Thread Rating:
  • 2 Vote(s) - 5 Average
Z Axis Height Calibration
#41
(11-12-2019, 12:42 AM)Differential_faw Wrote: So I am sooo confused. There seems to be mutliple ndifferent processes to fix things on a mechanical and a software level. How do you tell which problem.younhabe if any at all and which process to use. I also don't understand any of the processes. People keep saying home is not 0 but z is 0 or vice versa. Then folks are saying need to hit the level plate button then move up a couple mm??? This would seem to leave plate above fep as far as I am mis understanding. Can someone please help or am I too beyond lost to help?

First thing first, print off a calibration cube to see what the offset is.
Reply
#42
After going through the information and possible solutions here and in https://blog.honzamrazek.cz/ I finally got to an acceptable calibration. It is not perfect, and cannot be as demonstrated by putting the 5mm cube in the center and at each corner of the printing area: the center one came out 5.06mm high, and each cornered one was 4.93mm. This shows that the FEP flexibility is also involved as a variable, makingan absolute correction impossible. As in any process, the matter is to determine a tolerance, in this case +- 0.05 to 0.10mm seems realistic and should work fine for most applications. A smaller tolerance can be achieved working in a reduced central area only.

Following the many different recommendations, I went through the following procedures.

(1) as described in hhttps://blog.honzamrazek.cz/, I got the spring washer in the Z motor replaced by an equivalent size washer and had the "silencer" flexibility cancelled (in my case, by epoxy gluing a 3mm thick acrylic spacer around the rubber, so as not to take the part out and alter the motor position).
(2) detecting the FEP is not tightly pressed against the LCD, I leveled the build plate using a thick 0.23mm sheet of paper instead of a regular 0.10mm bond. That took care of it, and avoided having to rehabilitate the set zero option in Gcode and manually adjusting it every time. Instead, the original and easy Elegoo setting Home as Zero works fine.
(3) following a comment on drilling the build plate, and inspired by the already drilled one on the Kalend D100 (compatible with the Mars, by the way, available on Aliexpress), I drilled 6 holes at ø3mm and 20mm apart, centered along the width, the purpose being to help the resin flow. As the holes are precisely aligned along the central line on the ChihTuBox graphic, at the intersection of 1, 3 and 5cm on left and right from the center, I can place the models either avoiding those spots or helping the resin to flow out of trapped areas.

I also feel there is an issue with some layers sticking to the FEP and pulling it up enough to not be released and then coming back down, eventually being released in a later pull up and down but then leaving a couple of layers missing, or taking the piece apart or pulling it off the build plate and stuck to the FEP. So the next thing will be to print after altering the Gcode to increase the total demoulding raise to 8mm instead of 6 (M8070 Z8), and the slow raise to 5mm instead of 3 (M8070 S5). Hopefully this may resolve a problem I have with a particular model design which started failing on the Photon after updating the firmware and Gcode. Apparently it had reduced this raising distance to shorten raise and fall time from 6.5 to 4.5 deconds, and all my prints started to separate in pieces where the exposed surface increased. Previous to the upgrade, that model had printed without issue. This same problem also happens on the Mars, and I caught it during a visual inspection being stuck already at that same point as the Photon, but casually raising the plate for inspecting forced it to release with a strong popping noise. I'll later mention if that change was positive and worth the printing time increase.
Reply
#43
Actually what's causing the slight variation in z height across different locations on the build plate is the build plate itself. When holding the build plate upside down, the center is slightly lower (can be verified using the straight edge of calipers).

Search the Facebook group about honing the build plate. It takes about 15 minutes or so to do. Doing this will result in more consistent z heights across the entire build plate.
Reply
#44
(10-21-2019, 09:45 AM)Blackbird2016 Wrote:
(10-21-2019, 07:48 AM)mckeemorology Wrote: Thank you so much for this. Process makes total sense and worked perfectly. One of my mars was losing 1.4mm, the other 1.0mm, and both print *perfectly* now. I'm so glad we finally have a good solution that doesn't involve "just print on supports" or "get a hex wrench..."  Big Grin
Glad to be of help!
And still.... If you get in contact with Elegoo they provide a modified motor in a very friendly manner.

