Why are my Heightmaps always too high?

fcwilt

@nurgelrot

OK.

When you do HomeAll>G32>G29 you are setting the Z=0 datum near the edge of the bed (the second G30 in bed.g).

When you do HomeAll>G32>HomeZ>G29 you are setting the Z=0 datum at the center of the bed.

Given the overall shape of bed as shown in the height maps the Z=0 datum in each case is going to be different.

So the three height maps make sense to me.

Notice in the first the one the deviations are from -0.095 to 0.094 for a range of 189

Notice in the second one the deviations are from 0.009 to 0.191 for a range of 182

Notice in the third one the deviations are from 0.015 to 0.204 for a range of 189

The range is consistent given the the nature of the probe to not provide perfect readings.

And the second and third height maps appear to have the middle near 0 which is to be expected since the Z=0 datum is set at the center of the bed.

So I would say things are working as expected and the first height map is incorrect because the Z=0 datum was not set at the center of the bed.

Frederick

Nurgelrot

@fcwilt So with that logic (and it makes sense) the second example should be a correct height map since the homeZ probes back at the center of the bed. Only it isn't. I just replace my probe and redid my zoffset and now if I probe the center of the bed 180,180 using a G30 and watching the output in DWC I see it trigger a 2.72 as expected... but if I run a G29 the same point 180,180 triggers at 2.78. I run through it over and over and my G29's are ~.05 higher than probing in any other method. and subsequent home/g29 cycles with drive the height map higher and is driving me crazy

fcwilt

@nurgelrot

Curious.

When creating the height map are you insuring that baby-stepping is cleared and that no height map is loaded?

I don't know if G29 does that "behind the scenes" so I have always done it as part of the process of creating the height map, just to be safe.

What version of firmware are you using?

I will have to perform the same check on one of my printers to see if I see the same sort of difference between G30 and G29.

Frederick

Nurgelrot

@fcwilt yeah i give up i just wont use a mesh. all i did was up the number of points. from 5x5 tp 9x9 and the whole thing jumped up nearly .1 mm

make no sense

Nurgelrot

@nurgelrot So if I home all. Then do a G28 Z it probes the center of the bed and I watch it trigger at the correct z height. 2.69 run a G30 by hand it triggers at 2.69. I run a G30 P0 X180 Y180 Z-9999 it triggers at 2.69. I run a G29 and the point at X180 Y180 triggers at 2.76.... with all the other point high as well. I just don't get it.
Firmware 3.3 earlier
and now Firmware: RepRapFirmware for Duet 3 Mini 5+ 3.4.0beta3 (2021-08-24)

fcwilt

@nurgelrot said in Why are my Heightmaps always too high?:

@nurgelrot So if I home all. Then do a G28 Z it probes the center of the bed and I watch it trigger at the correct z height. 2.69 run a G30 by hand it triggers at 2.69. I run a G30 P0 X180 Y180 Z-9999 it triggers at 2.69. I run a G29 and the point at X180 Y180 triggers at 2.76.... with all the other point high as well. I just don't get it.
Firmware 3.3 earlier
and now Firmware: RepRapFirmware for Duet 3 Mini 5+ 3.4.0beta3 (2021-08-24)

One difference I see between you probe setup and mine is the P parameter for G31.

Yours is 500, mine is 25.

Create a macro with this code. When executed it will generate a result showing each probe trigger height.

One run on a printer of mine produced a mean reading of 0.006 and a deviation of 0.003.

Be sure and edit the second M558 to restore your normal probe settings.

M558 A1 S0.01         ; set probe to take a single reading

; take 10 readings and report results

G30 P0 X0 Y0 Z-99999
G30 P1 X0 Y0 Z-99999
G30 P2 X0 Y0 Z-99999
G30 P3 X0 Y0 Z-99999
G30 P4 X0 Y0 Z-99999
G30 P5 X0 Y0 Z-99999
G30 P6 X0 Y0 Z-99999
G30 P6 X0 Y0 Z-99999
G30 P8 X0 Y0 Z-99999
G30 P9 X0 Y0 Z-99999 S-1

M558 A? S?             ; restore probe settings for A and S

G90                    ; absolute moves
G1 Z5 F600             ; move Z to known height

gloomyandy

@fcwilt Is there any chance that your Z axis is losing steps/binding/slipping? That would explain both the "step" you have been seeing in the heightmap and also the change in reading at your homing point.

