Posts made by dwuk3d

@o_lampe thanks good idea about doubling up - that slip ring has 6 connections and I only need 4 - so I can double up the heater ones - 20A 230v would be masses more than I need.

I haven't done enough actual printing on the SV08 to find any heat bed issues - but it seems to maintain temps ok - although I did get a failure with the PID tuning,

Point taken about the project approach.

One of the advantages of not having a sponsor is not being restricted to specific goals, dates or budgets.

But you might be right that there is a risk of me losing interest in the project - especially as I've just booked a QM2 cruise - so really want to do a model of that ship before I go - so that I can get some comparison photos.

So far all of my changes are just reordering things that were already in the plan (or when better technology like INDX vs Stealthchanger comes along).

On this subject though I am about to change around my bed slinger approach a fair bit to not have a longer bed - but instead just move about 150mm - and add the ability to link multiple SV08's with this ability together - to get 350x700 or 350x1050 for example - with the build plates slightly overlapping the heated beds.

Will probably do dual gantry - but only single heads on the second SV08.

That will then deliver my original plan (and most popular video so far) - see Prusa Multi Mini.

My immediate plans (which I may change)

1. Get 3 headed build working well - with fully parallel, faster printing
2. Rotating bed
3. Angled extruder on rotating bed
4. 150mm bed slinging
5. 2nd linked or independent SV08 with 150mm bed slinging
6. 4th print head - probably on stealthchanger to allow for offloading.
7. iNDX - which may slot in earlier if previous steps go past December.

@o_lampe the bed is already mains with an SSR - so I was planning on putting the SSR output through the slip ring, plus the temperature sensor.

Are you saying that would be ok - or that I should look at a better quality slip ring?

I suppose continuous rotation would only be a novelty feature - I wouldn't need a slip ring for 90,180,270,180,90,0 rotation - which is the main reason fir me adding it.

So I could drop it if a good quality one is going to be massively expensive,

I could always disconnect the cables and use a bambulab cool plate for continuous rotation demo's.

Think I might need a Roto toolboard for when I add an INDX to my front gantry.
Nb/ my picture of the INDX toolhead is slightly bigger than the real thing according to Bondrech.

Might get one now as it could also help with reducing the width on my rear right extruder too - although I suppose I could just rotate the 1LC upwards so that it doesn't take up so much space.

Quite pleasantly surprised to find that slip rings not as expensive as I was expecting - this one for example looks to me more than enough for my needs

Re -print head alignment / dimensional accuracy- I'm think I might get a second ball probe to have on the same arm - with the camera in the middle.

I could then probe each head 4 times - with the arm in two positions.

Should from this be able to get pretty accurate dimensional accuracy /adjustment info.

All of the probing points will need to be in the central area where all of the heads can reach,

Might even be able to do two lots of probing at the same time if I can make the heads as thin as possible in the X direction.

Will also switch over the servo to a stepper motor.

Youtube response about probing.

there is quite a lot more work required to get really good bed levelling.
At present it does 3 stages all only using inductive probes at the moment.

1. Initial Z homing on the front extruder inductive probe only.
2. 4xZ axis motor adjustment using front inductive probe only.
3. Rear right inductive probe based Micro Z alignment vs front micro Z alignment using the two inductive probes.

What I need to add is

1. Use of the inductive probe on the new 3rd print head
2. Use of the Z height values from the ball probe
3. Bed mesh probing from one or all of the inductive probes.
4. I also have an eddy probe for fast bed meshing which I haven't deployed yet.
5. I can also do 'voron tap' type probing by lowering each head to actually touch the bed and then use the optical sensor on each head. I did this for one of the heads - but it all got a bit too confusing - but ultimately I think that type of probing is going to be more accurate than the inductive probes.

Also all probing is currently one probe at a time - whereas i think there are some opportunities for doing some things in parallel.

@o_lampe I do a lot of my long prints without heat on a Bambulab cool plate - mainly to allow for long overnight or out of the house pauses. Bed adhesion for PLA is actually so good I can't use it for small fragile things.

I'm going to make mine heated though from the outset - and just use the standard SOVOL square build plate - perhaps with the corners chopped off if it won't quite rotate.

I notice now on relooking at the video that he has a very wide brim around the chimney - and some of his other prints look ok - so I wonder if the artifscts you noticed are due to wobble because of the very thing chimney.

