English Amiga Board

English Amiga Board (https://eab.abime.net/index.php)
-   Hardware mods (https://eab.abime.net/forumdisplay.php?f=105)
-   -   29f800/160 custom ROM (https://eab.abime.net/showthread.php?t=75507)

kipper2k 22 October 2014 16:56

29f800/160 custom ROM
 
Hi All,



just a peek of a work in progress, i finally had a chance to play with my flash eprom board.

Here is a 1.3/3.1 kickstart installed in an a500, it is low profile and actually sits lower than an original kickstart ROM. It is reprogrammable, i just place it in my programmer adapter (i built custom) for my Wellon programmer and program it.

It is not programmable in the Amiga itself (would be nice if it could be made to be). This is just the beta version, changes to be made include putting the jumper block which controls the ROM selection on the other end of the board so my fastmem board doesn't block it. I can actually use a 3 pin extension cable so that i can relocate the switch/jumper block to wherever i want. I also have a multirom board designed for the A1200 which actually sits over both ROMs and will have the 2mb 29f160 flash on it.

The board itself can be set up by pushbutton on start or toggle switch, i may give both options.

http://kipper2k.com/amigaforsale/a500adapter.jpg

and a close up

http://kipper2k.com/amigaforsale/29f800croppedsmall.jpg


Thoughts etc ?

Franchute13 22 October 2014 17:31

uauu!
Good job!

jbenam 22 October 2014 17:51

Since it can't be directly programmed on the Amiga, how feasible would be to add an USB port along with some kind of IC to allow flashing it on PeeCee without additional hardware? :D

Toni Wilen 22 October 2014 17:52

I am still waiting for adapter that can be programmed by m68k code, without extra hardware. (Yes, I know it is not going to happen without some kind of Gary adapter/hack)

kipper2k 22 October 2014 18:32

Quote:

Originally Posted by jbenam (Post 982254)
Since it can't be directly programmed on the Amiga, how feasible would be to add an USB port along with some kind of IC to allow flashing it on PeeCee without additional hardware? :D

Ultimately it will need some kind of programmer, be it willem or another highend programmer/adapter to be able to program it. My programmer has a 48 pin ZIF on it (wellon VP-380), but i still had to make an adapter board in order to be able to program the 29f800 TSOP.. most cheaper programmers have 40 pin ZIFs which need their own adapter to be able to program it.

I would love for someone to make the software/hardware mod in order to make it in-system programmable on the Amiga but i dont have the smarts for that.

pandy71 22 October 2014 21:18

Quote:

Originally Posted by kipper2k (Post 982261)
I would love for someone to make the software/hardware mod in order to make it in-system programmable on the Amiga but i dont have the smarts for that.


We need Gary replacement ;)

NMI 23 October 2014 09:24

In the previous thread: Custom Flashrom design for Amigas RedskullDC said the software was almost done, and that was close to 2 years ago. Time to check with him if it's done by now? :)

Totally agree about the need for a gary replacement. Shouldn't be too hard to reverse engineer it or maybe Toni has it mapped out in software already? :)

Toni Wilen 23 October 2014 11:45

I think easiest way to make it incircuit-programmable is to have some kind of remap jumper that maps the flash in some unused address space which is read/write compatible. (F8-FF region is not writable, Gary won't allow it), for example upper fast ram region (assuming there is no 8M fast ram expansion).

Quote:

Originally Posted by NMI (Post 982333)
In the previous thread: Custom Flashrom design for Amigas RedskullDC said the software was almost done, and that was close to 2 years ago. Time to check with him if it's done by now? :)

Writing to flash chip is really easy and simple, modern flash programming is very high level. (erase chip/sector, write unlock sequence, write write command, write byte to correct address, poll for ready bit, repeat next byte)

Quote:

Totally agree about the need for a gary replacement. Shouldn't be too hard to reverse engineer it or maybe Toni has it mapped out in software already? :)
Emulation won't help and there is nothing mysterious in Gary. It is more or less A1000's logic glue chips in single package.

Knight2011 23 October 2014 12:53

To me (owner of an A500 with 1.2 ROM and MB rev5), it would be good to have this board to be able to switch to 1.3 (it seems that some games/demos work exclusively with 1.3). And please don't forget people with MB rev5 or earlier to have something just "plug and play".

PS: I'm not user of workbench so having 3.1 ROM would be probably useless for me. No?

kipper2k 23 October 2014 17:21

Just a quickie to mention i programmed V1.3 and V3.1 in the A1200 so now i can dual boot either Kickstart in my A1200. Took a while and some help from Zetr0 to get it sorted but it's definately worth it.

heres a pic...

http://kipper2k.com/amigaforsale/a1200romboards.jpg

next step is to use the 2mb 29f160 flash and set up 4 kickstarts, i'll try to tackle this today/tomorrow

Yulquen74 23 October 2014 17:22

To make a rom programmable by Amiga software without changing any hardware,
or without sacrificing RAM space, it could be possible to do it this way:

Make a board that plugs into ROM socket with 3-state bus-drivers for address and data lines, an SRAM chip, a flash chip, and a PIC microcontroller hooked up to all address and data lines, kickstart #OE, and all control signals for bus drivers, flash and SRAM.

