Custom Flashrom design for Amigas

[alexh]

In circuit programming is a must for a FlashROM design.

Would it not be a better idea to add a FlashROM to a new Kipper2k RAM card? And also add CFIDE68k?

Then could it not work in the same way that the FlashROM works in the Elbox eFlash, Individual Computer's KickFlash and E3B's Deneb / Algor Pro? Non of those physically plug into the ROM socket.

[kipper2k]

Originally Posted by alexh

In circuit programming is a must for a FlashROM design.

Would it not be a better idea to add a FlashROM to a new Kipper2k RAM card? And also add CFIDE68k?

Then could it not work in the same way that the FlashROM works in the Elbox eFlash, Individual Computer's KickFlash and E3B's Deneb / Algor Pro? Non of those physically plug into the ROM socket.

Adding it to one of my boards is an option, but i wouldn't want to force people to buy one of my boards just for the Flashrom portion. I do have plans to add the IDE68K to the A500 fastmem/adapter board with a little extra option (more because i want it more than anything else). but i digress

[Toni Wilen]

I'd like to have ROM socket flash ROM adapter because I don't want to change hardware configuration and it is compatible with all models.

ADDED: I predict that soon we have suggestions that this should become 100MHz 68060 board with 2G of RAM and IDE DMA controller and USB and MORE!

[kipper2k]

I had to join A1k.org in order to find a couple of links. I think i found the previous project, the translater is not the best.

Heres the link

http://www.a1k.org/forum/showthread....light=flashrom

[mfilos]

kipper2k my friend I gave the link as well in my #6 post

[kipper2k]

Originally Posted by mfilos

kipper2k my friend I gave the link as well in my #6 post

I just had to do it the hard way lol

[johnim]

this sounds interesting

[Amiga1992]

Originally Posted by Toni Wilen

ADDED: I predict that soon we have suggestions that this should become 100MHz 68060 board with 2G of RAM and IDE DMA controller and USB and MORE!

Hahahaha!

[cosmicfrog]

kipper can I have mine with Toni`s crystal ball on plz
oh plz make this happen, plz, plz, pleas

[Geijer]

Originally Posted by mfilos

Writing ROMS to Flash media was made in the past but stopped due to legale issues.
You can see a full thread by Reptile on a1k.org here: http://www.a1k.org/forum/showthread.php?t=13078

If it requires original Kickstart "piggybacked" (that can be used as fallback) maybe that can solve the legal issue.

[alexh]

So how many sideband signals (including voltages) would you need to support in circuit programming of a flash plugged into a Kickstart ROM socket?

[Toni Wilen]

Originally Posted by alexh

So how many sideband signals (including voltages) would you need to support in circuit programming of a flash plugged into a Kickstart ROM socket?

Gary based Amigas need following:

Chip Enable: hardwired to active (ground). This signal needs to be connected to address decoder, active when ROM address is selected and write or read. (Need extra logic because there is no Gary signal that can be used, see below)

Output Enable: Gary /ROMEN signal. Active when ROM address is selected and access type is read. Write never enables this signal. AFAIK No modifications needed.

Write Enable: This pin does not exist in ROMs. Extra logic needed.

I think thats all. Modern flash chips don't need any external programming voltages.

[pandy71]

How this sound:

SRAM + uC + DataFLASH (and necessary glue logic) - yes - it will bit larger than plain adapter but can provide required safety (signed code, passwords etc), perhaps preboot capabilities (ie selecting required Kick.configuration/programming from preboot screen).

Signal used for control can be based on OVL signal that already exist.

[kipper2k]

here is a pic of Redskulls design using original kickstart and 2 extra flash. Chip enable is used by manual switches



with just the one flashrom, the logic would be easier and it would be a case of setting the address lines to boot the required area of flashrom memory into the Amiga. CE (chip enable) must be logic low for the chip to be enabled and WE (Write Enable) must be logic low to enable flashing of the rom. Switch 3 determines whether one of the 2 512k images appears or if the board will appear as a 1MB kickstart image

[pandy71]

Yes, this implementation is straightforward and natural but there is no protection against accidental/malicious reflashing, whole flash can be erased by just sending few bytes/words in particular order.
IMHO FLASH adapter shall be designed in that way to prevent accidental/malicious corruption.

(jumper to prevent WE can be good protection but this is hack)

[kipper2k]

Originally Posted by pandy71

Yes, this implementation is straightforward and natural but there is no protection against accidental/malicious reflashing, whole flash can be erased by just sending few bytes/words in particular order.
IMHO FLASH adapter shall be designed in that way to prevent accidental/malicious corruption.

(jumper to prevent WE can be good protection but this is hack)

Yah, agreed. A simple jumper can force WE to remain high to prevent any writes to the flashrom. It could also be done by making it more complicated but for most people once the flash is programmed there is no need to move the jumper. For those people who are trying beta images etc then a simple switch could be attached to a jumper block to give them control

[pandy71]

Originally Posted by kipper2k

Yah, agreed. A simple jumper can force WE to remain high to prevent any writes to the flashrom. It could also be done by making it more complicated but for most people once the flash is programmed there is no need to move the jumper. For those people who are trying beta images etc then a simple switch could be attached to a jumper block to give them control

Then yes, perhaps accepting hack nature (jumper/switch - don't like this idea but in Amiga community solving configuration problems with hardware switches/jumpers have long history) this can be OK.

[NMI]

The more I think about it the more I like the idea of having the original ROM piggybacked on top of the flashrom.

It means that the flashrom can be sold unprogrammed, avoiding license issues.
Also it should work for both the "casual flasher" and the "hardcore developer". Those who use the flash as a "KS3.9" and only seldom do a reflash can easily switch back to a working KS and boot AmigaOS (almost) as if nothing happened to do a reflash from their normal WB environment. The "hardcore developer" on the other hand can easily replace the KS-ROM with an EPROM (or similar) containing the track/serial-loader of choice.

A jumper is OK, anything to elaborate and complicated will add cost and development time, or worse, may result in the project never getting finished.

[phipscube]

I've been interested in doing something like this for a while but never had the time (or energy) to really get going. I recently bought some of those premade chips from Poland and just this week I got them running in my A1200 with a custom 3.9. I've been trying to get vanilla 3.1 running on the upper part of the chips but its not working ATM, just get black screen when I switch on pin 1. Not sure if my original roms were dumped right so tried the WHDLoad versions split up and byte swapped in remus, they didn't work either. At some point next week I want to test if its something to do with the +5V line in my machine been too low for the flash and will try put 3.1 in the lower part (where 3.9 sits now).

I'd like to be able to flash them form the amiga at some point. My idea was to have 3.1 permanently in the upper part of the flash chips, then when switched into 3.1 mode have a simple CLI boot up and allow flashing of the lower part via software (but I'm no software designer so a friend may look at this at some point if I get the hardware bits going)

Anyway its nice to see someone with a bit more hardware design clout is on the case

Here's a couple of sneaky pics I was going to put in my A1200 project thread: -

[kipper2k]

can you post pics (same size) of the underneath of those boards please . One thing about the way i want to design mine, i want to leave the pins open so a ROM chip could be piggybacked, it would probably need to be reworked a bit though in order to accomodate this option