Can I ask what you received? They sent me a replacement z-axis assembly, but it looks identical to the existing one, down to the part number. I have fitted it and am running a test print, but it feels like they  may have sent me an in modified replacement. Is there any visual difference?
Reply
#45
(11-14-2019, 10:12 AM)SpaceDandy Wrote:
(10-21-2019, 09:45 AM)Blackbird2016 Wrote:
(10-21-2019, 07:48 AM)mckeemorology Wrote: Thank you so much for this. Process makes total sense and worked perfectly. One of my mars was losing 1.4mm, the other 1.0mm, and both print *perfectly* now. I'm so glad we finally have a good solution that doesn't involve "just print on supports" or "get a hex wrench..."  Big Grin
Glad to be of help!
And still.... If you get in contact with Elegoo they provide a modified motor in a very friendly manner.

Can I ask what you received? They sent me a replacement z-axis assembly, but it looks identical to the existing one, down to the part number. I have fitted it and am running a test print, but it feels like they  may have sent me an in modified replacement. Is there any visual difference?
I didn´t get my motor, yet. Seems to have stalled in German customs for some reason.
As far as I know you get mainly a new motor which internally features a different harder spring-shim for less play/air.
Makes no difference for me, when it eventually arrives I will mount it.
And still I am considering to do the leveling before every print, even with the new one since it makes the prints more accurate und just needs a few seconds for every several-hour print...
Reply
#46
Guys, the replacement stepper is the same stepper with an additional spring washer installed, that's it.

I've talked to elegoo directly and this was confirmed by them.
Reply
#47
(11-14-2019, 01:52 PM)admin Wrote: Guys, the replacement stepper is the same stepper with an additional spring washer installed, that's it.

I've talked to elegoo directly and this was confirmed by them.

Thanks for the confirmation.

The replacement motor has greatly reduced (but not entirely eliminated) the issue for me -- down to 0.3mm from ~1.0mm.  I'm happy with that, I think.
Reply
#48
Yep, 0.3mm to 0.4mm seems to be the norm with the right stepper. Now redo the z calibration and you'll be able to get it to 0 Smile
Reply
#49
I want to thank the admin as I used instructions for adjusting the z-axis and it made a tremendous difference as can be seen in the pic.  While not a perfect 20mm cube yet, I am working on it.  It went from 18.06 to 19.87 by just using the instructions in this thread.

Snassek

[Image: mhHd91Z.jpg?1]
Reply
#50
Either I am over simplifying things, or this thread is being way over complicated. I will test this afternoon to find out. From what I see, the machine has a spring in the z axis to prevent damage to the LCD screen should a part become dislodged and get mashed under the build plate. When pressing the autolevel button, the machine uses the auto limit optical switch to move to the zero position according to the optical sensor and then you lock the build plate. My understanding is the problem is that depending on the mounting of the optical sensor, the z axis can be off by a small amount and the spring already slightly compressed resulting a a few initial layers all printing in the same space until the spring tension runs out.

This thread details the solution as unlocking the set z=o parameter and then manually adjusting the variance each time you level the bed and resetting the z(home position). Why would you want to do this each time when you can just set this statically in steps in the configuration?

Example: I already printed a 20mm cube on my printer because i noticed detail parts were off a little. My cube is 19mm. you can see a bit of squish on the bottom too. Instead of doing this every time can you not just put this directly into the gcode configuration via parameter M8083 and M8084? This is what you are configuring already, but doing it each time, it is not static.

and plug that into parameter M8084 ( as negative(-) to move the build platform up 1 mm at print start):

;; ¡¾Z limit point and zero limit difference, the value is usually positive value¡¿

M8084 Z-1  ; If you want to set the offset in the configuration, please remove the semicolon in front of the line command, and you can also automatically configure the offset by setting Z              ;to zero on the int


I'll test another cube today, but if i am understanding these parameters correctly, you machine is now calibrated and all you have to do is level like normal each time and print. Am I missing something?

EDIT: This is the correct way to change the calibration and works PERFECTLY---however I have the sign on the z access wrong. The printer prints in -Z access so the correct Z access offset should be -1 for 1mm. I have adjusted it above to avoid confusion. Looks like the setting is in fact in mm. Just tested and confirmed by pressing print and measuring the starting point with a caliper with the resin vat set aside. Values updated above. In summary, the leveling routine always uses the 0 position read from the sensor, but parameter M8084 tells it to start the print job with the build plate moved  relative to the mechanical 0 position. (-) is up!!, This gives you back your layers that were missing (squashed) at the base.. This is similar to the method described above, except we are just putting the setting statically into the configuration using parameter M8084.

Proof is in the pictures. All I did was measure the first cube, put the difference (-.87) into the machine definition(M8084), install the definition and reprint the cube.

[Image: jVoNASG.jpg]
Reply


Forum Jump:


Users browsing this thread: 1 Guest(s)