Question about an Oval Filament
I'm writing this post because I'm facing difficulty.
I tried to extrude PA12 (Nylon) powder to make a filament of 2.85 mm in diameter but the filament I obtain is oval and not round.
I read the troubleshooting article about oval or flat filament and I have followed the advice:
- Reduce the temperature at the end of the extruder.
- Reduce the RPM
- The barrel is clean
- Reduced the output speed
(The only thing that I wasn't able to do yet is to optimise the cooling but we are working on it)
Even after that, I wasn't able to improve the shape of the filament.
It seems that when the PA12 is coming out of the nozzle it's already starting to be oval.
Do you have another solution to fix this problem?
Good to hear that you already tried a few things based on the article. Based on the things you already tried, your observation that the PA12 coming out of the nozzle is already starting to become oval, as well as the typical properties of PA12, it becomes quite clear that the cause is due to non-uniform shrinkage, caused by crystallization.
The only way to combat this is by modifying the cooling. Something that often helps in this situation, is making use of a circular cooling add-on (I think it is also mentioned in the article). I have added a zip file with all necessary STL files in there, so you can 3D-print one yourself :).
I hope this helps reduce the ovality!
Good afternoon @timo,
We used the circular colling you provide.
The diameter of the filament as improved. Now the filament is almost always round.
However, we are still facing some difficulty.
First, the filament tends to wave. When we increased the speed of extrusion, the filament is straighter. But then the filament isn't cooled enough.
Secondly, the size of the filament diameter won't stabilise. The size varies between 2.70 and 2.90 mm at best (Measured by the sensor and with a calliper).
On a side note, it can happen that the sensor won't measure the diameter of the filament (0.00mm display on the screen). I supposed that do to the wave of the filament.
Have you ever heard of this kind of default?
And do you have an idea to solve them?
Hello @francoisnicol ,
Thank you very much for this detailed explanation.
I think that the main issue here, is the "wave" pulling. This usually happens when cooling is excessive. The fans (or single fan) blow a lot of air, push the filament to a side, freeze it into a wave shape, and then the filament is bent toward the other side, freezes......
According to what you observed, there seems to be a very narrow window for optimal cooling.
Thickness fluctuation :
This could be influenced by the first point, which is why achieving straight pulling has higher priority. The sensor only works well when the filament is falling vertically (straight) through it.
Whenever the filament gets out of the sensor's range (wwhen it's waving around, for instance), the sensor reads a zero and starts its calculations again --> it impacts the entire pulling process.
In addition to the reading/moving issue, the flow of molten plastic coming out of the barrel might be inconsistent by itself. This would cause a diamter fluctuation that cannot be 100% fixed by the autotuning of the puller.
There is one trick you can try, which might solve most of these issues :
- Go to Main menu / Speed tuning, then switch to Manual.
- You can accelerate the puller enough to pull the filament straight vertically
- Then slowly decrease to get closer to 2.85mm.
- If the filament moves like a wave at one point, then increase the puller speed again to have it straight, then decrease the fancooling slightly, then try to decrease the puller speed a little again.
The objective is to find a compromise between RPM, cooling, puller speed.
In addition to this, once you leave the puller speed on one single manual value for 30min, the thickness graph represents the stability of the flow very well, without the influence of the automatic adjustments.
Does that help ?
This topic is quite hard to tackle because it involves many concepts 😉
I would be glad to discuss this again (in French if you want).
Sorry for the late reply,
I apply your advice, and I was able to obtain a straight filament. But I wasn't able to improve the calibration of the filament diameter.
As you say, it's a complicated topic and the fact that we use a PA12 powder and want to make a diameter of 2.85mm doesn't help.
If it's not a problem for the forum, we can continue the discussion in French.
Hello @francoisnicol ,
I thought that my answer could be of interest to other users, so I decided to write it in both languages, for maximum clarity 😉 Please find the French version below.
If I summarize the situation, you have managed the roundness of the filament and pulling straightness by adjusting the cooling. The remaining problem is the flow stability (diameter fluctuation).
I can think of two causes :
- imperfect settings (temperatures, RPM)
- feeding issues. Because you work with a powder instead of pellets, certain feeding issues may occur in the hopper, causing flow fluctuations, as illustrated here
It is difficult to assess the true stability of the flow because you have been facing pulling issues and the diameter is influenced by the automatic adjustments of the puller.
I therefore suggest you to run another test (+ log the data) with the puller speed set on manual : Main menu / Speed tuning / Manual / [change the speed with the black knob ; the higher the value, the faster ; choose a value that gives 2.85mm approx]
You can then send me the datalog + short video of the output. The Filament Thickness graph will allow us to look at the flow stability, unmodified by any automatic pulling adjustment.
En français :
Si je résume la situation, vous avez maîtrisé la rondeur du filament et le tirage, en jouant sur le refroidissement. Le problème restant est l'instabilité du flux (fluctuation du diamètre).
Je vois deux causes possibles :
- paramétrage imparfait des températures et RPM
- mauvaise alimentation. Car vous travaillez avec une poudre et pas des granulés, certains phénomènes peuvent empêcher une alimentation fluide, comme illustré ici
Car vous avez eu des difficultés de tirage, il peut être difficile d'observer la stabilité du flux, car le graphe du diamètre est influencé par les ajustements automatique de la poulie.
