exxos's DFB1 trials

[exxos]

Been doing some more experiments and hit a bit of a weird problem

I did not realise that the original code was booting from the MB ROM then switching to flash. Nothing necessarily wrong with doing that. However, while trying to specifically boot from the flash only, it doesn't In fact is even more weird because it I press the reset button it then boots up normally anyway.

I have changed the reset caps in my Falcon and even added huge values on there. With a 47uF it did boot up once, but not again. I cannot increase the value any further because I don't think the 12K can charge up 100+uF.

I am kind of scratching my head at the differences between a power up reset and pressing the reset button. I don't think there really is any

[stephen_usher]

The difference between power-up reset and the reset button is that the +5V will be stable with the reset button, there's no guarantee of this on power-up. You may need something like an MCP-100 on the power line and your own power-on reset circuit.

[exxos]

The difference between power-up reset and the reset button is that the +5V will be stable with the reset button, there's no guarantee of this on power-up. You may need something like an MCP-100 on the power line and your own power-on reset circuit.

True, though it was held in reset for about three seconds, it should really be enough for the voltages to be stable in that time.

Though it also does not explain why the on-board MB ROM works.

The only possible difference between flash and ROM decoding is that BERR isn't set after the first few addresses are decoded. I know the ST doesn't care, but no idea if the Falcon or TOS404 does..

[Rustynutt]

See the same screen on the AB at power up, after a crash a lot of the time.
Only thing I've came up with, is theres enough capacitance on the board to keep it "corrupt".
Not only on the AB, CT as well. Found with power off, grounding the board to bleed off voltages usually results in a clean reboot. Otherwise letting it set for a period before applying power.
Doug noticed this back in the 90's, and just wrote it off as weird Falcon behavior.
On the CT, when a screen like this appears, grounding out the board (rail to power supply case) causes the fan to intermittently spin even after power has been removed . So it bleeds off over time, or can be bleed off "manually".
Maybe scop it and see if you see the same.

[Rustynutt]

Or disconnecting the power cord

[exxos]

set for a period before applying power.

I shorted all the power rails, but made no odds for bootup. I am at a loss to explain it.

[exxos]

I tried a different board and this one behaves slightly different again.. This when I have to press the reset button about 15 times before it boots up. Pressing reset after that results in resetting perfectly well.

It all seems to circulate around the flash chip itself. For some reason the Falcon not like booting from it.

If I put the Falcon MB ROM in, and use the "disable flash" jumper, it all works fine.

[exxos]

So there are 4 accesses to flash before it dies. So at least it's trying to boot up at least .

IMG_0745.JPG (94.42 KiB) Viewed 1060 times

Not sure how long I can spend on the flash problem. I may have to consider just shipping with disabled until the problem is resolved.

Indeed all the original problems I found with the DFB1 were fixed and tested on the DFB1X. Though at no point have I ever tested the flash out until this week. So obviously there is some problem with that somehow which needs some time to diagnose.

[exxos]

Measured D30 and all this crazy sh*t going on. "normal" for Atari's I guess.

IMG_0748.JPG (283.82 KiB) Viewed 1058 times

IMG_0747.JPG (266.18 KiB) Viewed 1058 times

IMG_0746.JPG (258.18 KiB) Viewed 1058 times

[exxos]

Some progress but

It works if I jumper the flash bank select to use the lower bank.. even though I programmed it on the high bank