Calibration tolerances?

[cope413]

Installed my shiny new E3D V5 hotend on Friday. Played with it all weekend to get it calibrated and dialed in.

Printed the thin wall cal cube and am pretty pleased with my results.

Question:

Is it possible/reasonable to get the tolerances any tighter? If so, what should I be looking at?

Pretty pleased with .04 on the width and .07 on the height, but I'd love to print perfectly if it's possible...

[lordbinky]

I believe it is DELTA_DIAGONAL_ROD that determines your scaling given that all your movement settings are correct. So fine tuning that will change your X/Y scaling to be 20mm when you are intending 20mm. If you fine tune it for ABS keep in mind you might be large on PLA, ABS shrinks when cooling a few hundredths of a mm while PLA does not. Personally my X was smaller than my Y on square box prints, I went ahead and did an overhaul on the printer and re-aligned towers and adjusted cheapskates just right... and my effector plate (I printed it no worries for stock folks, magnetic arm upgrade and all)bites the dust. Luckily I got spares but that means it's down until I get time to play with it, which I don't for awhile.

[626Pilot]

Try to print this: http://www.thingiverse.com/thing:139113

Does the print succeed? I had to modify PRINTER_RADIUS to get it to work. Using Trick Laser arms.

Also, open and shut your calipers a few times. Does it land on 0.000? Mine never do. After you measure, shut the jaws and see how far it lands from 0. That's how much error there is in that measurement. (Also, you probably already know this but never use calipers to measure filament. Get a micrometer.)

[geneb]

Also understand that ABS and PLA have some amount of shrinkage as it cools. Don't mess with your printer settings until you know how much material shrinkage you can expect.


g.

[cope413]

I will print that now. Why did it fail for you? I've printed that high with no issues before. I have the TL arms too.

I actually do often get 0.000 on my calipers. I'm sure the error is +/- .02 but they're pretty good. I don't have a micrometer, but it's on my list. The guys I buy my filament from have one and I measure it when I buy it.

So it sounds like I'm right up against the tolerances... Yes?

I would love to print something exactly to spec, but my machine is printing beautifully right now, so I don't really want to mess with it if it comes down to filament shrinkage calculation. I change filaments all the time, and just recently added t-glass to the mix. Way too much work.

[cope413]

Printed darn near perfectly. One little layer misalignment - might not even be visible in the pics. I may have been printing a bit too quickly...

Now I have a 25cm track and field baton? A 250mm sleeve for loose quarters?

[lordbinky]

tubular

[dpmacri]

I'm in the process of printing the cylinder because I was curious how my machine is performing. I had already done the single-wall 20x20x10mm cubes and those came out fairly nicely: Wall thickness pretty close to 0.55 (0.56, 0.57) which is what I setup in Slic3r, dimensions reasonable (the height is actually 10.25 or so -- any thoughts on that?).

I don't have the hot-off-the-presses version of Slic3r, so it didn't have the "spiral vase" setting. But I figured I'd print the cylinder anyway. The interesting thing is that where the "seam" in the cylinder is, I would have expected a perfectly vertical line. But it's not It started twisting a bit (clockwise) after about 110mm. Now, maybe this is part of the problem, but I did set the speed multiplier to 500% via the LCD But that still should only be 150mm/s since I have 30mm/s set as the speeds in Slic3r. Has anyone tried something similar?

Also, on a separate issue, I just noticed that the "top" is starting to look poor (on the same side as the seem). Current height is about 210 mm and the last 15mm or so have gotten worse, with a hole appearing in one spot. I'll post some pictures once it's done.

[cope413]

I have a feeling that 150mm is pretty fast for this print. I printed mine at 90mm/s. I likely could have gone up to 100, maybe 120, but after that, I think I would have had layer cooling issues.

As for the "seam". Mine is straight, but it has 3 distinct segments. Not sure why it shifted slightly every 7-8cm, but I'm guessing it has to do with the slicer spiral vase code.

Did you choose, "randomize starting points"? What's your retraction set at?

[dpmacri]

I have a feeling that 150mm is pretty fast for this print. I printed mine at 90mm/s. I likely could have gone up to 100, maybe 120, but after that, I think I would have had layer cooling issues.

Yeah, I think 150mm was too fast. Right near the end (last 20 layers or so), I slowed it back down to 60mm/s and the problems went away. So I think, like you said, that the speed was too high as I got too far from the bed.

As for the "seam". Mine is straight, but it has 3 distinct segments. Not sure why it shifted slightly every 7-8cm, but I'm guessing it has to do with the slicer spiral vase code.

I don't know how well it's visible in this iPhone photo, but you can kind of see the seem going up and to the left. Now that it's off the print bed I realize that it started shifting sooner than I thought. In fact, I think the shifts started right after I had increased the print speed via the LCD. One other thing I think *may* contribute to the problem is that I have my cables from the print-head and fan zip-tied to the bowden tube. My printer is behind me and occasionally I heard a funny sound. But I turned around and didn't notice anything -- except once when I saw that one of the zip-ties was clicking against the mount where the extrude is. I'm wondering if the cables bumping against things are causing shifts of some sort.

