Notifications
Clear all

First layer issues? Look here!  

Page 4 / 6
  RSS
stahlfabrik
(@stahlfabrik)
Honorable Member
Re: First layer issues? Look here!

I already had contact with support. He said he brought my pull request to the attention of the developers. Still, I have not heard from the devs if I am right, or wrong - some kind of feedback would be HUGELY appreciated...

I think maybe there are other parts of the code that they have to check to see if there are any more bugs related to this. And of course the MK2, where I think the goal is to unite the firmware versions, uses that line of code IMHO too. Maybe for this machine the sign of the operation is valid? I really do not know.

But I really would like to receive some kind of feedback...

If you experience that my bug fix solves the problem for you: Please comment my pull request and please write an email to Prusa support!

Posted : 08/03/2018 2:07 pm
stahlfabrik
(@stahlfabrik)
Honorable Member
Re: First layer issues? Look here!



I did not try the G80/G81 commands yet. But I was wondering: during xyz calibration the PINDA searches not only in x/y direction but also the optimal z-height where it can detect the build plate targets. I assume this z sensing is equally affected by your patch, no?
Anyhow, maybe I just got a wrong impression during xyz calibration. The end result is an improved first layer consistency, so I'm happy with it 😉

I do find it a bit suspicious that PR did not give a reaction on the bug patch yet. A simple "yes, you're right - the sign of the temperature correction was wrong" or "no, you made a mistake" would be appreciated.

with the G80, G81 you make a mesh bed leveling and print the results. So you can compare point by point.

A pinda with a temperature under 35C must be used for this analysis. Because the temp compensation kicks in above 35C and pushes the mesh level up or down (depending on the firmware patch you use;-))

A non broken pinda will show nearly identical values for each of the leveling points from one measurement to the next.

About the xyz calibration. I am not 100% if my bugfix comes into play there. Maybe in the last phase where steel sheet is put on the printer. I would not do a XYZ calibration with a PINDA that is warmer than 35C...

Posted : 08/03/2018 2:13 pm
stahlfabrik
(@stahlfabrik)
Honorable Member
Re: First layer issues? Look here!

I do not want to spoil the effect of two more Pull Requests (not for bug fixes but for calibration tools) and a longer posting, but my further experiments show proof to me that my bug fix works 100%. 😎 😎 😎 😆

Posted : 08/03/2018 10:14 pm
Brigandier
(@brigandier)
Reputable Member
Topic starter answered:
Re: First layer issues? Look here!


I do not want to spoil the effect of two more Pull Requests (not for bug fixes but for calibration tools) and a longer posting, but my further experiments show proof to me that my bug fix works 100%. 😎 😎 😎 😆

Made me look at Github, saw 3.1.3 RC1 just went up on Prusa Downloads page. The firmware links I posted just a few posts back have the exact same commits plus your fix (or your fix+LA). If anyone is interested in 3.1.3 RC1, those firmwares should offer the same thing. 😉

My MK3 Parts: [Bowden] [New Shoes] [TPU Micro Springs]

Posted : 08/03/2018 10:19 pm
stahlfabrik
(@stahlfabrik)
Honorable Member
Re: First layer issues? Look here!



I do not want to spoil the effect of two more Pull Requests (not for bug fixes but for calibration tools) and a longer posting, but my further experiments show proof to me that my bug fix works 100%. 😎 😎 😎 😆

Made me look at Github, saw 3.1.3 RC1 just went up on Prusa Downloads page. The firmware links I posted just a few posts back have the exact same commits plus your fix (or your fix+LA). If anyone is interested in 3.1.3 RC1, those firmwares should offer the same thing. 😉

Yeah I already rebased my Pull Request. Still my bug fix has not been acknowledged. But as I just hinted and I have figured out two tools (software wise) to do perfect temperature calibration. With this and my bug fix I was able to print PLA without touching Live Z with a PINDA temperature of 35 and of 55. This proves to me that my offset now compensates the PINDA physics instead of boosting the effect. Now is pay back time - I spend so many weeks of my free time on this. :mrgreen:

Posted : 08/03/2018 10:25 pm
Brigandier
(@brigandier)
Reputable Member
Topic starter answered:
Re: First layer issues? Look here!




I do not want to spoil the effect of two more Pull Requests (not for bug fixes but for calibration tools) and a longer posting, but my further experiments show proof to me that my bug fix works 100%. 😎 😎 😎 😆

Made me look at Github, saw 3.1.3 RC1 just went up on Prusa Downloads page. The firmware links I posted just a few posts back have the exact same commits plus your fix (or your fix+LA). If anyone is interested in 3.1.3 RC1, those firmwares should offer the same thing. 😉

Yeah I already rebased my Pull Request. Still my bug fix has not been acknowledged. But as I just hinted and I have figured out two tools (software wise) to do perfect temperature calibration. With this and my bug fix I was able to print PLA without touching Live Z with a PINDA temperature of 35 and of 55. This proves to me that my offset now compensates the PINDA physics instead of boosting the effect. Now is pay back time - I spend so many weeks of my free time on this. :mrgreen:

From all of us who are going to enjoy your hard work, Thank You.

