Not a Voron - VZBot resurrection

[mvdveer]

Converted a TronxyX5SA to a VZBot just over 2 years ago now.

 

Has been a good printer, but the build was not to the standard that I would expect it to be now. Also wanted to convert to the AWD version.  Some other upgrades - the list may grow:

AWD Gantry (2 x X-motors and 2 x Y motors) 

48V power supply in addition to the 24V supply

Moons Motors all around - replacing the Tronxy ones

Octopus Pro Board with TMC5160Pro drivers 

Beacon Probe

Rapido hotend (I have one ) or buy a Goliath hotend.

Single motor Z- drive. (The previous build had dual motors on the Z, synced with a timing belt.

Upgraded from 8mm to 10 mm smooth rods

Upgrading black metal screws/bolts to stainless steel.

 

So I disassembled the printer, reprinted all the parts I was unhappy with, and started the reassembly. Not much of a manual to work from. For that reason I would suggest building this printer from the CAD files. Another oddity is that the build uses M4x10 SHSC for most of the plastic to frame attachments.

Squared the frame - don't know how this printer managed to print functional parts.  Frame was out of square by 2-3mm.  

Progress so far.

 

Next - The Gantry

[Buurman]

Im jealous! looking forward to the build and any advise is appreciated since this is one build on the wishlist for me too!

[claudermilk]

The need to mod in this one is strong... 

[mvdveer]

Some progress, but frustrating. We do not realise how much work the Voron Design Team has put into the build manuals until you build a non-voron printer.  

Fitted the four motors for AWD. Tensioners build into the rear mounts. 

  

Now this is where the frustration kicked in - routing the belts. To say the instructions are not the clearest, is an understatement, but got it figured out in the end.

Just a tip for those doing the AWD version. To route the belts for the front motors, a partial disassembly of the drive units may help. Space is tight on the printed version. (From the CAD, it seems the CNC version has more space - @Penatr8tor may be able to verify)

I ended up removing the idler in order to see what I am doing and route the belt around the motor pulley. This idler completely hides the motor pulley.

 

Also note that when assembling the pulleys on the motors, the gears of the left front pulley sits topside whilst the gears on the right sits towards the motor:

 

 

Belts done.

 

 

 

Next: Toolhead

 

[cjkennedy72]

Watching closely... This is my next mod once I finish my Voron build, which should be by early next week. BMG gears due today, per shipping, and I am stuck replacing a cheap builders tub/shower with a nice corner shower install. I know... off topic, but I finished the demolition and cleanup yesterday. Today it's time to install the waterproofing and drain. Tomorrow, tile and the shower walls/doors. LOL

[Penatr8tor]

9 hours ago, mvdveer said:

Just a tip for those doing the AWD version. To route the belts for the front motors, a partial disassembly of the drive units may help. Space is tight on the printed version. (From the CAD, it seems the CNC version has more space - @Penatr8tor may be able to verify)

I didn't need to loosen and/or move components to install the drive belts on my CNC kit 330 so that probably was an issue that was resolved when they when to CNC parts. Of course, I have yet to install belts on a core-xy without spewing a good number of expletives in the process.  If I remember correctly, I did have to use a zip tie to help fish the belt thru a couple tight areas.

BTW... Upgrades are looking good. Excited to see your next update.

[mvdveer]

Toolhead.

Spent some time rebuilding the toolhead. Had a choice between the Rapido UHF or the Dragon UHF. Could not find a mount for the dragon, thus went with the Rapido. Also changed the shroud (Shonky Shroud) for one that can mount LED's like the stealthburner.

Will stay with the Bondtech LGX extruder but also have a choice of an Orbiter 2 or Mini LGX. (I am not really that interested in EXTREME speeds).

Belts attached to the toolhead and gantry squared.  Still have to mount the X and Y limit switches and the Beacon Probe. (The two Beacon Probes I have are the "normal" variant and I have ordered an unassembled probe to improve my installation options). The hotend cooling fan needs to be fitted.

 

Next : Finish up the switches and fan - then on to the panels.

 

[mvdveer]

Got some work done on the Electronics bay. Mounting at the back of the printer as intended by the designers of the machine. Ordered six 24V fans for the Mellow Fly TMC5160Pro drivers, only to find out it uses 12V. Waiting for the order to come in. Very slow going at this stage but doing a bit whenever I can.

This toolhead is a beast!

Whilst waiting, programmed the Raspi, Octopus Pro and Mellow Fly SHT36 Canbus. Had difficulty creating a USB-Can bridge on the Octopus Pro and decided to take the easy way out with a FLY_Utoc.

 

 

[mvdveer]

12V Fans have arrived and installed.

It is alive - and no magic smoke. (My fans are spinning on the 5160pro stepper drivers)  @Penatr8tor had me worried with his experience. Big sigh of relief when after an hour all was still well.

Now for the initial tests on movements, then tuning, enclosing the printer, tidying up the wiring, the umbilical and then testing speed on this thing.

