Thinking out loud….

  • Deleted

  • administrators

    In theory this is the same as using FSRs (which is already supported) - AFAIK they use an interface board to make them look like an endstop.

    most people who use them use 3 at 120 degrees on a delta

  • administrators

    The potential problem I see with that is that the bed will already be very heavy if the printer is large, so the nozzle will have to press down rather hard to register on the load cells.

  • Given the weight of a large bed with a 6mm tool plate, I am assuming its location will be fixed. My build is similarly large with a bed consisting of a 6mm tool plate hard mounted to 8020 extrusions and a PEI top layer. Total weight is roughly 37 lbs. Instead of building a leveling mechanism for the bed, I decided to hard mount it square and level to the frame and then use 4 point leveling of the XY carriage using the corner lead screws with independent motor control for each. I'm relying on the upcoming firmware revision dc42 is planning for the DuetWifi to control and sync a 4 motor Z axis. I have yet to learn what type of probe will work best for that application but I'm guessing there are a number that will. I have a BL Touch mount I want to use for mapping bed flatness, I assume it can perform the carriage leveling operation as well. Best of luck with your design Calvin….

  • I'd really consider using a Piezo sensor in the hotend which is something I've recently done and it works. I've had fsrs they worked of a fashion, but trying to measure nozzle contact from the bed is IMO the wrong way around, the heat affects the sensors, the mass of the bed has inertia, it's more complex to build. Measuring nozzle contact from above the nozzle is the way forward. The only thing holding this back was that the nozzle would be wobbly, which my setup is not. There is a post on here about it and a link to the design on thingiverse.

    It's the best z probe I've used so far.

  • It's only really a proof of concept. The intention to make something neater and smaller, probably built into an effector. However it works very well. I was quite surprised when it did.

Log in to reply