Auto Bed Leveling (ABL) Sensor Comparison
One of the best upgrades you can add to a 3D printer to improve both performance and ease of use is an auto bed leveling sensor. Although they technically do not level the bed, they create a topological map of your bed and adjust the Z position of the nozzle to follow imperfections of your build surface for a more consistent first layer.
The first upgrade we do to all of our printers is solid mount the bed (remove the bed springs) and add an auto bed leveling sensor. This allows us to level the bed once and never have to worry about it again. When you have a print farm, the last thing you want to have to worry about is adjusting bed leveling knobs on a large number of printers.
Disclaimer: We currently sell a physical, hall effect type probe. We want to be upfront that we have not skewed our opinions based on this. The fact is, we tested each type of sensor extensively in our print farm and came to the conclusion that for our needs this was the most well-rounded option. Because of this, we decided to partner with a sensor manufacture to develop the 3DM Touch. We could have just as easily went with a capacitive or inductive style sensor which typically is much cheaper to manufacture.
Sensor Types Available:
If you Google “3D Printer Bed Leveling Sensors” you will find an endless number of manufacturers selling sensors of all shapes and sizes. To help simplify things we will break it down by sensor type instead of sensor brand. If you can understand the pros and cons of each of the technologies available it will help you determine the best one for your application.
Capacitive Sensor (ex. EZABL):
A non-contact sensor that can sense both metallic and non-metallic surfaces. This sensor works by monitoring the capacitance (how much energy the onboard capacitor can hold) which changes when an object is placed near its sensing face. It can detect any surface that has a die electric constant greater than air which should cover every build surface you will use (including glass). Another advantage of this sensor type is that it can probe extremely fast, saving you time during each print. Below is a basic diagram of how the sensor works.
One drawback is that while it can sense almost any surface, the distance at which it reads it will change depending on build surface type, temperature, and humidity. This is no problem if you only have one build surface and do not use an enclosure. If you plan to switch between different surfaces (ex. garolite, glass, PEI, polypropylene, etc.) like us, you will find yourself having to continually tweak your z-offset. Additionally, if you print in an enclosure where ambient temperatures vary, you will again need to adjust your z-offset.
Inductive Sensor (ex. P.I.N.D.A.):
This is another non-contact sensor, however, this type can only detect metallic surfaces. This means when the probe is analyzing a PEI spring sheet, it is actually mapping the spring steel and not the PEI surface you are printing on. This is usually not an issue, however, as the two surfaces should be relatively parallel with each other. If you are someone who likes to print on glass, garolite, or polypropylene—you may want to look for another option as these materials will not be detected. Although these sensors look very similar to a capacitive sensor, they work based on a completely different principle. They use the electrical principal called inductance. In a nutshell, an inductor coil in the sensor creates a magnetic field that changes when a metallic object is within its sensing distance. Non-metallic surfaces do not affect the magnetic field which is why the surface must be metallic.
Similar to the capacitive sensors, their readings are affected by temperature and humidity changes, so be prepared to make adjustments if you expect to see large temperature shifts. Some printer manufacturers, like Prusa, have done their best to use onboard sensors to automatically adjust for ambient temperature changes but in our experience it’s hit-and-miss. We have MK3S printers where we calibrated the z-offset at room temperature, put them in a heated enclosure, and found that the z-offset was then completely incorrect. If you are not printing in an enclosure, the temperature compensation seems to work fairly well for minor changes in room temperature, which is what we think it was really intended to do.
Physical-Hall Effect Sensor (ex. 3DM Touch):
This is the only sensor on the list that actually makes contact with the bed as a means of detection. Because of this, it is unaffected by temperature and humidity changes like the other two contenders. It functions by using a plastic plunger (pin) and a hall effect sensor to detect your build plate. As soon as the plastic pin makes contact with the bed, it is retracted and registered by the hall effect sensor. Also, there is a built-in solenoid that allows you to extend and retract the probe via g-code which gives you plenty of clearance while printing. Below is a diagram of how a basic hall effect sensor works. Please note, the plastic plunger has a magnet embedded towards the top which is what the hall effect sensor is actually monitoring.
The main drawback of this sensor type is that it is slower compared to the other non-contact types. While there is only a few seconds difference per probe point, this can add up if you are doing a 7X7 (49 points) or greater grid pattern. The good news is that in Marlin 2.0 and newer, there is a feature called HSMode which allows you to probe nearly as fast as the others. The next issue is that it contains moving parts so, in theory, there is more that can break on the sensor. Luckily, the mechanism is extremely simple so there is almost no chance of it actually failing. The new plastic probe tips are also made to bend so if you accidentally crash it, you can simply bend the plunger back in place and get back to printing.
