[Solved] MiST Hardware Revisions? MSX-Core-Problem

[guidol]

Does anyone got informations about differnt hardware revisions on the MiST?

I got a MiST-MIDI which I did buy in April 2015 and one MiST-MIDI which I did buy in September 2016.

Booth have now the FW (Firmware) s/w ver. ATH160721 - but
the MiST out of 2015 does load the MSX2-BIOS for the MSX-Core and the MiST out of 2016 keep hung at loading the MSX2-BIOS
and Iam using the same SC-Card and Cables.

Now I have sent a Mail to Lotharek - asking if he changed the mainboard or mainboard-components... if I get an answer you will read it here

[guidol]

a small Info in MiST-GitHub
https://github.com/mist-devel/mist-boar ... ingStarted


The pre-installed firmware version

The MIST board uses a permanently installed firmware in on-board flash as well as the called FPGA core file on SD card. Since both parts evolve over the time they need to be from roughly the same time. If the versions of the firmware and the core differ too much then the core may not boot (but no harm will/can be done).

Up to now two firmware versions have been installed in MIST boards before shipment.

If you purchased your board before May 2015 or if you got an older board from the third party you need to use the early core mentioned for your first tests.

Since May 2015 Lotharek started to flash boards with a firmware from April 2015. That version requires a more recent core. You need to rename this core to core.rbf before placing it on the SD card.

[Newsdee]

Firmware isn't hardware though? I'd expect both hardware to behave the same given identical fw?

[guidol]

I'd expect both hardware to behave the same given identical fw?

I did also expect the same behavior, but the newer one doesnt do that - with the same firmware - what the older one did/does.

[sebdel]

Might just be an unstable core. From my experience the smallest glitch in certain clock signals can make the core startup completely random. I would not be surprised if the NES core never booted on a certain number of MiST for example but I can't prove it and I also don't know if I fixed it
The RAM is also not always in the same state at core startup, that can have an effect. Did you try cold boot on both ? And by cold I mean at least a minute without power!

[jotego]

Might just be an unstable core. From my experience the smallest glitch in certain clock signals can make the core startup completely random. I would not be surprised if the NES core never booted on a certain number of MiST for example but I can't prove it and I also don't know if I fixed it
The RAM is also not always in the same state at core startup, that can have an effect. Did you try cold boot on both ? And by cold I mean at least a minute without power!

I agree. I wouldn't be surprised if timing is not constrained with enough detail in this core, or maybe constrains are not being met at all... Sadly, implementation is somehow disregarded in open source hardware projects.

[guidol]

Lothrek did answer "nothing has changed since beginning..bill of materials is the same".

But MasterOfGizmo (Till Harbaum) did answer me also, telling that core like the MSX-Core do overclock some parts of the FPGA.
Some FPGA-Chips could be more overclocked as other ones (like some Motorola-Chips in AMIGAs or the CPU ony some Raspberry Pi) and run stable.

The overclocking will not damage the FPGAs, but the core will hang.

Her did also write:
"Building the same core another time results in slightly different cores.
In simple words: Sometimes the software can choose which of the many wires
inside the FPGA to use for what and it randomly chooses one. And if a core is
close to the edge then these minor differences may again make a difference."

So there would be a chance that one of the other versions of the MSX-Core will run your MiST.

The 3 MSX-Cores are at
https://github.com/mist-devel/mist-bina ... /cores/msx

ocmsx_141010_r356.rbf
ocmsx_141012_r359.rbf
ocmsx_141022_r364.rbf

[Asmith]

Tried the 3 Msx cores , still no luck , it seems the MSX core is not for my MIST ...


A_s

[DrOG]

The same result at me...

[ManuFerHi]

I´m made some PCB mist boards, the problem is DRAM memory, the original reference MT48LC16M16A2P-75 is obsolete and replacement like MT48LC16M16A2P-7E or other similars are differents timings, all cores running OK, except MSX core crashes before load bios file.

[DrOG]

Actually besides the MSX core the Acorn Archimedes is problematic too, dont't know if it's for the same reason or not, but the Archie is very picky in aspect of SD-cards...

[dhester]

I don't know if it will help but I got an msx core to work on a ZX Uno with various sd cards and loading the core on the ZX Uno, I also have used various SD Cards and got it working on the MIST.

[ManuFerHi]

Actually besides the MSX core the Acorn Archimedes is problematic too, dont't know if it's for the same reason or not, but the Archie is very picky in aspect of SD-cards...

Can you see reference of DRAM memory?
MT48LC16M16A2P-75 >> MSX core works perfectly
MT48LC16M16A2P-7E o other memory don´t load bios file.

[DrOG]

Nothing similar:

hynix 948EA
H57V2562GTR-75C

On my Turbo Chameleon 64 the MSX core works flawlessly.

The ZX Uno is not bad either, but it has only one joystick port, i need more!

[ManuFerHi]

Nothing similar:

hynix 948EA
H57V2562GTR-75C

On my Turbo Chameleon 64 the MSX core works flawlessly.

The ZX Uno is not bad either, but it has only one joystick port, i need more!

That is the cause, that memory has different timings to the previous ones and it is necessary to modify the core.
Zx-Uno only supports MSX1.

[ManuFerHi]

Nothing similar:

hynix 948EA
H57V2562GTR-75C

On my Turbo Chameleon 64 the MSX core works flawlessly.

The ZX Uno is not bad either, but it has only one joystick port, i need more!

That is the problem, for some strange reason the MSX core only works with the memories MICRON MT48LC16M16A2P-75
Zx-Uno only support MSX1.

[ManuFerHi]

New core published today, fix SDRAM memory problem, i´m tested in various sdram models and work in all fine. Problem fixed.

[Sinclair]

This is very good news because it gave me the impression that the project was dead for the time it took without updating.

[guidol]

i´m tested in various sdram models and work in all fine. Problem fixed.

Also on my newer 2016 MiST the new MSX-Core does load the MSX2-BIOS without problems

So sometimes we have only to ask the right people
Thanks to Alastair Robinson and Till Harbaum

[Asmith]

What a great Xmas present ! Thanks everybody for this effort !

A_s

[DrOG]

Thanks everybody who was involved to fix the memory bug!

It runs now flawlessly on my new MIST, boots for the first time with MSX3BIOS.SYS!

It would be nice to improve the Acorn Archimedes core in the same aspect...

Thanks again this nice Christmas present!

[vebxenon]

Great news!

[vebxenon]

Well, new core doesn't work well to me. I'm using again previous version.

[Sinclair]

The latest version gives me problems accessing the disk.
I went back to the previous version.
I am one of those who have the original RAM.

[vebxenon]

The latest version gives me problems accessing the disk.
I went back to the previous version.
I am one of those who have the original RAM.

Same here