exxos's Atari, Amiga & retro forum

8 Bit Dreams wrote: 02 Dec 2018 19:08

terriblefire wrote: 02 Dec 2018 17:49 Please tell me where already :)

Please don't laugh...on Ebay... :oops: :oops: :oops:

Oh man, that’s not a good thing

alenppc wrote: 02 Dec 2018 19:29

8 Bit Dreams wrote: 02 Dec 2018 19:08
Please don't laugh...on Ebay... :oops: :oops: :oops:

Oh man, that’s not a good thing

Those arent chips.. they're black labels. :P

teasing screenshot..

Something that has been mentioned a few times is the ram latency. Its not as simple as 2 wait states etc..
I dont think anyone has done SDRAM on the Amiga before? Would be interested to know.

I use aliexpress quite a bit and you can get the XC95288XL-10TQ144C for $18 for qty 5 ($3.60 each) with free shipping. Just takes about two weeks to get them on average. I have gotten most of my parts there to build the TF530 and parts are on the way for the TF534 so I will have them and just have to wait for the PCB.

I know the ACA1233-55 uses sdram (improved sdram controller noted in specs) and I assume the other ACA123x cards do as well.

Both Jens and Matze are using SDRAM in their designs, we can't see Jens cpld code, but You can take a look on Matzes code in 68EC020TK accelerator card..

Where is Matze’s SDRAM implementation? I could only find SRAM designs.

I can’t see why you’d make a 8Mb SDRAM card. EC020 only supports 8Mb expansion space on the Amiga.

EDIT: it seems the Matze card is 16Mb. Must fill in all the holes in the memory map and do maprom. Fair enough.

terriblefire wrote: 03 Dec 2018 13:31 it seems the Matze card is 16Mb. Must fill in all the holes in the memory map and do maprom. Fair enough.

Yes, he even occupied MAPROM space, so total of this card is 10.5MB

he's also fighting with 64MB SDRAM expansion for 68030TK card,
There was same chipram issue You have atm, solved it three weeks ago.. See here:
https://gitlab.com/MHeinrichs/68030-RAM-IDE
however, he has some kind of IDE trouble, Systest runs all day long without a single crash, but as soon as IDE activity is present - card freezes...

seems that Commodore custom chips are very specific, these was made different as a default standard expected, there was datasheets for internal use for sure, back in the days these leaked from Braunschweig and helped designing accelerators and stuff in Germany..Now it's more complicated to do it without this specific knowledge, that exlains why so many projects was started but never worked and was dropped..
i still belive You will find a way to solve these issues, and this will be a huge step into accelerated Amiga world, this will help to design next generation of TF5xx just like You did with INT2...

I have the clock issues solved.

Its working now. but no burst yet.

I've just popped 2 pcbs in the post to superduper for some initial build reports from him.

I know it seems like i've done this fast but i've done SDRAM and even DDR before.

I had a look at Matze's code. He's not being nearly as sneaky as I am. Its possibly there are some remaining issues there. Hopefully not.