Nurgelrot

@gloomyandy Anything is possible at this point. I'll look into it. seems to only be a problem with G29. I had to print some parts last night so i just manually leveled and printed everything turned out fine.

I feel like something changed with how the mesh is made in 3.3+ This printer always had a mediocre bed but I was always able to get a pretty decent mesh made to make up for a few bumps/twists.

Nurgelrot

@fcwilt So running test probe in the middle of the bed (where Zoffset s calculated and Z=0 is set) gives me constant results.

9/25/2021, 9:38:35 AM	G32 bed probe heights: -0.002 -0.001 -0.003 -0.001 -0.001 -0.002 -0.001, mean -0.001, deviation from mean 0.001
9/25/2021, 9:37:41 AM	G32 bed probe heights: -0.002 -0.002 -0.003 -0.003 -0.002 -0.002 -0.001, mean -0.002, deviation from mean 0.001
9/25/2021, 9:36:48 AM	G32 bed probe heights: -0.003 -0.002 -0.002 -0.003 -0.001 -0.003 -0.001, mean -0.002, deviation from mean 0.001
9/25/2021, 9:35:37 AM	G32 bed probe heights: -0.001 -0.001 -0.002 -0.003 -0.001 -0.003 0.001, mean -0.001, deviation from mean 0.001
9/25/2021, 9:34:39 AM	G32 bed probe heights: -0.001 -0.002 -0.001 -0.002 -0.003 -0.002 0.001, mean -0.001, deviation from mean 0.001

not as good if Imove it to run over the front left bed screw (with z=0 still calculated from center):

9/25/2021, 9:23:28 AM	G32 bed probe heights: 0.067 0.072 0.069 0.071 0.071 0.071 0.072, mean 0.070, deviation from mean 0.002
9/25/2021, 9:19:38 AM	G32 bed probe heights: 0.071 0.068 0.064 0.069 0.072 0.061 0.077, mean 0.069, deviation from mean 0.005
9/25/2021, 9:17:31 AM	G32 bed probe heights: 0.069 0.071 0.071 0.068 0.068 0.071 0.069, mean 0.069, deviation from mean 0.001
9/25/2021, 9:16:27 AM	G32 bed probe heights: 0.071 0.069 0.078 0.071 0.069 0.068 0.071, mean 0.071, deviation from mean 0.003
9/25/2021, 9:15:09 AM	G32 bed probe heights: 0.068 0.067 0.067 0.059 0.074 0.068 0.062, mean 0.066, deviation from mean 0.004

As you can see this thin bed has a sag from the leveling screws to the center. which is why I like to use a mesh.

But was going on here this is at the left front leveling screw again and it just drifted down as i ran them

9/25/2021, 9:04:38 AM	G32 bed probe heights: 0.043 0.047 0.072 0.069 0.069 0.072 0.072, mean 0.063, deviation from mean 0.012
9/25/2021, 8:57:31 AM	G32 bed probe heights: 0.007 0.014 0.014 0.012 0.013 0.012 0.012, mean 0.012, deviation from mean 0.002
9/25/2021, 8:56:00 AM	G32 bed probe heights: 0.004 -0.001 0.003 0.002 -0.002 -0.001 -0.003, mean 0.000, deviation from mean 0.003
9/25/2021, 8:55:01 AM	G32 bed probe heights: 0.012 0.011 0.014 0.011 0.006 0.007 0.006, mean 0.009, deviation from mean 0.003
9/25/2021, 8:51:14 AM	G32 bed probe heights: 0.018 0.017 0.016 0.019 0.017 0.023 0.016, mean 0.018, deviation from mean 0.002
9/25/2021, 8:50:21 AM	G32 bed probe heights: 0.017 0.019 0.016 0.019 0.018 0.021 0.022, mean 0.019, deviation from mean 0.002
9/25/2021, 8:49:09 AM	G32 bed probe heights: 0.022 0.021 0.023 0.017 0.018 0.017 0.026, mean 0.020, deviation from mean 0.003
9/25/2021, 8:47:44 AM	G32 bed probe heights: 0.076 0.077 0.077 0.077 0.079 0.076 0.076, mean 0.077, deviation from mean 0.001
9/25/2021, 8:46:02 AM	G32 bed probe heights: 0.076 0.066 0.083 0.083 0.072 0.074 0.081, mean 0.076, deviation from mean 0.006

