#V2.4 300 AWD with lot of open topics

109 messages · Page 1 of 1 (latest)

frail hawk
#

One happy sunny day, during a usual G28 I discovered an annoying noise in my printer moves (see video attached).
During my research for the root cause of the noise I conducted the voron community and discussion went from identification of broken bearings (due to too much tension),
wrong tooth idlers and defect rails (due to wrong grease), to a slicer profile that would make every snail proud.
It looks like I have made a lot of mistakes and would like to correct these and reach higher print speed & quality (preferably one that matches with an AWD build)
As spitzbirne recommended during the discussion, I need to take a broader view on my printer health.

Voron_Stripes Printer Family & Size: Voron 2.4 300 ~ 1400 print hours

g2gfast Which ADXL was used: mellow345 NozzleADX
-> I have ebb36, cartographer & mellow345 NozzleADXL available.
Which one should be preferred?

🔬 How are you measuring absolute belt tension - by feel, app, or tension tool (which tension tool)?
-> Blue Digital gauge at target value 8.0 (higher than ususal due to AWD; value from some AWD chat)
--> I have just printed and assembled PfMakes GT2 Tension Meter, which as i understood should be used to tenstion at ~ +3mm after 3 with AWD drives
(not done yet for shaper pictures)

📷 Pictures:
-> find attached (I do not have an exhaust cover; so I added the bowden pass through picture instead)

🖨️ Current Slicer settings as screenshots - Speed and acceleration
-> find attached

0000_beltscomparison_20251209_200544.png 0000_front-idler_left-back.jpeg 0000_front-idler_left-front.jpeg 0000_front-idler_right-back.jpeg 0000_front-idler_right-front.jpeg 0000_full-front.jpeg 0000_full-side.jpeg 0000_KlipperAddOns.png 0000_KlipperVersion.png
#

V2.4 300 AWD with lot of open topics

#

🛠️ List of all relevant mods and options:

  • base was a formbot kit
  • aTinyShellScript AWD; CNCParts - 6mm belts
  • all X/Y = 24V Moons MS17HD6P420I-04 1,8° stepper with TMC2209 stepper driver
  • DoubleT CNC XY joints from funssor
  • CNA/RDBB rails Z1/medium preload on X & Y
  • XOL (2024) Toolhead with WristWatchG2 (WWG2)
  • TheFilter direclty under bed
  • Ellies BedFans left&right of bed
  • Cartographer CNC Mount with custom 6.5mm stainless steel spacers to match XOL & Dragon UHF size
  • CanBus for Ebb36, Cartographer, MellowFly ERCF
  • Zero offset Clamped GE5C Z-Joints
  • 5mm Panel left, right, back, front, top
  • Top Panel PTFE/Bowden Passthrough
  • Annex TradRack MMU (not directly attached with printer frame)
  • stock voron feeds have been replaced with high quality, screwed magnets (printer cannot moved up without the whole metal plate); these magnets stick to a metal plate slightly bigger than the printer.
    This metal plate has a weight of ~120kg. Below the plate there are foam mattress parts to disconnect printer+metal plate from desk. (desk is screwed into wall)
    Rational: ensure the printers center of mass is as deep as possible.
0000_ptfe-exhaust.jpeg 0000_toohead-left.jpeg 0000_toolhead_right.jpeg 0000_voron-cube_1.jpg 0000_voron-cube_2.jpeg 0000_voron-cube_3.jpeg 0000_voron-cube_4.jpeg 0000_xy-left.jpeg 0000_xy-right.jpeg
#

Some Warning about me:

  • I absolutely appreciate and respect the knowledge and experience you guys have and share here.
    Thus if I am asking "why should I do A/B/C?" - please do not take this as an offend or contempt of your qualification.
    I am asking to learn and raise within the shadow of you wisdom.
  • SysAdmin with small number of dev skills - means if you aks for adjustment from software side -> feel free to just throw it using a few short words
    But if mechanical adjustments required ... well tread me like you would try to explain to your toddler
  • Unfortunately due to private reasons the time I can afford for this is very limited - mostly not more than about one hour per day.
    So please expect process to take time.. unfortunately 🙁
  • While there is a second printer V0 from printing quality point of view it's better to assume it does not exist. - Maybe one day I will find time for my little friend.
    But at the moment the maintenances I do have to be planned proper in advance to not bring myself into trouble with need of a rescue raven 🙂

So the tasks I have already identified (and have according questions) from reading though other help resonance_basic_help topcis here are:

