Dataset from a random sampling of 30 out of 6 bundles:
0.02, 0.01, 0.02, 0.07, 0.09, 0.06, 0.02, 0.09, 0.08, 0.06, 0.08, 0.02, 0.02, 0.06, 0.24, 0.08, 0.00, 0.06, 0.06, 0.4, 0.01, 0.07, 0.06, 0.05, 0.01, 0.00, 0.23, 0.02, 0.08, 0.06
Percent fallout of inspection for bends: 10% (3 of 30)
Percentage exceeding supplier's spec tolerance: 0%
Updated on data to reflect an additional 216 units inspected after that initial 200 units:
Ct 1 unit: Dented end, mushrooms out
Ct 31 units: 0.44, 0.22, 0.4, 0.22, 0.32, 0.58, 1.06, 0.62, 0.56, 0.58, 0.34, 0.38, 0.42, 0.36, 0.44, 0.3, 1.46, 0.11, 0.95, 0.28, 0.14, 1.46, 0.15, 1.38, 0.32, 0.68, 0.34, 0.42, 0.31, 0.3, 0.36
Mode: 0.44, 0.22, 0.32, 0.58, 0.34, 0.42, 0.36, 0.3, 1.46
Percent fallout of inspection for bends: 7.4% (31 of 416)
Percentage exceeding supplier's spec tolerance: 2.4% (10 of 416)
Just in case you need it! 😄
I did this one as a high speed to save time but I see your earlier images of models have thinner layer heights, I can adjust to thinner layers heights if needed. Just let me know.
@logan, Give this one a try: 😄
The passing units were released Monday and Tuesday and are being used in production. QC inspections are generally a Go/No Go process, so we collect specific data on non-conforming units, but not on conforming units. Passing units are categorically passed, not evaluated further than having met passing criteria.
Sounds great thanks.
I spoke with @AOJAS and requested the raw data for the 400 extrusions already measured. This is plenty of data and there is no reason to measure more. I requested the data from @AOJAS. When I get the data, I will do capability analysis for straightness per their specification and per our requirement. Then we can make an informed decision.
I got with Levi and the Phil and Dave from framing to show them how to do the check. When unchecked parts are stocked out, they will have someone check the whole batch so that the work of individual framers isn't disrupted. I've left two samples of failing units at their granite surface temporarily so they have a reference for training people to identify the issue.
Converting that into metric and applying it to the the length of the EX-HD0086, at 0.3175mm per foot, and the 530mm extrusion being 1.738 feet(20.866 inches), that means that the supplier's tolerance on these is 0.55mm.
@kent and @AOJAS : We need to keep production running. We can have them verify straightness on the line for now. Please keep them supplied and let Levi know that they need to prescreen for straightness until we resolve this issue.
Should I release do a partial release for production personnel to do flatness checks on during production tomorrow?
Tue, Aug 20
It's a start, but what I was looking for was to add all the information that a supplier would need to the freecad drawing. That way, its all in one document and consistent with our other drawings.
@kent those were modeled more so for visual assembly purposes. not to be exact to the true part
Also, those models didn't match the supplier's drawing.
I didn't want to have multiple versions of the same part in the repo.
@kent what was the reason for removing the two assemblies? production_parts/assemblies/10mm_leveling_washer_spacer.fcstd
Any further updates or feedback, push updates to T9723 branch of Redgum/Quiver
I dont think changing the actuator is going to help much as this is consistently the E1 side. if it were bad actuators we would see a mix between E1 and E2 not moving far enough down. Seems like something with the assembly
@logan i too am wondering how this is being observed and rejected in production.
On the tool heads that are red tagged for this issue are they causing problems with the machine? like not clearing the blower shroud, etc?
This would add some complexity to the test and the test stand, but I imagine we could rig up a mount for a digital indicator to measure that travel. However, determining what that should be and what tolerance that spec would have would take some testing I think.
As far as remedies for the assemblies, a few things come to mind.
- Wire routing - if the wires aren't routed correctly they could be preventing the full range of motion
- excess drag in Z - we've seen this from a few different causes but with all 4 rods installed and the actuators removed, both sides should move freely from top to bottom.
- bad actuator
I did some thermocouple testing on a regular heatsink and a die cast to see the differences by placing thermocouples in the fins both at the top and bottom. On average they are very similar
This is the v8 version of the drawing. The 150 micron nozzles no longer have a question mark after the length of the orifice which is nice. There are a couple additional radius callouts, and the angle of the outside cone is called out.
@kent, Ill give that a try.
@logan I think so if it helps the situation
@west Do you think we could do with a 3in*lbs torque spec vs 5in*lbs?
There is a setting called outer wall wipe distance which is set to .25mm. This inserts a travel move at the end of the outer wall that might be contributing to this artifact.
My thought process is that there isn't a big ol gap in the model, so we should be able to print it without one.
Have you tried a different wall printing order? I think those would both be treated as outer walls.
@david.hall Since I haven't had any requests and we've moved past the development stage I'm going to go ahead and close this ticket. If anything arises please feel free to open a new ticket.
I improved what I could (image posted earlier in ticket). What quality are we trying to achieve? Do we have a better print to compare to?
No I haven't been working on it, I already tried everything I could think of
@logan, any updates here?