When printing the sails of my very tall Sea Cloud model (see youtube) I had to add braces between the sails at a few heights to stop the wobble.

You are right about that filament hiding print quality defects - I was actually thinking of purchasing 3 rolls of it for my printer for when I get on to some big parallel prints - as I thought it might hide the joints a bit.

Re printing in the centre - I think it is possible - but just extremely slow. I got the impression that the non planer solution demonstrated is pretty slow - even for the outer parts.

Some inspiration for adding further features to my printer after looking at this very impressive video

https://youtu.be/M51bMMVWbC8?si=kyLhSQ2Y6XhIFs34

It does look very impressive and I can't imagine how that benchy stayed upright when it was just being held in place by that tiny chimney and filament was be added to the benchy in a horizontal direction,

I'm going to try and simulate his printer with his slicer in blender - should hopefully result in some interesting animations.

What interests me probably more than the non planer aspects - is the concept of printing onto the side rather than top of objects - which I think could bring a lot of efficiencies to multi colour printing.

As you only really need the colours on the outer surface - and so would be able to print most of the structure of an object in a single colour like white - and then colour it in with a few passes of the different colours applied to the top, bottom and sides.

Perhaps using a multiaxis printer like in the video, or maybe with something to rotate the print around in more dimensions instead.

I'm going to be adding rotary movement as the next stage of my multi head SV08.

This video inspired me to maybe allow at least one of the print heads to pivot up to a bit of an angle - rather than always facing downwards - and then use it to print onto the side of objects - rotated to the appropriate angle.

Another issue I have is with the probe moving a bit - I've changed a few things - but it is still not perfect. I have been considering changing from using a servo to a stepper - I think a fairly big Nema17 might hold the probe more steady in the Y direction or I could try a Nema23. The magnet is probably better at holding it in place in the Z direction. I might try also trying so somehow interlock something into the indentation of the magnet when it pulls down - to also prevent Y movement.

Re rotary bed
The next challenge after that is bed coverage by each print head.

My current thinking is once per layer rotate the bed 90, 180 and 270 degrees. - I will be rotating a square rather than round bed.

That should allow all 3 heads to access the whole bed, plus allow parallel printing.

If i only rotate it like that and then rotate it back again then I won't have any wiring issues - but I would also like to try making in continuously rotate capable like yours - so that I can do 3 headed vase mode prints.

I think i will need a 4 way slip ring for this capable of mains power - any thoughts on this.

I suppose I could go 2 way - just for the bed heater - and try using something like an IR temperature sensor for temperature - but not sure whether the PWM or whatever stuff would handle that.

Made a bit of progress on alignment issues - i think its a dimensional accuracy issue - and may actually be a problem with my Dual Markforged kinematics when moving in the X and Y direction at the same time.

Or might be due to the backlash compensation I put on the old V axis - which is now the X axis (W) on my third toolhead - which is the one that seems to be squashing up its dimensions a little bit when compared to the other print heads.

i need to take a break from this printer for a few days - so will address the issue later,

One big discovery I did make though while trying to resolve this - is that what I previously thought were strange hangs on things like M598's, M400's or M596's - which led me to completely stop using M598's were probably more like the fact that echo commands sometimes seem to stop working on forked multiple motion systems.

I put in loads of echo's before and after commands thinking I could see where control on each file was hanging - but worked out today that processing was actually still running in loops and commands quite a long way further into the file - usually on G4 500's in loops.

When this happen's M122 still seems to tell you the G4 or other command each file is currently on - so I have changed every G4 to have a slightly different millisecond delay time value - so that it is then possible from the M122 to work out which delay loop each forked file is currently executing.

Might start using global variables to trace current position - rather than echo's to see if that works more reliably.

Nb/ I am still on 3.5.4 - will test out some of this on 3.6 at some point too.

@o_lampe yes you are right about needing to sync layer times.

I will probably do something symmetrical like a sphere first - but to really demonstrate the concept I want to do a really complex non symmetrical shape, with layer times syncronised by slowing down some of the heads at appropriate times.

Will want to do this on normal segmented parallel prints too - to avoid heads going idle and needing to be cooled down, parked, primed and reheated as much as possible,

One of my earlier youtube shorts demonstrated the multi head vase mode principle.

Also Nathan Builds Robots massive printer does something like I want to do - although his design is restricted to only having in and out and up and down on each head - so would suffer from the 'centre' speed issues you mention.

