Different layer thickness problem

[VladimirK 1]

I'm using Cura 4.6.1, Duet Wi-Fi, G-Code flavor RepRap, E3D Titan Aero direct extruder with Volcano,

M572 D0 S0.07 ; Pressure advance

My problems are on photos:

first one Nylon 0.4mm nozzle "Calibration cube"

 

Second is PLA 0.8mm nozzle, "Line Width" set to 0.8

 

I checked and those horizontal lines not at the same heights.

 

And it seems other models have no such problems (Polycarbonate, rotated 180 degrees)

 

Is it something wrong with Cura?

[burtoogle 513]

Could you please provide the project files for those calibration cubes? Do File->Save and attach the 3mf file(s) to this thread. thanks.

[VladimirK 1]

There 3 files attached: original stl I used, 3mf saved from Cura and gcode file used (first picture).

20mm_calibration_cube.stl A4_20mm_calibration_cube nylon.gcode A4_20mm_calibration_cube.3mf

[burtoogle 513]

Thanks for the files, could you please also provide the gcode and 3mf files for the PLA cube, thanks?

[VladimirK 1]

I did PLA at 0.8 several weeks ago, was disappointed and changed back to 0.4 nozzle. But PLA settings was the same except line width settings, I opened Cura with stl, changed line width to 0.8, saved 3mf and made yet another gcode. It shows the same 17 minutes expected vs 23 minutes with 0.4, however not tested to print it right now yet again. If necessary I will test tomorrow. My 3D printer mechanic is "Anycubic 4Max Pro" (bed moving only up and down), 24V power supply, 400 steps per rotation, 256 microstepping, so most likely it is not related to mechanic.

A4_20mm_calibration_cube 0.8 PLA.3mf A4_20mm_calibration_cube 0.8 PLA.gcode

[burtoogle 513]

Hi, I wonder if it is related to the heated bed temperature. Could you please try a PLA print with the bed much cooler, i.e. 30 rather than 70 and see if the spacing of that pattern changes. If I am right, I would expect the problem layers to be further apart compared to the part printed with a bed temp of 70. Thanks.

[VladimirK 1]

I will try today with 0.8, different bad temp etc, but if you think this happens due to bed power consumption - it is powered from 220V through solid state relay.

[burtoogle 513]

41 minutes ago, VladimirK said:

I will try today with 0.8, different bad temp etc, but if you think this happens due to bed power consumption - it is powered from 220V through solid state relay.

 

Oh, right, well try it anyway, you never know!

[VladimirK 1]

Really it is related to hotbed, it is from 30C (failed, unstuck from Ultrabase), 50C and 70.

 

Only I cannot figure how exactly, hotbed powered through SSR MGR-1 D4825 25A DC Control AC DC-AC relay, it works even when I tried to limit control current with 1-5.6 kOm resistor, but print had the same issues. Also I checked power supply on board with multimeter - it is 24.0V when hotbed turned on or not.

Edited July 27, 2020 by VladimirK

[burtoogle 513]

Ha! I wonder if it is interference from the AC hotbed causing the hotend temperature sensing to go awry? i.e. when the hotbed heater is on it causes the hotend temperature to change?

[VladimirK 1]

Is it happens through thermistor? I have PT100 also (except it shows room temperature several degrees higher), can it fix it? Also when hotend further from hotbed difference looks the same, it doesn't diminish with height. Maybe I will try with PT100.

[DivingDuck 104]

The problem shouldn't come from the SSR. 220 V Silicone heating pad? This looks more like the temperature is over- and undershooting too much. How is your setup for temperature control?

[VladimirK 1]

Yes, silicone heated 600W. So far I tried to change hotend thermistor from Semitec 104GT-2/104NT-4 (96 kOm at room temperature) to PT100 (110 Om) due to it less vulnerable to interference. Results are the same (first thermistor, second PT100)

 

In settings for bed I have only

; Heaters
M305 P0 T100000 B4138 R4700                        ; set thermistor + ADC parameters for heater 0
M143 H0 S120                                       ; set temperature limit for heater 0 to 120C

 

Another options was (if it is related to hotend temperature to try copper plated Vocano due to higher inertia and twisted pair as thermistor cable). Not tried yet.

[DivingDuck 104]

There is even more to do for controlling the bed heater. Bang-Bang-Control or PID-Control and setting PWM frequency for the SSR.