As you can see from the times I actually ran the prob over the bed adjustment screw first. was really confused until the coffee kicked in and I probed over Home. to see the Touch was actually accurate. (except for that odd bit with the decreasing height that I havent been able to reproduce)

fcwilt

@nurgelrot

I was just about to respond to your post with the test results and they are gone.

I ran that test on one of my printers six times and the results were consistent.

The mean values ranged from -0.002 to 0.004 and the deviation was always 0.003.

Frederick

Nurgelrot

@fcwilt yeah irealised i was looking at the data all wrong so I changed the post sorry

fcwilt

@nurgelrot said in Why are my Heightmaps always too high?:

@fcwilt yeah irealised i was looking at the data all wrong so I changed the post sorry

Well the results at bed center look just fine.

The readings from the corner which kept changing are strange.

Is the Z axis on your printer moving the bed or the gantry?

Frederick

Nurgelrot

@fcwilt This one started out life as an Ender5+ so its a moving Z bed on 2 independent screws.

Nurgelrot

@fcwilt So what should I be doing to clear out the probe -- Something very wrong is happening at least to the way Im thinking.

I would think I could probe the Middle then the Front Left and them the Middle again and since we are not chaging any setting just reporting that it would be okay. but somplace data is being retained between probe runs. how do I need to clear this?

Read it bottom to top
Middle again

9/25/2021, 10:10:34 AM	G32 bed probe heights: 0.063 0.062 0.062 0.059 0.062 0.062 0.061, mean 0.061, deviation from mean 0.001

Front Left

9/25/2021, 10:09:40 AM	G32 bed probe heights: 0.054 0.057 0.061 0.061 0.061 0.062 0.064, mean 0.060, deviation from mean 0.003

Middle

9/25/2021, 10:08:46 AM	G32 bed probe heights: -0.001 0.001 0.002 0.001 0.001 0.001 0.003, mean 0.001, deviation from mean 0.001

Nurgelrot

@nurgelrot home Z each time?

fcwilt

@nurgelrot said in Why are my Heightmaps always too high?:

@fcwilt This one started out life as an Ender5+ so its a moving Z bed on 2 independent screws.

Without being able to examine your printer I'm just guessing here but one possible problem is the Z axis guide rods.

I had some typical Chinese ones and they seemed OK until I compared them with higher quality ones I got from McMaster-Carr.

The MC bearings fit the rods with much less play compared to the Chinese ones.

So perhaps your bed guide rods also have too much play leading to variations in height away from the lead screws as the bed is moved up and down during probing.

It's a largish bed and any issues with the guide rods would be more noticeable near the edges of the bed.

Just a guess.

Frederick

fcwilt

@nurgelrot said in Why are my Heightmaps always too high?:

@nurgelrot home Z each time?

Well not homing, since I don't home with my printers with the Z probe, but I am setting the Z=0 datum at bed center as part of the test.

Frederick

Nurgelrot

@fcwilt It just something I'm doing wrong or not understanding in RRF. Marking it resolved for no since I'm sick of working on it. Klipper is on the printer right now and all the probes are exactly where they should be every time. Too bad I cant stand klipper

Thanks for your help with this I did learn a lot about how things work even if I didn't solve whatever is going on with this printer. (I have several other with no issues so who knows)