For anyone searching, InterLink Electronics makes the FSR® 404
20mm Donut with 5.5mm hole. It’s available at digikey.
The data sheet lists sensitivity to 20N, which, if my calculations are correct, should still be within the typical 300% safety margin of deformation during a 5KG extrusion force.
I ordered one and hope to get to experiment with it a little later this month…
I had an idea…
I really want a toolhead load cell concept that can be manufactured with a minimum of processes so its economical to test out. We will have to do a lot of iterations. Ideally just a flat block of laser cut aluminum with minimal to no machining. SendCutSend can do these for minimal cost.
I was idly sketching a 3D Printer concept in CAD and going over where I could mount the load cell and came up with this idea:
Its a plate that attaches to an MGN12H carriage via the 4 central holes. A 3D Printed part would be attached to the back to grab the belt teeth on a Core XY printer. The hole in the middle allows the belts to exit.
There are 4 outer 3mm holes to attach the toolhead to. So this is a “toolhead adapter plate”. You can use whatever toolhead design you can adapt to this.
There are 4 “dog bone” flexures but you only need to put a strain gauge on 1 of them. Its just symmetrical to carry the load symmetrically.
Negatives:
Riffing on this idea and trying to DFM for CNC process I came up with this:
I wanted really tight tolerances at those split lines to save as much space as possible. The easiest way to manufacture that is to just split the entire thing into 2 parts. Then you have easy access to all sides in 2 separate processes:
It also means your free to modify that part to suit your needs. E.g. adapt it to MGN9, or some other mounting.
The whole contraption goes together with 4xM3 bolts from the top/bottom. That adds some possibilities, like using these holes to mount the Extruder as well.
This looks manufacture-able and adaptable to existing hot ends and extruders.
interesting idea.
But if I understand this right, this works only if the x-rail is mounted sidewards on the x-beam like a voron style printer. If the x-rail is mounted on the top of the x-beam like a Vcore for example this design needs a change or implmented in the toolhead design between the frontplate and the top part that is attached to the rail carriage.
You could make a different plastic block to go in the middle that attaches to a top mounted rail. But I think you are right. For best space efficiency you could probably design something different for a top mounted rail.
I don’t like the top mounted designs because the cantilever distance to the nozzle tip is increased. (center of rotation is the center of the rail)
Yeah I know what you mean, but mind that the part fan is mounted on the backside of the toolhead and the motor of the extruder is placend above the x-rail, so the overall center of mass of the toolhead is more in the near of the middle of the x-beam, so the summary load ont the x-rail is more “pull downwards” compared to front mounted rail there you have summary more a “hanging mass”, that pulls the rail carriage sidewards if you know what I mean…
But no matter for sure we can create two version for these different designs
@garethky Look what a cool thing has appeared on the market.
https://aliexpress.com/item/1005008945830049.html
From what I can tell:
Overall, it looks like a big missed opportunity. I hope I’m wrong.
hx717 were not supported at first either, this is not a problem at all. The cool thing here is that there is a free implementation and a factory-made hotend with an integrated cell.
I think the hotend with attached load cell makes a very good impression. Might be worth a try to hook it up with the garethky electronics.
ADS131M02 ADC supports an SPI interface, nothing else. The electronic board from FLY is just a bad joke. But the ADS131M02 ADC looks like a nice candidate for load cell applications.
Adding support for an additional ADC type is not too difficult. But if there is no SPI then it can’t be done. But maybe there will be some other board that you can plug the load cell into for full integration.
Their doc page doesn’t exactly look promising: Introduction | FLY Docs
Still taking 3 samples and a tolerance of 50 microns! I’ll wait for some PROBE_ACCURACY results.
I haven’t tested the actual load cell yet, but I have this in hand and it’s machined really nice. Even has chamfers! Here’s a CAD if anyone needs.
That looks like a great part! Let me know how it goes when you set it up!
