Over Extrusion

  • Ok, there is an old thread on here about this for deckingman's core-xy but his machine seemed to fix itself! Alas, despite giving my machine a rough time it appears to have a stubborn personality and so far not decided to fix itself!

    Ok machine spec:

    Core-xy with 1.8 steppers from the Reprappro Ormerod 2.
    Duet v0.6 with 1.19.3.
    No z-hop on retraction.
    Titan 1.75mm universal extruder.
    LONG bowden tube (currently printing parts to move titan onto the carriage.)

    Extrusion steps calculated by running 100mm of filament out of the extruder with bowden removed. Procedure:

    • Remove bowden.
    • Clip filament 10mm from extruder.
    • Press a sharp blade into the filament to mark it using the outlet of the extruder as a guide.
    • Drive 100mm of filament at 5mm/sec
    • Press a sharp blade into the filament to mark it using the outlet of the extruder as a guide.
    • Drive 10mm of filament at 5mm/sec
    • Clip measured sample from extuder.
    • Press into a steel rule, aligning the start mark with 100mm and measure length with a target of 200mm on the rule.
    • Correct against that value in the normal way.

    I measure the filament in 3 places using a 3d.p. external micrometer with a ratchet on it. Zeroed to two clicks shut then measure the filament the same way. It doesn't compress the PLA of ABS visually at all.

    I've seen significant over extrusion, particularly on the first layer with PLA and ABS. I've just run a build and given up and dropped the extrusion multiplier to 0.95 as a temporary bodge - but I don't like this!

    So the over extrusion is at least 5% maybe 10%. This means if I'm measuring the filament as 1.75 it is really somewhere between 1.80 and 1.84mm I don't think my measurements are that bad, but I will try to see if the lab technician here can help.

    Anything I'm doing wrong?


    Nozzle: 0.4
    Layer thickness: 0.35|0.3

    Could this be an over temp extrusion issue?

    Currently running PLA at 180|175C

  • Looking at the steps as to potential source of error I'd need to be more than 0.5mm out, which again I don't think I am. It's around 0.002g per mm of filament so doubt weighing would me more accurate. Will consider a longer sample length though when material isn't so tight.

  • administrators

    I find that I need to adjust the extrusion factor differently for each roll of filament. With most filaments I need to reduce extrusion, sometimes by 5%, but with the most recent roll I bought from rigidink I need to increased it by 5%. I don't usually bother measuring the diameter (the rigidink filament is supposed to be accurate to within +/- 0.03mm), I just do a test print when I start a new roll and adjust the extrusion factor in DWC or PanelDue during the print to find out what works well for that roll.

    However, I suggest you measure the filament going into the extruder when calibrating the steps/mm, not coming out of it.

  • I don't think you are doing anything wrong

    The problem I had, which made no sense at the the time (and it still bothers me that it mysteriously fixed itself) was lot more severe as I was using extrusion multipliers in the order of 70%.

    I usually run 95% extrusion as a default because it gives me the best finish. I don't consider this to be a bodge though. My reasoning is on the lines that one can never fill a cuboid of X mm^3 with the same volume of cylindrical filament because however you pack those cylinders together, and however much you try and squash them, there are likely to be gaps. An analogy might be to look at the end of a pack of drinking straws and observe the gaps where the straws aren't if full contact with each other.

    One thing I did learn was that what appeared to be over extrusion of the first layer, was merely that the first layer height wasn't spot on. At that time, I also had issues with getting good first layer adhesion and dropping the probe offset effectively squashed the first layer and made it stick better but this of course can show as over extrusion (because the filament gets pushed out sideways). I found a better way was just to use a wider first layer width and put the probe values back to what it should be. Nowadays, I use 3DLac and set Z=0 homing as being 0.05 mm away from where the nozzle just touches the bed. So my first layer height is 0.3mm but the nozzle is actually 0.35mm above the build plate. This works very well for me but maybe wouldn't work with other print surfaces.

    There was also an issue with Slic3R at that time which caused over extrusion when infilling small areas, so slicers play there part too.

    You mention layer thickness of 0.35/0.3. Did you mean layer height? If so, that's fine but if you mean extrusion width then a better setting would be the nozzle diameter (0.4mm) or greater.

    IMO running an extrusion multiplier of 98 to 95% is quite acceptable and "normal".

  • Fair enough. Thanks for the feedback. The overbuild is enough to the nozzle to drag over the surface leaving a divot where it has tracked. I was hoping to be able to prepare stock builds and then change filament diameter prior to build to suit allowing to run repeat builds on new spools. Is there a way (other than fudging diameter) to supply an extrusion multiplier to the duet with gcode?

    Also any way to check nozzle temps? IR sensors aren't really designed for small shiney metal targets! I'm assuming because this is one of the thermistors with high resistance rather than a PT100 that the length or wiring or connectors are unlikely to have much of an effect on reading unless extreme?

  • @DocTrucker:

    "…..............Is there a way (other than fudging diameter) to supply an extrusion multiplier to the duet with gcode?"

    Not really. You can use the sliders or you enter M221 via the console or embed an M221 in the start gcode. There are other ways to change the amount that gets extruded. Some people alter the steps per mm but I prefer to keep that as calibrated and adjust the extrusion multiplier (which is what it's for). Some people print a small test piece when ever they get a new reel of filament, ascertain the extrusion multiplier then right that value on a sticky label and attach it to the spool.

    "Also any way to check nozzle temps?"

    Whatever you do, you are only comparing one measuring device (the one built into the hot end) with another. Generally, it's difficult to find a better place to measure the temperature than where the hot sensor is installed. You might use a more accurate sensor but still end with a measurement error if you try and read the temperature on the outside of the nozzle instead of inside the hot end. TBH, I used to use PT100s on my hot ends but discovered that they read within about 1 degree of a thermistor so I've gone back to thermistors because they are smaller and I was having problems with the very thin wires breaking on the PT100ss that I was using. I've also found that with my Diamond hot end, the best print temperature is always about 10 degrees less than the minimum recommended by the filament manufacturers. I suspect this is due to the long melt chamber but for whatever reason, that's what I find best. So there isn't a lot of point in trying to get a highly accurate measurement, because you'll likely find that you end up adjusting the temperature to suit your machine.

    That's my twopence worth anyway…..

Log in to reply