Dead pixels / solved

[veekoo]

I had two ROM chips with v5a and my machine showed 3 dead pixels (black pixels) at start screen with both. One of them is in the ROM texts O-letter. Two of them is below program texts r-letter. I used to think this might be related tolerances with components. These black pixels don't show up in any other screens and computer functions perfectly.

Recently I reprogrammed one ROM chip to DEVROM and the black pixels are gone with this. So now I am thinking there was some bug with memory address in video area and now it's fixed with DEVROM.

[at67]

Much more likely that your ROM was not burnt completely correctly, (verified), or had not been erased fully before programming. Re-erasing and pre-programming it with DEVROM probably just overwrote a bad ROM image with a good ROM image.

I have never seen ROMv5a exhibit artifacts like that in emulation or on real hardware in many hours of testing.

[veekoo]

I got these two v5a's from different source and I verifyed the chip againts file

[at67]

Yeah I take back what I said, in Discord about a year ago those same black pixels were seen by others in the same positions. This was my best guess back then:

the screen corruption is most likely the SDCard handling code for the unfinished expansion card Marcel was working on, there are a number of bugs in that code and ROM, as well as the annoying Audio bug , I personally don't run that ROM, I run DEVROM instead, (you can build that at the command line with 'make devrom')...the only reason i see to run it is to finish the work Marcel started on the expansion card, which in some aspects is close to being done

So that makes more sense why DEVROM solved your problem.

P.S. Thanks to bmwtcu for reminding me.

[lb3361]

the screen corruption is most likely the SDCard handling code for the unfinished expansion card Marcel was working on, there are a number of bugs in that code and ROM, as well as the annoying Audio bug

When I read this, it made me think of https://forum.gigatron.io/viewtopic.php?f=4&t=329 which describes an expansion card related glitch. Except of course that this should not happen without an expansion card. Without an expansion card, the ctrl instruction writes some random data at the address given by its argument. However, in ROMv5a, the ctrl instruction is only called twice: as the very first instruction after a hard reset, and at the beginning of Reset.gcl to determine there is no card, long before drawing anything on the screen. When there is no card, [1f8] is set to zero and the remaining checks are skipped. None of this is compatible with a reliable and repeatable problem like this. I find it very mysterious.

Glad to know that DEVROM fixes it. But I do not understand why....

[NewTechPhobic]

Hi for what it's worth I had exactly the same dead pixels in exactly the same place with rom 5a it comes and goes depending on which version of Babelfish I use.