Je vous propose donc de refaire un test (et enregistrer les données) avec la vitesse de poulie en manuelle : Main menu / Speed tuning / Manual / [changer la valeur avec le bouton rotatif sachant que valeur haute = vitesse rapide ; vous pouvez choisir une valeur qui donne environ 2.85mm en moyenne].
Vous pourrez m'envoyer le datalog ainsi qu'une petite vidéo en sortie de buse (sur le forum ou à email@example.com ). Le graphe Filament Thickness nous permettra de voir la stabilité du flux non 'modifiée' par les ajustements de tirage automatiques.
Pensez-vous avoir un problème d'alimentation ?
Que pensez-vous de la solution que je propose ?
Good afternoon @louis
I was aware of the phenomenon of ratholing and bridging.
To overcome this difficulty, I agitate a pièce of filament in the hopper to mix the powder.
At first, I thought that I was doing a good job, but it seems that our powder is so thin that it wasn't sufficient.
Therefore, we are designing a new hopper that will be more suitable for the powder (steeper wall, integration of a stirrer).
I will let you know when it's done.
As you can see in the picture I send, the speed of the extruder is not stable.
As you say, it can be one of the reasons why the diameter of our filament won't stabilise.
When the new hopper is in place, I will look if there is an improvement and, If necessary, I will try to find better settings for the temperature.
On a side note, when I open the extruder to remove the grid, I found that powder went under the extruder and is pouring out inside the chamber.
I don't think it's dangerous, but I would like to know if you ever face this.
I think the powder went between the joint and the screw fairing.
Thank you for your help. We appreciate that you take the time to guide us through our research.
I won't be available for three weeks, so it will take time before I try again.
Hello @francoisnicol ,
Thank you very much for this new honest and detailed answer !
It is a pleasure helping you 😉
Indeed, inconsistent feeding can cause RPM fluctuations. The reason is that the RPM control system is quick but still reactive. A variation of input density causes a variation of pressure, and therefore rotation.
Constant feeding would make things smoother.
RPM fluctuations can also be an indicator of inappropriate settings (when the temperatures are excessive or insufficient). But to diagnose this, we must first be sure that the feeding is consistent.
We do offer a vibration add-on device that you can place in the hopper : https://3devo.com/product/the-feeder/ . But please feel free to do some DIY work ; I would be curious to see your prototype 😉 Their are two ways you can tackle the ratholing issue : vibration, and stirring.
Powder inside the machine :
I have seen this issue happen with one fine powder before. The powder was slipping between the top of the hopper, and the top panel (exactly where the grid is). In that case, I protected the inside of the machine with tape.
I must say I never witnessed any negative impact on the electronics, but on the long run, this could have consequences.
Does that help ?
I would be glad to continue this conversation
We are changing our plan for the moment. We will try to use the Feeder to improve the flow.
To order it, we need to register the product in our database and fill in the reference number.
Could you provide it?
Powder inside the machine
I was able to remove the powder that had fallen inside. I will apply your technique for the next run and we will see the result.
Good afternoon @louis,
I hope you are doing well.
It took us a little bit of time to order and use the Feeder.
It seems that the Feeder improves the feeding rate of the extruder.
I've been following our test with the Devovision app and, I'm facing some difficulty.
Firstly, the graph about the filament thickness doesn't register the actual size of the filament.
We try to create a filament of 2.85 mm in diameter.
On the graph, the size is always under 1.5 millimetre.
However, the size displayed on the extruder screen was almost always above 2mm. I also measured the filament with a calliper and, the filament had a diameter above 2mm.
Secondly, the graph of the RPM of the extruder seems strange to me for the following reason:
- The value of RPM chosen was between 4 and 7. On Devovision, the value measured is multiplied by five.
- A lot of time, the RPM fall to zero. Is it an error of the captor? Does it have an impact on the extrusion?
I couldn't find any explanation of why the screw would stop. The feeding rate seems fine; the PA12 flow smoothly out of the extruder.
I hope you will be able to help me.
I will continue to work on my side to find a solution.
I think I know what is happening here. It seems to be an issue with Devovision using the wrong data (Filament diameter is showing in the RPM graph for example). So it is not the screw that is going to 0, but it is probably the filament moving out of range from the optical sensor.This issue should be fixed by updating the firmware on your Filament Maker to 1.3.2.
On this page you will find this firmware version: https://support.3devo.com/firmware-updates/
If the filament is moving around alot and moving out of range from the sensor, this makes it a bit harder for the puller mechanism to adjust the diameter accordingly. What I sometimes do then is put something with a metal/smooth surface on top of the sensor block, to keep the filament steady within the sensors range. In the attachment I have added an example.
Please let me know if it works or not, I would be glad to help :)
As you suggested, I download the last version of the firmware and, since then, the result makes more sense.
I also used an Allen key to block the filament and, it works remarkably. (we will try to create a tool that is easier to use) Now the measurements are way more stable.
The quality of the filament has improved drastically since the beginning of the test.
I've never been this close to a perfect filament.
The only thing that I need to resolve is the size calibration. It's way better than it uses to be, but the fluctuation is still out of the limit if we want to use the filament on our 3d printer.
If you can suggest what parameter to look at for the fine tuning, I would appreciate it because I tried different settings without improvement.