#
  1. Replace X & Y Rails
    ISSUE:
    • as said they are defect due to wrong grease (NLGI Class 0).
      STATUS:
    • New grease "Fuchs Lagermeister BF2" was already delivered & attached to current rails.
    • New X & Y Rails have beend orderd - waiting for delivery (again CNA/RDBB Z1/medium preload)
      QUESTION:
    • no question here; feel free to add thoughts
#
  1. Replace wrong tooth idlers
    ISSUE:
    • The powge 20T tooth idlers at front ATSS mounts are too big.
      They slightly rub wit the Z-belts (see pic).
    • Based on current knowlege (proof me wrong) I would prefer not to change tooth idlers with bearings.
      As of feedback I read throughout the community it seems a bit like gambling thing, if belt teeth on bearings cause VFA or not.
      STATUS:
    • After discussion witin the community the idea is to replace the front 20T ones with 16T ones.
      These have smaller diameters and as long as XY-joint stay on 20T no issue should appear.
      16T ones have been ordered at powge.
    • Current 20T (front & xy joints) seem to have very little space at top & bottom.
      I have ordered 3x5x0,5 ; 3x5x0,3 ;3x5x0,1 stainless steel shims to add them below & above the 16&20T as needed.
      QUESTION:
    • Any reasons agains the 16T?
    • Is adding shims to not allow the tooth idlers go up&down a good or bad idea?
0300_Z-rubbing.png
#
  1. X/Y joints clarification
    ISSUE:
    • My input shaper graphs are not as good as I would wish
    • As I understood from other topics the plastic/ABS printed ones can bring better results as they seem to decouple X & Y better.
    • with my PIF (eSUN ABS) printed XY-joints I was not really happy.
      They had to be tighten every 2nd or 3rd print. I knew at the time already about the tightening instructions
      (there was a picture more/less equal to reth github "XY joint screw guide.jpg" in voron discord pinned already)
      STATUS:
    • still in decision taking.
    • I think powge 20T idlers are too big (as they have bigger shaft) for printed parts
      QUESTION:
    • I was thinking about 5mm ABS printed part (with 42% gyroid infill; 3top; 3bottom layers; 4 perimeters) in size of the rail block to be added between the Y rail block and Funsor CNC part.
      & do the same with a printed part same size like the CNC joint and height between CNC part & X exrusion & bearing/idler shafts.
      The rational would be to have a printed part which is not "pressed" by screw only but full metal plate and thus too much/too less tensioning of the printed part woul d
      https://github.com/3DPrintingMods/VoronTrident-AluminumG
    • If going back to printed ones -> toughts about the Pins mod? https://mods.vorondesign.com/details/C8XJJ0fBGfBFKQZctjKeA
      Any knowledge if they can be used with toothed idlers instead of bearings?
0400_xy-spacer.png 0400_xy-spacer_bottom.png
#
  1. Carto CNC mount, XOL & Ebb36 mount
    ISSUE:
    • as of my understanding from other topics here in teh CNC mount + XOL are not optimal combination
    • the ebb36 toolhead board mount is known to possibly to "wiggle" around with this
      STATUS:
    • At the moment I am able to screw the ebb36 mount with 3 screws into WWG2
    • The XOL can be mounted using 4 screws to carto CNC Mount
      -> 2 lower behind the fans from front
      -> 2 upper at the wwg2 mount from front
      -> 2 from the back into wwg2 mount (where tap sensor is sitting with tap)
    • I designed an ebb36 mount which screws 5 times into the WWG2
      -> existing 3 ususal scews
      -> 2 additional lower scews at the back (where tap sensor is sitting with tap)
      These 2 screws go from ebb36 mount wth spacers through cnc mount into the wwg2 mount
      QUESTION:
    • would it make sense to drill 3mm through the 2 cnc mount holes (where tap sensor is sitting with tap) to remove thread and make them a "go through hole" only?
      Rational behind is that at the moment there is a thread wihtin the cnc mount & a (heatset insert) thread within the WWG2 mount.
      So with two threads after each other it will never be really possible to drag both (3 with the ebb36 mount) parts together.
0500_ebb36-mount_cad.png 0500_ebb36-mount_print.jpeg
#
  1. PTFE bowden & toolhead canbus cable path
    ISSUE:
    • as of "Umbilical Routing for Trident and V2" from Reth there are two issues with my current routings:
    • PTFE is going to top panel
    • canbus cable is going to A-Drive with PG7 which is known as one potential source of bad IS
    • I do not have panels with exhaust fan
      STATUS:
    • I want to go through back panel with both (bowden & canbus)
    • for top panel I designed a simple screwed covering cap which will replace the top ge4c passthrough
    • I will need to drill two holes into the back panel
    • details about solution still in design thinking (how to mount / where to drill etc.)
      would use a PUG connector for CanbusCable
      QUESTION:
    • as I need to screw holes anyway - I have seen exhaust covers with GE4C like bearing on bowden path (e.g. https://www.printables.com/model/958818-pug-and-pivoting-bowden-exhaust-grill/files)
      is this useful or would not be recommended?
    • If no moveable GE4C bearing - is it better to go straight or in some angle (30°/45°) through the back panel?
    • Is a Pneumatic connector / ECAS4 recommended for PTFE pass through?
    • as no holes have been drilled yet - is the middle of panel the best place or better at the back right/left corner ?
    • is this arm ( https://www.printables.com/model/978123-voron-bowden-ptfe-tube-guide-arm-and-canbus-cable ) recommended over a badge retractor solution ( https://mods.vorondesign.com/details/avSsiCFH7Vte3kYhqt4atA )?
0600_PTFE-CAP.png
#
  1. front doors:
    ISSUE:
    • my current doors have a very small space between each other (even tough my chamber temperature seems fine to me)
    • I see no way how to close that little gap except using a bar in the middle (which causes a less good view on the prints)
      STATUS:
    • I am using split doors like the stock voron, but 5mm panel, screwed hings & no magnets.
      Instead of magnets I am using Annex Panel Clips to open/close the doors
    • I think about replacing them with single panel (even though I liked the two half doors)
    • I am not a big fan of clicky clack door
      QUESTION:
    • can the split doors with annex clips have negative impact on IS graphs?
0700_front-door-space.jpeg
#
  1. Slicer profile from InputShaper
    ISSUE:
    • think my slicer profile does not really match the repupation a typical voron has
      STATUS:
    • not fully understood Reth "Portland Vice - Create a Slicer Profle.pdf" yet
    • might need to investigate more deeply
    • would focus on this as soon as hardware & input shaper side is done
      QUESTION:
    • pretty sure there are a bunch of questions to come.. sorry 😛
0900_slicer-settings.png
#

