New gantry joints, old resonance problems

[shrewd pawn]

LDO v2.4 rev C 350
clicky-clack door (dismounted)
Ti gantry backers
Cartographer adxl + CNC mount
Galileo 2 extruder & HF Revo
x-joint mounted xy endstop switches
Fysetc CNC AB, XY joints & tensioners
CL Z-tensioners
(a felt strip taped to the back of the gantry to quiet/prevent belt slap)
(tension: Spectroid app)

I just completed replacing the printed carriage and gantry joints with cnc parts; every screw received a liberal application of VC-3 vibra-tite and firmly tightened [1 to 2 gronks 😄 ]
This is just a starting measurement as I've only returned to the point where a rough tensioning has been done and the tool moves and homes properly--It still needs to have a proper long heatsoak and maybe more de-racking/squaring etc. (Then comes the fun of running through all the printing calibrations: first layer, EM, PA, skew.)
I didn't hear any obviously-metallic noises during the testing so I'm curious how much is just the cable chain being itself, maybe toolhead wiring or even belt slap...
The shake tune app has seen improvements since the last time I used it so I can't really compare to old data; I've been staring at the documentation, but I also wanted to get input before I try to "improve" these results.

In addition to the fun new shake tune app, I also noted there's also a stepper tune? I installed it and verified which ones I have installed and where, but it wasn't entirely clear how I should edit my stepper configs. (My printer.cfg is the suggested one from LDO edited for a 350, with various stuff pasted in, and a lot of commented stuff from replaced hardware (like inductive probe and tap).

[thorny lichen]

👋

please ping me or reply to this message so we know u are ready 🙂

no rush, take your time finishing

[half birch]

@shrewd pawn Are you using a toolhead adxl for the measurements?

[shrewd pawn]

yes. it's part of the Cartographer 3d module that bolts to the bottom of the carriage. I'm using the usb version...

[half birch]

No, do you have an adxl in the toolhead board?

If use that one please

[shrewd pawn]

the CNC parts interfered with the LDO kit that bolted on to the back of the hotend module, so I uninstalled it.

[half birch]

This should have a Nhsb correct:

Also how are you measuring belt tension

[shrewd pawn]

This should have a Nhsb correct:
(I had to look that up. ) No. 2-year-old rev C LDO kit with Stealthburner and hotend-mounted sensor. Since it's now redundant, picture shows it on a spare hotend. (If I'm reading correctly, nitehawk came with new LDO rev D kits.)

Also how are you measuring belt tension
Spectroid app on the smartphone. might get the tension tool with my next West3D order...

[half birch]

Use the ldo adxl and remeasure stuff

[shrewd pawn]

For what reason?

[thorny lichen]

https://github.com/Frix-x/klippain-shaketune/blob/main/docs/is_tuning_generalities.md#note-on-accelerometer-mounting-point

[shrewd pawn]

fair enough, but since I can't leave the sensor module installed during printing, its measurements don't reflect the printer in its normal operating state (ignoring the unsecured fcc ribbon cable draped over the print bed during measurement.)

[thorny lichen]

U can measure cold for now

Preheating to print temp is also possible with the adxl connected 🙂 many have done it already

[shrewd pawn]

You've missed my point. As I said before, the old sensor mounting block crashes into parts of the gantry, so it can't remain installed during normal operation. There would also be the matter of how to disable/ignore the cartographer adxl during measurement if both are connected.

[thorny lichen]

The printed part or what „mounting block“ are u referring to?

U only install the printed part for the ADXL and the adxl if u need it - like now the case is. Otherwise u remove it

I dont see the issue how it would crash into the gantry.
X homing goes to the right. Adxl sits on the left

As well as i dont see the issue with the carto adxl

[shrewd pawn]

If the point the exercise is to take measurements of a system in its normal operating configuration, then adding or removing anything from what's being measured defeats the purpose.

Setting that aside, you're saying having more than one sensor active won't be a problem?

[half birch]

I would like you to run the test again with the ldo adxl. The purpose of this is to confirm what the carto is reading.
I have a suspicion that your y-values are not correct because of carto.

[shrewd pawn]

fwiw, here's the last run from the old version. no carto, no cnc joints

new version after carto install, no cnc joints. carto sensor but before I saw the notice about sweeping

-# I'll have to dig the fcc cable from the depths of the wiring bay to run the old sensor again

[shrewd pawn]

huh. apparently a reboot was missed after turning off sweep in the config, so it's actually off for this run. Both sensors connected

-# man does that fcc cable add a lot of buzz

[half birch]

Loosens your belts by 1/8 of a turn.

