Good to know, thanks for the info. I’ll do some testing at different speeds.
The Lulzbot Single Extruder Calibration Workflow says to set extrusion length to 100mm and extrusion speed to 100mm/min, so that’s what I used. I assume if this was too fast for an accurate reading, they would revise the speed in their workflow?
I’ll try slower next time, and see if the number is significantly different.
If your extruder Idler tension is set properly and it’s feeding 1 hobbed bolt notch of filliament for every notch worth it turns and not stripping out or anything, you should technically be able to calibrate at any speed. In practice if you run it faster than normal, there tends to be the occasional slight slip in the feed which doesn’t really affect your prints all that much, but can be really important if you are trying to figure out calibration.
I just noticed the units for extrusion speed in Pronterface are mm/min, but the units in Cura are mm/sec. I can’t believe that both of these are correct, or we’d be hearing people talking about wildly different speeds, one 60X faster than the other. Does one of these have a typo? I don’t have the printer at hand or I’d just time it and see.
Both units are common in 3D printing. And yes, often people are talking about 60x different units 
Cura and Slic3r uses mm/s, Simplify3D and Pronterface uses mm/min. You have to look which values you enter where 
I like mm/s more because 40mm/s sounds more logic than 2400mm/min.
OK - So the extruder calibration workflow (https://ohai.lulzbot.com/project/extruder_calibration/calibration/) says to set:
Extrusion Length: 100
Extrusion Speed: 100
There are no units given. Since they are referring to using Pronterface to set the speed, and Pronterface is in mm/min that’s actually quite slow: less than 2mm/s. So there should be little to no chance of the speed causing a problem when calibrating (I hope). The first time I tried to set the e steps, I was just hitting the 10mm extrude button 10 times from within Cura. I have no idea what that speed was, but I’m betting it was quite a bit faster than a couple of mm/s. I was getting inconsistent results, so I downloaded Pronterface and tried the calibration from there.
Please choose one thread for this discussion, I don’t wont to write everything in two places. As I wrote before, 100mm/min is quite fast for extrusion speed, not slow.
Quote from the Z-Offset in Cura Thread:
By the way, 100mm/min for extrusion speed isn’t slow, it’s realy fast! As I calculated in the the other thread, values around 40-50mm/min are a common speed. With 100mm/min you get nearly 12mm³/s of filament, that’s not even feasable with some printers. I testet the TAZ5 up to 10mm³/s and it’s OK with this printer, but don’t think you can do this with every one…
I used a speed of 50 in pronterface to do my calibration. Seems like a pretty normal extrusion speed. I’ve also done it at 75 and gotten the same number.
As long as your filament doesn’t slip in the extruder, it shouldn’t matter.
At any rate I strongly suggest you follow my method for determining Esteps, which is carefully controlling all slicer variables to be certain you are printing the volume you are expecting, and adjust Esteps to get a good surface finish.
The recommended calibration requires measuring filament while it is in the extruder, and this is awkward and difficult to do accurately. For me anyway. That’s why I changed the method. Plus I was never able to get good results in the prints.
If you are altering flow in your slicer, you’re really just changing Esteps in a round about way. I just prefer to leave flow at 100% and make sure everything else is correct.
If you’ve done the Calibration and have to monkey with flow % to get a proper finish, then your calibration is wrong, bottom line. (assuming the remaining variables that alter volume of filament extruded are correct)
Billyd - I’d like to do as you suggest, but I’m too much of a Noob at this to really know how to “carefully control all slicer variables” or to be able to judge what a good surface finish should look like. I have only really looked at the few parts I’ve printed, an ddont really know what I’m looking at or what I should be looking for.
I know I’ve got the filament diameter correct (several decades of working in a steel wire mill taught me how to properly use a micrometer). It seems logical that if everything else is set correctly, there should be no reason to adjust the flow rate from 100% (I could see using it as a temporary tweak if you needed a quick way to fix a problem).
I need to do some experimentation with the e steps and the z offset, both of which I think can use some tweaking. Am I correct that the two may have some interaction with each other when trying to lay down a good clean first layer? That is, if I think I’ve got my z offset set correctly, that may need to change if it turns out I am over-extruding.
