Sovol SV08 Multiple Motion System Upgrade.

[dwuk3d]

Nema8 update - spent quite a lot of time yesterday (in between looking at Bambulab H2D videos) trying to get my improved Nema8 design to work.

Either the motors are getting weaker or my design has introduced more resistance - as the motors no longer seem to be able to lift the extruders despite a few changes to the 3d printed components and oiling up the linear rail.

Also managed to mess up firmware on Mini5+ and 6HC board again as I thought 3.6.0 Rc1+3 might be having some impact on Nema8 motor power -

Accidentally downgraded to 3.4.0 instead of 3.5.4 - which the Mini5 board doesn't like moving up from.

Had to manually update the Mini5+ again via SD card, and erase a reflash the firmware on the 6HC board too. Luckily I followed my own instructions on how to overcome the Mini5+ 3.4.0 upgrade issue earlier in this thread.

Same Nema8 lifting power issue with 3.5.4 though - so will move back to 3.6.0rc1+3.

I think I am probably wasting too much time on Z hoppers trying to keep the extruders as light as possible - will drop the Nema8's for now and try to over-engineer instead by trying some of the heavier more powerful motors and thicker support structures next.

Probably best to get everything working functionally and with good print quality up to at least the double IDEX phase - then worry about optimising weight as part of later speed tuning phase.

[o_lampe]

@dwuk3d The NEMA8s can be pretty quick (eg. 3000RPM vs ~300RPM of a NEMA17)
They also have a long leadscrew. You could place them vertically and use an 90° arm to translate speed into force (long leg /short leg ratio) The arms anchor point could be above the center of the package to reduce sideload to the linear rail.

[dwuk3d]

@o_lampe thanks i hadn't spotted that nema8's are faster.

I have been toying with the idea of doing some sort of lever based force amplification as you suggest - perhaps backed up by some sort of springs or counterbalance.

Will probably be more efficient if I set up some sort of off gantry force testing mechanism to better investigate the options.

One concern I do have with adding lever type mechanisms though is play in the extra linkages / pivot points.

Could also be an idea for a fairly interesting video comparing different linear motion options - with different motor sizes / gearing . movement translation mechanisms - to show how they stack up in terms of speed, strength and weight.

[dwuk3d]

Created short slideshow about the Bondtech INDX vs my Parallel printer - whether it is the answer for going beyond 4 colours depends a lot on the price....

https://youtu.be/PhH6jiCuCCA

[droftarts]

@dwuk3d said in Sovol SV08 Multiple Motion System Upgrade.:

Created short slideshow about the Bondtech INDX vs my Parallel printer - whether it is the answer for going beyond 4 colours depends a lot on the price....

Get enough views and they might send you one... or two!

Ian

[dwuk3d]

@droftarts Will be interesting to see if the video gets much pick up as it is certainly a very elegant design and could completely remove the need for me to add single nozzle multiplexing on top of tool changing - plus also remove the need for lots of electronics and motors etc. for the extra tools and multiplexer

Might do a better version of the video with a few animations of the different options next.

[droftarts]

@dwuk3d Yes, I really like the look of it, would be great on my Sovol SV08! Though I'm expecting it to be pricey.

Ian

[o_lampe]

@dwuk3d The way you mounted the rear INDX tool bar will not work, because it blocks Y-motion of both IDEX arms. You'll need overlapping motion to have full access to the center of the bed.
You could turn around the rear IDEX arm and put the INDX tool bar to the very rear of the frame....
There would also be a problem with the number of filament-holders you've places on each tool bar unless you want 4 filament holders for each extruder + the center one available for both.

Should Bondtech be aware of your virtual design, they might be convinced by an animated simulaton. If not, try to start a kickstarter campaign...

Edit: as a sidenote, if you had a moving bed (in Z) instead of a flying gantry you could even print while you change tools. That would be a huge timesafer

[dwuk3d]

@o_lampe thanks for the comments - you are correct about the Y axis issue.

My initial plan to overcome this is for both of my INDX gantries to have their own independent Z movement (a bit like the MissChanger). I was already planning to do this with the normal toolchanger option.

