Fly Swatters

17 views
Skip to first unread message

Tom Blevins

unread,
Oct 16, 2016, 11:33:59 PM10/16/16
to Arch Reactor

Sunday Night there was a couple of flies bugging me so I found a fly swatter file on thingiverse.
The first (1) printed just fine. It killed the offending flies in quick order. Morgan then suggested I print a couple more out to put out on the tables. So I fired another one up without glue. Out came (2) the second on and it's not near as good as the first one. So I reapplied glue and tried again the third attempt did not print hard at all. So I cleaned everything off and reapplied the glue and ran it again and the fourth result (3) is shown. Any thoughts on why it would degrade so much over a short time with no real variables changing.  Same Machine, Same settings, and Same filament but seriously degraded prints.

secondr...@gmail.com

unread,
Oct 17, 2016, 10:40:30 AM10/17/16
to Arch Reactor
If this is the Griffen printer it may be a couple of things. One is it may be loose on the sleds and or the extruding gear assembly. Either way it may be time for a PM on the printer to tighten it all up. Delta styles just have more moving parts then Cartesian and because of this are more prone to print failures.

Derek Sigler

unread,
Oct 17, 2016, 11:11:52 AM10/17/16
to arch-r...@googlegroups.com
I'd watch the temp real close.  On the Griffins I have best luck preheating to the desired temp and then running the gcode every print.  In many cases that I haven't invested time to research, I look down and the temp has dropped 5 degrees.

Myles Farrell

unread,
Oct 17, 2016, 11:20:22 AM10/17/16
to arch-r...@googlegroups.com
This doesn't explain the variation between prints, but the griffin does need some overhaul that is affecting the quality of prints. I tried printing something large the other day and noticed that the back corner wasn't right. The nozzle would come off the glass by a millimeter or two toward the 9:30 to 11 o'clock arc and drizzle out a zig zag pattern. This also affected the infill quality and layer adhesion. I tried running the calibration program and resetting the Z zero with no effect. Sam and I talked about it and suspect this is because one of the rear arms was broken and glued back together. The thought is that when the cracked rod was glued back together the length was changed slightly from the others.

On Mon, Oct 17, 2016 at 10:11 AM, 'Derek Sigler' via Arch Reactor <arch-r...@googlegroups.com> wrote:
I'd watch the temp real close.  On the Griffins I have best luck preheating to the desired temp and then running the gcode every print.  In many cases that I haven't invested time to research, I look down and the temp has dropped 5 degrees.

--
You received this message because you are subscribed to the Google Groups "Arch Reactor" group.
To unsubscribe from this group and stop receiving emails from it, send an email to arch-reactor+unsubscribe@googlegroups.com.
To post to this group, send email to arch-r...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/arch-reactor/1106027535.1426898.1476717107572%40mail.yahoo.com.

For more options, visit https://groups.google.com/d/optout.

Reply all
Reply to author
Forward
0 new messages