exxos's DFB1 trials

[Badwolf]

So come on then, I'll show you mine if you show me yours.

What did you do & what's the current config?

FWIW I decided to take the CPLD out of the equation and basically DSACK[x] now go high-Z on FPU access allowing the two bodge wires I showed previously to drive the CPU pins directly.

So this is exactly the datasheet implementation.

It made no difference on my board. My FPU still tops out at 30MHz.

So next step was to experiment with a separate oscillator -- I disabled the FPUCLK entirely and fitted a second socket for the FPU osc. This at least feeds a proper 5V signal to the FPU. Again, no difference. Tops out at 30MHz.

I think I've a 30MHz chip.

So I've not learned much here other than going high-Z on FPU does work, so I can avoid the complications of routeing via the CPLD if I do another board revision.

BW

[exxos]

So come on then, I'll show you mine if you show me yours.

Tune in tomorrow for the next thrilling episode...

Show Spoiler

HINT- Measures the PLD clock from the master OSC

[Badwolf]

Show Spoiler

HINT- Measures the PLD clock from the master OSC

??

Too opaque for me.

BW

[exxos]

Too opaque for me.

The clock was basically somewhere between 1-7V. We actually had this type of issue on the H5. Some OSC worked, others didn't. Despite them all having a proper 32MHz output. It would basically cause the machine not to reset.

But in this case it must be causing some really weird internal glitches in the PLD itself. Or just not registering a logic low properly half the time.

Coincidentally the keyboard going nuts in the diagnostic cartridge has also disappeared now

[Badwolf]

Too opaque for me.

The clock was basically somewhere between 1-7V. We actually had this type of issue on the H5. Some OSC worked, others didn't. Despite them all having a proper 32MHz output. It would basically cause the machine not to reset.

But in this case it must be causing some really weird internal glitches in the PLD itself. Or just not registering a logic low properly half the time.

Coincidentally the keyboard going nuts in the diagnostic cartridge has also disappeared now

Aha well done. I have one 50MHz oscillator like that -- won't work for toffee in this board but will drive my VGA card fine.

It is bloody odd that that's the problem, though. The board wasn't working with FPU here either and I'm not using the same oscillator as you! So what else have you done?

I can abandon this direct DSACK-DSACK firmware until next board spin, then, I take it?

BW

[exxos]

Aha well done.

I have one 50MHz oscillator like that -- won't work for toffee in this board but will drive my VGA card fine.

Yeah it is pretty odd they can oscillate up to higher voltages like that in the first place.

It is bloody odd that that's the problem, though. The board wasn't working with FPU here either and I'm not using the same oscillator as you! So what else have you done?

running your latest released firmware ( regular one that is)

I did change the pullups to 1K on DSACKx but not sure that made any difference. I think you had 2.2K in there ? So should be fine anyway.

I can abandon this direct DSACK-DSACK firmware until next board spin, then, I take it?

I would say so yes. I will do some loading testing on the oscillator which I will look into more tomorrow. I would put some series resistance in the DSACK lines from the FPU to PLD though. Assume 68R be about right. May need some experimentation..

I would definitely address all the capacitors issues we talked about earlier in the new spin. I would probably suggest just going with a separate oscillator for the FPU. As we cannot really assume what speeds the CPU or FPU can run at currently.

Incidentally it still does not run at 50MHz, it crashes on the first test But I need to put maybe a 33R in series with it yet, its hardwired to the OSC currently. But I think this particular FPU is just simply a 40MHz part.

[exxos]

Output of the OSC without loading.

IMG_0252.JPG (290.66 KiB) Viewed 977 times

With 133pF.

IMG_0253.JPG (275.96 KiB) Viewed 977 times

[stephen_usher]

Aren’t you supposed to pass the output from crystals through a capacitor to remove any DC offset anyway?

[dml]

Did you try the mandelbox raytracer yet? It should be a decent burn-in

[exxos]

Aren’t you supposed to pass the output from crystals through a capacitor to remove any DC offset anyway?

I have done that in the past on various projects when needed. Like on my RSO boards for the CT60 and even some Falcon clock patches. Though there doesn't seem to be DC offset in this case. Just seems to be over-volting more than anything.