Excel spreadsheet to determine workable T-O-M feed rates
Description
This is a work in progress, I cannot be responsible for you totalling your TOM as some basic assumptions I have may be wrong, and my math might be crap. Thanks to Dave Durant who steered my thinking in a different direction, and gave me some impetus to think of an easier way to calibrate my 'bot.
Instructions
*Green fields are values you need to come up with, orange are values the spreadsheet generates*
First, measure the speed at which the filament is pulled INTO the extruder. I simply measured 20mm intervals along the filament, marked and then started the extruder and timed it. After 6 intervals I averaged and divided by 20. This is your feed rate going in.
Then measure your filament diameter (Mine was 2.92mm), extruded filament diameter (Mine was 0.6) and plug these values in. This will then give you a calculated extrusion rate (flow rate).
Now this is where some funky stuff is going on. It seems based on my observations that ~13% of the volume is lost during extrusion. There are posts I've read on the MBI forums about it, and people seem to feel it's moisture absorbed by the ABS being cooked off, either that or some nasty volatiles.
Regardless, this is why I repeated step one, and marked out intervals on the ingoing filament, then started extrusion and chopped a length off every time I reached a mark. 6 or so lengths of filament, average their length and then put that value in the appropriate field (Mine was 221.7mm coming out for 10mm of filament going in) This then gave me my actual, observed flow rate. Using actual and predicted flow rates I came up with the ratio (0.8716). This ratio needs refinement, as it's based on one observation at one speed and one temperature. So, subject to change. It's there for those that want to skip determining actual flow rate and just use predicted.
*****EDIT: 0.85 seems to be the current best ratio for determining input volume vs output volume, in my own testing 0.87 produced values for feedrate that were a little too high, I will be doing more testing to determine, but for now if you don't wish to directly measure the extrusion rate of plastic coming out of your toolhead, use 0.85 (on the unmodified spreadsheet, look for the orange field with 0.8716 and replace) ****
Now, plug in your desired Layer height and W/T values, plus your (hopefully) approximate flow rate and the last box will be, in theory, the Feed rate you need to set to get a nice print.
I threw away my current SF profile and started with a fresh one (TOM, HBP). The only modifications were the spreadsheet values, and Raft BASE and INTERFACE integers were set to 0. Outline was enabled, and First Object layer infill feed rate was set to 0.8 (bit too fast) and First object Perimeter feed rate was set to 0.55 (bit too slow). Temperature was 225C.
From the look of the object, I need to fix my Z starting height ever so slightly, and probably reduce my feed rate a fraction. But it's real close!
Please be nice if I've screwed up something good and proper here :P
You must be logged in to post a comment.
Thanks for this spreadsheet, I'm about to start my measurements now. You mentioned in a previous comment that the calculated flow rate is too high, is this still the case? If so, roughly by how much is the figure too high?
I've also had a few thoughts on this, have you looked in to whether the extruded flowrate vs filament draw flow rate varies with temperature?
I would like to set this up as a feedback model. With a stepper extruded it should be possible to judge the volume of filament entering the extruder if the
diameter of filament can be consistently measured. Would it then be possible to recalculate feed rate on the fly and make adjustments as a print progresses?
I'm still testing the ratio, from the few tests I did I came up with a ratio for volume loss that was around 0.87, but others may notice differences. From other posts on the Makerbot forums the general consensus is that the ratio is more like 0.85. I initially used the ratio I'd calculated from observed flow rates in Skeinforge, to generate my feed rate, but found that with a small amount of tweaking it should have been a little lower. At the moment my good prints are using feed rates that when dumped back into the spreadsheet do in fact match 0.85!
I've only just received more ABS so I can finally go on testing! This chart is not intended to be a 'perfect' setting, it's just to get you in the ballpark for a 'workable' print. And then from there you should only need to alter settings slightly to get to the sweet spot :)
I've not tested yet at different temperatures, it's on the list of things to do :)
As Dave pointed out in this thread:
http://wiki.makerbot.com/forum...
Once we can get an accurate idea of the exact volume of plastic being extruded for a given rate of stepper
speed, and that information can be fed to skeinforge... Well, shit will get real :P I think that it's definitely going to be the defining goal of extrusion control!
For some reason when I click the download link in firefox it wants to save as "TOM" with no extension. I don't know if this has to do with how it was uploaded or just having spaces in the name or what.
I can open the file with OpenOffice for linux.
I can't open it with KSpread.. but that's not surprising ;)
I understand what is this for but.. what is the meaning of TOM, W/T and HBP?
Sorry, using but not explaining acronyms is a bad habit of mine!
TOM refers to the Thing-O-Matic but this sheet should be good for any extruder running at a fixed speed.
W/T refers to Width over Thickness, one of the most important variables in Skeinforge.
HBP refers to a Heated build platform.
I'm testing more values, and it seems that the spreadsheet will
return suggested feed rates that are slightly higher than they should be, so until I find a correlation that I can apply a ratio to, bear that in mind :)
I've not done much testing at this stage, so please let me know if it's working for anyone. And if people need the spreadsheet in different formats.
License
More from benipk
I'm working through the calibration process using your excellent spreadsheet now, and I find I don't see the 0.85 volume output:input ratio that you and others have found.
How exactly do you measure the length of extrusion?
I'm getting pretty much spot on (1% out) observed flow rate versus calculated feed rate, using a method which I think represents the normal dynamic operation of the extruder pretty well, and eliminates start/stop (ooze/compression etc) effects o
n the test. I put a mark on the filament about 10mm up from the extruder, start extrusion and wait for the mark to hit the top of the extruder body. This gets the extrusion up to speed and operating in normal continuous fashion. Then when the mark hits the top of the extruder, I put a kink into the
extrusion with tweezers and then let it run for approx 10 seconds or so. Then the tricky part, stop the extrusion and simultaneously put another kink in the output thread. It will still ooze a bit after that, but the kinks give you a definite set of points to measure the length of extrusion between.
The clear top plate of the mk5 lets you measure the distance of feedstock from the original mark (now inside the extruder body, but visible) and the top of the extruder.
Doing the measurement this way seems to make output volume pretty much equal input volume (probably to the extent that the test
is accurate anyway).
I marked out a series of intervals on the ingoing filament, and then started the extruder. Then, every time a mark on the ingoing filament reaches the acrylic on the top of the mk5, I then chop off a length of extruded plastic from the nozzle. I discarded the first piece as that will be a little off due to the start lag in the DC extruder, and then I measured the diameter of the extruded filament in several places, generated a volume based on that and the length (averaged over 6 or more pieces) and contrasted this with the volume of plastic I know is going in.
A lot of people claim not to see any volume loss, but I'm not doing anything to do with measuring the drive gear diameter etc etc. All my calculations show is that the volume of plastic going into the head does not match the volume coming out in the same time interval.
I might do a little more wo
rk with this in future, but I'm now running a stepper extruder which means I'm going to need to come up with a more complex spreadsheet soon based on being able to achieve variable flow rates.
Interestingly enough, when I altered the RPM in skeinforge to compensate for the gearing I'm using for my
stepstruder (51/7), there was still too much plastic being laid down, it wasn't until I applied the 0.85 constant that my prints became perfect ;)