My MK3 Parts: [Bowden] [New Shoes] [TPU Micro Springs]

Posted : 08/03/2018 10:27 pm
stahlfabrik
(@stahlfabrik)
Honorable Member
Re: First layer issues? Look here!



Made me look at Github, saw 3.1.3 RC1 just went up on Prusa Downloads page. The firmware links I posted just a few posts back have the exact same commits plus your fix (or your fix+LA). If anyone is interested in 3.1.3 RC1, those firmwares should offer the same thing. 😉

Yeah I already rebased my Pull Request. Still my bug fix has not been acknowledged. But as I just hinted and I have figured out two tools (software wise) to do perfect temperature calibration. With this and my bug fix I was able to print PLA without touching Live Z with a PINDA temperature of 35 and of 55. This proves to me that my offset now compensates the PINDA physics instead of boosting the effect. Now is pay back time - I spend so many weeks of my free time on this. :mrgreen:

From all of us who are going to enjoy your hard work, Thank You.

Thank you for helping me out to raise awareness! Awesome.

Posted : 08/03/2018 10:51 pm
Crunch
(@crunch)
Estimable Member
Re: First layer issues? Look here!

3.1.3 RC1 says it only for MK2.5 Do the MK3 and MK2.5 use the same firmware?

The Latest Firmware can be found here https://github.com/prusa3d/Prusa-Firmware/releases
Open Firmware Issues https://github.com/prusa3d/Prusa-Firmware/issues

Posted : 09/03/2018 12:45 am
Brigandier
(@brigandier)
Reputable Member
Topic starter answered:
Re: First layer issues? Look here!


3.1.3 RC1 says it only for MK2.5 Do the MK3 and MK2.5 use the same firmware?

Good catch. The MK3 branch is updated to 3.1.3 RC1 in the files, but they haven't posted it to the downloads page yet. I got confused as they have rearranged the download page columns and the MK3 download is on the far right instead of center now. 🙂

My MK3 Parts: [Bowden] [New Shoes] [TPU Micro Springs]

Posted : 09/03/2018 1:38 am
Crunch
(@crunch)
Estimable Member
Re: First layer issues? Look here!

Is the EINY04 variant for the MK3?

The Latest Firmware can be found here https://github.com/prusa3d/Prusa-Firmware/releases
Open Firmware Issues https://github.com/prusa3d/Prusa-Firmware/issues

Posted : 09/03/2018 1:46 am
Brigandier
(@brigandier)
Reputable Member
Topic starter answered:
Re: First layer issues? Look here!


Is the EINY04 variant for the MK3?

I'm not really sure the relevance of these codes. I assumed they were different versions of the EINY and that it may need to be adjusted, but Prusa only ever releases one version of the .hex file on its download page. Thus far I have always left it to whatever it is set to in the MK3 branch on Github and haven't had anyone complain that my builds don't work on their MK3.

My MK3 Parts: [Bowden] [New Shoes] [TPU Micro Springs]

Posted : 09/03/2018 3:10 am
gorkish
(@gorkish)
Eminent Member
Re: First layer issues? Look here!

PR doesnt get very engaged with community contributions honestly. Look at all the pulls for very simple stuff like translations to new languages that have been sitting for years. Still, a sign bug would be a rather egregious problem, and I'm really unsure of why if it is a bug it has not been noticed before. I have tried to examine it but I do not know enough about how it is supposed to work to be conclusive.

I have tried printing with this firmware and notice that I do have less irregularity over the bed surface than before; I still have to let the bed warm up for 10+ minutes before I can trust that my live z distance from a previous calibration will work. The print head tends to be too high if I jump straight into printing after the bed gets to temp. If I preheat and leave it for about 10m the first layer will be good.

Posted : 09/03/2018 6:09 am
stahlfabrik
(@stahlfabrik)
Honorable Member
Re: First layer issues? Look here!


PR doesnt get very engaged with community contributions honestly. Look at all the pulls for very simple stuff like translations to new languages that have been sitting for years. Still, a sign bug would be a rather egregious problem, and I'm really unsure of why if it is a bug it has not been noticed before. I have tried to examine it but I do not know enough about how it is supposed to work to be conclusive.

I have tried printing with this firmware and notice that I do have less irregularity over the bed surface than before; I still have to let the bed warm up for 10+ minutes before I can trust that my live z distance from a previous calibration will work. The print head tends to be too high if I jump straight into printing after the bed gets to temp. If I preheat and leave it for about 10m the first layer will be good.

I will come up with a new forum post where I detail the problems that you mention here.

I can say and confirm now: Try to do the first layer calibration with 35 degrees PINDA temperature (I will provide tools to make that easy). And then you have to do manual temperature calibration, which is impossible to do now with the stock firmware (I will provide tools for that as well).

Currently all that helps is to start prints with exactly the same PINDA temperature to get the consistent first layer height. That is especially true for the stock firmware...

