Cartographer Probe

[Squibo]

I'm thinking of replacing my TAP with a Eddie Current Probe, I like the tap but the once novel probing is getting boring and marking my pei plate and the , but I don't really know much about them

I use CAN so think the Cartographer is the obvious choice for now anyways.

What's the difference between the 2 boards and when do you need the standard vs low profile connections?

I run a pretty standard 250mm V2 machine with an afterburner tool head & SB2209 CAN board, I also haven't "got round to" input shaping yet, if that makes any difference?

[claudermilk]

Are you using KAMP and the fuzzy setting? That should help with the minor marking. While I've been using a textured plate almost exclusively since installing Tap, it isn't the probing that's worn the plate out.

[Squibo]

I am not, I'll look into it, but I was going to invest in a better TAP carriage mount, like a chaotic labs one as mine is a not the best quality cheapo Ali Express one.

But after seeing the recent news on the Beacon and Cartographer it looked like that was probably the future of z probing, so rather than spend more money on TAP, just bite the bullet and change systems.

 

 

[Dirk]

2 hours ago, Squibo said:

What's the difference between the 2 boards

You will need a standard version, CAN.

The lis2dw version has some issues because of how klipper treats it currently. So I would recommend the adxl version. 

I can definitely also recommend the mount, because it will save you from printing mounts which are less sturdy and can have issues.

The 90 degrees and flat packs are for different systems such as the vzbot and the loose ones are for the creality K1.

 

[Dirk]

3 hours ago, claudermilk said:

KAMP

KAMP? TAP? Those are so 2023. Do not get me wrong, I used them and was extremely happy to be able to mesh selectively before every print... But take a look at the video. This is my 2.4 doing a QGL and after that a klipper adaptive meshing. The whole thing is faster than half-a-round of TAPping  Forget about the (possible) dents

 

[zav3nd]

8 hours ago, Dirk said:

KAMP? TAP? Those are so 2023. Do not get me wrong, I used them and was extremely happy to be able to mesh selectively before every print... But take a look at the video. This is my 2.4 doing a QGL and after that a klipper adaptive meshing. The whole thing is faster than half-a-round of TAPping  Forget about the (possible) dents

I hate you. (not really)

 

Now I need this....

What are your QGL and Mesh move speeds set to? Assume the home speed is still 'slow'?

[Dirk]

4 minutes ago, zav3nd said:

What are your QGL and Mesh move speeds set to?

I do sensorless homing, so they are indeed slow. around 50 - 100mm/s

Do not forget that this is the cheapest Fysetc voron 2.4 350 package I could get. And the most significant change that gave me more speed and acceleration, was dropping the cables / drag-chain by converting to an umbilical (going CAN). 

On this video my QGL and bed mesh speeds are 800mm/s and acceleration is 17.000. This is klippper's fast mode QGL where  horizontal_move_z value is set to 3 or lower.

 

[claudermilk]

19 hours ago, Dirk said:

KAMP? TAP? Those are so 2023. Do not get me wrong, I used them and was extremely happy to be able to mesh selectively before every print... But take a look at the video. This is my 2.4 doing a QGL and after that a klipper adaptive meshing. The whole thing is faster than half-a-round of TAPping  Forget about the (possible) dents

... Maybe eventually. If I really get an itch to make a major update to the Trident. It would be nice on the V0, but I don't see that there's room to mount one.

[netzwerg]

I have a question regarding those probes.

Do they measure the surface of the steel sheet underneath the PEI coating (similar to an inductive probe), or do they somehow measure the top surface of the PEI coating like a TAP does?

[Dirk]

19 minutes ago, claudermilk said:

It would be nice on the V0

When you do mount it on the zero, not much space remains from the 12cm y. That is why it is not recommended. However, I have seen a printer that uses the new pandorra gantry, which has mounted one, without much loss of the printing space.

So it would be doable, however, you as a Nero fan, should know that he says: set the bed once on the zero and then never touch anymore

I am about to start building one, so I should document it.

 

21 minutes ago, claudermilk said:

... Maybe eventually.

Even if this won't convince, you, I do not know what will  Maybe @mvdveer?

 

[claudermilk]

Yeah, Nero and Steve both. I finally have mine levelled, so I'm not really messing with it now. Since I'm pushing the limits on what a glorified hot glue gun can do as to pushing tiny beads of plastic, mine is ultra sensitive to good leveling and a clean bed.

[Dirk]

11 minutes ago, netzwerg said:

they measure

Good question!

The Eddy Current currents, seem to be picked up when they run through metal only. so you need a metal sheet below the pei surface. A calibration process is required where the "real" surface height is set.

However, the latest developments in the Beacon and Cartographer probes show that they are one step further. According to news released, the probes do tap the surface now too. No details are known, but I think from (1st may 2024) more information will be available about tapping the surface with a nozzle...

This will be a great development because it would make the magnetic beds, FR-4 sheets, glass beds also accessible to such scanning. I am still waiting full expectations.

Edited April 30 by Dirk

[mvdveer]

8 hours ago, Dirk said:

Even if this won't convince, you, I do not know what will  Maybe @mvdveer?

You gotta be kidding - despite my best efforts - NO CAN @claudermilk (Pun intended)

[claudermilk]

Yes, but I am an extremely stubborn GOM (just ask my wife). 