Can you also install tmc_autotune

[shrewd pawn]

done.
Like I said at the beginning, autotune is installed, but it's not clear what can stay or go in the printer config.

[thorny lichen]

u only need the part in my screenshot 🙂 with your motor model No.

we only care about X and Y here. the rest is up to u

[thorny lichen]

u dont need to run the axes map macro, its currently in beta and therefor useless 🙂 Klipper knows how the ADXL is oriented

[thorny lichen]

if we tell u something, like with autotune, and something is unclear, just ask us and we will help, explain - whatever.

Im german, i can work best with direct and clear questions 😄

[shrewd pawn]

that's why I included my printer.cfg as LDO put a lot in the stepper sections

[thorny lichen]

not sure if this is a question, a seek for help or just a comment 😄 as i said, i prefer direct and clear questions so i know its indeed something where im supposed to help 🙂

"im trying to setup autotune, however, im not sure i understand the README correct as it states .... and i have ..... . do i understand it correct that im supposed to.... ?" - this would personally help me know where to help and whats not clear 🙂 to give an example

[thorny lichen]

i just checked, LDO puts the same in the TMC as well as Stepper config as the default VORON config in terms of parameter amount 🙂

LDO: https://github.com/MotorDynamicsLab/LDOVoron2/blob/afbbbdb8fe50ef59eb51f46d23a0a462b21c89d3/Firmware/octopus-printer-rev-c.cfg#L75
VORON: https://github.com/VoronDesign/Voron-2/blob/a192410e27ea345644ae5c4b29b4c9c40cbe1a73/firmware/klipper_configurations/Octopus/Voron2_Octopus_Config.cfg#L48

[shrewd pawn]

good, now you have additional examples along with my actual config.

There's only so many ways one can say "I was supplied with this configuration, I haven't had to work with these settings before and am unsure what settings are relevant and which ones are being replaced by this new tool." gestures at the entire stepper-related section

[thorny lichen]

If I got you right, u dont know which changes u need to apply to your current config to accommodate the autotune prerequirements as well as what to configure to setup autotune. did i got that correct?

[half birch]

what motors are you using?

[shrewd pawn]

a/b: ldo-42sth48-2004mah(vrn), z: ldo-42sth48-2004ac, e: ldo-36sth20-1004ahg(9T)

[thorny lichen]

X and Y config is inline with autotunes requirements, no changes needed. You just need to add what i posted in my screenshot a few lines above with your currect motor model No 🙂

autotune knows all motors that are part of the database https://github.com/andrewmcgr/klipper_tmc_autotune/blob/main/motor_database.cfg
so once u know which X and Y motor u have, u can use CNTRL+F on that database to copy paste the correct name 🙂

[half birch]

add this to your config

[autotune_tmc stepper_x]
motor: ldo-42sth48-2004mah

[autotune_tmc stepper_y]
motor: ldo-42sth48-2004mah```

then tmc_autotune should just work

[thorny lichen]

well or Reth is so nice and provides literally the copy paste config 😄

[shrewd pawn]

😅 Yes, I get that, but what about everything that's already in all the existing various [stepper] and [tmc2209 stepper] sections?

[thorny lichen]

doesnt matter 🙂 thats as i wrote all fine with the requirements

its just copy pasting

[thorny lichen]

u just have to set interpolate to true 🙂

pins are contained in your config, run_current is also in the config and u do not use sensorless homing nor any register settings

[shrewd pawn]

and all the other motor settings like microsteps and rotation_distance?

[thorny lichen]

They stay as is

They are not mentioned in the README of autotune

If u would need to remove or do whatever to them, we would have told u 🙂

We are here to help you, not make your printer worse 🙂 😉

[half birch]

if you are using sensorless homing you will need to comment that stuff out and follow the directions in autotune for setting the values

[thorny lichen]

No sensorless :). i checked

[shrewd pawn]

I run into this issue in my job, where documentation is written by people familiar with the subject/terms to people who are also familiar; as is the case here, it leaves significant ambiguity for those who are not familiar or presented with the material for the first time--the absence of documentation is also a source of ambiguity. My interaction with the stepper settings to this point has been removing irrelevant blocks intended for other kit sizes, the changes for the G2E install, and adjusting extrusion distance.

[thorny lichen]

I hear you, sometimes the docs read like they’re written for people who already know the secret handshake. The great thing is, that’s exactly why this community is here. If something’s unclear, you can keep asking questions until it makes sense. We’re all happy to help, if we weren’t, we wouldn’t be hanging out here in the first place.🙂

Also, as someone from Germany who doesn’t speak English as a first language, I might not always catch certain American customs or phrasing. That’s why I really appreciate it when questions are clear and specific, it makes it so much easier for everyone to understand each other and help out.