@o_lampe Can't find the @CNCModeller post - although I suspect you are talking about rotary printers - with the rotary motion is one of the Axis - and when you are right in the centre you need to speed the spinning up very fast to get the same amount of motion as you would get from the edge.

I'm not proposing that - I am still proposing a normal XY type AXIS - just having rotary as mainly a way of moving parts of the bed to different places within the print head envelope to allow better coverage / more parallel printing.

Continuous rotation would be nice too - to allow things like parallel vase mode prints.

@o_lampe Thanks - yes some of the tools did start priming a bit early - I need to tweak my timing algorithm - its based on calculating the required delay within the 20 seconds preheat window, then just kicking off reheat and then prime as soon as heat up is complete. I should delay prime until the last few seconds. - also the 20 seconds pre-heat the slicer sends isn't always completely accurate - might just be a bug, or might be that I haven't got the printer completely configured exactly as its defined in the slicer in terms of speeds and accelerations.

I was really pleased though at how quick the tool changes are coming - will be even more apparent when I do prints with less of some of the colours per layer.

I have considered investigating simulations - but I might just try making the algorithm monitor itself and auto adjust - at least when it repeats the same print.

Yes I can't parallel prime when printing between the heads on the same gantry - and I do need to sort out some of the details - like lifting and lowering.

I will hopefully stay at this phase for a few months to get the parallel printing working really well, - especially with a combination of Independent and IDEX -

The beg slinger approach I am not sure about - it might be a step backwards - and does bring space issues - so might consider extending the whole frame - at least in the Y direction.

Thinking about getting full coverage of the bed for all print heads.

My current plan for the next phase is occasionally moving bed in the Y direction - like this
https://youtu.be/479ecxP-sOE?t=194

I'm now wondering whether to go straight to phase 6 - the rotating bed approach might work better for the next phase.
It doesn't even need to be infinitely rotating - as that would require slip rings etc for the wiring.

Could for this phase just have it rotate 90, 180 , 270 degrees and back which the wiring should be able to handle - and I think that would mean the whole bed would be reachable by all 3 print heads.

Biggest challenge would be ideally carrying on printing while the rotation is happening, plus where to put the probe.

https://youtube.com/shorts/rL_nntNSEIA?si=CgKk240gnOwPgb8z

@dc42 when I had the X axis on T2 as 7 - which was C, when I tried to move on the X axis it would move only in 0.1mm steps - and it went in the wrong direction.

If I tried to move in bigger increments it would stop.

Changing over to W not a big issue - I think it makes more sense having rear gantry as UVW - than UVC

I still have a gap between B & D - but doesn't seem to be causing any issues so far.

First proper three headed print.

Some work required on alignment, purging and speed - but quite pleased with results so far.

Gantry 1 and 2 are running on different motions systems - and very slightly in parallel.
https://youtu.be/4TYAep9QcUE

Three colour build area a little bit restricted - so need to get the print heads a little bit smaller...

NB/ Flag on printer is 70% of this scale - so could print a slightly bigger one.

Changed over C gantry to be W as missing out a letter caused problems with tools definitions.
Also put V motors on correct Axis

Kinematics now simply the following

M669 K1 ; configure CoreXY kinematics
;     X:Y:Z: U: V: W:A:B:D
M669 V0:0:0: 1: 1:-1:0:0:0

Rough scale of Bondtech INDX toolhead overlaid over my current SOVOL toolhead.

Almost ready to start trying homing.

Decided to home the left extruder on the right extruder - mainly so that can use an identical design for two extruders.

So right extruder will need to be homed first on the gantry.

Also doubled up on end stops again - so single optical sensor left extruder for U axis (X) (homing on to other extruder) , D Zhopper, then another single optical sensor for right extruder - C axis (X) and A Zhopper.

Will probably have to rationalise the axis letters at some point.

Used an M4 bolt again for lead screw on left extruder as need to order some more lead screw nema 11s.

Updated my CAD for INDX a bit based on latest pictures, plus created a Voron Zero version.

@dwuk3d All wired up

Starting on wiring up the 1LC on the last extruder.

Horizontal placement of the board above the extruder is not that neat - but I guess the for left hand extruder the overlap isn't a big issue.

For the other two extruders I made up adaptors - to make the upgrade reversible - I think for this one I will take a different approach - and cut and replace the connectors on any cables that are long enough - as likely to have two spare extruders once I do the INDX upgrade even if I end up trying a Klipper version later.