Hey, that's a really great site for grabbing heightmaps!

Oh I see, it's drifting off to the side indefinitely. It could be that it's somehow not switching to absolute coordinates, so each absolute X coordinate just sends it off further and further down the X axis, in spite of the G90 that the Mach3 post includes at the beginning of the G-code. Maybe the controller is expecting it with some other G/M codes somehow. It seems like it could just be something finicky with the controller like that. Whatever it is I'd really like to know what it is and work in a fix - understanding the what/why of it.

It could very well be something else too, like the G40. There's no telling what the issue is without some trial-and-error just messing around with it. It could even have something to do with the lack of a "start read" character, the % line. I'd suggest trying removing that first line from the  working G-code and see if it still works, or if its presence is inconsequential.

PixelCNC's posts are relatively simple, I tried to keep all the cruft out and avoid over-complicating anything with features and vernacular nobody ever uses 99.99% of the time. The most complex part are the codeblock definitions at the bottom where it references different variable names/states within the block's string of characters. You could really just directly replace PixelCNC's program prefix codeblock with the one that works for you, but knowing why the Mach3 post isn't working with your 6090 controller would be much more useful to others in the future. That way I can work in a fix without breaking Mach3 support for other controllers that the posts already work for.

Thanks!

That's rather confusing. You say that you commented the CodeRelative "G91" line from the post-processor but the G-code you're outputting doesn't even have a G91 in it - it's using the CodeAbsolute/G90. There's a G91.1 that's being included via the PrefixBlock but that's a different code altogether and only applies to arc/helical motion command coordinates (which PixelCNC doesn't generate yet, just linear feeds and rapids right now).

I have a feeling that what "worked" was something else other than commenting out the CodeRelative line, because that should have no impact on the G-code that's output. I'd investigate further and compare the G-code that is actually being generated. Compare a program generated with and without the CodeRelative line commented out in the post, because they should really just be identical  due to the CodeRelative variable not being referenced anywhere in the actual command block definitions. I'm inclined to  think that the problem might be the G91.1 that's included in the PrefixCodeBlock and that Mach3 controller doesn't know that it's a G91.1, seeing it as a G91 instead. ...but that doesn't explain how "initializing" the machine with a different program causes it to work even with the G91.1 still in there.

Hrmmm.

Life's been pretty good and busier than usual over here for us too :) 

Hey that's a good idea! Go ahead and start a thread and I'll sticky it at the top of the posts and add my own pics to it too. I aim to get the price up a bit further once we hit beta and move to a more professional website, and have better learning materials (tutorials, videos, etc.) Software without learning materials or "proof" of what can be done with it is a bit of a hard sale but judging by existing sales (without any legit advertising/promotion yet) there clearly there are people who can take a look at a piece of software and imagine what the possibilities are and know it's well worth the price.