|
3D touch level sensor fix, turn it in a reliable and high accuracy sensor March 31, 2021 11:36AM |
a bed leveling sensor is one of key tools to achieve consistent printing efficiency and it is a must when first layers play an important part of overall part shape and dimension, this is even more the case with rather small parts and thin layer thickness (<100 um).
Among many sensors, I prefer the HAL effect sensor over all others (like BL touch / 3D touch) as it should be rather accurate (sd ~10um) and both temp. and hum. have a neglectable inf;uence. There is another big advantage of a HAL type sensor with probe pin but this is out of scope of this subject.
2 months ago I purchased a 3D touch v2 of Geeetech. Rather soon after the installation I was confronted with sensor failures randomly: pin did not deploy and red blinking led (at both 50% or 80% duty cycle).
I checked everything: connectors, wiring, firmware (delay time, and all other settings), set scew on back of sensor to position the probe pin, but nothing resulted in a steady flawless operation. I monitored the voltage and signal pin with a scope 24/24 to see if power fluctuations due to bed or extruder or any other source was influencing the 5V line or if the pwm signal was not steady or corrupted with noise or false due to software errors.
Nothing was helping, sometimes it worked for 3 days continously and suddenly the failure returned.
Taking a closer look at the sensor probe pin taking it out of the sensor housing (unscew the set screw on the back side), I discovered the issue with this sensor:
(see pic 1)
In the back side of the probe pin is a small cylindrical magnet, that triggers the HAL sensor when it passes going upward. This magnet was not fixed!, it moved about 0.5 mm when using a metal micro screwdrover on the back of the sensor pin, as the magnet is pulled towards the metal...
(see pic 2 and 3)
Due to this the magnet moved up too close to the metal set screw and the magnetic force between pin and set screw became too high to be released when Gcode gives a M401 command.
Most probaly this magnet was initially fixed with acrylic glue to hold it at the bottom position of the probe pin tube zone, but the repeating pin deploy and stow movement shocks detached it's fixed location.
I used a 0.6 mm cut of a 1.75mm flexible filament and positioned it in the gap under the set edge of the probe pin:
this flexible small disc will ensure the magnet fixed position and prevents the set screw to come too close to the magnet when it is scewed in the sensor housing and settles in the probe pin back end. As it is flexible and not compressed this will give a firm and reliable fixation of the magnet.
(see pic 4 and 5)
Before mounting it back in its position, I did a 200 times deploy/stow loop using pronterface and it worked flawless.
A sensor accuracy measurement (10 times 10 measurements) using M48 command with random repositioning in between sensing revealed an average standard deviation of 7.4 um (sd : 2 um). the measurement was done is HS sensing mode (high speed)
So, this 3D touch of GEEETECH is actually a very good HAL sensor but poorly assembled.
If you suffer from the same issue, check the probe tip !
Among many sensors, I prefer the HAL effect sensor over all others (like BL touch / 3D touch) as it should be rather accurate (sd ~10um) and both temp. and hum. have a neglectable inf;uence. There is another big advantage of a HAL type sensor with probe pin but this is out of scope of this subject.
2 months ago I purchased a 3D touch v2 of Geeetech. Rather soon after the installation I was confronted with sensor failures randomly: pin did not deploy and red blinking led (at both 50% or 80% duty cycle).
I checked everything: connectors, wiring, firmware (delay time, and all other settings), set scew on back of sensor to position the probe pin, but nothing resulted in a steady flawless operation. I monitored the voltage and signal pin with a scope 24/24 to see if power fluctuations due to bed or extruder or any other source was influencing the 5V line or if the pwm signal was not steady or corrupted with noise or false due to software errors.
Nothing was helping, sometimes it worked for 3 days continously and suddenly the failure returned.
Taking a closer look at the sensor probe pin taking it out of the sensor housing (unscew the set screw on the back side), I discovered the issue with this sensor:
(see pic 1)
In the back side of the probe pin is a small cylindrical magnet, that triggers the HAL sensor when it passes going upward. This magnet was not fixed!, it moved about 0.5 mm when using a metal micro screwdrover on the back of the sensor pin, as the magnet is pulled towards the metal...
(see pic 2 and 3)
Due to this the magnet moved up too close to the metal set screw and the magnetic force between pin and set screw became too high to be released when Gcode gives a M401 command.
Most probaly this magnet was initially fixed with acrylic glue to hold it at the bottom position of the probe pin tube zone, but the repeating pin deploy and stow movement shocks detached it's fixed location.
I used a 0.6 mm cut of a 1.75mm flexible filament and positioned it in the gap under the set edge of the probe pin:
this flexible small disc will ensure the magnet fixed position and prevents the set screw to come too close to the magnet when it is scewed in the sensor housing and settles in the probe pin back end. As it is flexible and not compressed this will give a firm and reliable fixation of the magnet.
(see pic 4 and 5)
Before mounting it back in its position, I did a 200 times deploy/stow loop using pronterface and it worked flawless.
A sensor accuracy measurement (10 times 10 measurements) using M48 command with random repositioning in between sensing revealed an average standard deviation of 7.4 um (sd : 2 um). the measurement was done is HS sensing mode (high speed)
So, this 3D touch of GEEETECH is actually a very good HAL sensor but poorly assembled.
If you suffer from the same issue, check the probe tip !
|
Re: 3D touch level sensor fix, turn it in a reliable and high accuracy sensor April 02, 2021 10:37AM |
Dear,
I just wanted to help out those who suffer too from a possible manufacturing glitch which seems to impact not only the GEETECH 3D touch....
Regarding the um range , my sd of 7.4um means that I can assure a first level thickness deviation is within + or minus 22 um with a certainty of 99,99 %
The why, I did not discuss here on this forum, but such weird remarks about my accuracy aim is a bit strange because you have no idea what I am doing.
your sensor seems piezo tech, but is completely unusable for what I am doing.
Have a great day!
Alex
I just wanted to help out those who suffer too from a possible manufacturing glitch which seems to impact not only the GEETECH 3D touch....
Regarding the um range , my sd of 7.4um means that I can assure a first level thickness deviation is within + or minus 22 um with a certainty of 99,99 %
The why, I did not discuss here on this forum, but such weird remarks about my accuracy aim is a bit strange because you have no idea what I am doing.
your sensor seems piezo tech, but is completely unusable for what I am doing.
Have a great day!
Alex
Sorry, only registered users may post in this forum.