exxos's Atari, Amiga & retro forum

Yeah you can use addmem, this was trixters mem range....addmem 08000000 0fffffff pri 40

Yeah the mediator doesnt seem to obey the CFG_IN/CFG_OUT pins ... (probably because most accelerators dont use them) So instead it allows accelerators to have one autoconfig card before it starts spamming the bus at 0xE80000. Which meant that we could either have ram or the speed control device as autoconfig. Other accelerators have dirty hacks to put controls up at 0xFFxx..... but i just wanted to be compatible.

So for A1200 (and probably CD32 too) the autoconfig device will be the speed control device and the ram will go in the A1200 fastmem area.

Im pretty underwhelmed with the Mediator hardware.. its extremely picky with its timings.

supaduper wrote: 12 May 2021 07:54 Yeah you can use addmem, this was trixters mem range....addmem 08000000 0fffffff pri 40

Works for me as well (thanks), curiously, I can't manage to get exec along with the libraries depicted in 32-bit RAM, MuFastRom PROTECT ON in startup-sequence didn't produce results. Just to note, I pulled the 680x0 libraries and mmu.lib from BestWB and sysinfo reports a 68060 with MMU and FPU present (as it should, it's a rev. 1 full 68060).

I've noticed screen draws/graphics tests in AIBB show it on par or below an A4000, (likely because almost 50% of the libraries are in ChipRam). Still, this works surprisingly well (I don't have the timing fixes for 123c and 125c as verified tonight, but I see no instability at this stage)

I had quite a few problems trying to get it running stably with 3.0 and eventually gave up and put 3.1 roms in there. All problems then vanished! I blame the roms, not the accelerator

Finally got the bugger up and running after the insertion mishap, explained here: https://www.exxosforum.co.uk/forum/viewt ... =84&t=4257

Changed out the 150 pin connector, talk about a task in patience :!: and beveled the A1200 edge.

Gingerly flicked the power switch and was very happy to see first this


Then


The memory CPLD is certainly pumping especially the top corner of the chip was about 5-10 degrees hotter than the middle.
Probably will put a couple of passive heatsinks on the 2 CPLDs.


The bus CPLD was around 43 degrees and the 060 Rev 6 did even break a sweat at 50Mhz (mid 30s)

Thanks @alenppc for the nice build and thanks @terriblefire for designing this great upgrade

Now it is time to have a little play :D

I put a low profile 5v raspberry pi fan on that CPLD and powered it from the TF1260.
It keeps the chip nice and cool and has a side effect of cooling the cpu a bit with the breeze it creates. I'm sticking with this as my A1200 seems rock solid like this and never gets that hot.

Is there a TF Fan Speed program?
I think on one of Gadgets screen grabs on this TF build there was a folder that show this program.
My fan is currently powered off my Gotek, and mounted to it!

I don't think the speed control works at the moment! I just have a 12v fan connected on the 3 pin fan connector - middle is 12v, right hand side is the switched ground (I think).

Yeah not sure whats up with the fan stuff. TBH... unless you have a solas its not useful as there are no temp monitoring IC's on the TF1260.

Danoo wrote: 15 May 2021 10:13 The memory CPLD is certainly pumping especially the top corner of the chip was about 5-10 degrees hotter than the middle.

Do bear in mind that laser is just an aiming aid, and usually aligned nearer the top of the IR sensor range. From the distance/position you're holding the meter in that photo, you're likely measuring the whole CPLD, if not more.