Experiments in high speed printing.

As I have been attempting to do complete print beds I have been wondering just how fast can a Mendel go?

So today I thought I would find out as I needed a new extruder base.

This was printing at 60mm/s with a 60% slow down for the outer shell. Travel speed between printing was 80mm/s.



I had to overcome a few problems as the Z axis has to move independently of the X and Y or the electronics cant handle it and it causes non accelerated movement. This leads to missing steps. I solved that by enabling homing and putting G4 P000 pause command of 0 time.

Other alterations are you need a higher restart extra distance to make it pick up at this speed. This can lead to some blobbing but with a 50% build time reduction that's not bad. Now I just have to set all the parameters so it prints more accurately and I should be printing super fast.

---

Make your Mendel twice as accurate.
[www.thingiverse.com]

Are you using a firmware with acceleration?

I'm using speeder firmware on gen6 which has acceleration.

---

Make your Mendel twice as accurate.
[www.thingiverse.com]

It's interesting that when going around the holes (circles), it's doing the same thing as RAMPS 1.2. It doesn't go around the circle smoothly and sounds like it starts and stops. If you make a slightly larger circle with a radius of 10 mm, you would probably notice the blobs around the circle as I did. Take a look at my post.

The picture below shows a rounded corner of a square around 70 mm on each side. The fillet on the corner has a radius of 10 mm. I wonder if you can reproduce this with Gen 6.



I'm thinking that this can be fixed by installing Kliment's SD card shield.

Edited 1 time(s). Last edit at 06/17/2011 05:56PM by brnrd.

I saw the same issue during radii with Tesla's Tonokip firmware. Since switching to caruKlip, this issue has gone away entirely. That firmware buffers many lines of gcode so you don't get the micro-pauses between lines. You would have the same effect using an SD card shield with any other firmware, as the pauses between lines would be insignificant when reading directly from a card.

---

Help improve the RepRap wiki! Just click "Edit" in the top-right corner of the page and start typing. Anyone can edit the wiki!

I forgot to mention that I am using the latest Sprinter from Caru. For now, I have fixed the problem by slowing down the perimeter feed and flow rate to 1/4 the operating feed rate (80 mm/s). This seems to be working.

Quote

Other alterations are you need a higher restart extra distance to make it pick up at this speed. This can lead to some blobbing but with a 50% build time reduction that's not bad.

The extra distance should always be zero or else you will get too much plastic. What you actually need is a small delay after the fast forward to give time for the plastic to exit the nozzle and stick.

---

[www.hydraraptor.blogspot.com]

I'm not sure about that. It might work with your hot end but I'm curtain it wouldn't with mine. And I have spent much time observing and testing things out. Mine would just put a thin thread down until it started again. It has a delay so that it doesn't extrude until it has pushed a bit through so you need a little extra push or the tread is not the correct diameter. Maybe delaying it would work I just can't see how with the behaviour mine exhibits.


How do you insert the pause command for the start of every thread?

I think that what I needed is a dynamic system that adds extra advances based on how long the filament has stayed retracted. I have to compromise so I don't get big gaps from a long move but the blobs are reasonable from a short ones.

---

Make your Mendel twice as accurate.
[www.thingiverse.com]

Madkite Wrote:
-------------------------------------------------------
> I'm not sure about that. It might work with your
> hot end but I'm curtain it wouldn't with mine. And
> I have spent much time observing and testing
> things out. Mine would just put a thin thread down
> until it started again. It has a delay so that it
> doesn't extrude until it has pushed a bit through
> so you need a little extra push or the tread is
> not the correct diameter. Maybe delaying it would
> work I just can't see how with the behaviour mine
> exhibits.

Well yes if it oozes when it is supposed to be stopped then you would need to make up for the plastic lost. I assumed your retract stopped it oozing.
>
>
> How do you insert the pause command for the start
> of every thread?

I use the python sleep() function ;-)

>
> I think that what I needed is a dynamic system
> that adds extra advances based on how long the
> filament has stayed retracted. I have to
> compromise so I don't get big gaps from a long
> move but the blobs are reasonable from a short
> ones.

The basic problem is that you are not retracting enough, or fast enough to prevent ooze.

---

[www.hydraraptor.blogspot.com]

I found the problem.

I didn't need a delay nor should I have had to add extra advance like you said nophead.

It was all because of the pinch wheel. My old one was straight and brass. The sharp edge had worn off the teeth and it had never been really good.

So I made a new one which I turned a curve in that was the tapping size for 3mm by .5 tap and hobbed it with the tap. I made it of silver steel so I hardened it too so it would keep its sharp edge.

Now the settings I change seem to have much more tangible effects. And the the grip is fantastic while only distorting the filament by the tiniest amount. It also only has very fine teeth so its still quite smooth after passing through.

Perfect for my next improvement.

Thanks for the help everyone.

---

Make your Mendel twice as accurate.
[www.thingiverse.com]