Where I’m from, we like our communication the way we like our engineering: precise, clear, and preferably with a well-labeled diagram. That sometimes means I skip a bit of small talk and get straight to the point, not because I’m unfriendly, but because efficiency is kind of our love language. If I ever come across as too direct, please know it’s not my intention at all, that’s just the German wiring, and I truly appreciate your patience with it.🙂 👍

[shrewd pawn]

😆 "Theory Of Operation" has been replaced with "Just Do It."
new cold runs

[half birch]

can you loosen both belts by a quarter fo a turn

[shrewd pawn]

forgot to mention I put the door back on last run. on or off, the system *really *likes to hum from 50-60Hz

-# if it didn't mean figuring out wiring and mounting xy endstop switches, I'd go umbilical to get rid of most of the chain buzz

[half birch]

Can you tell me what your belt tension is at?

[shrewd pawn]

spectroid says 106Hz for both

[thorny lichen]

we have not seen a VORON Cube yet. please print it with the slicer settings mentioned in the Guidelines/how to post here 🙂

[half birch]

@shrewd pawn https://github.com/The-Original-Reth/Resonance-Stuff/blob/main/How to read input shaper graphs.pdf
it looks like you might have

1. a potential bearing issue page 25/46
2. bleed through on xy joints

[shrewd pawn]

here's what it looks like after just a few hours of heatsoak

[half birch]

Can you do all the test cold please. We can’t account for changes to the printer when test are performed hot

[shrewd pawn]

I haven't changed anything except temperature.

[half birch]

temperature changes everything. The heat causes everything to expand. Which throws off troubleshooting as it adds more variables.

[shrewd pawn]

(🧟 zombie thread revival) I've been poking and prodding as time permits...

some "fun" findings:
--the most-alarming vibrational noises were coming from the toolhead. Completely uninstalling the cable cover resolved the most egregious buzzing.
--the leads on my HF Revo hotend are too short and use a microfit connector. I made an adapter rather than remove the connector. I have a better soldering iron and ferrule/pin crimper now, so I can properly address another source of unnecessary bulk & buzzing.
--The Ti backers lifted up the middle cable chain enough to make light contact with the underside of the anchor--was not apparent until the resulting "sawdust" collected in the chain joints. Added a washer under the screwhead so the chain isn't auto-centered at the backer/rail end.
--The rail-mounted end of the toolhead chain had warped into the rail channel, raising the entire chain off the rail at an angle and making firm contact with the underside of the lowest parts of the toohead (cable door, E motor.) Used m3x8 screw and washer sandwich to fix the chain warp and skew.
--the vibra-tite (vc-3) has been working as-advertised--it was applied to every fastener on the gantry, from the z-carriage joints in to the toolhead and extruder assemblies; only the screws mentioned above might actually come loose. Even after a fair number of heat-cycles and plenty of vibration, everything is "gutentight" 🙂

These cable chains really are (right proper) annoying, and I've committed to a partial rev D conversion by ordering a Nitehawk kit, doing a toolhead umbilical along with sensorless homing (and reprinting the toolhead parts with Polylite ASA.) Also considering lightweight/simple options to arrest/dampen toolhead wire vibration, even if it's just a very poofy pipecleaner. This overhaul will also include swapping in an sd card with Trixie (current build is Bullseye). I'll try to do testing while most of the toolhead is removed for the process...

[shrewd pawn]

Trixie & sensorless didn't work out, so leverless limit switches are attached to the toolhead carriage--the cable door still buzzes, but the wires underneath are better-managed. Top hat, Nitehawk + umbilical conversion completed. Belt tension rechecked with Diyshift's belt tension meter (printed by West3D.) Felt strips added to suppress belts buzzing against extrusions. Now that everything broken by the klipper update is fixed, here's a proper update...

[half birch]

Go ahead and set the shapers to EI and lets run a test print

If the test print comes out bad, I am thinking it is your umbilical, and you might have to find a way to support it since you cannot use a galvanic

[shrewd pawn]

tbh, the VFA (belt/gear noise and edge ghosting) are still present, though I should probably revisit PA settings

[thorny lichen]

4 days have passed, whats the current state here? Any update for us? Do you need further help?

Graphs look fine, print looks good

not seeing too much for us here to work on, especially considering the name of this channel #1308016330326413393 "basic"

[shrewd pawn]

🤷‍♂️ Unless you know of a way to keep the cable door from buzzing badly at those low frequencies... (and yes it's securely fastened.)

[thorny lichen]

i only had SB for a few months, so i dont have much to comment on that.

question is if thats actually causing issues or not