Page MenuHomeAleph Objects Inc

Workmanship or quality standard needed for vernier print
Closed, ResolvedPublic

Description

We could use something that says whats passable as far as the looks and quality of the vernier print. Are infill imperfections ok? Text imperfections? X and Y offset line imperfections ok? Sometimes there are lines that have parts missing from them. Sometimes the letters of the text might smudge together. Etc

Event Timeline

There are a very large number of changes, so older changes are hidden. Show Older Changes
kent added a comment.Mon, May 13, 8:34 AM

Next step is to induce offset and backlash errors and take a set of photos of them so we have examples of z offset too high, too low, etc. and add them to the calibration OHAI.

MikeR added a comment.Mon, May 13, 9:46 AM

@kent I am still getting some ooz with the v8 model you made.

@MikeR does this print also have @olivers changes as well?

@DaniAO It does not, you can tell from the skirt.

Ive done 2 test prints with @oliver gcode. So far I haven't seen any oozing. I am going to do a couple more prints to make sure its consistent but its looking good so far.

MikeR added a comment.Mon, May 13, 3:22 PM

With the v8 gcode, once I got both of my z-offsets to be the same, I started to notice that the E1 was printing very squished once the actual objects was being printed.

@kent had said that there was a mistake with the model when it was made and I was told @oliver was going to correct this. @DaniAO was seeing the same thing when she was testing.

I updated the branch with a gcode sliced with the most recent change.

MikeR added a comment.Tue, May 14, 1:21 PM

After trying a few prints with each adjustment @oliver has made, Im still seeing the E1 nozzle printing to squished.

MikeR added a comment.Tue, May 14, 1:24 PM

@DaniAO I know you have been doing some testing with these gcodes. Are you seeing the same thing in your prints as me? Maybe I'm having an isolated issue specific to my machine but I don't think there is.

@oliver @kent
We need to get this print figured out and completed asap.

Thank you!

MikeR raised the priority of this task from Normal to High.Tue, May 14, 3:47 PM
oliver added a comment.EditedWed, May 15, 7:10 AM

@DaniAO @MikeR These are pretty much the default gcode, the only thing I changed was copying and pasting 2 skirts into a third gcode with initially no skirt. The rest of it is default settings.

I also realized I may have forgotten to include the backlash compensation on, fixing that part now.

@oliver interesting.... something did change though as @MikeR and I are seeing the same problems, that extrduer 2 doesn't always print everything.

I know you have been focused on helping QC stay a float, so I would like to ask @kent to look at this and fix this. We need to get this figured out asap.

Thank you

@DaniAO Working with @EricNugent to adjust some of the settings in the profile to try and improve it first.

Give this one a shot. At the very least it is much faster.

kent added a comment.Wed, May 15, 9:31 AM

@EricNugent ok testing it now. I just did a default standard polylite PLA print of the smol stl and it took 31 min 11 s (cura 3.6.9 marlin 2.0.0.116). It did still have a blob in it.

@kent the one I just ran was approximately 15 min. There is a slight chance of a goober happening when it starts the black, but nothing we can't fine tune.

@EricNugent while this print is much faster, (14 minutes) I noticed the skirts are smaller- We need those to stay bigger to help get Z calibrated.

I also noticed that I had the same issue printing where x for the 2nd extruder doesn't print all that well, nor is it a clean print.

But to add we are moving in the right direction!

@DaniAO For sure, it will need some fine tuning. But the overall shape of the part can be done in half the time, which will help us troubleshoot better. If we have a checklist of everything we are trying to hit I can fiddle with it more but this was pretty much a fist attempt.

kent added a comment.Wed, May 15, 10:17 AM

Standard PolyLite profile 31 min 11 s


whatever profile was used for vernier_quick_5-15-19.gcode 18 m 19 s

I like the skirt, first layer, and second layer speed bump but I think it's a little too fast on the cylinder though. I like the way the seam is handled in the default profile better than the new one. It didn't have that characteristic blob, but it didn't form the vernier lines very well. We want something that will print a bigger skirt with only one pass for z-offset adjustment to be on the same gcode as the vernier print so the techs can dial in the z-offset manually during the skirts.

kent added a comment.Wed, May 15, 10:19 AM

I think we could make this look a little prettier with some linear advance, it's definitely building up at the direction changes on the second layer infill and tips of the vernier lines.

I adjusted the skirt distances to separate them out a little. Tried to fix the missing details on the front black part. One solution may be to add a black perimeter to the outside of the print to act as a built in prime tower.

