Where I attempt to build two Raven060!

[Mikerochip]

ROM3 and ROM4 are not exposed anywhere.
They could possibly be exposed on some of the free header pins next to the ATF1508 but I'm hesitant to do so since they'll be going away on rev.A2 (pins are needed for other stuff)

But regardless, it'll need some bios+firmware effort to make it work.
ISA cycles needs to be initiated on cartridge access else you're going to be reading off the wrong databus.
In terms of bios it needs to PMMU-map the cartridge space as seen by the cpu to some ISA-RAM range.

I think the more elegant solution is to create a real ISA card that decodes and reacts to that isa-ram range.
I've always just assumed this is what the Hades cartridge-port ISA card does, and now that I looked around and found the schematic I see that yes, that is exactly what it does. It has a gal16v8 on the card that is responsible for "talking ISA" and generating the rom3/4 signals.

Ah. I remember looking at that, long ago.
Yeah, a neat solution.
The reason I'd forgotten about it, or, just dismissed it, I think, is the blasted cartridge sockets aren't available anymore.
I don't think anyone has any spare.
And of course, they're non standard 1mm pitch, or something.
I suppose, we should look at either getting some made... 3d printed??
Or bodging an existing socket instead...

Tomorrow's problem

Actually, tomorrow's problem is finishing my bga psram routing! I'm nearly ready for it.

Just ram, cpu and vga left, after the ATF1058 is programmed.

[agranlund]

The reason I'd forgotten about it, or, just dismissed it, I think, is the blasted cartridge sockets aren't available anymore.
I don't think anyone has any spare. And of course, they're non standard 1mm pitch, or something.

Yeah I remember there was some talks about these getting super hard to find nowadays :/
I've got one to grab from my for-parts STFM but that's about it.

[Mikerochip]

OK! Great success!!

I finally got my ATF1508 programmed, and without soldering, thanks to some clever jtag/board layout from @agranlund

Code: Select all

Open On-Chip Debugger 0.12.0+dev-snapshot (2025-07-16-14:15)
Licensed under GNU GPL v2
For bug reports, read
        http://openocd.org/doc/doxygen/bugs.html
Warn : TMS/SWDIO moved to GPIO 8 (pin 24). Check the wiring please!
adapter speed: 400 kHz
Info : BCM2835 GPIO JTAG/SWD bitbang driver
Info : clock speed 400 kHz
Info : JTAG tap: ATF1508AS.tap tap/device found: 0x0150803f (mfg: 0x01f (Atmel), part: 0x1508, ver: 0x0)
Warn : gdb services need one or more targets defined
svf processing file: "/home/mikerochip/raven060/nessi_250714.svf"

Time used: 0m13s956ms
svf file programmed successfully for 3569 commands with 0 errors
shutdown command invoked

I installed Raspberry Pi OS "bookworm" and then openocd. (sudo apt update and then sudo apt install openocd)
So, it should work quickly and easily, on any raspberry pi.


Raspberry Pi JTAG pinout.

Raspberry Pi OpenOCD 12 JTAG Pinout.png (69.17 KiB) Viewed 166 times

I wired the jtag pins on the pi, and gnd, to the corrosponding jtag pins on the raven board.
I didn't connect 3.3v or 5v, and the raven board then, needs to have an atx psu connected, and the board powered on.

One custom cable later, and I could send jtag commands from the Pi to nessi!

OpenOCD commands.

Code: Select all

openocd -f /usr/share/openocd/scripts/interface/raspberrypi-native.cfg -c "adapter speed 400" -c "transport select jtag" -c "jtag newtap ATF1508AS tap -irlen 3 -expected-id 0x0150803f" -c init -c "svf /home/mikerochip/raven060/nessi_250714.svf -quiet" -c "sleep 200" -c shutdown

The only problem I had, was, when I tried it first, I got nothing but errors.

Code: Select all

svf processing file: "/home/mikerochip/raven060/nessi_250714.svf"
Error: tdo check error at line 21
Error:     READ = 0x0000000
Error:     WANT = 0x150803f
Error:     MASK = 0xffffffff
Error: fail to run command at line 711
Error: tdo check error at line 21
Error:     READ = 0x0000000
Error:     WANT = 0x150803f
Error:     MASK = 0xffffffff
Time used: 0m3s283ms
svf file programmed failed

The CPLD was second hand, and, obviously used, before. And, the previous owner had disabled the jtag pins.

So, reading up, you can re-enable JTAG pins, by applying 12V to the /OE1 pin.
In this case, Pin 84 is the /OE1 pin, which anders had connected to J105, pin 2.
So, I did just that, via a 2K2 current limiting resistor.
I connected a one end of a pin header/jumper cable to a 12v fan connector (eg J206 Pin2) and the other end to a 2K2 resistor, and that to pin 2 of J105.