[mvdveer]

9 hours ago, claudermilk said:

Yes, but I am an extremely stubborn GOM (just ask my wife). 

That's why you're in the GOM club - a prerequisite. 

[therm_virtual0y]

I have the Cartographer probe on my SB and after getting it dialed in, am getting great first layers without having to do any adjustments.  I still watch to make sure because I can’t believe it!  So many versions of probes and macros that worked and then didn’t sometimes.  . Has anyone put this on the Ender 3?  I have a Microswiss NG and found a modified BL touch adapter that looks like it could work.  Also, unsure if it would need a low profile or standard with ADXL of course. 

[BeerMedic522]

On 4/29/2024 at 1:33 PM, Dirk said:

KAMP? TAP? Those are so 2023. Do not get me wrong, I used them and was extremely happy to be able to mesh selectively before every print... But take a look at the video. This is my 2.4 doing a QGL and after that a klipper adaptive meshing. The whole thing is faster than half-a-round of TAPping  Forget about the (possible) dents

Dirk, I can’t see your posted video…

[Dirk]

14 minutes ago, BeerMedic522 said:

Dirk, I can’t see your posted video…

Oh... I had to super duper compress it to fit into the limit of 5mb here. So it's possible I selected some unknown codec.

Whatever it is, I am sure VLC mediaplayer will run it

[xyleth]

21 hours ago, Dirk said:

Oh... I had to super duper compress it to fit into the limit of 5mb here. So it's possible I selected some unknown codec.

Whatever it is, I am sure VLC mediaplayer will run it

 

Can you share the config for that QGL setup?  That's amazingly quick!

[Dirk]

6 hours ago, xyleth said:

Can you share the config for that QGL setup?

Of course. 

One cave-at though. The QGL you see there makes use of klippers fast mode. If you define the horizontal_move_z to low enough (3 here) it stops with lifting the Z up and simply moves without lifting.

If you have a crooked gantry, the move WILL scratch your pei plate. That is why I share here a macro which one round of 'slow' QGL to take the biggest gantry inconsistencies out of the equation.

It also contains a conditional homing from my 'sensorless' setup. Use your own.

Please please use it carefully. And share it as much as you want

Part of the reason the moves are fast, is that my QGL and [printer] speed is set at 800mm/s. You need to verify first if this is going to work for your voron (Ellis' speed_test is a good place to start). 

 

[gcode_macro Carto_QGL] # safer and faster QGL for Cartographer
gcode:
    _CG28 # conditional homing
    QUAD_GANTRY_LEVEL horizontal_move_z=10 retries=0 retry_tolerance=1.000
    QUAD_GANTRY_LEVEL horizontal_move_z=3
    G28 Z

 

[Yezariael]

On 4/29/2024 at 7:33 PM, Dirk said:

KAMP? TAP? Those are so 2023. Do not get me wrong, I used them and was extremely happy to be able to mesh selectively before every print... But take a look at the video. This is my 2.4 doing a QGL and after that a klipper adaptive meshing. The whole thing is faster than half-a-round of TAPping  Forget about the (possible) dents

nice colors Thought on first sight that you used my video

[chrisalbertson]

Tech moves fast.  Now in May 2024, the eddy current sensors can to "tap" and sense to phyical top later of the PIE plastic plate.      THis will establish the "zero" at one point.  Then it can move up a couple mm and fly ovoer the plate at 800 mm/second

 

The ability to "tap" the plate is a very recent software-only development that could be applied to any eddy current sensor using Klipper.

 

The tap force is less than the older "Tap Sensor", so not markls are make on the plate.

[netzwerg]

15 hours ago, chrisalbertson said:

Tech moves fast.  Now in May 2024, the eddy current sensors can to "tap" and sense to phyical top later of the PIE plastic plate.      THis will establish the "zero" at one point.  Then it can move up a couple mm and fly ovoer the plate at 800 mm/second

 

How accurate is this compared to TAP?

I mean, the eddy current sensor will tap one point, that will be pretty accurate, but then when it "flies" over the plate it will measure the distance to the steel plate below the PEI coating, where accuracy will depend on the evenness of the PEI coating? Or am I missing something here?

For now I am super happy with the TAP probe, which is slow but gives perfect first layer every time. Since I like trying new things, eddy probe is surely on my list.

[Penatr8tor]

30 minutes ago, netzwerg said:

How accurate is this compared to TAP?

TAP is accurate to 0.0004mm per Voron Design.

Beacon probe accuracy is 0.0005mm per Beacon. I would suspect that the Cartographer is similar.

But here's where the rubber meets the road and also where these new eddy current probes leave TAP and inductive probes in the dust...

Two full runs @500mm/s taking only 20 seconds and 22 thousand points versus what... 25 points for TAP or inductive? I mean who even needs KAMP when you can probe the entire bed twice in 20 seconds?

And if that wasn't reason enough, they now have an ADXL on board and can also touch probe.

In my humble opinion... If you're building a printer... Adding one of these is a no-brainer.

[netzwerg]

I understand, it's quite accurate, but that accuracy is against the steel sheet, not against the surface of the PEI coating?

I mean if PEI coating isn't very even, first layer can still be tricky? Or am I missing something here such as PEI coatings are usually quite even, even on cheap chinese plates?

Speed is definitely impressive.

Edited May 16 by netzwerg