So here’s another question for you calibration gurus:
I just looked at the gcode for some of the STL files (including STL files for parts downloaded from the Mini parts files, as well as the test disks in mhackney’s very good “A Strategy for Obtaining Great Prints”). Every one of these files starts with the line: M92 E850.000000 So what’s the point of fine-tuning the e steps number and putting it in the EEPROM if the first line of gcode just overrides the e steps to 850?
I know I must be missing something here. Is there another command buried in the STL file that retrieves the e steps number that we so painstakingly measured and stored in EEPROM and makes that the active value?
Hm?? The test disks are stl files, not gcode. If your Slicer has M92 in the start gcode and you looked at your self created gcode files, it would be absolute logical that you see it in every file 
But then the question is, why is M92 E850 in your start code…
Edit: I think you also missunderstand M92. This command sets the current axis position to a given value. In this case, the extruder is set to 850. That has nothing to do with esteps!
I loaded the test disk into Cura, then clicked “Save Gcode” and looked the the file that was created (I just used MS word to view it). The first line in every Gcode file I save this way includes the M92 E850. If I look in Cura, turn on “switch to full settings…” under the expert menu, then click on the “Start/End Gcode” tab, there is no M92 command in the code shown (and I have NOT made any changes to the code shown here. It is as it came when I downloaded the software). In any of the Gcode files I save after loading an STL file, all of them have the M92 E850 as the first line. After that, the code looks the same as what is shown in the “Start/End Gcode” tab (the only difference is that some of the values shown as variables in the tab have been filled in with numbers in the saved Gcode file - for example, actual temperatures based on the type of filament selected in Cura).
The Mini printer is not hooked up, since I’m not with it now. Is it possible that Cura Lulzbot Edition tries to look up the e steps from the printer and defaults to 850 if it can’t find the printer? But if it’s trying to LOOK UP the setting from the printer, why would it then need to SET the setting as a first step in the print process (seems kind of like looking up a page number in a book, then writing it down so you’ll remember it … ON the page where you just looked it up).
I’ll try it again when I get to the printer to see if it changes, but the logic of this is just escaping me right now. (Maybe somewhere it’s checking to see if the esteps is even set, and finding no printer it’s thinking “oops! no e steps have been set. I better put in something close just to start with.”)
Attach the gcode of a small part to you post, I will have a look at it. At leat my Cura with TAZ5 doesn’t do that.
Here is the gcode generated from the STL file for one of the 20mm test cylinders in the “Strategy for obtaining great prints” thread. (pt5mmx20mm-cylinder.gcode)
I also included a file where I cut and pasted the Start Gcode shown in the Cura “Switch to Full Settings…” window. It’s an MS word document, since I couldn’t figure out how to export it or print it directly. (Cura Start Gcode.doc)
This just seems to be the default Cura LB Edition behavior - at least when the printer is disconnected. I’ll see what it does when connected when I get a chance to mess with it.
pt5mmx20mm-cylinder.gcode (10 KB)
You are right, I mixed up M92 and G92. Your M92 is realy setting the esteps…
I think I found the reason: Click Machine -> Machine settings. There you find an option to set the steps per mm, set it to 0 to use the firmware settings. Here is a picture:
No idea why it is set for you, maybe it’s a default for the Mini? Very bad if it’s so, because it will be very hard to find this if there is a problem with extrusion that needs calibration… 
That was it! I think I remember it came set for 0. I must have put it in when trying to set the e steps, before I downloaded Pronterface, and I hadn’t spotted where to enter Gcodes in Cura. I reset it to 0, and it’s not leaving that M92 line anymore.
THANK YOU! I never would have figured that one out on my own. It would have been extremely frustrating when I went to fine tune the extruder this weekend
That’s a good catch. It might be something in the start script of your slicer.
I don’t see the M92 command in the g-code files I’m generating through Simplify3D. Just checked my version of LB Cura 17.10, and the genertaed g-code doesn’t have the M92 command either…
Cura only adds it to the g-code if you have something other than 0 in the “E-steps per 1mm filament” field in the Machine Settings page that Sebastian referenced
Lowering the extrusion speed while calibrating e steps made a noticeable difference in the e steps setting.
Using the 100mm/min speed recommended in the extruder calibration OHAI gave me an e steps of 866. Slowing to 40mm/min dropped the e steps to 825, and it was much more repeatable.
Thanks for the tips. It should get me closer to where I want to be.
John
That sounds much better than 866 
I think your prints will look much better now.