Nothing can be written to kickstart space, so the data that we want to write will have
to be put into the 16 LSB bits of the address during read accesses to ROM area. The micro-controller will have to run much faster than the 7MHz bus, and #OE accesses for kickstart will trigger interrupts which detects what addresses are accessed.

A spesific series of accesses scattered in ROM space that will never occur except during flash programming, can act as a command for the PIC, to start buffering the data to SRAM. When the buffering is finished, another command may have the PIC write the data to a desired block in flash.

The Amiga software doing the transfer will not really know when the PIC is finished
with this, so it would have to wait for an amount of time it is guaranteed finished,
having interrupts and regular rom accesses blocked.

Changing kickstarts could be done with long resets, or commands initiated before short resets to do a change to a certain image, even the image in RAM could be tested.

The real difficulty would be to make this mess as small a footprint as possible.

kipper2k 23 October 2014 17:26

Quote:

Originally Posted by Yulquen74 (Post 982389)
To make a rom programmable by Amiga software without changing any hardware,
or without sacrificing RAM space, it could be possible to do it this way:

Make a board that plugs into ROM socket with 3-state bus-drivers for address and data lines, an SRAM chip, a flash chip, and a PIC microcontroller hooked up to all address and data lines, kickstart #OE, and all control signals for bus drivers, flash and SRAM.

Nothing can be written to kickstart space, so the data that we want to write will have
to be put into the 16 LSB bits of the address during read accesses to ROM area. The micro-controller will have to run much faster than the 7MHz bus, and #OE accesses for kickstart will trigger interrupts which detects what addresses are accessed.

A spesific series of accesses scattered in ROM space that will never occur except during flash programming, can act as a command for the PIC, to start buffering the data to SRAM. When the buffering is finished, another command may have the PIC write the data to a desired block in flash.

The Amiga software doing the transfer will not really know when the PIC is finished
with this, so it would have to wait for an amount of time it is guaranteed finished,
having interrupts and regular rom accesses blocked.

Changing kickstarts could be done with long resets, or commands initiated before short resets to do a change to a certain image, even the image in RAM could be tested.

The real difficulty would be to make this mess as small a footprint as possible.

even a daughter board connected by a cable could be possible, i'm willing to try to build something to someones design

if anyone wants to buy one of these pcb from me, it would be $2 each + shipping (nothing on the board). i have a few of them kicking around

pandy71 24 October 2014 16:23

Going further - serial NOR flash + uC + SRAM - then various images can be stored in flash, uC can perform authentication and flash reprogramming and SRAM will provide ROM emulation. As some glue logic is required anyway then perhaps easiest is to use softcore uC that can fit in small CPLD/FPGA.
SRAM will allow to modify KickROM on the fly (and also fast access).

dirkies 24 October 2014 17:02

Drool... upto 4 kickstarts readily available with a simple switch? Hopefully you can pull that one off!
That means the death sentence for Relokick but who cares?! :)
Good work!

Arnie 24 October 2014 17:05

I presume you still need to programme these yourself?

kipper2k 24 October 2014 18:21

These can be preprogrammed for people, the only caveat is that people must prove ownership of the rom in order for these to be sold legally. I believe this could be done by the buyer sending the files to the seller for writing. Once written it can be treated like a regular read only rom for most people that have no intention/need to rewrite it.

A while back i did actually do a small logic circuit that will select a specific rom to be loaded from a multi kickstart flash. i posted it somewhere on here with a short video, I'll have to go looking for it.

I am just working today on the 29F160 2mb chip, i have proven that it will multiload 4 x 512mb kickstarts by selecting address lines A18 and A19 so it's just a case now of tidying up the ROM selection at boot process

Arnie 24 October 2014 18:30

I would want to program them myself for 2 reasons. One of the rom images is a KS1.3 with SCSI and the other is a 3.9 custom image. I wouldn't be able to prove I own the roms. (unless you count eeproms)

kipper2k 24 October 2014 18:58

Quote:

Originally Posted by Arnie (Post 982654)
I would want to program them myself for 2 reasons. One of the rom images is a KS1.3 with SCSI and the other is a 3.9 custom image. I wouldn't be able to prove I own the roms. (unless you count eeproms)


if you own the original rom that was used to create the scsi version then i don't think that falls foul of the law, not sure though... let other people have their say

Arnie 24 October 2014 19:04

Proving you own the roms aside, I doubt that you would be able to tell from a file whether it's been altered or original, would you?

kipper2k 24 October 2014 20:45

Quote:

Originally Posted by Arnie (Post 982673)
Proving you own the roms aside, I doubt that you would be able to tell from a file whether it's been altered or original, would you?


Yup, every file has a checksum or hash value, and you can take a look at the boot screen and the actual first part of the file which will show the original file version number


All times are GMT +2. The time now is 16:59.

Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2024, vBulletin Solutions Inc.

Page generated in 0.04840 seconds with 11 queries