Trying but failing to update my TF536

[Scrappysphinx]

Hi all, i am trying to update my TF536 with the newest firmware that was released in October.

I have tried it with a pi first and it finds the card and identifies the cpld but fails with a mismatch at fuse 9: 0vs 1

i then tried using the ISE Impact via virtual machine which is how i programmed it in the first place and it finds the cpld and identifies it but gives me an error about "Can't find suck instruction IDCODE" and all attempts to erase or program it fail with the following error

INFO:iMPACT - Current time: 2 Jun 2023 15:04:36
// *** BATCH CMD : Erase -p 1 -o
PROGRESS_START - Starting Operation.
INFO:iMPACT:451 - The device 'xc95288xl' is in 'in-system programming' mode.
Maximum TCK operating frequency for this device chain: 6000000.
Validating chain...
INFO:iMPACT:1206 - Instruction Capture = '0001'
INFO:iMPACT:1207 - Expected Capture = '1001'
INFO:iMPACT:2129 - Boundary-scan chain test failed. Please check tdi->tdo connection between the cable and device:'1' ( 'xc95288xl').
A problem may exist in the hardware configuration.
Check that the cable, scan chain, and power connections are intact,
that the specified scan chain configuration matches the actual hardware, and
that the power supply is adequate and delivering the correct voltage.
PROGRESS_END - End Operation.
Elapsed time = 0 sec.

Anyone have any tips on what i might be doing wrong or how i can reprogram it?

Thanks

ISE.jpg (133.27 KiB) Viewed 3025 times

Pi.jpg (287.17 KiB) Viewed 3025 times

[Steve]

I followed this guide to flash mine, your commands look different:
(With a Pi)
https://linuxjedi.co.uk/2020/12/01/prog ... pberry-pi/

I think the fuse thing means the chip is locked from being erasable and then of course it's failing to write because it cannot erase.

[Scrappysphinx]

I was also following the LinuxJedi post but followed his updated guide located here as pi os has been updated now (not that i understand Linux all that well)

https://linuxjedi.co.uk/2021/11/25/revi ... pberry-pi/

[Scrappysphinx]

Just for a bit more info, i have checked continuity from the CPLD to the Jtag header pins and that checks out fine. The card does boot and IDE works but locks up shortly after booting into workbench with Kickstart 3,1, OS 3.2 lasts a bit longer but again locks up eventually.

Both OS boot fine with a standard 68000 cpu (from ADF via gotek)

System is a Denise (A500+ ITX Form Factor) with 2mb chip ram

[matt020]

Check the JTAG wiring to the Pi...

Pin JTAG
1 3V3
6 GND
7 TMS
11 TCK
13 TDO
15 TDI

In the Pi, issue this command, no need for sudo, just do it as it is...

xc3sprog -c matrix_creator -j

and you should get a list of the CPLDs information from the TF card. If you don't, you may have a cable problem. If you get a list of CPLD information then issue this command:

xc3sprog -c matrix_creator -v -p 0 tf536blahblah.jed

I use this above method, I use a dupont cable (has individual pins on each end of the cable and I have to squint my eyes in order to stick the pins through the holes on the card). Sometimes it fails to write, and it is because the cable is not making a good connection, even thought it looks fine. I wiggle the cable and try again. Sometimes I do this over and over and eventually it makes a good enough connection and it works.

I think you have a problem with the cable, it's just my guess.

[Scrappysphinx]

Thanks for your reply

I have double and triple checked the cabling using dupont cables from the pi to a soldered header on the TF card because i didn't want to have to wiggle the pins to get a good connection.

Using your xc3sprog -c matrix_creator -j command gives an unknown cable

pi1.jpg (28.38 KiB) Viewed 2952 times

however using the updated sudo xc3sprog -c sysfsgpio_creator -j command from linuxjedi (because of the updates to pi os)

does indeed locate and identify the cpld so i'm pretty sure the cabling is fine

pi2.jpg (32.54 KiB) Viewed 2952 times

I plan to reflow the CPLD today and try again as that's the only other thing i can think of trying

[DoG]

The 3,3V on the JTAG is just for reference. You probably need to power the TF536 itself?

[Scrappysphinx]

Thanks DoG, i understand that its just a reference and the TF needs powering. I do have it installed in an amiga for power, the same as when i programmed the firmware after building it

[supaduper]

Hey Scrappy I have had the odd Xilinx cpld that would not program using the pi and having the same the fuse error, the only way was to use the Xilinx impact utility as I found that the ERASE was protected on the CPLD, so In IMPACT I just went into propeties and unchecked ERASE protect and then I could erase the IC and program it

[Scrappysphinx]

Hi buddy, i have tried with Impact too, i have tried an erase but it fails, i then tried the override write protect check box and apply but again it fails with the following error.

INFO:iMPACT - Current time: 6/3/23 12:07 PM
// *** BATCH CMD : Erase -p 1 -o
PROGRESS_START - Starting Operation.
INFO:iMPACT:451 - The device 'xc95288xl' is in 'in-system programming' mode.
Maximum TCK operating frequency for this device chain: 6000000.
Validating chain...
INFO:iMPACT:1206 - Instruction Capture = '0001'
INFO:iMPACT:1207 - Expected Capture = '1001'
INFO:iMPACT:2129 - Boundary-scan chain test failed. Please check tdi->tdo connection between the cable and device:'1' ( 'xc95288xl').
A problem may exist in the hardware configuration.
Check that the cable, scan chain, and power connections are intact,
that the specified scan chain configuration matches the actual hardware, and
that the power supply is adequate and delivering the correct voltage.
PROGRESS_END - End Operation.
Elapsed time = 0 sec.

erase1.jpg (40.28 KiB) Viewed 2923 times