-> due to issue https://github.com/Frix-x/klippain-shaketune/issues/259 with "Axes Shaper Calibration" command I cannot create input shaper graphes with shaketune.
-> I will try to create with klipper shaper_calibrate command and ~/klipper/scripts/calibrate_shaper.py tomorrow

frail hawk
#

As promised here are the shaper graphes and my current printer.cfg
(looks like i have applied wrong axes map x,z,y - default xyz would have been fine)

Think that's now everything one needs to know about current status 🙂

0000_shaper_calibrate_x_20251210_103115.png 0000_shaper_calibrate_y_20251210_103115.png 📎 printer_20251210_110000.cfg
cursive root
#

okay this is going to take me a bit to go throuhg

#

first can you install tmc autotune and shake tune.
Then rerun the graphs including compare belts and input shaper in shaketune
For testing purposes I would use the ebb36 adxl

#

In addition I saw that your xy joints are upside down. This may be an issue when it comes to belts and input shaper. Lets run the test first

#

I would also get gates toothed idlers. They do not have as large a flange as the ones you are using and are less likely to hit the z-belts

frail hawk
frail hawk
cursive root
#

what is your goal?

#

if it is to diagnose mechanical isses then you want to use the ebb36

#

if it is to try and see what the nozzle is doing during printing then you might have luck with the nozzle probe.

#

we do not diagnose issues here with the nozzle probe. This is becuase the nozzle acts like a pendulum and can create multiple peaks

frail hawk
# cursive root what is your goal?

ok - as I have to identify mechanical issues first ebb36 adxl is the better choice
do I understand correctly that (at a later stage) after no mechanical issues are visible and slicer profile can be created the nozzle probe is the way to go (because then I want to know what the nozzle does during printing)?

cursive root
#

That is a decision in the later stages you will need to make for yourself.
I do not use nozzle probes for any of my printer. This is because in my experience nozzle probes tend to create double humps which throws off the math for input shaper

frail hawk
#

klipper downgraded to v0.13.0-403 - here are the graphs with ebb36

IMG_2290.png IMG_2289.png IMG_2287.png
cursive root
#

what is your goal

frail hawk
#

Several goals:

  1. I would like to identify mechanical issues
  2. correct these & mistakes that are non-IS-friendly I have made during my build/mods
  3. create a printing profile that best matches my printer in regards to speed & quality
  4. learn during all of these about the printer itself
cursive root
#

I am not seeing mechanical issues in your graphs.
Since you are using AWD it is not unheard of to get the 125hz peak. This is generally due to an unblanced toolhead

#

can you set both shapers to EI

#

then run vibrations profile up to 300 mm/s

#

just to verify you stated your belt tension is about 2-3mm past 3.0 on the pf meter?

frail hawk
cursive root
#

This is the reason you y-axis is so low

#

you need to go 2-3mm above 3.0 for awd

#

if your motors are double sheered

#

if they are not double sheered then you are not getting the most out of awd, cause you cannot tension them more than say 2.0 on the pfmeter

frail hawk
cursive root
#

yes

#

it means a support on the top of the motor shaft

#

this is becasue the motor shaft already has support in the motor

frail hawk
#

hm.. okay - then I have the wrong motors as my shaft is to short for implementing this

cursive root
#

you are getting 7k accel for y with EI. which is about a 3k improvement over 2wd.