 

https://duet3d.dozuki.com/Wiki/Choosing_a_bed_heater#Section_Bed_heater_driven_using_a_Solid_State_Relay

[VladimirK 1]

Yes, it was hotbed settings. First I checked

 

M307 H0
Heater 0 model: gain 90.0, time constant 700.0, dead time 10.0, max PWM 1.00, calibration voltage 0.0, mode bang-bang, inverted no, frequency default

 

It was default settings. I tried to use PID before but turned it off due to errors with copper plated Volcano (heating it takes too long).

 

after tuning:

 

M307 H0
Heater 0 model: gain 237.8, time constant 782.6, dead time 13.4, max PWM 1.00, calibration voltage 24.3, mode PID, inverted no, frequency default
Computed PID parameters for setpoint change: P43.8, I0.427, D411.2
Computed PID parameters for load change: P4

 

and saved in config-override.g as

M307 H0 A282.2 C416.3 D1.6 S1.00 V24.3 B0

 

That fixed problem, just as good as printed on cold bed. Further tuning PT100 to show proper temperature and tuning H1 show no significant difference.

I wonder why it happened at all, some precompensation in Duet?

[burtoogle 513]

Glad you have solved the issue. Even though I suggested that it was the heated bed that was the cause of the problem, I don't really understand the mechanism of failure here. What did the gross over/under shooting of the bed temp actually do?

[VladimirK 1]

By measuring with caliper it is both, slightly more plus slightly less. In extruder. Related to bed. No idea.

Edited July 28, 2020 by VladimirK

[DivingDuck 104]

33 minutes ago, burtoogle said:

I don't really understand the mechanism of failure here.

That's a simple story: There are two points to consider . The first one is the fast heating capacity of silicon plad's. For fast heating  you calculate between 0.33 to 0.9 W/cm², optimal is ca. 0.5 W/cm². That is for an 300mm x 300 mm heater plate 450 W and this will heat up a plate within 3 to 4 min from 20 to 80 °C. While a usual 12V/24V heater plate need much longer, the controller have more time to control the heat up process and that is what the usual firmware setting is. In that time the silicone heater will easily overshot by 20 to 40 °C what will cause a significant expansion for the heat  bed. The second point is how printer manufacturer fasten the heat bed. They usually fix the bed on all 4 corners with an spring loaded screw. Well, this ask for even more deformation as this is a stupid design with problematic impact (as you can see). 😉

 

@VladimirK, it is ok to have a slight overshooting, as this is how adaptive control works in combination with thermal inertness of the aluminum plate . Check if you can increase the PWM slightly. (I use 20 Hz for my SSR's with 600 W pads)

[burtoogle 513]

So you're saying that the layer variation is caused by the print bed thickness varying greatly due to the wide bed temperature swing?

[DivingDuck 104]

Correct but not only thickness, it also start to wrap. You can easy check this for your printer. Level your bed for PLA (maybe 60 °C) and then heat up the bed to 80°C and check the distance between bed and nozzle. That's what happen.

Edited July 28, 2020 by DivingDuck

[burtoogle 513]

In this instance, it looks like the layers that are indented correspond to when the hot bed was active and the layers that are not indented correspond to when the hot bed was inactive. So the layer height (distance from nozzle to last layer) is greater when the bed is on and reduces as the bed/print cool down. Therefore it can't be expansion of the bed/print that is causing this effect, it has to be something else. As you say, the bed could warp. It's not clear what the mechanism is here.

[DivingDuck 104]

You maybe overlook the impact of bed mounting. When you fixing a aluminum plate on all 4 corners with a screw so that it can not expand/move evenly in all directions, what will happen? Usually it will put expansion forces to the screws and try to bend them, screws it self will put forces against the plate too because they can't move too. And not to forget that the thermal expansion of the heat bed isn't consistent everywhere too. In reality it will bend both, screw and bed. In the end this will always have a measurable effect on distance between nozzle and bed. This effect will be more noticeable as bigger a bed will be and even more if the plate is thin.

Edited July 29, 2020 by DivingDuck

[VladimirK 1]

It works with 20Hz. From what I understood it makes temperature changes smooth: instead of jumping 6 degrees up and down it changes slowly.

[DivingDuck 104]

Nice. Please make a photo from your calibration cube. I like to see how it looks now :)

[VladimirK 1]

After changes. Lines there from "X" on the left, bottom at the bottom of photo. From left to right pressure advance 0.07 (was great with 0.4 nozzle), 0.035, PA disabled. It seems for 0.8 nozzle I need negative pressure advance?