New Rostock Troubles

[cyber.shifter]

So yay for life taking us on wild journeys. I finally getting back to the printer. I've moved twice, graduated college, and become single. lol. Anyway, so I wanted to upgrade my delta arms to ball joints so I went ahead and got the kit for the ball joint arms from SeeMeCNC, as well as the auto-tensioning carriages. Got those installed, resquared the frame of the machine, and re-calibrated as best as I could. I also calibrated my e-steps, which I hadn't done before. Printed before and after the e-steps thing because I was trying to see if that was my issue with printing, but it was not. My infill was not reaching my borders properly. From what I've read this sounds like an under-extrusion problem. So I did e-steps, and then thought of temp calibration. Which is when I remembered this post! So I tried the method mentioned in the above forum post. I was able to successfully make all the measurements. However, I have a question:

Can I use the same formula if my thermistor reads under 1000 ohms while at temp? Also, is this a problem? Mine is reading somewhere around 236 ohms while at operating temp, vs. ~96 kOhms while at room temp.

[cyber.shifter]

By the way, when I conducted the measurements, my thermistor showed 230 C but my calibrated meter showed only 203.7 C. I suspect this is a good bit of my problem.

[Polygonhell]

I did notice the two times that I unclogged it that I had melted ABS in the PEEK tube. However, the thermistor read that it was at the temperature that I wanted. If the thermistor is reading off, is there a way to correct for it in the firmware or something?

This is indicative of too aggressive a retraction setting.
What can happen is in areas of the print where there is a lot of retraction, the end of the filament becomes compressed as it cools and it jams in the tube.
It's much more common with PLA, but if your retraction settings are too long, or too fast you can see the same thing with ABS.

I always recommend the same thing when trying to isolate problems with extrusion, disable retraction entirely and see if the print completes, if it does then go back and tune it.

[cyber.shifter]

I'll check on that. Just to clarify, I'm not having the clogging issue I had before, I'm having an issue with the infill not meeting with the walls. However, I did see a bit of leftover plastic in the tube when I took it apart to do the temp measurement this time. After I have finished with the temperature settings, and conducted a test, I'll begin testing the retraction settings. One thing at a time! lol! Otherwise I don't know what fixed things. Thanks for the input though! I'm new to this, and I wouldn't have thought to check retraction settings. I really need to learn more...

[cyber.shifter]

Ok, here's an update. I did the temperature calibrarion and the nozzle was off by almost 30 degrees Celsius. So I corrected that following the mentioned method. Then I did a test print with the same material and profiles as before. It came out better. First layer stuck well and everything. So I decided to go ahead and do the bed temp calibration as well. I got lucky, it was already pretty spot on, so I left it. Then I wanted to make sure my positional calibration was good so I went ahead and did that. I used a 0.038 mm feeler for Z-height and 0.203 mm for towers and center. Got that finished, re-glued, and started another test print. Same object. Results were terrible. The had filament wouldn't stick and when it finally did, I had the same infill problem. I feel like I'm back to square one... ugh



*I was trying to load pictures onto this post, but for some reason cannot. The attachments upload like they're going to work, but then they don't do shit. Not sure why. I've attached tons before. Anyway, suffice to say, I closed the gap between outline and infill in the improved model, and then it went to shit after the positional calibration, worse than before.*

[lightninjay]

If you have not tried to, place your cursor wherever you want a particular uploaded photo to go, theb click "Place Inline" next to the photo upload section files when editing your post.

[cyber.shifter]

That button didn't even appear. I think it's a size issue. The old ones I uploaded were from an 8 MP camera and these are coming from one that's 13 MP and the pictures are a good bit bigger. I tried uploading some random picture and it worked fine. So I'll have to change the resolution when I get home on the pictures.

[cyber.shifter]

Ok, here are the pictures.

The first one shows the print before the temperature calibration on the right, and after on the left. The next picture is after the first one, but with a tower calibration in between. Since then, I have tried recalibrating, with no good results. I discovered my endstops were the tiniest bit loose, so I took care of that. Recalibration seemed like it would work because there was a noticable change. Alas, i get worse results akin to the second picture. This is ridiculously frustrating.

As a side note, I should mention how I do the calibration. I start at Z tower, do calibration until it's good on a 0.203 mm feeler gauge. Then I go around. After I do all three, I go back to Z and repeat until they all feel the same. Then I do bed center, make adjustments if necessary, and repeat. Is this correct?

[lightninjay]

I would chalk this up as potentially being a slicer setting issue. Do you mind posting your Gcode and possibly the STL?

[cyber.shifter]

I don't mind at all. I'll be home from work in an hour or so, I can do it then. Meanwhile, I am using MatterControl, because I wanted to start at the beginning again. Apparently its not very good in the slicer territory. But I didn't want to change too many variables at once, so I was loathe to switch until I was sure it was the problem.

[lightninjay]

Right, I am not sure how much you have fiddled with mattercontrol, and its three different slicing engines.

http://wiki.mattercontrol.com/Category:Slice_Engines

That is a link to an explanation regarding the three different engines for Slic3r, Cura, and MatterSlice.

With that in mind, they also have links to the different settings you are allowed to modify within those slicing engines, within mattercontrol's settings.

Were you wishing to have finer control over your slicer settings and parameters, you could download and install a standalone slicer, like craftware, KISSlicer, Cura. Slic3r has even made some decent advancements in terms of infill and better toolpathing. The idea here is to be able to calibrate your slicer settings just as much as you have calibrated your printer. The main things you will want to look for within any slicer are extruder nozzle width, and any overlap settings which adjust gaps between extrusion paths.

[cyber.shifter]

Yeah, at one point I was using Slic3r and RepetierHost but I wanted to go back to basics while fixing my mechanical problems. I guess I should have just stuck with Slic3r. I've heard good things about KISS though.

