fine tuning print

Hey all,

Finally got an 'almost' perfect print from my printer, I am still struggling with the bottom few layers not comming out as well as the other layers. I am looking for some suggestions regarding this. This was printed .31mm with a 4mm nozzle at 185c at 3mm/s (really slow i know, i am redesigning a drive system with the printer which is why I need this part in the picture).

I am also getting slight variations in the levels (you can see them on the right edge) where they should be tapered they are not tapering correctly)...

Any thoughts?

Thanks in advance!

one thing i came across recently was the Z axis calibration and particularly that your layer height has to be set at a value that your Z steps can handle = a whole number

Check out > [richrap.blogspot.sg]
look for the section accuracy

you may also want to watch the temp of your first layer(s) . Looks a little hot

Good luck

bigfilsing

I never read anywhere that is has to be a whole number, could you tell me where thats being said?

Maybe that is also my problem with some areas.

Edit:
Nevermind, I found it, but I don't think it really matters... Even though I will try it.
I read elsewhere that I just should do an 80% layer height from my nozzle width?

Edited 1 time(s). Last edit at 05/02/2013 06:33PM by Ohmarinus.

did you even look at the blog link ????

Click the link and scroll down and look for the section "Accuracy"

Also there's a post in this forum called "layer height based on Z axis steps"

Edited 2 time(s). Last edit at 05/03/2013 12:33PM by bigfilsing.

Haha, yes I did, but you didn't read my complete message obviously




bigfilsing Wrote:
-------------------------------------------------------
> did you even look at the blog link ????
>
> Click the link and scroll down and look for the
> section "Accuracy"
>
> Also there's a post in this forum called "layer
> height based on Z axis steps"

thats what happens when you read a post ..wait to reply and the poster edits it :-)

So did it help or make any difference at all ??

I remember a few weeks back i was printing some 50 mm dia thin walled (1.5 mm) pipe and was getting problems every 10th or so layer where the layer wasn't bonding to the previous layer well. Making sure i set my layer thickness to a "whole number" of Z steps helped with that.


Ohmarinus Wrote:
-------------------------------------------------------
> Haha, yes I did, but you didn't read my complete
> message obviously
>
>
>
>
> bigfilsing Wrote:
> --------------------------------------------------
> -----
> > did you even look at the blog link ????
> >
> > Click the link and scroll down and look for
> the
> > section "Accuracy"
> >
> > Also there's a post in this forum called "layer
> > height based on Z axis steps"

Yes, it works now set it to 0.325 and works great.

Thanks for sharing that link again I actually had read it twice before, but didn't pay too much attention to the accuracy part, also because I was very new to printing back then and was still battling other problems



bigfilsing Wrote:
-------------------------------------------------------
> thats what happens when you read a post ..wait to
> reply and the poster edits it :-)
>
> So did it help or make any difference at all ??
>
> I remember a few weeks back i was printing some 50
> mm dia thin walled (1.5 mm) pipe and was getting
> problems every 10th or so layer where the layer
> wasn't bonding to the previous layer well. Making
> sure i set my layer thickness to a "whole number"
> of Z steps helped with that.
>
>
> Ohmarinus Wrote:
> --------------------------------------------------
> -----
> > Haha, yes I did, but you didn't read my
> complete
> > message obviously
> >
> >
> >
> >
> > bigfilsing Wrote:
> >
> --------------------------------------------------
>
> > -----
> > > did you even look at the blog link ????
> > >
> > > Click the link and scroll down and look for
> > the
> > > section "Accuracy"
> > >
> > > Also there's a post in this forum called
> "layer
> > > height based on Z axis steps"

Great to hear it worked

This is also why i make sure my steps /per mm for all axis are whole numbers . I've seen examples where people have 2 or 3 digits after the decimal point. How they expect the stepper to perform "zero point something something something" of a step is beyond me. Baring in mind 16 micro steps already gives 0.1125 of a degree of rotation (or 0 degrees 5 minutes 20 seconds)

Can somebody please explain why these should be whole numbers?

The dimensions in the STL are arbitrary floating point numbers. They should be multiplied by the steps per mm (which are also floats) and then quantised to motor steps, preferably by rounding to nearest, but even truncation will at most give one microstep error.

Also, why do people say layer heights should be a multiple of Z steps, when they are typically only 0.390625um. Who can see such a small distance? I see people say they get banding with certain layer heights, but why?

---

[www.hydraraptor.blogspot.com]

Miss read topic, once I read it properly it is very interesting,
another thing to think about

Edited 1 time(s). Last edit at 05/07/2013 06:08PM by Myles.

Ill be the first admit i simply go with what i read and or makes sense , certainly when trying to solve issues. Im not sure how the firmware or each individual type of board deals with anomalies like being given an instruction it or the steppers cant actually deal with. > a fraction of a micro step

I guess my theory with my problem of bonding ( or lack of it) every 10th layer or so was an accumulative error or something along those lines.
How is a request for a "partial" micro step dealt with? Is it ignored? Rounded up or down? Buffered then adjusted at some stage ( layer) in the future?
If we then take tolerance , wear and tear on components and mechanical accuracy into the mix this may amplify or reduce the net effect.
All i really know for fact is that it helped ( or appeared to) with an issue i was having.

Cheers

And I have been thinking about this.
I think I understand the mathematics of dividing steps, but as Nophead said, "1 step is 0.000390625 mm on the Z axis " ,
and extruding molten plastic from a 0.5/0.4 mm nozzle , I would think this is unlikely to make much difference.
I shall try some experiments.

Whole numbers for X and Y steps however, does not make sense to me. It means that over a print of 100 mm there is a
difference of 100 steps, at a setting of say 71 or 72 steps per mm that could be a difference of over 1.3 mm in 100 mm , and that
means 100 mm could come out at 99.4 mm for 71 or 100.8 mm for 72.
For example, 71.37 steps per mm means that over 100 mm it moves 7137 steps as opposed to 7100 at 71 or 7200 at 72.
Please correct me if I am wrong.
I had a lot of fun calibrating the X and Y steps after replacing the printed pulleys with metal ones,
those decimal places do make a difference .

Edited 1 time(s). Last edit at 05/08/2013 06:19AM by Myles.

When using absolute gcode there shouldn't be any cumulative rounding error. It's a firmware bug if there is.

---

[www.hydraraptor.blogspot.com]

Quote

I guess my theory with my problem of bonding ( or lack of it) every 10th layer or so was an accumulative error or something along those lines.
How is a request for a "partial" micro step dealt with? Is it ignored? Rounded up or down? Buffered then adjusted at some stage ( layer) in the future?

The Firmware converts requested positions to steps and tracks those internally, since all requests during a print are absolute there is no accumulation of error without a bug in the firmware.

---

___________________________________________________________________________

My blog [3dprinterhell.blogspot.com]