Skirts are better but print still is same x issues.

@EricNugent can we add a perimeter with extruder 2? I think this will help.

Thanks,

Third times the charm...or at least that's what I have heard.

There will need to be some quick edits. Just realized it is printing top layers for E0 before printing E1 on the tower, which explains why the seam looks different/caused the droop for others

This one is ready for review.

logan removed a subscriber: logan.Wed, May 15, 2:47 PM

Looking better!

kent added a comment.Thu, May 16, 8:00 AM

@EricNugent I like the skirts. I think it would help to put a pause command in there so you could inspect the skirt and make adjustments then resume. The new edge in the model doesn't prevent the blob and does this high-ish frequency infill on the first layer that is kind of slow. The tower is too fast and melty. It took 26m 16s to print for me.

Last edit I can do for today, hope it at least helps.

MikeR added a comment.Thu, May 16, 9:01 AM

Testing commencing

MikeR added a comment.Thu, May 16, 9:52 AM

Done a couple prints with this new gcode. Its looking very good to me. @DaniAO have you had a chance to print this one yet?

I was in a meeting but will get this tested now.

@karrad @youngmrcarlson @DRobertson @TyTh can you please test this file with PLA/PLA combo?

Thanks!

@DaniAO new code looks good on this end. Not sure how it compares to the other machines however

kent added a comment.Thu, May 16, 11:21 AM

The removal temp is too high at 50c and it is set to keep heating. There's no blob though.

kent added a comment.Thu, May 16, 11:22 AM

also it's not forming the far right leg of the x vernier on any of the these tests mine included.

TyTh added a comment.Thu, May 16, 11:37 AM

gcode on mine looks good as well

@TyTh any chance you could post a picture?

Agree the bed is a bit hot, so if we could cool that down.

After getting my offsets fixed, the print looks good.

@kent are you able to fix this or do we need to wait until @EricNugent is back?

kent added a comment.Thu, May 16, 11:54 AM

He hasn't pushed the stl of the model he's working on or the cura profile that I know of so I'm limited to editing the gcode. I can do that easily for the bed temp thing at the end, but not so much for the speed settings.

@kent I apologize, I thought @EricNugent was out for the day.. but he was just headed to lunch and then helping out in production.

I just talked to him and he said he can make those adjustments.

Thanks!

Adjusted the bed temp and removed user specified z seam. Bed cooling from 60-45c eats up 2-4 minutes depending on environment, but should be good. I can push the stl's, gcode, and cura profile if that will help.

Looking good!

I ran this code twice, as the first one had a mistake:

v6 round one

but the second run of the same gcode came out great (v6)

I like the way this looks.

@kent @EricNugent @Steven think we are good to push this?

v5 (average diameter=9.469mm)

v6 (average diameter=9.522mm)

I did print this over here, but I don't have a camera to take a picture with. I didn't watch if it happened on the second one, but on the first gcode it seemed to do a "boop for good luck" after the print finished.

@youngmrcarlson can you take it to Brent to get pictures please?

Did this "boop" leave anything behind?

karrad added a comment.EditedThu, May 16, 3:10 PM

Looking at these photos, it appears the lower +/- seems to be connected on all the prints but it is not connected on the right hand side. Something we want to look into? (I didn't notice at first glance, so a customer might not. But seeing all these photos side by side, it looks to be gcode)

@DaniAO The boop didn't seem to leave anything, just a number of [seemingly] unnecessary travel moves after the last layer.

@youngmrcarlson Would the "boop" be the nozzles raising and lowering after the print once in park position? If so, this is the nozzles re-priming themselves after the print completes

DaniAO reassigned this task from kent to Steven.

I feel like this is good to go.

@Steven @kent please review so we can get this on the USB for Production.

Thank you!

kent moved this task from Needs Review to Report Issues Here on the Quiver board.Thu, May 16, 4:13 PM
kent claimed this task.

I don't think this is ready.

MikeR added a comment.Thu, May 16, 4:26 PM

@kent do you have an idea on how to get rid of the alleged bloop?

kent added a comment.EditedThu, May 16, 5:19 PM

@MikeR try this:

MikeR added a comment.Fri, May 17, 8:01 AM

Testing commencing.

MikeR added a comment.Fri, May 17, 8:32 AM

Im trying the gcode @kent just pushed.

MikeR added a comment.Fri, May 17, 8:50 AM

The gcode @kent made is fast and has no bloops or goobers or ooz or whatever you wish to call it.