All testing was done with three brand new Prusa MK3S printers modified to accommodate the different sensor types.
We wanted to test how well the three probes could continually reproduce a 0.24mm, single layer print without any other variables changing such as temperature or build surface types. We tested this by using micrometers to adjust the z-offsets until all three printers were printing a true 0.24mm thick first layer. Once dialed in, we printed the test print 5 more times and measured the thickness in several spots using the same micrometers. All of the probes were able to reproduce a 0.24mm thick first layer consistently without any deviation. If the inductive and capacitive probes are more accurate, they did not translate into any real-world results. Our conclusion: in a controlled environment all three options will provide accurate readings.
Winner: Tie (Inductive/Capacitive/Physical-Hall Effect)
Our expectations: set the z-offset once and continually get a perfect first layer, regardless of the build surface type or ambient temperature. When you have several printers, you don't want to have to think about tweaking z-offsets every time your printer setup changes. This saves both time and frustration which, in our opinion, is the most important function of a bed leveling probe.
On paper, the two non-contact sensors should be the most accurate. Their lack of moving parts and the sensing method give them a leg up (and was shown by Thomas Sanladerer on his YouTube comparison). However, when you throw in real-world variables such as different build surfaces and drastic temperature changes present in an enclosure, the Physical-Hall Effect type sensor comes out head and shoulders above the rest.
Similar to the previous accuracy test, we set our first layer height to 0.24mm and then printed an object that was only one layer thick. We then removed the print and measured it with micrometers and made adjustments until it was exactly 0.24mm. Once the z-offset was adjusted to produce a true 0.24mm first layer, we then heated the enclosure to 35°C and hit print without adjusting anything. The 3DM Touch was the only one that was able to reproduce the same 0.24mm first layer while the other two failed to have a successful first layer. We did this 4 more times and the results were the same each time. The non-contact sensors just weren’t able to give the same z-offset reading with the large change in temperature.
The next test was about switching between build surfaces. We did not test the inductive sensor as it can only sense metal and knew it wouldn’t be able to handle this task. For this test, we again adjusted the z-offsets on the two printers to produce a 0.24mm first layer using a spring steel build surface. After this, we then swapped out the spring steel build surface for a polypropylene plate. As expected, the 3DM Touch printed the same 0.24mm first layer while the capacitive (EZABL) sensor gave us a first layer height of 0.17mm which is an error of 30%. This is because even though the capacitive sensor can detect polypropylene, it detects it at a different distance than spring steel. If you plan on using different build surfaces, we recommend staying away from the inductive sensors as they will fail to detect everything except metal.
Winner: Physical-Hall Effect Sensor
Both non-contact sensors are considerably faster than the physical probe. They don’t have any moving parts so no time is required to cycle the probe up and down. As mentioned earlier, Marlin 2.0 has introduced HSMode for the 3DM Touch which has helped increase probing speed, however, even then it will still be ~10% slower to probe the same number of points. If probing speed is extremely critical in your application, an inductive or capacitive sensor might be the best choice for you.
Winner: Tie (Inductive/Capacitive)
After testing each sensor type (multiple of each, over several months) none of them showed any signs of having issues with durability. Some argue that the 3DM Touch sensor has moving parts so it, in theory, could wear out. We are confident that your machine would need to be replaced long before the plunger mechanism wears out. The old BLTouch probes had metal tips that would get damaged if you crashed your machine but the newer ones all use “breakaway” plastic tips that deflect if crashed.
Winner: Tie (Inductive/Capacitive/Physical-Hall Effect)
Because the 3DM Touch is a lot more complex in its design, it tends to also be the most expensive. There are also very few manufacturers who produce them which is also a driving factor of cost. Quality capacitive and inductive sensors, on the other hand, can be found for less than $10. This is approximately a third of the cost of 3DM Touch-type probes. If you are on a tight budget and still want to add a bed leveling probe to your machine, both the capacitive and inductive sensors will be a great choice.
Winner: Tie (Inductive/Capacitive)
All three sensors seem to do a more than adequate job as long as you don’t plan on switching build surfaces or changing ambient temperatures drastically. Casual users who are on a tight budget will likely be extremely happy with either a conductive or inductive type sensor.
If you are a 3D printing “super-user” and have the extra budget, we would recommend the 3DM Touch as it requires the least amount of calibration in the long run. It handles everything from build surface changes to large temperature variation without ever needing to do any additional calibration.
Each user is going to have a different set of requirements for their printing needs. For us, the 3DM Touch made the most sense; for others, it may be a capacitive or inductive sensor. We hope that this article helps clear up any uncertainty about the various options available and will help you make an informed decision for your own needs. If you have any questions about areas we did not cover, please feel free to contact us and we will do our best to provide answers.