Unable to calibrate better than .18

[nohc]

So not sure what has happened. I have a v3 with e3d v6 that has been working just fine. I auto calibrate with the Octoprint plugin and always come in around .08 to .02. Here is the series of events that occurred recently:

- I started noticing leveling issues on prints though I was still getting good numbers when running the plugin.
- Then the hotend started crashing into the bed and running the plugin would result in something normal like .08 but I would re run the plugin and it would correct again from something much worse like .2 or so. Its like the calibration was not saving when it ran. I didn't check the eeprom numbers during this to confirm unfortunately.
- I cleared the eeprom, re flashed the firmware, ran the manual delta calibration wizard a couple of times and got to .18. I tried again and same results. I can get .18 but nothing better using the calculator or the plugin (should be the same results anyways but figured I would try).

I have had this issue before when I upgraded firmware but running the manual delta wizard on the seemecnc site resolved it for whatever reason.

Not sure where to go from here. I looked the machine over and nothing obvious stands out mechanically. Any suggestions are appreciated.

[nohc]

So after a lot of reading I decided I would give a Duet FSR setup a try. I have finished installing a Duet wifi, Ultibot FSR kit, and Tricklaser FSR plate.

The best I can come up with is .08 using a 100 probe radius. If I try and back out the radius to 135 I get more like .16 deviation which is close to where I was. I have gone over the machine dozens of times mechanically and just cannot find anything wrong with it.

Here are my probe heights

G32 bed probe heights: 0.148 -0.100 -0.087 0.096 0.011 -0.160 -0.052 0.103 0.098 -0.104 -0.140 0.050 -0.136 0.015 0.101 -0.081, mean -0.015, deviation from mean 0.101

G32 bed probe heights: 0.136 -0.075 -0.137 0.059 0.061 -0.047 -0.027 0.178 0.111 -0.104 -0.103 0.025 -0.111 -0.035 0.051 -0.056, mean -0.005, deviation from mean 0.093

G32 bed probe heights: 0.073 -0.075 -0.137 0.109 0.048 -0.072 -0.052 0.128 0.136 -0.067 -0.077 0.025 -0.061 0.040 0.013 -0.106, mean -0.005, deviation from mean 0.084

I generated a heat map using the 135 radius. Any suggestions?

[geneb]

Something I noticed on my v3 the other day might be affecting you - my X carriage shell is broken at two of the attach points - you should check all your carriages to make sure that hasn't happened.

If those are ok, I'd try flipping the glass to see if your bed map changed. You also need to make sure that there's NO filament at all on the nozzle. I calibrate hot which gives me a chance to retract the filament and wipe the nozzle clean before i run the probe series.

g.

[dc42]

That pattern of ridges and valleys along the X axis suggest backlash in joints or carriages, making the nozzle height above the bed depend on which direction you approach the XY point from.

[nohc]

My carriages do have small hairline cracks near the screws but I have had these since I first assembled it. Maybe its becoming an issue now.

I also have noticed that on some probes my effector and hotend seem to tilt/wobble when they hit the bed. Should the assembly stay very solid during probing and does that sound like more symptoms of backlash? I have checked that all the barbells are snug and the effector assembly is all tight. It only does this under the force of a probe move.

Thank you for looking at this.

[geneb]

If the carriages don't wobble on the towers, they're likely ok.

Try this - home the printer then move the nozzle (G0 Z100 F3500) - this will put it at 100mm above the bed. Now grab a carriage and see if you can move it. If you really haul on it, it will move, but moderate pressure shouldn't show any movement. If it does move, you may need to tighten the belt a /little/ bit. The idea is to keep the belt tight enough so that the carriage won't move, but not so tight you're creating drag on the motor.

g.

[nohc]

I didn't see any significant play or movement in the carriages. I went back through the pulleys and belts today and nothing seems loose or anything. I switched out the effector assembly back to the stock one to rule that out and got the same results and same basic map. I switched two carriages around and the map did change some but deviation stayed the same. I am thinking maybe just order new carriages and give that a shot?

[geneb]

Could it be possible that you're getting mechanical binding at the FSRs? I don't know how the TL FSR plate works, so I'm just guessing.

g.

[hectori]

The new skates need alignment with the 120º tower and center of the bed.
They also need alignment with the level of the bed.

Hectori

[tjschweizer]

You may want to check the sensitivity of your FSR board.

[hectori]

check this. Now I have a flat bed readings.

http://forum.seemecnc.com/viewtopic.php ... rs#p108652

[nohc]

So I swapped all my parts to a Tricklaser Max Metal frame. First calibration attempt was .06 and that is with me forgetting to tension the belts. I did use new FSRs but I was having calibration issues before I even moved to duet so I think I just had something wonky going on mechanically. Going to order a titan aero and I think this will be a very nice machine.