I’ve seen various auto-leveling probes like the one printrbot uses, delta maker, and others. We currently use FSR sensors. I was wondering if there are any other type of affordable probes out there that would be a good fit for mounting next to the hot end for a delta bot? Anyone have any ideas? Detection distance should be >2mm and for an aluminum bed.
one idea I would like to eventually see used, is a camera-based auto-level probe. Something like OpenMV (http://hackaday.io/project/1313-openmv) could be programmed to appear as an end stop for the reprap electronics. They are looking for ~$15 BOM (and I guess ~$25 retail price), which is comparable to an inductive probe, but would have less restrictions on the material (no need to be close to metal)
I’ve experimented with a merlin hot end a bit recently by connecting two wires to an endstop connection on the board. One wire would go to the brass of the hot end and the other would attach to the aluminum bed. When the hot end touches the bed, the endstop is closed. The only issue here is the bed has to be perfectly smooth for good contact. Tested this by moving the hot end around by hand, and it worked most of the time. Aluminum bed also can’t be anodized.
@Laird_Popkin Yes, that’s the one printrbot uses, already know about those 
@Ivan_Krasin Camera sounds cool, but adds a lot of unnecessary complexity and costs in my opinion.
I just got my probe. One of these: http://www.ebay.com.au/itm/141377046408
The same model printrbot use and it’ll be used with a heated aluminium bed. Detection distance is 2-3mm, it’s $4 and they seem very reliable. Interface to the electronics is going to be included in the RAMPS 1.5 boards too, but for now it’s easy enough to get working.
One thing that might be done would be to put raised ‘pads’ on the corners of the build plate. Moving to a pad and lowering a limit switch on to the pad could give you a good level detect without having to risk damaging the hot end once you have calibrated the different in distance between that switch and the hot end’s lowest point. That does not help you for the middle of the build plate though which can be slightly higher as some plates tend to bend upward when they are heated (expansion vs attachment stresses)
Any reason you don’t like the one we use (Printrbot). It’s so easy to implement, I’m surprised other companies haven’t adopted it yet.
Brook
@Brook_Drumm , that sensor requires a metal bed, detection distance is quite small. If you check out their kickstarter, you’ll notice they’re using an acrylic build surface, much like DM does. So the sensor you guys are using doesn’t quite work for some of us. 
How much accuracy do you need? microswitches are pretty darn repeatable and come very small.
@ThantiK is right…somewhat. @Brook_Drumm I do like the probe you guys use, but my post is about finding out new solutions to see what else is out there.
Endstops, FSR, Printrbot probe, all that are known ways.
I’m trying to find new ways to do auto-leveling, such as my second post above points out - having the hot end positive and the bed negative and “closing a circuit” once both touch each other. That is a new method.
This is not for acrylic, it is for aluminum bed that I’m searching a method for.
Another option that I haven’t seen tested yet is using an Infrared Sensor.
Here’s an interesting infrared sensor that detects 3-80cm (that’s a huge range). It’s also really small. The only downside (I think) is the diameter of it is rather large at 17mm. http://www.aliexpress.com/item/Adjustable-Infrared-Proximity-Switch-Photoelectric-Sensor-Obstacles-Avoidance-3-80CM-Detection-Range/1259877650.html
@Shai_Schechter that thing likely is not much more than one of the Sharp IR sensors in a fancy case. The bare sensors themselves might be something you’d want to look into, though they are not as precise or fast as other types and might fail once they are gunked up from the plastic’s fumes.
Best of luck. One of my life mottos: if it ain’t broke, don’t fix it. Sounds condescending now that I read it! Just trying to save you cost and trouble. Our solution works great with aluminum and is super cheap.
Brook
@Brook_Drumm , some of us idealistically do this for the invention-aspect of things. Business-wise, it makes the most sense to stick with what works. Maker-wise, it’s not always as satisfying.
Ideally, if I had my own business - it would be separated into 2 distinct parts. R&D, and the people who solidify the R&D into production-quality/manufacturable products. And I’d completely be part of the R&D group.
@Thomas_Sanladerer you bring up a very good point that I over looked about the fumes.
I’d be curious to see how the IR sensor works. Etch/mark/tape ‘locating corners’ on the bed and try? Things around the head get coverd in mess but more than a cm away it seems fine. And these IR sensors are remarkably hardy because of how they work. A mess covered lens would change it’s sensitivity, not accuracy.)
@Shai_Schechter Sure. The thing is, it’s cheap and works well. And because it’s a distance sensor (so no moving parts!), which works with an aluminum bed (the original question), and it’s cheap, the question is - why not use it? Same point @Brook_Drumm made later, and of course he’s got a bit of experience with it.
In contrast, contact sensors require moving in and out of position, or some sort of system of targets or holes (like MBot uses), so that the sensor triggers on hitting the print bed but doesn’t hit the print.
The IR sensors are interesting. Are the precise/repeatable enough at very close range for this application? I usually see them at a range of many cm rather than 1-2mm. Do they even focus that close? And how dependent are they on the surface they’re aimed at? I know an induction probe will detect a metal print bed even if there’s tape on it, and a mechanical switch will trigger on anything. But will IR get weird if it’s bouncing off of bare metal, or tape, etc.? It’d be interesting to test.
IR sensors need calibrating. They’re basically just measuring the level of IR coming back, which will vary at least a little according to the object. It seems to me that something with an on/off response is likely best. Hmm. with some optics you could do a nice thing with an angled laser and a very-narrow-field sensor.