That immediately enabled the jtag pins, and I could program, using an SVF file.
But! Anders doesn't have an svf file in the releases. Luckily, we can create one using ATMISP7.
(Which works both under windows, and linux via wine)

Loading the program, click edit, and add device.
Pick the ATF1508AS, and select "Program/Verify" from the command list.

ATMISP7 Add Device.png (6.26 KiB) Viewed 166 times

Once selected, you need to tick "Write SVF" and set a file name for the new .svf.

Clicking run on the main window, produces the .svf file.

So! Big success.

if you have a jtag programmer installed under windows, you can also use openocd from there.
My knock off usb blaster no longer works under windows 11, so, I went the raspberry pi route.

I had initially tried with that old usb blaster clone, and my st-link v2 cheapie clone.
The former blue screens windows! and the latter doesn't expose two of the jtag pins. (It's mainly used for SWD/SWIM on STM32 chips, I guess)

I've got some new st-link v2 clones coming shortly (with jtag and swd pins), and two new usb blaster clones also.
I'll report on their success rate, when they arrive, if anyone wants to know!

I use the xpack openocd binaries for openocd under windows, normally.
https://github.com/xpack-dev-tools/openocd-xpack

Anyone wanting a full tutorial, let me know!
Most of the salient points are above.

SO!
To that end, Anders, (In the A2 revision) can we have a 12v trace and a 2k2 resistor placed inline, connecting to another pin so we can jumper it to J105 pin 2? maybe to the right of pin2.

Might be a useful "force programming" jumper... ?? In case something goes horribly wrong!
ie jumper pin in 1-2 or 2-3 for normal use, and rotate 90 degrees, and connect pin 2 to a new "programming enable" pin?

it's not a big deal, now that I know what I'm doing But, is a useful to have feature, I guess!

But, much more usefully for you, doing so, would let you repurpose the four jtag pins on the 1508 for something else, should you need them...

Next;
pcb routing/layout for bga ram simms

*Edit, I've added the .SVF file, for anyone who wants it, along with an "Erase the ATF1508AS" .svf, to set it back to factory

[agranlund]

What a good writeup @Mikerochip , those official programmers are quite expensive so it's good that it can be programmed with other methods!

Putting a "proper" jumper for jtag restore isn't a bad idea at all!
I like the idea of being able to use those four pins for other thing

[Mikerochip]

What a good writeup @Mikerochip , those official programmers are quite expensive so it's good that it can be programmed with other methods!

Thanks! I wasn't sure if I should include more detail.
But, I was determined to not spend several hundred euro getting one of the official programmers.

It's especially useful to know, given a raspberry Pi zero is only a few euro, and can do the job just fine.

Putting a "proper" jumper for jtag restore isn't a bad idea at all!

Excellent, that's what I thought!

[dml]

Anyone wanting a full tutorial, let me know!
Most of the salient points are above.

Thanks for writing this up & detailing it so clearly, so we can try it without too much guesswork. I had seen the 12v trick described but wasn't clear this was safe to do on-board!

I have a locked ATF chip from my first attempt at flashing (IIRC think it was the only reason or the primary reason my Raven didn't fire up immediately) so I should be able to recover it the same way.

[Mikerochip]

Thanks for writing this up & detailing it so clearly, so we can try it without too much guesswork. I had seen the 12v trick described but wasn't clear this was safe to do on-board!

I have a locked ATF chip from my first attempt at flashing (IIRC think it was the only reason or the primary reason my Raven didn't fire up immediately) so I should be able to recover it the same way.

Excellent, glad it's of use to someone other than me!

It's only safely possible to do, because @agranlund cleverly broke that pin out to a jumper!
I presume he knew about 12V on /OE1 before today

Also, you just need to put the cable in place, power on the board, give it a second or three, then power it off.
As soon as the board powers on, the 12V enables the JTAG pins permanently, it seems.
You can then safely power off, and remove the cable. It doesn't need to remain connected during programming.

Then program as normal.

A raspberry pi zero is only €6 plus postage, so, it's cheap enough that one can almost be permanently left connected to the jtag, to make an all in one self contained Raven board

I made a molex to usb type A female, which I can use to power the Pi Zero. I think I'll put a switch on it, too.
You can buy them cheaply on ebay/Aliexpress premade, also.

With Anders creating an I2C port on the Raven, I imagine a small program could be written on the Pi, to wait for a command, and then start the flashing script. You could then, in theory, program Nessi from the Atari
It would require a reboot, most likely, but, it's doable!!
That, though, is another day's work

[agranlund]

It's only safely possible to do, because @agranlund cleverly broke that pin out to a jumper!

I wouldn't go that far
Lucky coincidence is more closer to the truth