Loading
NirDobovizki

Fast edge over/under extrusion test

by NirDobovizki Jun 12, 2016
Download All Files

Thing Apps Enabled

Please Login to Comment

On the third iteration at 78% flow rate as suggested I managed to achieve great dimensional accuracy (10mm x 10.04mm).
Then I noticed the wall separation which looks just like under extrusion.

Something is not right here...

Hello and thank you. By printing your test pieces and processing the results using the spreadsheet, my printer has been transformed from an amusing toy into a very useful tool. For the record it took 3 iterations to settle on a flow rate of 76%.

I appears to me there's some error in the calculator your provided. I'm no 3d printing master, but I'm pretty confident over/under extrusion is usually fixed by tweaking in terms of 1 or 2%. Adjusting it to 75% or 125% is just going to leave you with a shapeless blob/mesh of plastic.

Not trying to disrespect anyone, but I feel like if no one points this out, lots of makers are going to waste their time.

Hi, not disrespectful at all, what you are saying makes perfect sense.

I didn't write the calculator myself, I just designed the test model and I didn't test the calculations much.

The real test is, starting from a well calibrated print, print variants and multiple wrong extrusion multipliers (for example: 0.85, 0.9, 0.95, 1.05, 1.1 and 1.15) - and then check if the calculations correctly gives you the inverse of the bad multiplier.

I'm writing this here hoping someone other than me will do this and report his/her findings, if no one does this then I'll do it but I don't know when I'll get to it (sorry, I'm quite busy right now)

How firmly/loosely should the two pieces fit together?

Hi, it should fit very firmly - you should have to apply some force to get it in the first time - but very little force.

Both sides of this part are exactly 10mm, even in industrial CNC it's customary to leave at least 0.25mm clearance for plastic parts that should fit firmly in place (I'm not a machinist, this is information I found online, I know clearance is much more complicated then this).

So the parts don't really fit, but the part is small and plastic is flexible and you only have to bend it ~0.1mm for it to fit so it does connect relatively easily

Hope this explains it

Thank you for the response, I will try adjusting the flow so the pieces fit together tighter

Question: why 0 infill? You want top layers to be printed as a bridge?

The real reason is that I want it to print fast and it prints just fine with no infill

Even if it didn't print just fine, I designed this to test wall extrusion problems, so I only need the top and bottom in order to prevent the walls from flexing and moving when I measure them, I don't care about the top surface quality

This is ingenious!! Thanks!

My outer dimension is around .5 mm larger than the inside dimension. The calculator says I need to reduce the extrusion multiplier to 0.75 but then I am definitely under extruding since the layers start separating and there are holes everywhere. Any ideas?

I had no idea I was out by so much! I'm pretty new and my prints were acceptable but had some annoying things that made me wonder if I was over-extruding. I printed these and they didn't even come close to fitting together. Using that calculator, it told me to change from 1 to 0.75! I may have a few more iterations to go through and we'll see where things settle... but I wanted to say thanks for the Thing and the clear instructions.

[edit: Oh shoot! I just realized the instructions said zero % infill and I had it left from my last print at 45%. Would that throw off my measurements? Probably. Starting over]

Five iterations later... it settled on 80% and man, those fit so perfectly together I drool - just a bit.

important to note that the multiplier varies from:

  1. Extrusion rate (if you test measure at 1mm/s and try at 3mm/s there will be a difference) (reason isn't 100% exactly known)
  2. How tight/loose the extruder tension is pushing the gear against the filament
  3. Material / Brand

OK, so my outer extrusion is 9.8 mm, and my inner extrusion is 10.1 mm, that means I have a .3 under extrusion right?
Now what?

My other comment got flagged for moderation for some darn reason, but while I greatly appreciate @NirDobovizki producing this model (it is a GREAT calibration piece!), I strongly disagree with him that in your case you should not make adjustments to improve. In fact, if underextrusion is your only problem and you are using a 0.4mm nozzle and 0.42mm extrusion width (assumed), you are underextruding by a very large margin and need to adjust your extrusion multiplier to 1.56! There is a lot of room for measurement error here so you want to be very careful making yorur measurements that you are accurate.

I developed a calculator for use with this model available here https://goo.gl/kbuZMi and I have been receiving good reports from people. I have my dimensional accuracy at about 20 micron after using this to calibrate.

Hi, I actually agree with you.

