On Thursday 05 November 2020 09:01:53 andy pugh wrote: > On Thu, 5 Nov 2020 at 04:34, Johannes Fassotte > > <[email protected]> wrote: > > It seems to me that a regular file or small database is more than > > sufficient for a tool table with a choice depending on if pictures > > are wanted or not. I see absolutely no valid reason for messing > > around with NML buffer sizes to handle the tool table. > > The reason that the tool table is present in the NML file is that the > data is potentially required by the realtime layers. This often runs > in kernel space, and kernel space has no file system access. > > It does seem that a low-effort fix might be to only load the NML file > with tools actually used in the loaded G-code file. (Though this might > be impossible in some circumstances, for example tools or offsets > being selected by analogue inputs or computed from probe results) > > A better solution in my opinion is simply to accept that any tool > change or change in tool data "busts the queue". Maybe I lack > imagination but I can't see any need to blend paths through a tool > change.
Particularly since the tool changer is supposed to be at a known, fixed location. Blending those moves is just asking for a crash. I write 95% of my own code, and the count of broken tools because an xy move started while it was still retracting from a hole has negatively effected my tooling costs and wrecked several parts. So I'm getting into the habit of adding a G64 p0.001 to the preamble of my gcode. OTOH, I am not a production shop, making 10,000 copies of something, so time to me doesn't have a huge cost as it will get done whenever I get it done. Most of what I do is one offs. Very costly in terms of time. > This would allow a call back to Userspace for the new tool data, and > then a database query. The only sensible way to do it, Andy. If the machine has to stop and wait on the database query for 50 milliseconds, so be it. If it has a tool changer, it sits for lots more time than that waiting for the chain or carosel to bring the next tool into grabit position. Not as long as it waits for the human to do it. but still, that 50 ms while it fetches the new tool data gets lost in the other motions needed to change the tool. Cheers, Gene Heskett -- "There are four boxes to be used in defense of liberty: soap, ballot, jury, and ammo. Please use in that order." -Ed Howdershelt (Author) If we desire respect for the law, we must first make the law respectable. - Louis D. Brandeis Genes Web page <http://geneslinuxbox.net:6309/gene> _______________________________________________ Emc-developers mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/emc-developers