#

however, with double sheer you can increase the belt tension as described above, and likely get 10k accel

frail hawk
cursive root
#

default

frail hawk
#

ok - here is the vibrations profile

0000_vibrationsprofile_20251211_223313.png
cursive root
#

you printing abs

frail hawk
#

yes - only

cursive root
#

try for ext perimeter speed of 200+

#

what hotend you using

frail hawk
#

Dragon UHF - Undertaker 0.4 nozzle

cursive root
#

what is your linewidth

#

and what layer hieght are you using

frail hawk
#

for external it is 0.4
0.2 is the layer height (0.25 for first layer)

image.png
cursive root
#

I would set everything at 200 mm/s

frail hawk
#

you are calculating volumetric flow?

cursive root
#

your infill needs 25.6 mm^3 at 200 to print

#

your exteriors need 16 mm^3

#

so your external perimeiter can handle the flow, but your infill may not

#

so set all speeds to 200 mm/s

#

I do not know what the max flow is on your hotend, but I am assuming it is around 24-30

frail hawk
#

so I am using 7K for the external - 7500 for internal
All speeds at 200
correct?

image.png
frail hawk
#

here is the voron cube with above settings

IMG_2321.jpg IMG_2319.jpg IMG_2318.jpg IMG_2320.jpg
cursive root
#

Can you post the speed view from the slicer please

#

The cube looks good except for the bottom and top of the reverse y-side

#

You are also going to need to adjust pa/em with the higher speeds

frail hawk
cursive root
#

a picture like this

image.png
#

with the picture being taken on the reverse y-side

frail hawk
#

sorry - i was without pc acces till now 🙂

image.png
frail hawk
#

very big thank you at this point - I already learned that much from your speak & repo 🙏

cursive root
#

what I have started to do is to get my speeds and accels. Put it all in the slicer. Use a previous PA and EM and then yolo the print.
I then look at it, and see what corrections need to happen.

#

if the PA/EM was set with slower speeds I then correct those.
What I find when I adjust PA is some of the VFA will disappear.

#

if there is no bad resonance, I will then play with the speeds

#

and keep doing this, making adjustments till I can't get it to look better

cursive root
#

do you have slow down for layer time set?

frail hawk
# cursive root what I have started to do is to get my speeds and accels. Put it all in the slic...

yes - PA & EM are from previous settings
what about SCV / jerk - we just kept it at 5 during tests - should I try to get higher there (as long as accels dont't go down)?
From earlier recommendation within the community the following custom gcode in pic attached is active which sets the SCV higher for internal infill & gap infill to speed up the print (not really sure why exactly 12 ... not 10 or 15 or 16)

image.png
frail hawk
frail hawk
cursive root
#

when your scv starts to get hihger you have a greater chance of having rounded corners

#

So with layer time on as previously set you were not in full control of the speed.
This could be the cause of those lines on the top and bottom

#

you might also want to thikn about moving to a more recent slicer such as orca slicer, prusa slicer, or prusa slicer BOSS

frail hawk
# cursive root you might also want to thikn about moving to a more recent slicer such as orca s...

I will check out prusa slicer - at the moment I am on SuperSlicer 2.7.61.10 (which is a beta version)
... with MMU there are some additional things I have to take into account for switch of a slicer

May I ask about the rational behind your recommendation?
I agree that there was no process for a long time and there ar estill a lot of things to catch up on - but susl is a known evil for me at the moment 😛

frail hawk
#

PA definitely needs some love 😂

IMG_2328.jpg IMG_2327.jpg IMG_2326.jpg
cursive root
#

it has been many moons since I used superslicer. I think if you want all the current knobs orca slicer is leading the pack

#

@sage siren what are you thoughts on slicers these days

#

@frail hawk in your view are these current prints better than the previous prints?

sage siren
#

I don't know much about new slicers. I use Superslicer 2.5.59.13 version since that one does what I expect. I have not tried other slicers.

frail hawk
sage siren
#

But some nozzles have that effect even with good PA

frail hawk
sage siren
#

Yup

#

There it is

#

You need to lower PA but it will never be perfect because those nozzles are like that

frail hawk
#

ellies pattern is still the best way to tune this?

sage siren
#

With 0.001 steps

#

But you will always have inconsistent PA with Undertaker

frail hawk
sage siren
frail hawk
#

@cursive root - how exactly do you reach this screen (screenshot from portlan vice talk)?

I checked pursa slicer 2.9.4 but do not have this extended option (with jerk & cruise ratio) for accell speeds (even not in expert mode)

image.png
cursive root