I didn't mean to say you can't calibrate an hobby printer to less then 0.1mm error, I also didn't mean you shouldn't calibrate - yes my previous message very much say what I didn't mean to say, it was late and I should edit more before clicking "post reply", totally my fault.

What I wanted to say for most models a 0.2mm error is within the model tolerance - so, while it is a miss-calibration it is, most of the time, not a critical issue.

And I really like your calculator, I'm going to add the link to the main details text.

I disagree quote a bit with @NirDobovizki here. It is possible to control the extrusion widths to a much greater precision than 0.2mm on hobby level printers. I was referred to this model because I suspected some slight overextrusion and this was referenced as a good test object. My machine was out by approximately the measurements you specify and making a couple of small test iterations these objects now fit together perfectly and I am now out about 20 microns which is close now the tolerances of both my machine and my calipers.

Indeed I was having overextrusion by about 6% and fixing it cured my top surface roughness and stopped the printhead from chattering and catching when printing across infill layers.

But your calculation is not correct. I am not sure which you are talking about as 'outer' and 'inner' but if the measurement to the inside of the fork is 9.8 and the measurement outside the single protrusion is 10.1 then you have slight overextrusion. If my assumption is backwards, it wont really matter since the calculation for underextrusion is the same, just the numbers end up negative. Since both measurements should be 10mm, then your extrusions are (10-9.8)/2 = 0.1mm too wide on one end and (10.1-10)/2 = 0.05mm too wide on the other. So let's average the error to 0.075mm, correct for it and iterate again. Don't worry that they are different; the reasons are various and a bit complicated, but suffice to say it's easier to iterate a couple of times than try to solve for these errors.

To calculate your extrusion multiplier you have to know your slicer's target extrusion width and what (if any) extrusion multiplier is currently applied. Let's assume for the sake of argument that you are printing with a 0.4 nozzle with extrusion width of 0.45 on your outside perimeter and have a 1.0 extrusion multiplier currently applied.

So, target extrusion width is 0.45mm and actual extrusion width is 0.45mm + 0.075mm error = 0.525mm. So your extrusion multiplier should be 0.45mm desired / 0.525 actual =~ 0.86 times the current extrusion multiplier if one is in use so 0.86mm * 1 = 0.86mm

Let's say you then run a second run and for the sake of argument find that the pieces fit too easily. You might measure 9.95 mm on the outside measurement and 10.01 mm between the fork when printing with 0.86 multiplier. Feeding that all back in you have:

(9.95 - 10) / 2 = -0.025mm error on outside (underextrusion)
(10-10.01) / 2 = -0.005mm error on fork (underextrusion)
-0.015mm average error
(0.45mm - 0.015mm) = 0.435 actual extrusion width
0.45mm expected / 0.435 actual = 1.03 * 0.86 current multipler = 0.88 new multiplier

Hope that helps and is not too confusing. You should only have to make 2 adjustments unless your machine is really inconsistent. If you set it up in a spreadsheet where you just plug in your measurements you can dial in a new filament in 20 minutes using this object without having to go through the process of measuring filament feed distance, measuring filament diameter, etc and then wondering why you still have to guess at an extrusion multiplier. So while I disagree with @NirDobovizki on his opinion of what is good enough for a hobby printer, I am very glad he produced this model since it makes an absolutely perfect calibration object!

I'm probably having a senior moment (at only 5.7 decades old? surely not!) but I couldn't understand what the 'Perimeter Extrusion Width' value is meant for in your no doubt excellent 'Extrusion Multiplier Calculator'. Can you explain it to me? Ta!

Your lines are off by 0.1/0.2mm, that's perfectly acceptable for most things you can print on an home 3d printer anyway, it could be a slicing rounding error, it could even be the accuracy limitation of your printer's mechanics or the device you used to measure the part - I would leave everything as is unless there are other more visible quality problems.

If you suspect extrusion problems because of other issues:

  1. Check you slicer configuration, especially flow rate and any feature designed to make holes bigger (because bigger holes means thinner part)

  2. Get a new nozzle, they clog and wear out and you can get them for next to nothing on AliExpress (last set I ordered was $1.22 for 10 nozzles) I just keep spares and swap nozzles at the first sign of trouble, it usually helps.

  3. Google "3d printing under extrusion" for more information about diagnosing and fixing extrusion problems