[mvdveer]

In the thread -  Smol Printer Build: V0.2-S1r1 by @claudermilk,   @Penatr8tor said:"  I struggled all weekend getting the VzBoT configured to home/move in the right direction."   (Did not want to steal your thread Chris, which I love because you have your daughter involved. Kudos and father of the year to you! Thus the reply here)

Well, I will join that club of struggle. First had problems configuriing the Mellow Fly 5160 Pro drivers. Turns out the Octopus Pro board with the 429 chip has an issue where the max31865 chip interferes with the spi communication whether it is in use or not. (Unable to write tmc spi 'stepper_x' register GLOBALSCALER). Tried the TMC5160 BTT drivers with the same result. Nothing connected to the EXP 1 or EXP2 headers either. Also tried both hardware spi and software spi settings:

#HARDWARE
[tmc5160 stepper_x] 
spi_bus: spi1
cs_pin: PC4
#diag_pin: PA8       # for sensorless homing
interpolate: false
run_current: 1.9
sense_resistor: 0.033

#SOFTWARE
[tmc5160 stepper_x] 
cs_pin: PC4
#diag_pin: PA8          #for sensorless homing
interpolate: false
run_current: 1.9
sense_resistor: 0.033
spi_software_mosi_pin: PA7
spi_software_miso_pin: PA6
spi_software_sclk_pin: PA5

Gave up and installed TMC2209 drivers and got the machine to home. (Must say am impressed with the beacon probe) Finally got plastic flowing and printed the worse Voron cube EVER - will post pictures later when I get home. Layer shift city!. Checked all the bolts - tight. Checked belt tension - correct. Checked pulleys - tight.

Reprinted - same result. Only thing I have not tried is to slow the machine down, though was printing at settings I use on the vorons. (120ms speed). Might try to decrease the max acceleration in the [printer] section of the printer.cfg file.

[printer]
kinematics: corexy
max_velocity: 2500                      # MAY HAVE TO CHANGE THIS DUE TO LAYER SHIFT
max_accel: 10000                        # MAY HAVE TO CHANGE THIS DUE TO LAYER SHIFT
#max_accel_to_decel: 5000               # MAY HAVE TO CHANGE THIS DUE TO LAYER SHIFT
max_z_velocity: 20
max_z_accel: 1500                       # MAY HAVE TO CHANGE THIS DUE TO LAYER SHIFT
square_corner_velocity: 10

The question in my mind now - is it speed/accelleration, the belt path or the AWD conversion, or something else. 

Thus this slow build has been made even slower. Tasks at hand:

1. Fix the layer shift - Changing the speed and acceleration settings to see if that fixes it. If not, it means stripping the belt path, check for possible faults. (disassembling the four motor drives). Once that is fixed,sorted and printing without layer shifting, then:

2. Re-installing the Mellow fly 5160Pro drivers and applying a fix I found to see if it solves the communication error. The fix basically disable the max31865 function on the octopus pro;

 

[static_digital_output disable_max31865]
pins: !PF8

 

Obviously will change the pin to the appropriate pin on the board. But then if the layer shif is fixed by slowing the machine, what is the point of the pro drivers?  Mmmmhhhh

Don't you just love this hobby? Especially when wifey askes why oh why?

[Penatr8tor]

Sorry I'm so late to the party. @mvdveer, Sad that you had issue with 5160's and a bit relieved to find other humans having issues just like me LOL. Your VzBot 2.0 is looking really good though. I haven't setup my beacon probe yet. Going to do that tonight, then I have the right hand RSCS fan and panel to mount, a web cam and klipperscreen and then I move it to my office for much tweaking and experimenting and dialing in. Before the dialing in I'll snap some photos and finish up the build diary.

53 minutes ago, mvdveer said:

Don't you just love this hobby? Especially when wifey askes why oh why?

Because you get the equivalent feeling of accomplishment from solving the problem as the problem gave you in frustration.

Hahahaha

[mvdveer]

Layer shift City:

   

Worst Voron Cube ever.

[Razor1666]

The layers themselves don't look terrible... Once you fix that one thing, the layer shift and speeds, two....., two things, the layer shift, speeds and pressure advance, three......, three things....lol

Love this hobby! Hahaha

 

[claudermilk]

13 hours ago, Penatr8tor said:

Because you get the equivalent feeling of accomplishment from solving the problem as the problem gave you in frustration.

Hahahaha

This. So much this. This was me last night after seeing two serial IDs:

[Penatr8tor]

@claudermilk That's the number one message I give to young people.

I say,

Everything in this universe is bi-polar, meaning, everything has an opposite. Hot cold, light dark. left right, good bad, hard soft, and so on ad infinitum. So, if you're looking to get the good feels without first going thru the struggle... Good luck, it ain't gonna happen. You can't have cold without hot to compare it to and you don't get the feeling of accomplishment without first putting in the work. No matter what you try... You just can't get around it and you can't fight against it. It is the way the universe is set up, understand and accept the reality we live in and learn to work within its boundaries. 