It even does a short pause to inspect the skirts and make adjustments. I like this one the best so far. @youngmrcarlson @karrad @DaniAO @TyTh @anolen have any of you had a chance to print this out yet? This one looks like a winner.

kent moved this task from Report Issues Here to Needs Review on the Quiver board.Fri, May 17, 8:56 AM
kent added a comment.Fri, May 17, 9:11 AM

@west there is a v1.5 x double bearing holder on the T6988 branch and there is v1.4 on master. Do you know anything about that? It is commit 2643183a5c98ef7303a0c664f9cdd517f4f91724.

@youngmrcarlson print from today

The feel of it shows some gaps in top layer, but very clean lines/color transition

kent added a comment.Fri, May 17, 10:09 AM

I think the backlash setting could be dialed in a little better on that machine. gaps between the lines is a good indicator of that. What were they set to for that print?

I'm planning on intentionally setting some things wrong and then doing prints to get a good set of photos of what different offset and lash settings look like when they are wrong.

@kent X: .14, Y: .09, Z: .08, Smoothing: 2mm, Correction: 100%
I'll do the automatic calibration and retry the print.

Its kind of hard to see the gaps with translucent filament. If possible could we refrain from using clear or translucent colors?

kent added a comment.Fri, May 17, 10:39 AM

If you could post print times too that would be helpful. Getting calipers on the post for min/max measurements is a bonus. Use new polylite pla black and lulzbot green if possible to match production.

14m 29s

I have recorded the circle dimensions for the first 4 prints ive done with the latest gcode.

Circles are on the small side but would be withing the 50 micron tolerance.

kent reassigned this task from kent to Steven.Fri, May 17, 11:12 AM

@Steven I made a calibration-print branch with the source for this calibration print, which is going to help speed up production based on my testing. I think it's in a state that it can be merged in now and included on the next tag, but continuing work will need to be done on upadating the ohai to include images of limit samples for workmanship standards.

youngmrcarlson added a comment.EditedFri, May 17, 11:23 AM

@kent Well, backlash calibration didn't help. New values: X: .14, Y: .07, Z: .06mm.


Print time - 15m

kent added a comment.Fri, May 17, 11:26 AM

the thing about the autocalibration is that it does offsets and backlash at the same time. It failed pretty hard on the z nozzle offset calibration in your case @youngmrcarlson what is the z nozzle offset at now?

@kent Nozzle offsets: 44.58mm, 1.34mm, 2.34mm

kent added a comment.Fri, May 17, 11:37 AM

Yeah nominal on that third one is zero... Might be best at this point to set it back to factory defaults by looking up your machine here: http://devel.alephobjects.com/ao/manufacturing/documents/master_build_log_TAZ_PRO.ods

@marcio Maybe we should set a threshold that above a certain value it throws a error message. If you have an as-measured z nozzle offset value greater than 2mm something is definitely not working in the autocal process like it's pushing through the cube or something. @youngmrcarlson did you notice any deflection of the bed plate during the auto calibration?

@kent I have Quiver 15, so it's not on that spreadsheet. Didn't notice any deflection, the skirts were a bit high if anything. Z Offset is set to -.60mm.

@youngmrcarlson try to in the nozzle offset menu, set the z to 0.07. Thats been an average. In production, when they get a number thats way off looking they try to set it to what the average has been and that has been helping.

Steven reassigned this task from Steven to kent.Fri, May 17, 12:17 PM

@kent the calibration directory now has numerous version of the blend file and prints. We should only have the parts that are expected to be used in production in the repo. Can you please clean up the files that are older versions and/or not expected to be used in production?

Looking good on my end.

kent added a comment.Fri, May 17, 1:00 PM

@Steven ok this is done with a597043.

kent reassigned this task from kent to Steven.Fri, May 17, 1:32 PM
Steven closed this task as Resolved.Mon, May 20, 9:21 AM

This is now merged into master

Some of the calibrators are having trouble measuring the skirts and setting the nozzle heights because the E1 skirt gets covered up by the E2 purge skirt and measuring the E2 skirt may not get the E1 nozzle height right. Besides that it seems to be going well.

kent added a comment.Tue, May 21, 10:07 AM

@wolffchadd @robert @paulette you can also power cycle the machine during the dwell that has the "inspect skirt height" message on the screen and that's pretty much the same as the circle gcode but faster.

TyTh removed a subscriber: TyTh.Tue, May 21, 10:07 AM