[cyber.shifter]

Ok, I've attached the gcode for the object (which is what I guessed you meant) and the .stl

climbing.stl

(95.1 KiB) Downloaded 321 times

climbing.gcode

(216.81 KiB) Downloaded 310 times

[nebbian]

I've seen the same thing with slic3r. You can fiddle with the "infill/perimeters overlap" percentage to adjust this (in the "Advanced" tab).

In some cases the only way that worked for me was to lower the number of perimeters (vertical shells) to 2.

[lightninjay]

Ok, so here is what I mean by slicer settings...

The first below is your gcode file, unmodified, loaded into CraftWare for viewing.
The second is my file, but also take note that I am printing with a Prometheus V2 using a .3mm nozzle.
The third is a representation of printing with a stock .5mm nozzle using my CraftWare settings.
#1

#2

#3

The main thing that I notice is that your gaps seem to coincide with the small little gaps shown in the screenshot. You can click it to see the fullsize image. Any area that seems to show bed through it, will show bed on a real printer, simple as that.

Try printing the attached Gcode on your printer. I tend to run pretty conservative settings within the printer community, but CraftWare estimates your print to complete within 20 minutes.

The attached Gcode are set to print the first layer at a height of .2mm, at a speed of 10 mm/s and then 30 mm/s for the second layer and on, with a resolution of .2mm. Both are set to print at 230 because I usually only bump my prometheus to 240 to print with ABS. 230 to 210 is what I print PLA. If you don't like the temps, preheat your printer to whatever preferred temps you have found, and start the gcode from either the sdcard or your preferred printer host. Once the printer or host has engaged the Gcode and sets my preset temps, just go ahead and set them back to your preferred temps that you just preheated to. The print will continue at your temps and you will be able to try different temps to your satisfaction.

DISCLAIMER: IT IS NOT MY FAULT IF THIS GCODE HARMS YOUR PRINTER. I HAVE USED THESE SETTINGS RELIGIOUSLY ON MY PRINTER AND THEY HAVE WORKED FOR ME. PLEASE MAKE SURE YOU HAVE YOUR PRINTER CALIBRATED AND YOUR BED Z-HEIGHT PROPERLY SET.

[mhackney]

This is a well known issue with slic3r. One way to deal with it if you are stuck with using it is to rotate the part so the sharp angled leg (in your case) is either parallel to or less angled than the perimeter. Many times just a 5 degree rotation will miraculously clean up those first layer blues.

[cyber.shifter]

This is very good information. It looks like your infill in a good bit closer to the perimeters, are you doing more overlap? I don't mind running conservative settings. I prefer quality to speed. I've been printing at 20 m/s per mhackney's recommendations in his guide, but I wouldn't have thought of rotating it to improve the gaps. So I guess my next step is going to be double-checking my calibration, and then trying that GCODE you sent me lightninjay. Did my calibration measures sound acceptable? I keep reading mixed signals about Z-height. Some people say as close to touching the glass as you can get, others say 0.1mm or something else. Not sure what is ideal. I've been doing 0.038mm for Z-height and 0.203mm for towers and center.

[lightninjay]

This is very good information. It looks like your infill in a good bit closer to the perimeters, are you doing more overlap? I don't mind running conservative settings. I prefer quality to speed. I've been printing at 20 m/s per mhackney's recommendations in his guide, but I wouldn't have thought of rotating it to improve the gaps. So I guess my next step is going to be double-checking my calibration, and then trying that GCODE you sent me lightninjay. Did my calibration measures sound acceptable? I keep reading mixed signals about Z-height. Some people say as close to touching the glass as you can get, others say 0.1mm or something else. Not sure what is ideal. I've been doing 0.038mm for Z-height and 0.203mm for towers and center.

So when doing a Z-calibration, it is important to note how you are doing it. Some people will manually lower their nozzle to the center of the bed with the LCD controller and set the Z=0 this way. This is truly the 0 point and should technically be where the nozzle is touching glass, because the slicer will offset the Z by however much your first layer height is. Most calibration routines which are run as scripts from this site for non-auto-calibrating machines, tend to calibrate JUST off the bed in order to slide a piece of paper in there. This is calibrating the Z=0 with a calculated offset. Either way is acceptable, as long as they are performed correctly.

[cyber.shifter]

I've been setting the Z-height by using the controls on the machine, and the above mentioned feeler gauge. Would it be better to use that gauge and then offset the Z-height? The way I usually do it is following the Rostock manual, but using feeler gauges instead of paper.

[mhackney]

How thick is your feeler that you are using?

[cyber.shifter]

0.0015" or 0.038mm. Thinnest one I could get.

[mhackney]

I would compensate for the thickness of the feeler by adding that to the Z height. The other thing to keep in mind is that the "feel" for the snag test is completely different with a metal strip vs paper. The texture on the paper makes it easier to feel consistently. Whereas with metal feeler it is really more pressure feedback. I find it easier to be consistent with paper. I use rolling paper though since it is extremely thin.

[cyber.shifter]

I experimented with a caliper and the gauges and found that the measurement was correct when the gauge had enough resistance that it almost couldn't move. However, what I'm thinking is that I can just use a 0.1mm feeler, since I'm going to offset for it anyway. So I would add the thickness of the feeler to the Z-height in the EEPROM?

[cyber.shifter]

I'm also used to using feelers because of working in precision manufacturing previously. And all I have around is cheap printer paper. :-/

[mhackney]

Yes, that works perfectly well. I have a 1mm and 10mm gage block that I use to test Z height. The thicker blocks make it easier to see and feel I think.

I CNC machine. The rolling paper trick was taught to me by an old-timer. It works amazingly well and won't dull expensive carbide mills.