Posted : 09/03/2018 9:50 am
douglas.d2
(@douglas-d2)
Eminent Member
Re: First layer issues? Look here!



3.1.3 RC1 says it only for MK2.5 Do the MK3 and MK2.5 use the same firmware?

Good catch. The MK3 branch is updated to 3.1.3 RC1 in the files, but they haven't posted it to the downloads page yet. I got confused as they have rearranged the download page columns and the MK3 download is on the far right instead of center now. 🙂

I don't understand where to download this update on GitHub. I can't find it.

Posted : 09/03/2018 4:10 pm
neil.e
(@neil-e)
Estimable Member
Re: First layer issues? Look here!



3.1.3 RC1 says it only for MK2.5 Do the MK3 and MK2.5 use the same firmware?

Good catch. The MK3 branch is updated to 3.1.3 RC1 in the files, but they haven't posted it to the downloads page yet. I got confused as they have rearranged the download page columns and the MK3 download is on the far right instead of center now. 🙂

It's up now: https://github.com/prusa3d/Prusa-Firmware/releases/tag/v3.1.3_RC1 . Downloaded the source zip from that page and it appears the offending line is still offending:

mbl.set_z(ix, iy, current_position[Z_AXIS] - offset_z); //store measured z values z_values[iy][ix] = z - offset_z;

Posted : 09/03/2018 4:19 pm
Tiago
(@tiago)
Reputable Member
Re: First layer issues? Look here!




3.1.3 RC1 says it only for MK2.5 Do the MK3 and MK2.5 use the same firmware?

Good catch. The MK3 branch is updated to 3.1.3 RC1 in the files, but they haven't posted it to the downloads page yet. I got confused as they have rearranged the download page columns and the MK3 download is on the far right instead of center now. 🙂

It's up now: https://github.com/prusa3d/Prusa-Firmware/releases/tag/v3.1.3_RC1 . Downloaded the source zip from that page and it appears the offending line is still offending:

mbl.set_z(ix, iy, current_position[Z_AXIS] - offset_z); //store measured z values z_values[iy][ix] = z - offset_z;

They will never accept a pull request without heavy test, a language translation can be ok to accept just by reading, but such changes like that need proper test or else they risk to damage beds if someone make mistakes.
This is also not the fix for the problem, just a workarround that work a bit better, the problem still there, at least for me, it haven't help that much, and with ABS i will print on air if not ajust z live, so i think we need to wait a far more time util they solve this problem

Posted : 09/03/2018 5:08 pm
stahlfabrik
(@stahlfabrik)
Honorable Member
Re: First layer issues? Look here!

No. Your temperature calibration is wrong then. You cannot fix it yourself.

I will post later how I did it.

For the MK3 my fix works. For MK2S I really do not know. Maybe the code works perfectly for that machine. Then my fix has to be moved elsewhere.

So yes there might be more work to do for compatibility with both machines.

For MK3 I am very confident.

Posted : 09/03/2018 6:20 pm
AbeFM
(@abefm)
Member
Re: First layer issues? Look here!

Where is the "fixed" firmware?

It's based on which (i.e. no LA, no 3.1.3 fixes)?

I would certainly like to try it, with the LA back on. I'm certainly frustrated that my adhesion issues with PEI and a bone-stock I3 are worse than my old microswitch-and-glass Geeetech.

Thanks for catching/testing this!

I maintain an informal list of San Diego, CA 3D printing enthusiasts. PM me for details. If you include a contact email and I can add you to the informal mailing list.

Posted : 09/03/2018 9:17 pm
stahlfabrik
(@stahlfabrik)
Honorable Member
Re: First layer issues? Look here!

here is the write up that I teased about.

https://shop.prusa3d.com/forum/original-prusa-i3-mk3-f30/1st-layer-problems-in-depth-look-at-software-pinda-t14815.html

Posted : 09/03/2018 9:53 pm
Brigandier
(@brigandier)
Reputable Member
Topic starter answered:
Re: First layer issues? Look here!


here is the write up that I teased about.

https://shop.prusa3d.com/forum/original-prusa-i3-mk3-f30/1st-layer-problems-in-depth-look-at-software-pinda-t14815.html

Got it open in a tab, reading soon. 🙂

Decided to eat my own dog food and flash the latest firmware I posted here a night or two ago, got some of the best prints I have gotten off the MK3 thus far (sans having to print slowly on external perimeters, oh well). Bought some nice gold PLA that matched the steel sheet near perfectly, decided it would look better than the orange LCD panel printed at Prusa's nasty fast speed. The top surfaces and corner sharpness really highlight how amazing Linear Advance is:

First up, good first layer with this fix. 😉

Next up, look at how sharp the corners are with zero bulging, and look at how uniformly smooth that top layer is. I suspect this is combined effect of LA and the extrusion bug fixes:

More glamour shots with great top layer smoothness and sharp corners:

Finished and assembled:

Close up of bottom layer:

My MK3 Parts: [Bowden] [New Shoes] [TPU Micro Springs]

Posted : 09/03/2018 11:30 pm
Page 4 / 6
Share: