Klipper devs are developing a function that let all ldc1612 eddy current probes have tap functionality, cartographer is also developing an open source version.
I cant wait for the cartographer tap to come out. Richard has done amazing things at such an affordable price. Makes me wonder why beacon is so utterly expensive. Mine's worked flawlessly for 6+ months.
Can confirm Beacon contact is great. The printer auto calibrates its z offset for each print so I can change build surfaces, nozzles and even whole hotends with having to set anything.
very simple. Just do an upgrade and follow the doc how to modify your config. Takes some time, you have f.x. also to upgrade tthe firmware of your beacon. But when you are finished its very impressive. Super new functionality for zero coins. Great customer service !!!
How does that work, though? As someone who changes nozzle sizes relatively often - and has observed a difference of over half a mm in nozzle length between different sizes and brands - anything relying on a statically-set nozzle Z offset just isn't going to work for me.
I have become so used to printing on non-metalic surfaces (mostly G10) that it always surprises me how many people completely skip over the fact that this is a big downside to any eddy current or inductive probe. I guess there is a lot of users who have "print surface" so synonymous with "metal sheet" in their mind this does not even occur to them. But for me an Euclid probe keeps delivering perfect first layer with reasonably fast probing sequence on any surface so well that for now I am not even considering swithing to any of these "game changing" new probes.
@@ygk3d One point I did not mention to keep it brief: how much resolution on the mesh you need is also influenced by the type of print surface. On my Ratrig V-Core 3 500 I tried even pretty high mesh resolutions with different surfaces. It seems that if you forget some tiny debris on the magnetic sticker and put the standard spring steel over it you get a hill in that place that can screw up your first layer if no probe point is near. But you put 1.5mm thick G10 with a metalic sticker on the bottom onto the same debris on the same magnet and the hill may have similar height but it is so gradual the neighboring probe points make the mesh relfect the real shape of the surface well enough to have no issues with the first layer. I use 10x10 when printing 450x450 mm or bigger and with the adaptive meshing algorithm this ends up being 3x3 mesh for any small print. With lifting the Euclid only 2mm after probing and 10k travel accelleration the probing is done in a few seconds. So this makes it hard for me to see how the faster scanning instead of probing would be a benefit even if it worked on G10 surface. Either it saves me few seconds on an hour long print or it gets me higher mesh resolution that does not reveal anything new. I can absolutely see a scanner like the Eddy being much better than for example BLTouch that is quite slow with the pin deployment needed twice or more at every probe point. Compared to that it will be both faster and have better resolution.
@@hebijirik great perspective. Thanks for sharing. The use of adaptive meshing definitely makes a big difference. Really interesting about spring steel versus G10. You find the G10 has good adhesion and release for most materials?
Based on the beacon contact demo, it seems the contact/tap implementation would work at enough distance through the G10, so perhaps for those surfaces you still do nozzle probing?
@@ygk3d The G10 is absolutely excelent for PETG which I print very often. Unlike PEI powder coated surfaces there seems to be zero deterioration after any amount of printing. I once scratched it accidentally so I tried an orbital sander on it, cleaned and washed all the dust from the sanding and it was like new again. Still sticks like crazy while hot and releases easily when cooled down. For PLA you need to degrease and clean it well, even fingerprints can decrease the adhesion too much. I would say it it slightly less bed adhesion for PLA than the PEI surfaces. But still very usable. When the plate cools down the PLA prints just completely release on their own. ASA/ABS is adhering a bit better than PLA but not as much as PETG. I decided to try the Vision Miner Nanopolymer adhesive on it and it works great too. It removes the PLA sensitivity to having it perfectly clean and freshly degreased. Prints other than PETG still completely detach once cold and the surface is still smooth. I like the low cost (I got 4 500x500 for less than the cost of just 1 PEI sheet), the ability to easily fix up slight damage and the option to buy different thickness based on how stiff I need it for the size. I use 1.5mm on my 500mm V-Core but it felt too stiff for smaller printers. I tried 1mm thick for a Prusa MK4 and a 200mm V-Core and they work nicely. Disadvantages are: the glass smooth bottom side of the print might not suit every part, so I still use powder coated PEI when I need that look. If your printer has magnets for metalic beds you can put a metalic sticker (used to make a magnet-holding whiteboard on any surface) but then you only get to use one side of the G10. Or you have to mess around with some other way to clamp it to your bed. Depending on where you are buying it might be easy or not - I emailed some local company that makes custom insulation for electric instalations and they sell it in 1050x1025 pieces in any thickness and they can cut it to smaller pieces of dimensions I ask for. So I got it quarted for the 500 and I got 16 pcs for MK4 🙂.
P.S., eddy current distance sensors are not linear. I work for a sensors company, and temp stability and linearity are our biggest woes in all forms of capacitive and inductive sensing (eddy current sensing counts as inductive).
Nice move from BigTree, glad to see they have or are working on the three most common ways for talking to the rest of the printers, and that price point is very tempting. Will keep this in mind on my Ender5 to Mercury one upgrade I am working very slowly on.
We went from inductive probes to klicky-probes because PEI coating on magnetic build-plates is uneven. Eddy probe does not take into account non-conductive media (PEI). Lidar is the only way.
We went to klicky probes from inductive not because PEI coating is uneven. It's because the z-offset has to change anytime you change build plate and that is annoying. A good PEI build plate has a very consistent thickness across its surface. beacon and probably eventually all eddy sensors are now also touch probes. The new beacon software allows it to detect when the nozzle hits the bed with around 50 grams of force and klipper is eventually going to support this natively for other eddy sensors. This means the probe can now compensate for the thickness of a build plate on top of the bed while also giving automatic z-offset calibration. This only needs to be done for the initial bed leveling and z-homing, as any quality build plate that goes on top of the bed will vary by less than a fraction of a single layer height in its thickness. The bed should be the only thing varying by an amount significant enough to effect initial layer quality. Thus for the actual slow part the, bed mesh, the eddy sensor is used as normal and thus saves time. Also the increased accuracy from the denser mesh means that the build plate can vary more in its thickness before it could sum to a great enough deviance to cause a problem with initial layers. Also LiDAR actually struggles with textured PEI plates, because it picks up the texture rather than the average thickness of the plate which is what you actually want. This is why bambu printers for instance turn LIDAR off by default on textured PEI.
Unless you are making changes to your printer constantly, the clicky or tap mesh works great if you, do a large number of points, and save meshes for a set of typical temperature ranges you print at. for example. I heat soak my printer with panels off at 60c bed temp for 30min, then do a bed mesh and save. in my print template for PLA, i load that bed mesh before print starts. I do that for PETG, ABS/ASA, etc. so I have a few profiles for different conditions. If you have a thick enough bed plate, and don't change the screws or other things on your printer, and do the same length of heat soak each time, the bed etc will expand and change basically the same every time so that the saved meshes work great. At least they have for me over the past couple of years.
These are really time saving sensors, not something needed if you don’t care about time. Printing is a significant portion of my income and 30 minutes heat soaking the bed is money down the drain, as is probing a complete mesh every print. Being able to build a mesh for every print in 10 seconds means I can hit print and don’t need to stand around for 5(or in your case 30) minutes to ensure the layer goes down clean. Any printers running 24/7 are not gonna be able to produce reliable first layers every single time without building a mesh or having an expensive and thick machined bed. These sensors don’t offer anything that can’t be achieved with other probes, except saving a lot of time.
Inductive has been around in CNC for a long time. I remember using inductive sensing on an Amada Lasmac LC1212 Apha II way back in 1998. It's great to see this being introduced for bed levelling.
Beacon contact is already here and it works great and I definitely have perfect first layers. Beacon auto calibrate did it all on its own and I’ve never had a probe be able to “auto calibrate“ on klipper using their probe calibrate module without having to baby step it afterwards. I’ve used bl touch, klicky, tap, and now beacon and beacon is the best probe hands down especially with contact.
@@Detroit_Playa I was shocked the first time I set up Contact with my Beacon and the first layer was just perfect immediately without doing anything. I've used every other type of sensor and probe out there. Inductive, capacitive, servos, clicky, Tap, etc. Beacon was worth the $70 I paid for it all day every day.
As someone who had id for more than 40 days, i can say that it's has it's problems. They have beem fixed as of now and i am happy with it and it's fast, but i use it with tap for the perfect z offset and without tap it's another story.
How is this effected by thermal changes? PINDA was ultimately phased out to a loadcell due to the bed probing not being accurate when heated even with a thermistor
Very nice video! Thanks for your work. I hope my printer will have one of those eddies at some point but have a long way to go. You are doing a great job in helping me on my journey.
When I went to order a Beacon they contacted me shortly after saying "Hey, we're planning an updated version in a couple months. We know you placed the order, but if you'd like to hold off until then we'll send you the H no extra charge." Definitely won me over as a customer for that. Especially now that the H has auto z-offset.
_"The probe emits a electromagnetic field which induce a small electrical current in the nearby conductive medium"_ *But this is exact principle any inductive proximity sensor/switch used in 3D printers operates upon.* But you saying like it's something absolutely new and different which is not true. The actual major difference between the generic inductive proximity switches is that the first one is permanently set to a specific triggering hysteresis loop that obviously requires vertical movements to trigger whereas the latter one continuously register changes in the electrical current yet measures distance to a build plate surface.
AHHHHH! You are so right it hurts! Induced electric currents ARE eddies! In electromagnetism, an eddy current (also called Foucault's current) is a loop of electric current induced within conductors by a changing magnetic field in the conductor according to Faraday's law of induction or by the relative motion of a conductor in a magnetic field. Eddy currents flow in closed loops within conductors, in planes perpendicular to the magnetic field. They can be induced within nearby stationary conductors by a time-varying magnetic field created by an AC electromagnet or transformer, for example, or by relative motion between a magnet and a nearby conductor. The magnitude of the current in a given loop is proportional to the strength of the magnetic field, the area of the loop, and the rate of change of flux, and inversely proportional to the resistivity of the material. When graphed, these circular currents within a piece of metal look vaguely like eddies or whirlpools in a liquid.
So what you are saying is that it is totally the same, except for the meaningful difference that makes it night and day different and better, that is new to 3D printing. Thanks for that info. It has also made me realize that cars are just the same as stage coaches... I mean, wheels on a carriage have been around since before ancient Rome.
Does the Eddy current sensor work with all types of bed materials? How about using a BLTouch-like setup with a smooth, metallic ball that moves slowly and continuously registers Z-axis displacement? It could perform an initial probing to adjust for any wear and use a random path to cover the entire area, so it doesn’t always drag over the same spots. Plus, the slow movement helps reduce noise. Or maybe a laser sensor, though it might struggle with glass or weird texture/colored beds due to refraction issues. What would be the most precise system?
i've been planning to integrate eddy into all my printers, along with revo, so thanks for this. one of the things i've never been fond of with btt devices in particular is having to flash firmware to everything before it can be used, but eddy & similar probes are something special & i see as more than worth the effort. something i feel the need to point out, though, is btt has been touting for quite a while now that ebb36 & ebb42 both have accelerometer function integrated, so you shouldn't need an external accelerometer anymore. i'll be testing that when i finish building my micron+, but until then i cannot confirm.
hi! I also started using Eddy! I have a problem when I print parts with a grid from Eddy. The height of the printed parts increases by about +0.6 mm. What could it be??? FADE is off! If without a grid, all the dimensions are perfect!
Possibly why you never get perfect first layers: maybe more of an issue for v wheels, but may apply to linear rails too. If there is any z deviation in height between the nozzle and where the probe ends up that matches where the nozzle tip was, then that will be a shifted value. So this is usually partially fixed with zero-y-offset mods on cheap v wheel printers, where the y value is the same for the probe and nozzle, yet it does nothing for the x axis. If there is a slight imperfection in one of the axes within the distance between your nozzle and probe, then it will not be accounted for in the mesh. The only way to fix this is to nozzle probe in one point then move the scanner or bed probe to the exact same point. Do this across the whole bed, it will then give you a map of the variance between actual nozzle distance and what the probe measures. Realistically you need to subtract or add this variance for the whole bed from the live bed mesh. Hopefully that makes sense, and it would be something possible to test.
Or just have the probe on the nozzle like the "clicky" solution. This is a step back from anything that has the probe directly under the nozzle. Another solution would be to implement skew compensation in Klipper for the Y axis, we already have it for X axis and it solves the offset problem as well.
Based on my research technically you can run I2C over SPI bus which we have on BTT SB2209 but no one have done it so far and it could be an interesting setup.
The new LDO nitehawk-36 usb toolhead board has a usb expansion port for a probe to plug into, so you dont have to run a separate cable. Unfortunately the nitehawk-sb (two piece stealthburner toolhead style) didn't get the port. Also the beacon D is a little cheaper then the beacon H at only $65 if you don't want the adxl345 accelerometer that's built in. Which you might not considering a few tool head boards already have them integrated like the nitehawk, btt SB2240, ebb36/42 or if you have a bed slinger and need the sensor to be movable from the hotend to the bed.
Well presented information. I think the main reason why we can't get perfect layer is that we keep measuring metal plate which is still powder coated with uneven layer of PEI.
Do you know if I Could I use the I2C pin that is available on manta 8P V2 board for using Eddy Coil version? or the unique option that i have for connection is usb connection mode?
Note, this doesn't work on glass, it doesn't measure the surface, it measures the metal below the surface, so any variation in the for example the PEI thickness will NOT be detected
Great video and thanks for all the info. But did i understand that right? After all that, Klicky and TAP are still far more accurate, all be it slower?
When i type in bedmesh_calibrate method=scan etc in defaults to manual probe. When i use BTT_bedmesh_calibrate it probes individual points on the bed. Help i have eddy coil
Did you have to convert the Troodon 2 to canbus to make this work? That part wasn't clear in the video. I have three Troodon 2 Pro units now and may consider this upgrade
I've been using the BD Sensor's from Mark at PandaPi 3D, its cheaper than these, more compact and the sensor is remote from the PCB so you can install extreme close to the nozzle, I even managed to install one on a Voron 0.2, I have a video showing it in action.
14:00 THAAAAAAAANK YOUUUUU!!!! OMG for real i've been issues since beginning of 3D Printing to set the z-hight correctly and was everytime wondering why my bed is "flat" but at such an offset of the 'middel'... I never found this out upon your tiny advise 🙏🙏🙏 god thank you 🙏🙏🙏
I bought the BTT Eddy USB and it doesn't work with any applications that require a 100c+ bed. Can someone share their experience using other contactless probes in high temp applications?
@@RasTona_ After a few months of struggling with eddy I decided to give up yesterday. It is way too temperamental and buggy. Just another rushed copycat product. Cartographer is about to release a new high temp version and judging by their great track record with other probes it should be the holy grail
I have a bambu the only upgrades i did was the obsidian or DLC coated nozzle. I know DLC used a lot in gun industries so i figured it be good and it's great nozzle. Is this device domething you cash only use on voron style printers or could it make your bambu better? Thanks
Would love to see your take on a CAN BUS conversion walk-through/ install! After finalizing wiring my 350mm 2.4 i learned about how much nicer a CAN BUS setup would be so it's something I've been wanting to do since mid build 😅
Is there a reason for this that makes a lot of sense? Unless if its price is very low or speed is insanely high, probes, switches, and induction already have a lot more accuracy than what is needed. If you need more accuracy, you should be investing more in a flatter bed because its flatness will start becoming an issue due to the plastic settling. Then again, fdm just isnt accurate enough for that level of accuracy.
That's cool! I would like to use Eddy on my printer, but maybe my use case is a bit weird. I want to generate a mesh that I can view and adjust the bed leveling manually. Although it could still be used for compensation, I think that if I can make it more level and flat physically and reduce the compensation - it would be good. I have an EasyThreeD K9, which is a very cheap printer, and I am modding the hell out of it. Some may have seen a video on such subject, but my modding might be a bit more relaxed. No Klipper, no Octoprint, stock ET4000+ motherboard and customized Marlin 2 firmware. Do you think this will be possible to make?
This combined with strain gauges to get accurate nozzle offsets and pressure information sounds like the ideal way to scan a bed. First you tap like 3 points with the nozzle, then you get a detailed fast scan over the whole bed, then you compare those 3 areas with the offset of the eddy current sensor and bang, you've got insane accuracy across the whole bed and no z offset faff.
"bad edges" - one of the disadvantages of this style of probe is that its not taking a measurement at the nozzle. Its offset in x and Y from the nozzle. This means it cant capture axis twist as seen at the nozzle tip. This usually shows up as more error near the edges of the bed. I still haven't seen anything that makes me believe that a nozzle contact probe isn't the right solution. Eddy/Beacon scanning are just a faster way to gather bad data. Klipper can do thermal compensation as the first layer progresses (for frame growth). I've got a patch that extends this to nozzle growth compensation. With that feature plus some other tweaks I think we will be able to beat Prusa's first layer consistency.
This was an awesome video... I was wondering if I'd be able to use TAP for offset and Beacon/alternative for mesh... This video answered my question perfectly.
I saw on the discord for it that they are supposedly adding a tap feature like the contact one as well. Has worked perfectly for me for like 6+ months now even at 105c beds for abs
The BigTreeTech Eddy is currently not supported in mainline klipper yet, unless you wanna use Big tree techs fork( no one would want to). So they leave it up to the klipper developers to finish there product, unlike beacon or Cartographer3D that develop and support there products before and after releasing them.
@@SilentRush420 Not true. It is fully supported in mainline with only temperature compensation waiting for a merge. The merge is sitting in PR state at the moment and after a few small tweaks will be included in mainline very shortly.
Idk how my original ender 3 with a bl touch gives perfect first layers everytime. But these more advanced ones are simply worst with spots who are way too close to too high. People crapping on the physical probes. They got what they wanted and now they face another issue being temperature. M leaning into putting a bl touch on my sv08 at this point
Loving your content. I've recently just gotten into 3d printing amd the mechanics to me aren't a problem. However I've been trying to learn klipper and it's a process but your understanding and explanations are a great help!!
I disagree that it is more accurate then probing. You are scanning the metal and not the actual surface unlike something like voron tap. Tap probes the actual surface and not some theoretical material with an offset. Tap does not need an offset unless you want more or less squish on the first layer. After that you set it once and forget about the Z offset even after changing plates. Also it does not save much time when it takes a few seconds to prove when you compare to a min and a half on a 2+ hour prints..
Not to mention that as your nozzle wears it decreases in length - tap probe homing corrects for this by itself. There's a reason both Prusa and Bambu Lab are almost exclusively using tap homing.
Yeah idk I use a gasoline bed for Nylons petg, and glass with pei for everything else, so these fancy inductive probes aint gonna do shit for me. The best way is lidar, and it shouldn't be too hard to figure out as the K1 max and X1C already have lidar models on them. Just need to figure out the programming side of things...
Cartographer is working on the touch probing for z-offset. It also includes an accelerometer like beacon and has temperature compensation like eddy,it supports usb and canbus and is priced right in between eddy and beacon. Given that, it's got the best bang for your buck! I've had mine on my k1 max for a couple of months now and it's been amazing. Would have been interesting to see your findings if you had added that one into the mix too.
its not really a versus in my opinion. for cheap printers, go with eddy usb. for something like a voron i would go for a can board + tap/klicky + eddy coil. slightly more expensive than beacon all together and much more functionality and only one cable
@@KilianGosewischtap is quite possibly the worst thing you can put on a voron, it cuts accel performance in half and brings it down to ender level resonance response. and tap with revo is the equivalent of swapping a lawnmower engine into a lambo. this is one of the many many many many advantages of beacon over tap, not having the extra rail axis, weight, etc. canbus also has a slew of issues compared to usb and other bus systems like ethercat, but thats a whole other thing.
Cartographer seems to be the best probe, you can choose between CAN/USB and change it after the fact. IDM is also a Beacon clone and cost the same a Cartographer. I would like to see the difference between all those.
Good question. I’m not super familiar with processors and their specs but I’ll take your word for it that the RP2040 is overkill. I guess they’re cheap enough that it doesn’t matter.
Klipper has RP2040 support already as a generic controller, and the chips are very cheap. Never know what all you can truly add in the future with onboard processing for very little cost. My personal mainboards that I design are all RP2040 for similar reason, even if it's "overkill".
Why did you side-mount it? It fits right into the probe space behind the printhead on the mgn mount! Love my Eddy, best thing added to 3d printing since heated beds.
@@ygk3d I gotcha. That's why I went with the USB Eddy, to do the Z homing. I had to widen the wiring channel a little with my dremel, but it fits! Tap looks like a nice light/strong mount.
I have seen someone who got the cartographer probe installed on theirs so you might be able to. It would have to be the USB model of course. It will depend on getting BTT's version of klipper on the mainboard and toolhead boards. I think that might be difficult though because I believe the cartographer install for the K1 relies on using a custom install script just to get vanilla klipper on the K1. Maybe when the Eddy is merged into vanilla klipper it will work.
I'm sticking with my BLTouch while I wait for the laser rangefinder method that scans the bed while traveling at 500 mm/s, and achieves a resolution in tens of nanometers.
I wonder why beacon requires you to not have any metal parts in the proximity, it shouldn't matter, all you would need to do is to just run some zero calibration high above the bed, metal bits above the coil will just shift the readings. Eddy literally has two metal screws 15mm above the coil PCB.
Beacon is all about that reliability. After talking yo the creator you could have metal parts in the keep out but it'd impact performance and as such he advises against it so you can reproduce the results Beacon is showing on other machines
OK so these are (a little) quicker than an inductive probe and about as accurate. I'm sticking with inductive with Voron Z END-STOP, at least for now :)
![Felix "Unfair" | [Stray Kids : SKZ-PLAYER]](http://i.ytimg.com/vi/Oswujxm2Ag0/mqdefault.jpg)
Klipper devs are developing a function that let all ldc1612 eddy current probes have tap functionality, cartographer is also developing an open source version.
That’s amazing news!
I cant wait for the cartographer tap to come out. Richard has done amazing things at such an affordable price. Makes me wonder why beacon is so utterly expensive. Mine's worked flawlessly for 6+ months.
@@YourBuddyDinec richard didn't spend any money on rnd
Got a link to this "announcement"?
@@iitsalex It's pinned on the discord
Can confirm Beacon contact is great. The printer auto calibrates its z offset for each print so I can change build surfaces, nozzles and even whole hotends with having to set anything.
Nice!
How did you get it?
very simple. Just do an upgrade and follow the doc how to modify your config. Takes some time, you have f.x. also to upgrade tthe firmware of your beacon. But when you are finished its very impressive. Super new functionality for zero coins. Great customer service !!!
@@valent_t It's a regular Beacon with their new firmware downloadable from their site as of a few days ago.
How does that work, though? As someone who changes nozzle sizes relatively often - and has observed a difference of over half a mm in nozzle length between different sizes and brands - anything relying on a statically-set nozzle Z offset just isn't going to work for me.
I have become so used to printing on non-metalic surfaces (mostly G10) that it always surprises me how many people completely skip over the fact that this is a big downside to any eddy current or inductive probe. I guess there is a lot of users who have "print surface" so synonymous with "metal sheet" in their mind this does not even occur to them. But for me an Euclid probe keeps delivering perfect first layer with reasonably fast probing sequence on any surface so well that for now I am not even considering swithing to any of these "game changing" new probes.
That’s fair and totally true - spring steel sheets have become somewhat ubiquitous.
@@ygk3d One point I did not mention to keep it brief: how much resolution on the mesh you need is also influenced by the type of print surface. On my Ratrig V-Core 3 500 I tried even pretty high mesh resolutions with different surfaces. It seems that if you forget some tiny debris on the magnetic sticker and put the standard spring steel over it you get a hill in that place that can screw up your first layer if no probe point is near. But you put 1.5mm thick G10 with a metalic sticker on the bottom onto the same debris on the same magnet and the hill may have similar height but it is so gradual the neighboring probe points make the mesh relfect the real shape of the surface well enough to have no issues with the first layer.
I use 10x10 when printing 450x450 mm or bigger and with the adaptive meshing algorithm this ends up being 3x3 mesh for any small print. With lifting the Euclid only 2mm after probing and 10k travel accelleration the probing is done in a few seconds. So this makes it hard for me to see how the faster scanning instead of probing would be a benefit even if it worked on G10 surface. Either it saves me few seconds on an hour long print or it gets me higher mesh resolution that does not reveal anything new.
I can absolutely see a scanner like the Eddy being much better than for example BLTouch that is quite slow with the pin deployment needed twice or more at every probe point. Compared to that it will be both faster and have better resolution.
@@hebijirik great perspective. Thanks for sharing. The use of adaptive meshing definitely makes a big difference. Really interesting about spring steel versus G10. You find the G10 has good adhesion and release for most materials?
Based on the beacon contact demo, it seems the contact/tap implementation would work at enough distance through the G10, so perhaps for those surfaces you still do nozzle probing?
@@ygk3d The G10 is absolutely excelent for PETG which I print very often. Unlike PEI powder coated surfaces there seems to be zero deterioration after any amount of printing. I once scratched it accidentally so I tried an orbital sander on it, cleaned and washed all the dust from the sanding and it was like new again. Still sticks like crazy while hot and releases easily when cooled down.
For PLA you need to degrease and clean it well, even fingerprints can decrease the adhesion too much. I would say it it slightly less bed adhesion for PLA than the PEI surfaces. But still very usable. When the plate cools down the PLA prints just completely release on their own.
ASA/ABS is adhering a bit better than PLA but not as much as PETG.
I decided to try the Vision Miner Nanopolymer adhesive on it and it works great too. It removes the PLA sensitivity to having it perfectly clean and freshly degreased. Prints other than PETG still completely detach once cold and the surface is still smooth.
I like the low cost (I got 4 500x500 for less than the cost of just 1 PEI sheet), the ability to easily fix up slight damage and the option to buy different thickness based on how stiff I need it for the size. I use 1.5mm on my 500mm V-Core but it felt too stiff for smaller printers. I tried 1mm thick for a Prusa MK4 and a 200mm V-Core and they work nicely.
Disadvantages are: the glass smooth bottom side of the print might not suit every part, so I still use powder coated PEI when I need that look. If your printer has magnets for metalic beds you can put a metalic sticker (used to make a magnet-holding whiteboard on any surface) but then you only get to use one side of the G10. Or you have to mess around with some other way to clamp it to your bed. Depending on where you are buying it might be easy or not - I emailed some local company that makes custom insulation for electric instalations and they sell it in 1050x1025 pieces in any thickness and they can cut it to smaller pieces of dimensions I ask for. So I got it quarted for the 500 and I got 16 pcs for MK4 🙂.
P.S., eddy current distance sensors are not linear. I work for a sensors company, and temp stability and linearity are our biggest woes in all forms of capacitive and inductive sensing (eddy current sensing counts as inductive).
Nice move from BigTree, glad to see they have or are working on the three most common ways for talking to the rest of the printers, and that price point is very tempting. Will keep this in mind on my Ender5 to Mercury one upgrade I am working very slowly on.
We went from inductive probes to klicky-probes because PEI coating on magnetic build-plates is uneven.
Eddy probe does not take into account non-conductive media (PEI).
Lidar is the only way.
Exactly my thought. And about beacon - same, right?
@@vim55k yep, same
How thick is pei? Honestly pissing into the wind. It works flawlessly on pei.
We went to klicky probes from inductive not because PEI coating is uneven. It's because the z-offset has to change anytime you change build plate and that is annoying. A good PEI build plate has a very consistent thickness across its surface.
beacon and probably eventually all eddy sensors are now also touch probes.
The new beacon software allows it to detect when the nozzle hits the bed with around 50 grams of force and klipper is eventually going to support this natively for other eddy sensors.
This means the probe can now compensate for the thickness of a build plate on top of the bed while also giving automatic z-offset calibration. This only needs to be done for the initial bed leveling and z-homing, as any quality build plate that goes on top of the bed will vary by less than a fraction of a single layer height in its thickness.
The bed should be the only thing varying by an amount significant enough to effect initial layer quality.
Thus for the actual slow part the, bed mesh, the eddy sensor is used as normal and thus saves time. Also the increased accuracy from the denser mesh means that the build plate can vary more in its thickness before it could sum to a great enough deviance to cause a problem with initial layers.
Also LiDAR actually struggles with textured PEI plates, because it picks up the texture rather than the average thickness of the plate which is what you actually want. This is why bambu printers for instance turn LIDAR off by default on textured PEI.
I use a glass plate and the cartograph works just fine, its only kinda useless as the glas plate is quite flat. But this is just my experience
A walkthrough of the can bus upgrade would be amazing!
I would also love to see a CAN bus upgrade to the 2.4!
Your best video yet! The way you talk to the camera has drastically improved. It's no longer monotonic like before. Keep it up!
Unless you are making changes to your printer constantly, the clicky or tap mesh works great if you, do a large number of points, and save meshes for a set of typical temperature ranges you print at. for example. I heat soak my printer with panels off at 60c bed temp for 30min, then do a bed mesh and save. in my print template for PLA, i load that bed mesh before print starts. I do that for PETG, ABS/ASA, etc. so I have a few profiles for different conditions. If you have a thick enough bed plate, and don't change the screws or other things on your printer, and do the same length of heat soak each time, the bed etc will expand and change basically the same every time so that the saved meshes work great. At least they have for me over the past couple of years.
Good tip!
These are really time saving sensors, not something needed if you don’t care about time.
Printing is a significant portion of my income and 30 minutes heat soaking the bed is money down the drain, as is probing a complete mesh every print. Being able to build a mesh for every print in 10 seconds means I can hit print and don’t need to stand around for 5(or in your case 30) minutes to ensure the layer goes down clean. Any printers running 24/7 are not gonna be able to produce reliable first layers every single time without building a mesh or having an expensive and thick machined bed.
These sensors don’t offer anything that can’t be achieved with other probes, except saving a lot of time.
Why is it bolted to the side of the toolhead at 1:40
its formfactor was designed to replace omron probe in the stock stealthburner carriage.
If you keep watching you’ll find out 😉
Inductive has been around in CNC for a long time.
I remember using inductive sensing on an Amada Lasmac LC1212 Apha II way back in 1998.
It's great to see this being introduced for bed levelling.
Inductive sensors were common in automation in like 1960s.
Thank you Faraday and Lenz for making it possible to get the Eddy "probe" invented.
Beacon contact is already here and it works great and I definitely have perfect first layers.
Beacon auto calibrate did it all on its own and I’ve never had a probe be able to “auto calibrate“ on klipper using their probe calibrate module without having to baby step it afterwards.
I’ve used bl touch, klicky, tap, and now beacon and beacon is the best probe hands down especially with contact.
Good stuff! Impressive that you didn’t need to baby step whatsoever.
@@ygk3d very impressive this is the first time in 5 years of 3d printing on multiple printers I couldn’t believe it either.
@@Detroit_Playa I was shocked the first time I set up Contact with my Beacon and the first layer was just perfect immediately without doing anything. I've used every other type of sensor and probe out there. Inductive, capacitive, servos, clicky, Tap, etc. Beacon was worth the $70 I paid for it all day every day.
@@WhiteG60 I’ve got the same story it’s absolutely impressive
As someone who had id for more than 40 days, i can say that it's has it's problems. They have beem fixed as of now and i am happy with it and it's fast, but i use it with tap for the perfect z offset and without tap it's another story.
Interesting. It seems good in combination with Tap or an endstop pin. I haven’t used it for homing so can’t speak to how repeatable the z offset is.
How is this effected by thermal changes?
PINDA was ultimately phased out to a loadcell due to the bed probing not being accurate when heated even with a thermistor
Very nice video! Thanks for your work. I hope my printer will have one of those eddies at some point but have a long way to go. You are doing a great job in helping me on my journey.
Thank you!
When I went to order a Beacon they contacted me shortly after saying "Hey, we're planning an updated version in a couple months. We know you placed the order, but if you'd like to hold off until then we'll send you the H no extra charge."
Definitely won me over as a customer for that. Especially now that the H has auto z-offset.
_"The probe emits a electromagnetic field which induce a small electrical current in the nearby conductive medium"_
*But this is exact principle any inductive proximity sensor/switch used in 3D printers operates upon.* But you saying like it's something absolutely new and different which is not true. The actual major difference between the generic inductive proximity switches is that the first one is permanently set to a specific triggering hysteresis loop that obviously requires vertical movements to trigger whereas the latter one continuously register changes in the electrical current yet measures distance to a build plate surface.
I wondering same. My old prusa mk3 got proxmity sensonsor build in allready. 😂
AHHHHH! You are so right it hurts! Induced electric currents ARE eddies!
In electromagnetism, an eddy current (also called Foucault's current) is a loop of electric current induced within conductors by a changing magnetic field in the conductor according to Faraday's law of induction or by the relative motion of a conductor in a magnetic field. Eddy currents flow in closed loops within conductors, in planes perpendicular to the magnetic field. They can be induced within nearby stationary conductors by a time-varying magnetic field created by an AC electromagnet or transformer, for example, or by relative motion between a magnet and a nearby conductor. The magnitude of the current in a given loop is proportional to the strength of the magnetic field, the area of the loop, and the rate of change of flux, and inversely proportional to the resistivity of the material. When graphed, these circular currents within a piece of metal look vaguely like eddies or whirlpools in a liquid.
So what you are saying is that it is totally the same, except for the meaningful difference that makes it night and day different and better, that is new to 3D printing. Thanks for that info. It has also made me realize that cars are just the same as stage coaches... I mean, wheels on a carriage have been around since before ancient Rome.
Have you looked at the bd sensor?
will you be doing a vid on beacon contact when it comes out as it solves a lot of problems
Beacon contact is out already (filming must have been done a while ago). Contact literally >just works
@@Bogm8 lol had no clue it was already out
Does the Eddy current sensor work with all types of bed materials?
How about using a BLTouch-like setup with a smooth, metallic ball that moves slowly and continuously registers Z-axis displacement? It could perform an initial probing to adjust for any wear and use a random path to cover the entire area, so it doesn’t always drag over the same spots. Plus, the slow movement helps reduce noise.
Or maybe a laser sensor, though it might struggle with glass or weird texture/colored beds due to refraction issues.
What would be the most precise system?
@@Gambinaceo no, the surface must be metallic. Your other ideas sound great!
i've been planning to integrate eddy into all my printers, along with revo, so thanks for this. one of the things i've never been fond of with btt devices in particular is having to flash firmware to everything before it can be used, but eddy & similar probes are something special & i see as more than worth the effort. something i feel the need to point out, though, is btt has been touting for quite a while now that ebb36 & ebb42 both have accelerometer function integrated, so you shouldn't need an external accelerometer anymore. i'll be testing that when i finish building my micron+, but until then i cannot confirm.
Does this work on magnetic quick swap build surfaces & textured build plates?
Will it work with glass beds?
hi! I also started using Eddy!
I have a problem when I print parts with a grid from Eddy. The height of the printed parts increases by about +0.6 mm. What could it be??? FADE is off! If without a grid, all the dimensions are perfect!
Possibly why you never get perfect first layers: maybe more of an issue for v wheels, but may apply to linear rails too. If there is any z deviation in height between the nozzle and where the probe ends up that matches where the nozzle tip was, then that will be a shifted value. So this is usually partially fixed with zero-y-offset mods on cheap v wheel printers, where the y value is the same for the probe and nozzle, yet it does nothing for the x axis. If there is a slight imperfection in one of the axes within the distance between your nozzle and probe, then it will not be accounted for in the mesh. The only way to fix this is to nozzle probe in one point then move the scanner or bed probe to the exact same point. Do this across the whole bed, it will then give you a map of the variance between actual nozzle distance and what the probe measures. Realistically you need to subtract or add this variance for the whole bed from the live bed mesh. Hopefully that makes sense, and it would be something possible to test.
Or just have the probe on the nozzle like the "clicky" solution. This is a step back from anything that has the probe directly under the nozzle. Another solution would be to implement skew compensation in Klipper for the Y axis, we already have it for X axis and it solves the offset problem as well.
Based on my research technically you can run I2C over SPI bus which we have on BTT SB2209 but no one have done it so far and it could be an interesting setup.
The new LDO nitehawk-36 usb toolhead board has a usb expansion port for a probe to plug into, so you dont have to run a separate cable. Unfortunately the nitehawk-sb (two piece stealthburner toolhead style) didn't get the port. Also the beacon D is a little cheaper then the beacon H at only $65 if you don't want the adxl345 accelerometer that's built in. Which you might not considering a few tool head boards already have them integrated like the nitehawk, btt SB2240, ebb36/42 or if you have a bed slinger and need the sensor to be movable from the hotend to the bed.
Well presented information. I think the main reason why we can't get perfect layer is that we keep measuring metal plate which is still powder coated with uneven layer of PEI.
what about the issue with the z offset not saving from eddy from btt...
id be curious to find out if the usb/temperature version could work on a MagnetoX
what about bd-sensor ? i am really happy with it
@YGK3D Would you mind sharing the slicer settings (Gcode feeds/speeds/material settings) for your first layer bed-covering tests? :)
Do you know if I Could I use the I2C pin that is available on manta 8P V2 board for using Eddy Coil version? or the unique option that i have for connection is usb connection mode?
@@jhoanperez4742 maybe, but the included cable is only long enough to reach a tool board, not the mainboard.
Is the non usb version a direct replacement for the bl touch? Or do i need to get a usb version for the ender 3
Note, this doesn't work on glass, it doesn't measure the surface, it measures the metal below the surface, so any variation in the for example the PEI thickness will NOT be detected
Great video and thanks for all the info. But did i understand that right? After all that, Klicky and TAP are still far more accurate, all be it slower?
