So bumping into issues with inductive proximity sensors.
Started with Fotek PS-05N (pictured) on a 1mm steel bed, works well but the bed bends.
Moved to a 2.5mm Aluminium bed which is the same mass as the steel one, it’s now flat and works well, but the sensing distance of the “5mm” sensor is now ~1mm and if any corners start to turn up then the sensor gets broken off.
I got a PS-15N (15mm), which is a significantly larger package but has a good sensing distance to Aluminium. Though having been doing some ABS printing recently I’ve found the sensing distance can vary by about 1mm as the sensor itself is heated by the nearby hot bed.
Does anyone know how well the Printrbot sensor works with an Aluminium bed and if it changes as it is heated by radiation/convection from that bed?
@Thomas_Sanladerer
I dont know this Sensor, but some of the Round ones got the Option to set the Sensing Distance with a little Screw on the Top.
The Fotek sensors have no adjustment, but that’s not going to solve this issue. I need repeatability and a good sensing range to Aluminium.
I’ve got an 18x8 round that works great, picks up aluminum foil through glass at about 4mm
Fwiw, a capacitive sensor may be a better solution
I think I’m going to play with the PS-15N and the PS-05N a bit, measure the repeatability through temperature ranges, and if the PS-05N is repeatable (I suspect the 15 is not), then I’ll look at a more compliant mounting option, maybe a spring.
Looking at @Thomas_Sanladerer 's video, he’s running a sensor even closer to the bed than I am.
I know my nozzle ends up about 0.2mm closer to the bed if i measure before the bed is heated up, but i’d guess that’s due to the bed plarform and other stuff expanding and not due to the sensor getting messed up. These are industrial parts, after all, which are rated for pretty rough environments. Even the Chinese ones.
I wrote a python script to measure the repeatability of the sensor, which at each temperature point was very good, tiny range of readings.
Between a bed temp of 20 and 100, the bed moves down by 0.14mm, I knew this already, it’s due to how the bed is mounted on top of the heater and is part of the reason I have a sensor.
But, if I leave the X carriage and sensor away from the stable and hot bed for 5 minutes vs leaving it over the stable and hot bed for 5 minutes, the sensor’s sensing point moves down by 0.63mm.
So this is either the sensor itself moving as something in the X carriage warps from the bed heat, or the sensor changes. I could check which using the nozzle to bed distance, but the nozzle might also be moving with warp, so that’s not going to tell me for sure either way.
The sensor isn’t very solidly mounted, I’ll try and mount that more solidly (I think I know where), then repeat the test. Thanks @Thomas_Sanladerer , I hadn’t considered that it might be the frame.
I have never noticed heat affecting the inductive sensors I use in industrial automation however I am not looking for the high accuracy repeatability we are after in 3D printer calibration.
One thing I have found however which you have observed also is that the heater spreader plate will warp quite a lot when heating. However I also found that this is not necessarily an even amount across the plate. Since this is what you are sensing and not the glass plate on top of it this will produce errors. Another source of inaccuracy could be the plastic of the sensor and mount expanding during heating.
Do you have a dial gauge you can monitor this with when the printer is heating?
There’s no glass, as I said I use a 2.5mm Aluminium bed mounted on top of the sensor, it is this I’m sensing. It does not warp, the point grid shows that quite clearly, I’m dealing with a height change.
The sensor can barely see the copper in the heater through the soldermask, certainly not through the mirror tile I used to use.
Sorry, though you went for a longer sensing range to enable a glass bed and see alloy underneath. The copper traces on the heater do not have enough mass to sense being only around 35um thick.
See this vid https://youtu.be/Ntz-r8ECLTw shows that your sensor should be pretty accurate and this one is sensing through the glass to the heater spreader plate. The sensor also is stated to handle 80ºc in the specs however there is no mention of repeatability, just on/off hysterisis.
-How slowly are you traveling down in the calibration cycle? This will have a big influence on accuracy. Scan time can certainly make a difference as you are dealing with the sensor triggering, the transistor closing then sending an electrical signal back to your board that processes the signal and stops the steppers. There will be a delay here. So question is if you slow the printer right down during calibrariton does the repeatability improve.
-Is the accuracy repeatable if the bed is cold?
Again a dial indicator is a must for diagnosing these problems if you have one.
I zip tied the PS-15N under the middle of the carriage next to the E3D, a lot of the movement is gone. Now printing a carriage with built in mounts at that location.
Thanks @Thomas_Sanladerer , it looks like it was the carriage warping at the ABS temps, PLA prints were always fine, I do those at ~40C now.
I’ve got a new X carriage that very solidly mounts the probe now, but it’s not a lot better.
The Fotek PS-15N (15mm sense distance) DEFINATELY does drift. The PS-05N seemed fine.