Tall Cylinder with crooked seam

Did you choose, "randomize starting points"? What's your retraction set at?

No randomization. Retraction was set at 5mm with 60mm/s speed. Z-lift is 0.2 and Retract on layer change was checked.

[626Pilot]

I don't use Slic3r anymore but I'd guess that blobbing is the print head oozing either at the start or finish of drawing. Try switching to KISSlicer. There is a forum here just for that slicer and in that forum is a thread where you can download the ideal settings. You will get better retract performance. Slic3r doesn't have as many "tricks" as KISSlicer to avoid blobs and strings.

My prints were failing previously because the TL arms were just a hair different in length from the originals. I went through a few test prints to see what would happen if I altered the printer radius, and once I figured out the right way to go my problems went away. If PRINTER_RADIUS is off you will get lensing effects, i.e. the carriage will not be able to move in the XY plane without deviating up or down as it moves away from the center. The effect is tiny and doesn't show up in small prints, but it is multiplicative and after enough layers your print will turn into spaghetti.

[dpmacri]

Thanks, I'll have to give KISSlicer a try -- I'm definitely seeing problems with stringing if I make more than one object or I print something that moves around a lot (I printed a small, 2-part horse for my daughter and the legs all had little strings that I needed to shave off when it was finished).

I'm also using the TL arms but I haven't done much to the PRINTER_RADIUS. I adjusted it just one step during the calibration and my bed seemed level at that point.

It seems to me that someone needs to create a sequence of calibration steps with good/bad images to tweak various things. I know I fell a bit uncertain about which things to change and why. I've been trying to find the necessary information in the forums but certainly feel that my simple mind could use a sequence of "if print X looks like Y, do Z"

[dpmacri]

Hmm, I just downloaded KISSlicer and MHackney's .ini settings. But the extruder isn't putting out enough material There are so many settings, I'm not sure what to change I did tweak the filament diameter (I started with the PLA Translucent RED and made a copy for my own) but I'm not sure if that made things worse (the diameter in the .ini was 1.6 something, but mine is 1.78). I'll post a message over in the KISSlicer section, but if anyone reading this knows how to get KISSlicer to extrude more filament, please let me know.

[626Pilot]

Material tab -> Flow Tweak. If you didn't measure your filament with one of these, get one. Calipers are no good.

Also, if Flow Tweak has to be set outside the range of {0.9 - 1.1} your diameter measurement is off. A good test is to see if you can print the first layer in PLA on Kapton with no adhesive or bed heat. If you can't get it to put down the first layer without messing up, either the flow or the height setting is off. (With PLA, the point of adhesive/bed heat is to help it print AFTER the first layer.)

[dpmacri]

Thanks. A digital micrometer is on my "To Get" list

However, the problem has nothing to do with KISSlicer -- turns out my extruder is jammed Sorry for the false alarm.

[geneb]

tubular

Totally.

[dpmacri]

So after realizing that my extruder was jammed, I started to try to disassemble it. When I when to turn the nozzle to remove it, the metal print head twisted at the junction with the PEEK -- yikes! I waiting for things to cool down further, finished the disassembly, got the jam out and put it back together. I also switched that fan that I was using to cool the PEEK. Originally I was using one from Jameco but it started making a bunch of noise recently and I'm wondering if it wasn't spinning well. So I put in a different one.

I was able to print just fine after that (and KISSlicer did an awesome job compared to Slic3r -- Wow ).

The thing I'm now wondering is whether perhaps my temperature readings are off by a lot. I had bumped things to 190 in KISSlicer before I realized that the head was jammed.

Can anyone give me good instructions for how to calibrate the temperature of the hot-end? Do I remove the filament and stick my thermocouple probe down in there? When I read on the outside, with the Thermistor reading 200, the best I can get with the thermocouple is around 165. But I'm sure I'm reading a "cooled" down location.

[Eaglezsoar]

You are correct, you should stick the thermocouple into the hotend all the way to the nozzle to get accurate results.
If you can afford it, you should check out the E3D all metal hotend. You would still need to do the thermocouple test
but there is no plastic that can melt. One of the best hotends on the market.

[geneb]

There's not a thing wrong with the stock hot end. I've run 1.6km of filament through mine and have NEVER had a jam.

You should never try to change nozzles without gripping the aluminum hot section with a pair of slip joint pliers. (carefully!)

If you test the hot end with a thermocouple, you'll need to remove the PTFE liners in order to have enough room for the probe to fit. (unless of course you've got a small thermocouple lead that can fit inside the 2mm inner liner)

g.

[dpmacri]

Thanks, Carl and Gene. The E3D hot-end is probably on my wish list for Christmas, but at this point, my wife will probably balk at me spending anymore significant dough on the printer My thermocouple probe is only 1.6mm at the tip, but the wire to it gets to about 2.8mm, so I'll need to remove at least one of the inner PTFE tubes. I'll try that and get my temperature calibrated correctly!