When i type in bedmesh_calibrate method=scan etc in defaults to manual probe. When i use BTT_bedmesh_calibrate it probes individual points on the bed. Help i have eddy coil
Did you have to convert the Troodon 2 to canbus to make this work? That part wasn't clear in the video. I have three Troodon 2 Pro units now and may consider this upgrade
No. I used the USB version of Eddy on the Troodon.
I'd expect the heat bed underneath would interfere! As might the saturation from the neo magnets.
I don't have a gantry.
How does the probe help with z offset.
Is it best combined with a load cell nozzle?
probably dum question but do you have installed eddy on a rpi4 ? i have a ender 5 plus with merlin (klipper firmware) with rpi4
Will it work through a glass bed?
What is that green toolhead? Looks really good!
That’s the EVA toolhead on the RatRig V-Core 3
is this similar to the eddy sensors on the a1 toolhead??
I've been using the BD Sensor's from Mark at PandaPi 3D, its cheaper than these, more compact and the sensor is remote from the PCB so you can install extreme close to the nozzle, I even managed to install one on a Voron 0.2, I have a video showing it in action.
+1 Mark's BD Sensor works great ... I think it's a much better solution than Beacon or Eddy
Awesome, I have the 2.0 pro after watching your videos. I was already considering Eddy. Im nervous about installing can but it is a needed upgrade
14:00 THAAAAAAAANK YOUUUUU!!!! OMG for real i've been issues since beginning of 3D Printing to set the z-hight correctly and was everytime wondering why my bed is "flat" but at such an offset of the 'middel'... I never found this out upon your tiny advise 🙏🙏🙏 god thank you 🙏🙏🙏
I had already learned of the Eddy on BTT's site and ordered 2 USB versions on impulse... I'm glad that it was actually a good decision :3
Me too, I ordered my USB version two weeks ago, but they've yet to ship it, hmph!
I bought the BTT Eddy USB and it doesn't work with any applications that require a 100c+ bed. Can someone share their experience using other contactless probes in high temp applications?
@@RasTona_ After a few months of struggling with eddy I decided to give up yesterday. It is way too temperamental and buggy. Just another rushed copycat product. Cartographer is about to release a new high temp version and judging by their great track record with other probes it should be the holy grail
Marlin, Z-Probe offset?
So is there a version that will directly replace a BLTouch ?
I have a bambu the only upgrades i did was the obsidian or DLC coated nozzle. I know DLC used a lot in gun industries so i figured it be good and it's great nozzle. Is this device domething you cash only use on voron style printers or could it make your bambu better? Thanks
No. Not compatable with Bambu's ecosystem. Bambu has it's own system that works great.
Would love to see your take on a CAN BUS conversion walk-through/ install! After finalizing wiring my 350mm 2.4 i learned about how much nicer a CAN BUS setup would be so it's something I've been wanting to do since mid build 😅
On a voron 2.4 Dupont on the ebb36 will stick out past the x rail. It will catch on the back gantry chain. Can would be interesting.
Is there a reason for this that makes a lot of sense? Unless if its price is very low or speed is insanely high, probes, switches, and induction already have a lot more accuracy than what is needed. If you need more accuracy, you should be investing more in a flatter bed because its flatness will start becoming an issue due to the plastic settling. Then again, fdm just isnt accurate enough for that level of accuracy.
don't most canbus boards have an accelerometer built in?
How easy this to install on a K1 Max
How about LIDAR ?
That's cool! I would like to use Eddy on my printer, but maybe my use case is a bit weird. I want to generate a mesh that I can view and adjust the bed leveling manually. Although it could still be used for compensation, I think that if I can make it more level and flat physically and reduce the compensation - it would be good.
I have an EasyThreeD K9, which is a very cheap printer, and I am modding the hell out of it. Some may have seen a video on such subject, but my modding might be a bit more relaxed. No Klipper, no Octoprint, stock ET4000+ motherboard and customized Marlin 2 firmware. Do you think this will be possible to make?
No
This combined with strain gauges to get accurate nozzle offsets and pressure information sounds like the ideal way to scan a bed.
First you tap like 3 points with the nozzle, then you get a detailed fast scan over the whole bed, then you compare those 3 areas with the offset of the eddy current sensor and bang, you've got insane accuracy across the whole bed and no z offset faff.
"bad edges" - one of the disadvantages of this style of probe is that its not taking a measurement at the nozzle. Its offset in x and Y from the nozzle. This means it cant capture axis twist as seen at the nozzle tip. This usually shows up as more error near the edges of the bed.
I still haven't seen anything that makes me believe that a nozzle contact probe isn't the right solution. Eddy/Beacon scanning are just a faster way to gather bad data.
Klipper can do thermal compensation as the first layer progresses (for frame growth). I've got a patch that extends this to nozzle growth compensation. With that feature plus some other tweaks I think we will be able to beat Prusa's first layer consistency.
Hi Gary, I'm using Klipper, is there any chance you'd be willing to share that patch?
Hi will this work witht he sonic pad ona a ender3 s1 pluse ??
im trying to find out the same thing - for my E3 Max Neo....
i think we are out of luck right now. hopefully after the Klipper mod it will work...
This was an awesome video... I was wondering if I'd be able to use TAP for offset and Beacon/alternative for mesh... This video answered my question perfectly.
Thank you!
Great video of an exciting product. I had to jam a usb hub under my printer for the beacon. This thing is slick!
Until btt eddy has some some actual more features, I'm sticking with cartographer3d for 35$ lol
I saw on the discord for it that they are supposedly adding a tap feature like the contact one as well. Has worked perfectly for me for like 6+ months now even at 105c beds for abs
Ya, I can't wait for tap support on my carto. It's made my k1 max so much better!
The BigTreeTech Eddy is currently not supported in mainline klipper yet, unless you wanna use Big tree techs fork( no one would want to). So they leave it up to the klipper developers to finish there product, unlike beacon or Cartographer3D that develop and support there products before and after releasing them.
@@SilentRush420 Not true. It is fully supported in mainline with only temperature compensation waiting for a merge. The merge is sitting in PR state at the moment and after a few small tweaks will be included in mainline very shortly.
@@looxonline well it is now... that comment is a month old, and eddy's been released wayyyy longer then that.😂
can you run EDDY USB to RPI when using SB2209?
@@Liam-it7yu yes
@@ygk3d Thank you!
