Extruder E-step Cal Inconsistent Results

After getting my i3v 10" working very consistently with what I consider acceptable print quality (BIG thanks to Roxy and DACB for the autobed leveling code work), I thought it was time to do some fine tuning to see if I could get even higher quality prints. So, last night I started going through the calibration for the extruder following the ZennmasterM videos and after 3 hours of effort was getting wildly different results after each attempt. I started with extruding 97.6mm of PLA (used pronterface to extrude 100mm) and made adjustments from there using Zenn's calcs, and after making the estep adjustment to 861.7 by using the LCD controls wound up with a measurement of 95.6mm extruded. Huh?

Long story short, I reset estep to 841 and did 5 100mm extrudes and got varied results with every extrude, up to 8mm of difference between extrudes. I tightened the guildler screws and that had no effect. And like I said, I'm actually getting very good prints. I also made 10 measurements of the filament over the 100mm of PLA to be extruded and had a consistent 1.73mm. So I was wondering if anyone has run into this before and has any suggestions.

What are you using to measure the filament run through the extruder? I printed out this thing http://www.thingiverse.com/thing:114614 minus the slider, marked it in 10mm increments, then used it to mark and measure the filament above the extruder. I ended up using 900 for my esteps.

I too had an e-steps in the 900 range. See the Configuration.h in the github fork for what I ended up using. I will say that the kind of behavior you are seeing could be from a loose gear on the extruder stepper. Where the set screw engages on a forward extrusion but slips on a retraction. This would make calibration seem consistent (continuous forward feed) but prints are effectively uncalibrated.

I do need to recalibrate my extruder. I just noticed two days ago the large gear on my extruder was very loose on the hobbed bolt. I could turn it a good 5 degrees without the bolt moving at all. I've replaced it, but haven't recalibrated yet.

Thanks for the replies all. I have been measuring the filament above the extruder like in the Zenn videos. And with the extruder steps remaining constant at the default of 841 I am getting deviations of up to 5mm in back to back extrusions. I have checked the large extruder gear and it has no wobble, so is tight on the hobbed bolt and the gear mesh is not loose. The extrusion when viewed during the extrusion appears to be very smooth with no interruptions or gear chatter. Also, the extruded filament from the hotend appears to be of uniform thickness and extrudes straight down. Tomorrow night I will try removing the Hexagon hotend and extrude raw unheated filament and measure to see if it is a backpressure issue.

One interesting thing to note is Zenn said he was extruding PLA (as was I during the testing), but he had the hotend set to 230c. That is much hotter than I would normally print at with my temp at 185c, and that was the temp I did the testing at. Does this test require a higher temp to reduce any backpressure influence on the test, or is 230c a reasonable temp for PLA and I'm just doing it wrong? I suspect he misspoke and was extruding ABS, but I don't know.

To be clear, I am actually very happy with the print quality as it is now, especially after getting the ABL system working. I am just searching for even better if it's possible. Who doesn't want better right?

First, I've learned I get more consistent results if I toss out the first long extrusion. Make sure the guidler bearing spins free with no filament in the extruder, that the filament spool has little or no drag, large gear isn't rubbing on the mounting bolts for the top two delrin wheels, etc. Before you installed the extruder, did you make sure the filament feed hole in the base doesn't have loose filament stands from being printed on the side?

Temperature is relative on these repraps, not absolute. I can say that I have one spool in particular of PLA that will stop extruding if I tried to print at 185.

Good ideas Printbus. I'll check those items tonight and see if there are any issues.

OK, I finally have consistent results. The issue was the Greg's extruder/Hexagon hotend .4 nozzle combination can't deliver repeatable results at 100mm/min extrusion rate. I switched to 65mm/min and the results of 3 tests at the default estep setting of 841 was 24.95, 24.91, 24.95. After adjusting to an estep setting of 884.8 the next 3 test showed 19.89, 20.13, 19.93. Well with the margin of error for the width of the marker I was using to do the measurements. I haven't done any testing between 65mm/min and 100mm/min to determine what the maximum feed rate is for consistent extrusion for this extruder/hotend/nozzle combination, but I'm now confident that at 65mm/min I will get good results. This is why I was baffled by the early testing results and yet had by my standards very good prints. I have always had the perimeter FR set to 65, so it was within the consistent range of the hardware. I may have to get some different nozzles to see if this behavior is nozzle diameter limited, or if the higher feed rate deviations exist in the extruder or hotend. I'm going to do a few prints of items I have already printed at the old estep value to see if I can see a difference in print quality. BTW, the Zenn video showed a 100% FR on the LCD controller, but never showed what he had set as a FR in Pronterface. So I have no idea what FR he was using for the test.

I use an exacto knife to make a small cut in the filament to mark length measurements for calibration.

I prefer to calibrate my extrusion from prints. I print a small box with 0% infil and no top/bottom. I use a layer height of .3 mm and a shell thickness of .45mm. I measure the actual thickness of the wall and use that to adjust my E-steps.

I do that to adjust my flow rate, but not to adjust the esteps. What is your technique?

I'm not sure what you mean by adjusting your flow rate. Do you mean the flow % option in Cura (extrusion multiplier ration in Slic3r)? If so this does the same thing as multiplying your E-steps by the same amount. I prefer to leave it at 100% and make the changes in my firmware.

Hey Drone,

In addition to the normal extrusions to check the calibration of your Esteps, I've found that printing a small box with 100 percent infill has helped me really dial in my extrusion. I described this test in the link below.

http://3dprintboard.com/showthread.p...ll=1#post20845

GMay3- This makes sense also and I will give that a try. How long does that box take you to print and have you tried the box at different feed rates? I think at the end of the day it's about being able to deliver a known volume of liquefied filament over a given time period. This can be changed by a number of different things like nozzle size, motor accuracy, filament slipping, gear lash, and nozzle temp I would think.

