Falcon030: Installation of clock patch and DFB1X issues.

[stephen_usher]

Having tested the DFB1X board in my Falcon030 last night and found issues running software which accessed the DSP chip in a certain way I decided to go ahead and install the Exxos clock patch V2 board I bought a few years ago when I first got the machine.

So, this morning I dismantled the machine and started the upgrade. Firstly I floated the original resistors from the pads with a soldering iron laden with solder.

IMG_4123.jpeg

I then fitted the V2 patch board to the PAL legs. Instead of using the supplied Dupont header pins I used resitor legs as these can be fitted to the legs of the chip in situ easily and then fed into the board, sliding the board down and then soldering from above.

IMG_4122.jpeg

I then installed the patch wires.

IMG_4124.jpeg

Finishing things off I fitted the "mandatory fix" for the SDMA chip.

IMG_4125.jpeg

After putting things back together I ran tests. Without the DFB1X all tests worked as before. All the demos ran and so did the tracker program.

After fitting the DFB1X I ran the same tests as last night...

Unfortunately there's no improvement. The system fails in the same way with the Joy demo and tracker as soon as they hit the DSP chip. In fact everything works in the same way as it did before the installation.

So, I've now uninstalled the DFB1X board and put the machine back together.

[Steve]

You might want to try the delay / solder bridge for choosing sdma delay on/off. Might have been worth playing with that a little.

Here's a post explaining how to enable/disable the feature:
viewtopic.php?f=29&t=7267&p=119870&hili ... ma#p119870

The only other thing I can think worth trying is the newer v4 clock patch version. And/or making sure the expansion header pins are nice and clean with some Deoxit.

[stephen_usher]

The few times that I got AceTracker working and tried to play a song, after a few seconds it fails with this message:

IMG_4126.jpeg

When Joy and ProTracker fail they generally send the display into a weird mode that the monitor can't handle.

[smoore100]

I must be lucky with my Falcon, I can run the Joy demo and the tracker with the DFB1X fitted without doing any clock patches.

Is there anything I can tun as a test to help you identify what the problem may be?

Many thanks,

Simon

[stephen_usher]

I'm wondering if the firmware on the DFB1X is too old. Not sure how to update it though.

[stephen_usher]

I must be lucky with my Falcon, I can run the Joy demo and the tracker with the DFB1X fitted without doing any clock patches.

Is there anything I can tun as a test to help you identify what the problem may be?

Many thanks,

Simon

Not that I know of. I'm guessing that the clock patch wasn't actually needed in the first place and it's something else which is causing the timing issue.

[smoore100]

Have you tried this?

F030TEST.zip

[stephen_usher]

The stock machine passed all the tests I could throw at it.

With the DFB1X I could run anything as long as it didn't access the DSP in a certain way. Lots of demos which used the DSP would work fine, but others, such as Joy, wouldn't. I had Frontier running for hours with out issue.

This is why I installed the clock patch. But it changed nothing.

[Steve]

I suppose once you've done some of the basic troubleshooting options I mentioned, you'd want to probably get some advice from Badwolf.

[stephen_usher]

This afternoon I refitted the DFB1X, but before I did this I replaced the jumper headers with right-angle ones to make them lower profile and to enable me to fit cables to the enable, ROM select and Xilinx headers so that I could bring them out to the mouse/joystick bay and fit switches for the enable and ROM select.

IMG_4127.jpeg

IMG_4128.jpeg

IMG_4130.jpeg

I can now do everything I need to without re-opening the Falcon. Also, by lowering the header pins the metal shield rests directly upon the top of the 68030 in a way that allows it to become a heat sink, not that it needs it as at 40MHz it's only getting up to about 40C when fully loaded.