_ZX-81_clone for writing back you can mail me at mahjongg at xs4all.nl replace the at for the commercial at sign @, or you can use the forum of the raspberry PI organisation where I wrote about this clone in the "offtopic" section.
And I will upload kiCad files after I know revision 1.6 is working well, which can happen only if revspace reopens, and I have built and tested my own copy, and I have recovered enough from my current bad condition of what I think of as my "long-covid", only I don't remeber ever having a case of covid, but I do suffer from the same symptoms. I'm 66 now (2021/2022) so I don't see myself getting a job anymore, and consider myself retired.
I noticed that somehow I hadn't uploaded a correct BOM, so I added the Bill of material file in the form of and .ODS (open officie spreadsheet) file
The BOM contains exact component designators for all components used from either farnell or mouser, so you will know EXACTLY which components to buy.
I noticed that somehow I hadn't uploaded a BOM, so I added the Bill of material file in the form of and .ODS (open officie spreadsheet) file
The BOM contains exact component designators for all components used from either farnell or mouser, so you will know EXACTLY which components to buy. Use version 1.8 so that you don't use a mix of logic families.
Note that R21 (the series resistor for a power LED) is NOT in this BOM, as I added it after I created the BOM, but should be 330 Ohm (if you add a Power LED) That is if you built the keyboard I designed for the ZX81plus35 that also works with this ZX81plus38, in the schematic a value of 100Ohm is used, which is also okay, although much brighter.
on june 15, 2022 I uploaded a BOM for revision 1.7 of the ZX81+38 creating some corrections R20 should be 680R, R21 should be 330 R R29 should be 1M RP22 can be 10Kx8 but 22Kx8 works better with some keyboards, C7 should be 47nF, most importantly U21 should be a 74HC04 or 74HCU04 NOT A 74HC14!
on june 15, 2022 I uploaded a BOM for revision 1.7 of the ZX81+38 creating some corrections R20 should be 680R, R21 should be 330 R R29 should be 1M RP22 can be 10Kx8 but 22Kx8 works better with some keyboards, C7 should be 47nF, most importantly U21 should be a 74HC04 or 74HCU04 NOT A 74HC14! There is a strong possibility the crystal oscillator won't work when using a schmitt trigger inverter (74HC14) use a HC04 or better a HCU04.
On June 23, 2022 I finished the schematic for the latest revision 1.7 version of the ZX81+38 with several small corrections. and also updated the layout. The main correction is the rotation of the keyboard-data connector J3, so the ZX81+38 is now completely compatible with a ZX-81 keyboard foil, (I will create a specific keyboard PCB for the ZX81+35, or even a real ZX81 later on) other changes are that U21 is now a 74HC04 with R29 as an 1M resistor.
@@ -41,4 +43,4 @@ I also updated a component placement drawing for revision 1,7 and uploaded it he
Lastly for completeness, I also uploaded the ROM dump of the real ZX-81, which you are legally allowed to burn into a ROM and use with the ZX81+38.
On 5 juli 2022 I discovered that KiCad had changed back to the logic family of some libraries for wich I had changed the original logic family from LS logic to HC logic. I intended that all logic would be HC logic, just like in my ZX84+35 initial clone. But this may cause problems because HC logic and LS logic are incompatible. So the first thing I did was updating the BOM, to also see if there actually were HC replacements for the LS logic I had used by mistake, luckily all te HC logic I needed was available in DIP. Next up will be revision 1.8 of the schematic, which will also get a clock inverter using a spare half of a flipflop. This might be necessary to solve a timing error that can occur with EPROM's that are too slow to put pixel data into the shift register that results in all black character squares (due to the databus pullups that pull the databus high when nothing is driving it. Then I will correct the component placement diagram which also shows some wrong logic families. My last remark is that although theoretically the ZX81+38 should not work with mixed logic, it seems that in practice it still seems to work. I finished building up my own copy with mixed logic so I can check this for myself. Also note that I no longer recommend using sockets with machined pins, use dual wipe sockets instead. This is reflected in the BOM. Lastly perhaps thsi clone will not work with original N-MOS Z80 chips, its recommended to use modern (C-MOS) versions,
On 5 juli 2022 I discovered that KiCad had changed back to the logic family of some libraries for wich I had changed the original logic family from LS logic to HC logic. I intended that all logic would be HC logic, just like in my ZX84+35 initial clone. But this bug in KiCad may cause problems because HC logic and LS logic are incompatible. So the first thing I did was updating the BOM, also to see if there actually were HC replacements for the LS logic I had used by mistake. Luckily all the HC logic I needed was available in DIP. Next up will be revision 1.8 of the schematic, which will also get a clock inverter using a spare half of a flipflop. This might be necessary to solve a timing error that can occur with EPROM's that are too slow to put pixel data into the shift register, which results in all black character squares (due to the databus pullups that pull the databus high when nothing is driving it). Then I will correct the component placement diagram which also shows some wrong logic families. My last remark is that although theoretically the ZX81+38 should not work with mixed logic, it seems that in practice it still seems to work as I have reports that it does! I finished building up my own copy with mixed logic so soon I can check this for myself. Also note that I no longer recommend using sockets with machined pins, use dual wipe sockets instead. This is reflected in the BOM. Lastly perhaps this clone will not work with original N-MOS Z80 chips, its recommended to use modern (C-MOS) versions, such as the Z84C0008PEG.