TF360 Status

[theq]

I think thermals and space will be an issue with this card in the end.

My A3660 with a rev. 6 CPU is running at +/- 60°C when fully utilized, and that is without any heatsink or fan attached. With the TF330 being cramped up already and with the TF360 design going to be roughly the same, maybe it's time for a new riser which puts the TF-card outside of the cabinet? Just like the SX-1 riser does. The benefits of the SX-1 riser design is that you could also run a FMV card simultaneously to the TF-card (if the card allows it?). But then again you'd lose the small form factor in which makes the CD32 a great Amiga.

And for the record, I've got another 68060 rev. 6 in the mail so I can make myself a TF360 when you release it for the masses. :)

[terriblefire]

I expect it will need cooling but the CD32 is way better ventilated in this area than an A1200.

[theq]

I expect it will need cooling but the CD32 is way better ventilated in this area than an A1200.

I'm talking A4000 now, but I'll guess we'll see. :)

As you were asking for future request it would be awesome if you could put a 5V-GND and a THERM0-THERM1 pin-header (like what Chucky did on the A3660 card). This way I can connect my Arduino-temperature monitor the CPU's die thermistor.

[terriblefire]

I was planning to actively use those pins in a future version. It’s trivial to throttle the CPU to let it cool.

An A4000 version is a bit away yet

[terriblefire]

Current status is: mine works flawlessly.

Supaduper is sending me his as he can’t get it going now.

Chucky is having fastram issues.

[Higgy]

Hopefully they are simple fixes.
Interesting that the initial Warp1260 power up and TF360 have very similar Sysinfo numbers. TF slightly ahead apart from Warp1260 showing 7.2 on Chip Speed vs A600.
I guess for the same clock speed all 060 @50Mhz give 37*** ish. Are the main gains to be had over vintage accelerators IDE and memory speed? (Oh and the massive price savings TF cards bring :D )

[terriblefire]

Are the main gains to be had over vintage accelerators IDE and memory speed?

The support logic chips and memory tech have moved on since the 90s,

I have got @supaduper's TF360 working nicely. I am going to post it back to him tomorrow.

[Higgy]

Love the latest video. Worlds fastest CD32! Video output seemed to be Composite rather than RGB though this time.
Where do you get the CPU removal tool, a bit like a comb? Cant find those on Ebay.
Your comment about out of spec memory reminded me of those MiST FPGA issues with Archimedes and MSX Cores. Different boards had different spec memory.
Final thing, did you say on earlier videos you were hoping to get 128mb on later/final revision?
Thanks

[terriblefire]

Love the latest video. Worlds fastest CD32! Video output seemed to be Composite rather than RGB though this time.

Yup.. my testbed CD32 doesnt have working RGB. Its a very hacked up motherboard that has headers added to let me probe all the signals i need. I dont want to mess up another board.

Where do you get the CPU removal tool, a bit like a comb? Cant find those on Ebay.

ChuckyGang sent me to open some beers.

Your comment about out of spec memory reminded me of those MiST FPGA issues with Archimedes and MSX Cores. Different boards had different spec memory.

Yup. I did the Archimedes core for MiST and this code is derived from that code.

Final thing, did you say on earlier videos you were hoping to get 128mb on later/final revision?

No it will be 256Mb on the final verison. But no need to waste cash on that in a pure prototype.

[terriblefire]

Update:

My card now running at 66Mhz with some issues.

1. Noticed that I CLA pin was low and was causing the address bus to "walk". Tied it high with a dupont wire.
2. Bus CPLD was always driving the cards internal data bus during a read... this caused alot of the heat.
3. The glitch at 66Mhz is always a lock up at address $DFF096. Looks like the Bus CPLD doesnt register the TS. This will be fixed in Rev 1 by using TIP.