Idk how my original ender 3 with a bl touch gives perfect first layers everytime. But these more advanced ones are simply worst with spots who are way too close to too high. People crapping on the physical probes. They got what they wanted and now they face another issue being temperature. M leaning into putting a bl touch on my sv08 at this point
Does this work on Ender 3 s1 pro? :)
Loving your content. I've recently just gotten into 3d printing amd the mechanics to me aren't a problem. However I've been trying to learn klipper and it's a process but your understanding and explanations are a great help!!
Thank you very much! Glad to hear.
I disagree that it is more accurate then probing. You are scanning the metal and not the actual surface unlike something like voron tap.
Tap probes the actual surface and not some theoretical material with an offset.
Tap does not need an offset unless you want more or less squish on the first layer. After that you set it once and forget about the Z offset even after changing plates.
Also it does not save much time when it takes a few seconds to prove when you compare to a min and a half on a 2+ hour prints..
Not to mention that as your nozzle wears it decreases in length - tap probe homing corrects for this by itself. There's a reason both Prusa and Bambu Lab are almost exclusively using tap homing.
I've been considering/planning on setting up my printer to Klipper for a while now. I think this may seal the deal.
Yeah idk I use a gasoline bed for Nylons petg, and glass with pei for everything else, so these fancy inductive probes aint gonna do shit for me. The best way is lidar, and it shouldn't be too hard to figure out as the K1 max and X1C already have lidar models on them. Just need to figure out the programming side of things...
Cartographer is working on the touch probing for z-offset. It also includes an accelerometer like beacon and has temperature compensation like eddy,it supports usb and canbus and is priced right in between eddy and beacon.
Given that, it's got the best bang for your buck! I've had mine on my k1 max for a couple of months now and it's been amazing.
Would have been interesting to see your findings if you had added that one into the mix too.
why did you mount the eddy on the side of the SB? this is meant to go on the inductive mount
just watch the video dude
Pay attention
Its a nice concept but it just doesnt work with non conductive beds such as glass or g10
Can this be used on the ender 3 S1 ???
As long as it runs Klipper and has a metallic bed it should work.
will eddy melt like the induction probes? :D roasted several of them.
Sovol SV 08 compatible?
@@Ninjadave18 yes
@@ygk3d beautiful
You dont need to stop klipper when flash the sbc it stops and restarts automatically now
Beacon has two additional killer features: z offset and accelerometer. Just that it’s worth the price difference.
Max operating temperature too, beacon clears all
its not really a versus in my opinion. for cheap printers, go with eddy usb. for something like a voron i would go for a can board + tap/klicky + eddy coil. slightly more expensive than beacon all together and much more functionality and only one cable
@@KilianGosewischtap is quite possibly the worst thing you can put on a voron, it cuts accel performance in half and brings it down to ender level resonance response. and tap with revo is the equivalent of swapping a lawnmower engine into a lambo. this is one of the many many many many advantages of beacon over tap, not having the extra rail axis, weight, etc. canbus also has a slew of issues compared to usb and other bus systems like ethercat, but thats a whole other thing.
Please do a CAN conversion walkthrough.
Cartographer seems to be the best probe, you can choose between CAN/USB and change it after the fact. IDM is also a Beacon clone and cost the same a Cartographer. I would like to see the difference between all those.
Might be worthy of a follow up video!
what about lidar
Thanks for the deep dive video!
you can use TOF050C-VL6180X to read the distance ...
why does eddy have a whole rp2040 on board? isn't that a bit of a waste? couldn't they have used an ATtiny or something less powerful?
Doesn't seem to have its own mcu (eddy coil). Eddy usb and can have a rp2040 you have to define as separate mcu
Good question. I’m not super familiar with processors and their specs but I’ll take your word for it that the RP2040 is overkill. I guess they’re cheap enough that it doesn’t matter.
Maybe because it's easier to manage inventory? BTT has a lot of products all using the same rp2040 chip.
Klipper has RP2040 support already as a generic controller, and the chips are very cheap. Never know what all you can truly add in the future with onboard processing for very little cost. My personal mainboards that I design are all RP2040 for similar reason, even if it's "overkill".
Magnetic induction is not proportional to distance, it decreases by distance squared divisor.
Interesting. Thanks for clarifying.
Why did you side-mount it? It fits right into the probe space behind the printhead on the mgn mount! Love my Eddy, best thing added to 3d printing since heated beds.
I’m using Tap, so no mounting position for a probe.
@@ygk3d I gotcha. That's why I went with the USB Eddy, to do the Z homing. I had to widen the wiring channel a little with my dremel, but it fits! Tap looks like a nice light/strong mount.
@@ygk3d ...also, I should have watched 100% before commenting. Doh! LOL
My Orion touch probe and glass bed always works fine on the printer I built from scratch.
Anyone know if Eddy can be jiggered into a K1 Max?
I have seen someone who got the cartographer probe installed on theirs so you might be able to. It would have to be the USB model of course. It will depend on getting BTT's version of klipper on the mainboard and toolhead boards. I think that might be difficult though because I believe the cartographer install for the K1 relies on using a custom install script just to get vanilla klipper on the K1. Maybe when the Eddy is merged into vanilla klipper it will work.
Yes, I've already seen it. A user in the Devil Designs Discord has it working well. They combined btt repo with official +config.
Magnetic Beds, an issue?
Large fixed magnets are generally an issue for inductive probes. Magnetic stickers are fine.
@@ygk3d Thank You 2000%
Same issues as all non touch probing though, must be ferromagnetic, must be uniform surface
I wish there was something like the Eddy to upgrade from BLTouch on an older printer.
The BTT website clearly indicates that it is compatible with the Ender 3 although I thought it was for Klipper.
@@petercallison5765 I missed that part, I'll look into it, thanks
I'm sticking with my BLTouch while I wait for the laser rangefinder method that scans the bed while traveling at 500 mm/s, and achieves a resolution in tens of nanometers.
If you ever use a glass or garolite bed you will waiting tens of centuries. Besides, who needs resolution in tens of nanometers?
The can bus board has an accelerometer onboard.
Why not optical? It would elevate the need of magnetic surface.
I wonder why beacon requires you to not have any metal parts in the proximity, it shouldn't matter, all you would need to do is to just run some zero calibration high above the bed, metal bits above the coil will just shift the readings.
Eddy literally has two metal screws 15mm above the coil PCB.
Beacon is all about that reliability. After talking yo the creator you could have metal parts in the keep out but it'd impact performance and as such he advises against it so you can reproduce the results Beacon is showing on other machines
well explained and good research
OK so these are (a little) quicker than an inductive probe and about as accurate. I'm sticking with inductive with Voron Z END-STOP, at least for now :)