TF536 Rev 2 Design Complete (all gone)

[8 Bit Dreams]

I will build buffered IDE
got my prototype yesterday, started my build with non buffered ide first, to be on the same page with You guys, but used cheap Samsung RAMs instead. IDE is not working with firmware from rev.1, so moved to rev.2 fw that Stephen shared here. Card is a bit flaky atm, i assume it is due ram (these ram chips are doing nice on TF330 otherwise).. will swap these today and see how card behaves with original chips, also waiting on buffers from mouser - there is a delay in delivery, seems FedEx has some problems with shipping atm...

[terriblefire]

Guys i'm not sure how many times i have to say this. Firmware incompatibility is the reason for a revision bump. If boards are firmware compatible they get an A,B,C bump e.g if this board was firmware compatible with rev1 then it would have been rev 1b. Please stop trying to use the wrong firmware because sooner or later it will damage something.

[alenppc]

I will build buffered IDE
got my prototype yesterday, started my build with non buffered ide first, to be on the same page with You guys, but used cheap Samsung RAMs instead.

Which RAM is this?

[LADmachining]

Managed to get my board completed and it powered-up into DiagROM straight away. Currently about 3/4 of the way through a slow memory test. 64megs sure takes a long time to test!

Once it completes this test, it will be installed in my A500++ board for further testing including a soak test on AIBB, and will report back on how that goes.

Very pleased with how this build went, much quicker than previous boards. Thanks Stephen!

TF536Rr2 First Boot Mem Test.jpg (135.73 KiB) Viewed 4301 times

(oh, and the workbench is in need of a tidy!)

[terriblefire]

I probably wouldnt bother letting it complete. If it gets through the first meg its fine.

[PaulJ_2.0]

I ran frontier via WHDLoad for 3-4 hours and it was faultless.
The CPLD did get quite hot though in comparison to Rev1 which runs fairly cool, but no actual issues to report on stability.
I will run some more demanding demos and report back.

[LADmachining]

Well, it has been running various games and demos for a couple of hours without issue. I did note that when booting workbench, it gurus just as it starts booting from the CF card. Disabling caches in the early startup menu allows it to boot to workbench.

Disabling the startup-sequence, then loading WB, if you try to enable IBurst or DBurst caches in SysInfo or AIBB results in an immediate crash. Don't remember this behaviour with the TF534.

After the most recent reboot, it has now decided to stop reading the CF card. Leaving it shut-down for a while to see if it is a temperature related issue that may require reflowing the CPLD.

IDE speed in SysInfo with CPU Fastrom was around 4.5MB/s, a bit faster than my tests with the 534.

[PaulJ_2.0]

I get no such problems enabling and disabling iburst or dburst cache in AIBB or Sysinfo.
Infact I can't report a single glitch, everything has worked flawlessly for me.

A500+
2MB Chipmem
3.1.4 KS
CF card with OS3.1.4

[LADmachining]

I get no such problems enabling and disabling iburst or dburst cache in AIBB or Sysinfo.
Infact I can't report a single glitch, everything has worked flawlessly for me.

A500+
2MB Chipmem
3.1.4 KS
CF card with OS3.1.4

Thanks.

Looks like I have some more work to do then.

[PhilC]

Try a different cpu first maybe? But as Paul says, its ran great so far for us.