I think I am going to try this in steps and try an extrusion test without the hotend attached to test the motor and gears with just a measured filament extrusion. Then I will attach the hotend and try the same test again to see the comparative results. I will try at several feed rates to see why on my printer 65mm FR is repeatable and 100mm is not. Might be a waste of time, but interesting to me none the less.

The box takes maybe 10 to 15 minutes for me to print.

That's an interesting point about different feed rates. I have not tried these calibrations at different feed rates but that would be an interesting experiment to see if the box's volume changes.

Either way, I would recommend trying to dial in all of your final print speeds and then run the box volume calibration at that set feed rate. I think of the esteps as a course extrusion tuning and the box volume test as a fine tuning. Exactly as you say, the accuracy of the amount of plastic going into the print below is the most important thing. I just adjust the extrusion multiplier in slic3r until the boxes top is perfectly flat.

Yeah definitely try without the hotend if you're interested in doing it for science! I would be interested to see what you find!

To get my extrusion multiplier, I sliced a calibration cube to be hollow with one perimeter. I used the "vase" option in slic3r to make it easy and quick-printing. I set one up for ABS, and another for PLA. Both .gcode files state perimeters should be 0.4mm wide, so I compare printed wall thickness against that number, and set the extrusion multiplier for that filament spool appropriately. It's the first object I print from a new spool. Slic3r is set to treat all filament as 1.75mm, with different multipliers set for each different spool.

Quote:

---

Originally Posted by AbuMaia

To get my extrusion multiplier, I sliced a calibration cube to be hollow with one perimeter. I used the "vase" option in slic3r to make it easy and quick-printing. I set one up for ABS, and another for PLA. Both .gcode files state perimeters should be 0.4mm wide, so I compare printed wall thickness against that number, and set the extrusion multiplier for that filament spool appropriately. It's the first object I print from a new spool. Slic3r is set to treat all filament as 1.75mm, with different multipliers set for each different spool.

---

That's a great idea AbuMaia. I would say that's an even more fine tuned test than the solid box volume one. I might try this to see if I can dial my extrusion in even more. Sometimes it's hard to tell if the box is completely flat on top and this gives you direct measurement feedback!

Just curious, how close have you gotten to 0.4mm when you actually go to measure it?

So far my multipliers are between .87 and 1,22, over 5 colours and among both PLA and ABS. I don't know if I've yet gone back and printed the vase-box with the multiplier already set. I'll give that a try and let you know.

Oh, one step I forgot: measure the four sides, toss the largest measurement, add the other three together, and divide by 3 to get the average. Then compare that number to the perimeter width number.

The box walls came out between 0.39 to 0.41

Quote:

---

Originally Posted by AbuMaia

So far my multipliers are between .87 and 1,22, over 5 colours and among both PLA and ABS. I don't know if I've yet gone back and printed the vase-box with the multiplier already set. I'll give that a try and let you know.

Oh, one step I forgot: measure the four sides, toss the largest measurement, add the other three together, and divide by 3 to get the average. Then compare that number to the perimeter width number.

The box walls came out between 0.39 to 0.41

---

Great results! I'm definitely going to try this. Thanks for the step by step!

I can't take credit for it. I learned it here: http://zennmaster.com/random-things/...-2-fine-tuning

I am curious how you are getting .39 - .41 wall thickness. I also am running an E3D v6 for 1.75 mm and so far the best I can do is about .49. What settings are you running with? Are you using PLA or ABS?

Thanks.

The measurements I reported were with 1.75 ABS. Slic3r is set with a 1.75mm diameter and 1.02 extrusion multiplier for this spool, determined as described in the link I posted. Print Settings and Printer Settings have just two saved profiles, one for ABS, one for PLA. Filament Settings has many saved profiles, one for each spool. I use a 24mm hollow cube .stl sliced for ABS and PLA with 1 perimeter, 0 infill, 0 top layers, spiral vase option enabled, 1.75mm diameter and 1 extrusion multiplier, to provide the baseline object against which I determine the extrusion multiplier.

Abu- I'm curious what your gcode reported in relation to the actual measurement. At the beginning of the gcode it lists the "external perimeters extrusion width =" value. How does this value correspond to your measured values? Probably not important, but I would like to know if the correlation is even close to accurate.

; generated by Slic3r 1.1.7 on 2014-10-29 at 19:18:24

; perimeters extrusion width = 0.40mm

Yeah, I may need to experiment more. I looked at the instructions referenced earlier. That one used 3mm filament and the instructions say to leave the Slicr settings to standard as far as widths. He used a .4 mm end and Slicr reported a perimeter of .67. Like Abumaia, my code (printing one perimeter) shows a perimeter of .4, but the best I could get it down to was about .49 measured. That was all the way down to about 85% rate. I'm not sure how you get .39 to .41 out of a .4 head. It seems like there would be some expansion??

Quote:

---

Originally Posted by AbuMaia

; generated by Slic3r 1.1.7 on 2014-10-29 at 19:18:24

; perimeters extrusion width = 0.40mm

---

So the extrusion specified by Slic3r is tracking very nicely with the measured value after you set your extrusion multiplier. That's very encouraging for seeing the cal have the desired effect with Slic3r. Thanks for checking that ABU. Great info. Although I'm not sure if the calibration is affecting the Slic3r information (used as part of the calibration. I hope so since that would be dynamic and should track to other wall thickness requests), or if the slice3r information now matches the actual extrusion thickness because of the calibration (and would pertain to only that thickness that you calibrated). But either way it is good to see the calibration seems to make a difference in requested vs. actual in at least the calibrated wall thickness.