I could as you said reverse the 2nd gantry too - but I would like to try first having a tool gantry in a place where both Y gantries could occasionally reach. This would allow individual tools to be loaded onto any of the 4 print heads- which firstly think would look interesting in a video, but more importantly would mean the full build plate area is accessible for every filament colour.

In terms of the number of tools - I am aiming for 12 colours - so a minimum of 12 in total.

But I also want a few extra to allow some colours to be duplicated to allow them to be parallel printed.

Ultimately where I think I am aiming to get to is being able to demonstrate doing up to 12 colour prints of models faster than a single colour print of the same model on most other printers. Plus also have no waste other than a few prime towers. With the prime towers being shaped so that they can be stacked and joined to make useful objects.

The Bondtech INDX would help me particularly in keeping waste down - as before this I was expecting to have to do some single nozzle multiplexing for a few of the colours.

I've posted some stuff on the INDX discord - but Bondtech understandably seem much more interested in simpler, single head designs - Even if their prices are quite high I think I will purchase two of their extruders heads and a few Nozzles for my project in a few months,

Fitting a Duet 1lc toolboard into one of their print heads, plus a Z Lift motor outside is probably going to mess up the aesthetics of the INDX extruder toolheads quite a lot though.

I am planning to design some covers for my existing toolheads soon to make them look a bit neater - I think I might make them look a bit like the INDX!

[droftarts]

@dwuk3d Where did you find the CAD files for the INDX? Annoyingly, the Discord link on https://www.bondtech.se/indx-by-bondtech/ is not valid.

Ian

[dwuk3d]

@droftarts I have posted about the problem on the Discord and they say they will update the link with a permanent one - so hopefully it will be fixed soon.

Re the CAD - it's not to scale - I just knocked up a rough presentation of the components from a few screenshots from the video an the 3D Musketeers Podcast - so that I could play with different configurations.

[dwuk3d]

@dwuk3d Bondtech INDX Projected costs now out - $35 per tool head. Really happy with that.
$250 for the toolhead/extruder - a bit pricey - but worth it I think for the reduced complexity the solution brings.

Availability Nov25 - which is quite a while off unfortunately.

Tool change time 12 secs, with heat up only 4 secs.

Overall looks very promising.

I don't think I could justify adding 4 toolheads at that price to my system - but two I think would work quite well - one on each gantry - with the IDEX extruders on the gantries being single colour, or maybe with an MMU/Box Turtle type add-on (which wouldn't have to wait until November).

[dwuk3d]

Might have solved the problem with my Nema8's - which seemed to work before but then stopped working.

I first thought it might be my new improved design and secondly maybe the firmware.

With the help of Physics Teacher Anthony - who reminded me that I need to lubricate the bolt - that helped a fair bit and the lifting improved - but the motor still kept stalling.

The biggest issue though was an oversight on my part re motor speed. I was just using the Dashboard controls - and I'm not sure how to control the speed.

But when I do a G1 H2 A1 F100 for example it seems to works ok - so I think I was just running the motor too fast from the dashboard in my initial tests.

If I had just connected up the end stop, and ran the proper homea.g with lower feed rates- It probably would have worked ok.

Will still investigate over engineering the lift motors a bit - but a good step forward.

[droftarts]

@dwuk3d said in Sovol SV08 Multiple Motion System Upgrade.:

The biggest issue though was an oversight on my part re motor speed. I was just using the Dashboard controls - and I'm not sure how to control the speed.

You can change the feedrate in Settings > Machine-specific > Feedrate for move buttons (mm/min). However, it applies to all axis buttons, except extruder axes.

Ian

[dwuk3d]

@droftarts Thanks will try that.

I guess I should set the maximum feedrate for the motors in the stepper configs too - however the problem with that is gravity - i.e. They can go quite a lot faster downwards than up - and for Z hopping I should use the fully available speed in both directions.

[dwuk3d]

SV08 Double Gantry INDX option - with middle tool rack having independent Z

Tools can be moved between front and rear gantry.