[mvdveer]

The search goes on. Slowed the machine down - no difference.

Stripped the belt path, re-checked screws, re-assembled - no difference.

Interesting that all 4 Voron cubes printed (each time after an intervention), all have the layer shift at the same points. They look identical.

Something weird if going on.

Come on Sherlock - detect!

[atrushing]

Are your A/B motors getting hot? I see you have 1.9A set for the 5160s. I'm sure they are fine with the amps but maybe the motors aren't quite up to it. The fact that you are loosing steps on both axes also says something but I'm not sure what..

I tried running 5160 drivers on my Trident but it was a real headache so I put them aside for the time being.

[mvdveer]

53 minutes ago, atrushing said:

I tried running 5160 drivers on my Trident but it was a real headache

They are more than a headache, could not get the SPI bus to reliably work (interference by the max31865 chip on the Octopus Pro 429) - I switched to TMC2209 steppers running at 0.8 A to test printing. Once I get the printer running as it should, I will revisit the TMC5160 pro's . My current configuration.

[tmc2209 stepper_x]                     
uart_pin: PC4
interpolate: false
run_current: 0.8
sense_resistor: 0.110
stealthchop_threshold: 999999

Done input shaping and whilst I know just about nothing on interpreting the graphs, they do not look too bad. (There may be a mechanical problem on the X-Axis though I could find none)

 

1 hour ago, atrushing said:

The fact that you are loosing steps on both axes also says something but I'm not sure what..

Glad it is just not me - have been racking my brain for a solution.

Now thinking my filament path may be an issue - going to change that tonight. It has quite an acute angle where it exits the machine at the rear left.

It that doesn't work, I am going to disable the two front motors and run conventional front idlers to see if that makes a difference.

[atrushing]

6 hours ago, mvdveer said:

Done input shaping and whilst I know just about nothing on interpreting the graphs, they do not look too bad. (There may be a mechanical problem on the X-Axis though I could find none)

I don't know much about the graphs either but at least the spikes are pretty clean. I have seen noisier graphs so a mechanical problem wouldn't be my first guess.

Good luck troubleshooting!

[mvdveer]

On 9/2/2023 at 11:18 PM, cjkennedy72 said:

Watching closely... This is my next mod once I finish my Voron build,

Just a pointer before you start - the manual is a shocker. Do yourself a favour and read both the AWD manual for printed parts AND more IMPORTANTLY, the manual for the AWD All Aluminium. The latter has so much more information. And the solution to the problem of my layer shift city. More on that later - waiting for a Voron Cube to finish printing.

[mvdveer]

Problem 1 Solved:

No more layer shifting. AWD was the issue. Motors were NOT synced. Thanks to burgo_enemy_of_thelettuce  on VZBot discord for pointing me in the right direction.

Motors were not synced - quite important as you have two motors driving X and two driving Y. I followed the printed manual version which does not even mention syncing the motors. I know, I should have thought of that - science and all. The AWD aluminium version of the manual however does mentions it. 

@cjkennedy72 when you assemble your pulleys on the motors, be sure NOT TO align the pulley grub screw with the flat surface of the motor shaft as this will prevent the motors from syncing correctly. I learned that the hard way - Frustration Galore!!! - AND - LAYER SHIFT CITY

Synced the motors, sliced a Voron cube in Orca slicer, Printed AND..........

    

These were printed at the VZBot settings - FAST (250mm/s and 10000mm/s acceleration) Will slow it down for better quality and until I can get the 48V 5160 pro drivers working. 

 

 

Not perfect but a major improvement.  Dimensionally correct. At least now I can start to fine tune the machine and continue the build - fit the panels and RCFS, then work on getting the 5160 Pro divers to work. Have not even touched the Ellis guide - on to that next to see what I can improve. Looking at the cubes - the first layer too close, still some layer shift, pressure advance needs to be tuned. Re-look at the belt tension, especially Z-axis, re-run input shaper after tensioning belts.

 

 

[mvdveer]

Jumping the gun - but going to convert to CPAP (Wife says I need this because I snore!) Have received the fan and tubing. Will be printing the necessary parts and start the conversion. So many MODS!!! Please STOP. Noooooo - too much fun!

[mvdveer]

Benchy - Done (no tuning - 45min)

  

 

Enough play - let's get serious - Tune the damn machine will you

[mvdveer]

I was also told on discord that I had a lot of red flags for a VZBot:

1. Octopus Pro v1.1 - bad board

2. SHT36v2 - voltage problems

3. EVA3 toolhead - outdated - apparently 2.4 is better

4. Rapido hotend - will fail

Ah well, too late now - lets see how it performs with the red flags

(Not really worried - more a tongue in the cheek post)

[claudermilk]

Now for the sense of accomplishment and moving on to the next step!