Welcome to the SeeMeCNC Community

So I finally stopped procrastinating and installed the FSR kit from Ulitbots and Micheal's FSR mounts, on the whole they work really well and openDACT gives great calibration results in only a handful of iterations now.

The issue is the FSR's are more sensitive on the extremities on the bed (which I kind of expected) so much so that the calibration is thrown off enough to lift on the outer edges. The FSR's sit well in their mounts with no X/Y movement but slide effortlessly in the Z helped with some Vaseline. Am I missing something or is it just a case of altering horizontal radius to account for the slight difference in sensitivity relative to the origin?

You should not have to alter any of the delta parameters - that's why you installed FSRs to do auto calibration!

How do you know the FSRs are less sensitive in the middle or are you basing that on the results of opendact? If you are actually experiencing less sensitivity in the center of the bed, the only way I can think of that could occur is if there is flex in your bed. In this case, the bed acts like a big spring so more force is needed in the center to trigger. What is your bed stack composed of? It could also be an artifact of the calibration software itself. I have "0" experience with opendact but given the number of odd issues, constant changes and the fact that it actually mucks with the steps/mm does not give me a high degree of comfort with it.

When you tap with a finger tip on the bed watching the LEDs, does one or more trigger with the same tap pressure all over the bed? Also, while calibrating, watch the tip of the nozzle touching the bed and see if you can observe some movement. Another factor could be the the hot end or effector is deflecting in the center of the bed.

Using the very scientific method of finger tapping after noticing the slight error from opendact, I tried the finger tapping around the bed. On the perimeters of the bed all FSR's trigger with approximately the same tap pressure and in between towers illuminates 2 LEDs at the same time. However, progressively working into the center of the bed progressively requires more pressure, its not a lot more but its a noticeable amount. When the FSR's do trigger in the centre of the bed, all the LEDs trigger at the same time as well.

Since I first installed the FSR I have changed my hotend mount as hotend would flex towards the right when looking straight on at the machine before triggering, since changing to a new mount and effector this no longer occurs and the nozzle has a clean straight tap before triggering.

The bed stack is currently snowflake -> onyx -> trick laser heat spreader -> stock glass -> PEI installed with your guide. The sensitively to finger tapping in centre remains regardless of bed stack including just on the onyx with snowflake. Yes I was also a little concerned about opendact and the steps per mm change, mine was changed from 80 to 82.33mm.

Ok, good feedback, thanks.

Is it possible that there is a bow in the glass or aluminum in the center? Press on the center of the bed with something stiff - like a rod of some sort - and watch closely to see if anything moves.

David built in the ability to add a flex compensation into each probe point. I've not had to use it on any of my deltas but I know a few folks do. But these folks do not have a snowflake in their beds so it isn't surprising that they have a little flex. You do so I wold expect a nice rigid bed unless there is a bow as described above.

If there is no bow, I would remount the FSR holders on the base - clean them up with IPA and relubricate them. Maybe there is something just not quite right that would get fixed. If that doesn't work, then the compensation factor should take care of it.

I used a long screw rob for a TV stand I have spare and there was a small hint of movement, the glass and aluminium are level with a straight edge, however there is a pretty big bow in the snowflake. Would flipping it over and remounting help at all?

I'll have a go remounting the FSR holders and cleaning and re-lubricating them to see if that helps as well

You could try flipping it over. If the bow is up then flipping it over would prevent it from distorting the aluminum and glass and maybe allow them to sit flat.

I cleaned the plungers and mounts and re-lubricated as well as resetting their positions on the melamine top. I also flipped the snowflake and the difference was apparent immediately with the tapping of the bed, I would say the pressure required is now almost identical over the bed with only a tiny addition required in the middle.

I re-ran opendact, then ensured Z height was set appropriately with feeler gauges after opendact added the FSR trigger offset. The result of this is that all points I am unable to get a 0.05mm feeler gauge under the nozzle at all points on the bed with no triggering of the FSR at Z 0 as well.

All in all I am pretty damn happy! I shall celebrate by printer a case for a Raspberry Pi to create a gameboy

Thanks Michael for your help on this thread as well as your build thread for all the SSR, PEI, FSR and other joys I've pincked!