Falcon error: T4 Video Counter in Memory Controller

[Badwolf]

A candiate raises its head.

Valid XBERR signals go into the GAL U63 but EBERR is never asserted. HIGHZ is set at all times other than during the early reset.

Screenshot 2025-09-12 at 22.37.28.png (108.38 KiB) Viewed 249 times

This, is the only GAL that's not been swapped over from my working Falcon board as, on my board, it's the one with the clock patch piggybacked on it.

Time for that to change. Will it be the only problem? I doubt it, but I feel a new GAL will be in my future sooner rather than later.

@exxos, you don't stock replacements for these (any more?), do you?

Cheers,

BW

[dml]

@exxos, you don't stock replacements for these (any more?), do you?

I burned a set of these in the past - I think I just used the common TL866 thing. The JEDs should still be downloadable.

[exxos]

@exxos, you don't stock replacements for these (any more?), do you?

I burned a set of these in the past - I think I just used the common TL866 thing. The JEDs should still be downloadable.

I've never tried doing them. I could never work out if the jeds floating around were final or not. I tried the gals on a mste once and couldn't never get the thing stable. So kinda assumed the same with the falcon gals. So never took the risk in them.

Maybe @tOriman could do something to check original gals vs the ones online.. Unless someone knows 100% they are good jeds to start with..

[dml]

Maybe @tOriman could do something to check original gals vs the ones online.. Unless someone knows 100% they are good jeds to start with..

I have no way to know that the JEDs I used are 100% match - but I can say that I revived a dead machine with burned GALs and it is still here, still working. I haven't noticed any special problems with it.

[mikro]

Valid XBERR signals go into the GAL U63 but EBERR is never asserted.

Hmm.

And btw yes, I did program GALs anew with JEDs from Rodolphe Czuba's website, they work well.

[exxos]

I have no way to know that the JEDs I used are 100% match - but I can say that I revived a dead machine with burned GALs and it is still here, still working. I haven't noticed any special problems with it.

I just tried reading on of my Falcon GALs and it read data back! Assume its not code protected

Code: Select all

1111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111101010101010101010101010101010101010101010101010101010101010101011010101010101010101010101010101010101010101010101010101010101010010101010101010101010101010101010101010101010101010101010101010110101010101010101010101010101010101010101010101010101010101010101001100110011001100110011001100111001100110011001100110011001100110011001100110011001100110011000110011001100110011001100110011001100110011001100110011001100110001100110011001100110011001100110011001100110011001100110011001110011001100110011001100110011001001100110011001100110011001100111001100110011001100110011001100110011001100110011001100110011001110011001100110011001100110011001100110011001100110011001100110001100110011001100110011001100110011001100110011001100110011001100011001100110011001100110011001100000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111010001................................................................11100001....................$

[Badwolf]

Valid XBERR signals go into the GAL U63 but EBERR is never asserted.

Hmm.

And btw yes, I did program GALs anew with JEDs from Rodolphe Czuba's website, they work well.

You're simply a trend-setter, Mikro!

But yes, bizarre how I've ended up in a couple of the same pits as yours! First that reset diversion and now this!

I've got some GALs on the way. Which is good anyway as I've wanted to play with GALs or their more modern versions.

BW

[Badwolf]

Code: Select all

1111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111101010101010101010101010101010101010101010101010101010101010101011010101010101010101010101010101010101010101010101010101010101010010101010101010101010101010101010101010101010101010101010101010110101010101010101010101010101010101010101010101010101010101010101001100110011001100110011001100111001100110011001100110011001100110011001100110011001100110011000110011001100110011001100110011001100110011001100110011001100110001100110011001100110011001100110011001100110011001100110011001110011001100110011001100110011001001100110011001100110011001100111001100110011001100110011001100110011001100110011001100110011001110011001100110011001100110011001100110011001100110011001100110001100110011001100110011001100110011001100110011001100110011001100011001100110011001100110011001100000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111010001................................................................11100001....................$

That's what SHE said..!

BW

[Badwolf]

OK. I wasn't expecting that:

I removed U63 from my main Falcon (breaking a clock wire in the process. Grr.) and transplanted it...

...same symptoms!

I didn't have much time to investigate, but here are the options I can see:-

1) EBERR is shorted to VCC so U63 can't pull it low. I'm measuring 1k between them, so I don't think it's this.

2) EBERR is shorted to some other line that's being actively pulled high. Possible, but it's none adjacent to it on the GAL or the expansion connector. This needs further pursuit, however.

3) DFB1 is erroneously pulling EBERR high. Possible as currently my rev 4 board is in there, but this behaviour was also seen with DFB1X, so seems an outsider.

4) U63 is behaving correctly and EBERR is in the HIGHZ state because of some other failing condition. This seems the most likely to me, but I need to read the equations again and work backwards. The HIGHZ input appears consistently high apart from during reset, however.

5) Both GALs have failed in the same way upon insertion into this board. The nightmare option, but shouldn't be ruled out. Also why the preference is always to put suspect chips in working boards rather than the other way around, but in this case the clock patch complicated things.

6) ..? Open to suggestions.

To be continued.

BW

[exxos]

Lift the gal pin out out of the socket, and see if it tracks then?

Might be worth looking at the gal code as I think it was all a bit if a mixed bag of chaos of what's driving what.