Parallel Printing and Parallel tool changes

A bit messy - but probably what I might end up with -

INDX on 2 heads on both gantries - with IDEX too - using existing SOVOL heads, with option to offload both IDEX heads when not needed.

So this is making me think I will probably still do a tool changer upgrade on my printer if I complete the 4 head stage before INDX becomes available.

Some estimates of tool change/colour change times

INDX tool change is about 17 seconds in total including reheat and priming.

XL based on PrusaSlicer about 14 seconds (3 secs quicker).

Doing the same thing with Bambu Studio and an H2D - Intra Nozzle total change times seem to be estimated at 11.7 seconds (about 5.3 secs quicker). IDEX times likely to be similar

If you move all swaps to the same Nozzle then Bambu Studio estimates 76secs for a normal Multiplexing filament change.

Doing the same thing for an X1C comes out as 95 seconds per change -

An MK4 with MMU3 using the same method is estimating at 67 seconds per change.

Dual Gantry parallel INDX or Stealtchanger tool changes could be as quick as sub-second - and faster than Ratrig toolshift..

Parallel printing could then reduce single colour printing by appx 40% on dual gantry for some larger models, and by over 70% for symmetrical models on quad head printer.

[dwuk3d]

Don't really understand what is happening with my NEMA8 motor.

It started working pretty well with the slower speeds, and I created a sort of test bed - and it lifted 500g quite happily.

Then it stopped working - it might be electronic, the motor, or my mechanical design - I might be over constraining the bolt

Lead Screw Nema11 now arrived - the lead screw is a bit long - 50mm - but will try that - with the coupler for the lead screw not too constrained if I get any trouble.

This motor is rated at 1.0A - will be interesting to see how more powerful it is.

Nema 11 installed - sticks up more than it really needs to - might chop off the lead screw, or try and source one with shorter screw - but managing to lift 500g and the extruder so far

[dwuk3d]

Simplified Dual IDEX belt routing.

Decided it will be easier to put one 'Ratrig Hybrid' Y Axis motor above (for rear gantry) and below (for front gantry).

It will mean that the bed needs to be raised up a bit - due to the height of the bottom motor - but this will be needed eventually anyway when I move
to an occasionally moving bed.

Will have to redesign the extruder carriage belt couplings so that there is room one of the belts to fit through.

I think this should work well for the front gantry in particular as the belt fix point will be very near the Y axis rail carriage.

For the rear gantry they are further away - as the fix point is at the top - so there might be a slight pendulum effect.

Setup will be mirrored on the other side - could I suppose swap the top and bottom couplings - but I guess it will be better to keep both sides of each gantry the same.

[o_lampe]

@dwuk3d said in Sovol SV08 Multiple Motion System Upgrade.:

Don't really understand what is happening with my NEMA8 motor.

Their internal magnets might have degraded due to overheating? I noticed the same on some NEMA14 pancakes I used to use for extruders. They survived only when I reduced their rated current by 50%.
I replaced them with brushless gimbal motors and a selfmade simpleFOC-interface and the problem was gone.Torque/weight ratio improved a lot, too.

[dwuk3d]

@o_lampe Thanks - looks interesting - especially the gearing you have added.

I would like to investigate adding some nylon or metal gears if I can source some suitable ones.

I haven't really investigated driving normal motors from Duet - but it would be useful to be able to try something like your motors if I can control them ok with GCODE.

I switched away from servo's partly as they didn't play well with parallel gcode processing.

I'm wondering whether the type of Servo Motors that BambuLab have started using in the H2D extruder and AMS2 might be of use.

I guess the ideal solution would involve some sort of close loop feedback and temperature monitoring.

Think I will at least put some speed and reliability testing in my Macros - I think I should be able to home the ZHopper AXIS, do a few up and down movements and check the home point hasn't drifted due to missed steps.
Could also I guess use the same method to find the maximum move speeds up and down, plus also use the same method on my U & V axis with their optical end stops.
So far in my first steps my Nema11 seems to be able to handle F1000 moves up - which look pretty fast.