So ... 3d Printers....

[nitrogen_widget]

This is the hotend i've printed up.
https://www.thingiverse.com/thing:4236354

Here is the dual drive extruder.
https://www.amazon.com/dp/B094HRCSYF?ref=ppx_pop_dt_b_asin_title&th=1
Here is the smaller stepper for the extruder.
https://www.amazon.com/dp/B07TY4BFF2?ref=ppx_pop_dt_b_asin_title&th=1

I never saw a direct drive's components so was very confused at first on how the hotend heatsink mounted to the printer with me removing the the mount from the chiron.
I feel stupid now because i stared at the extruder and the heatsink and then the bulb went off and it was easy.

The new stepper is so small next to the original stepper.

after i got it all assembled i realized i forgot to press the nuts into the inside so had to take apart and do again. this is where the new bracket attaches to the carriage via the original hotend bracket.

painted up and almost ready to go on.
I put some rubber cement on the part cooling fan because even though there isn't much resistance i felt some kind of seal should of been between the fan and the part cooler.
I'm waiting for some heat sink compound to arrive since my last bit was rock hard. i want to use it on the threads for the copper heatbreak going into the heat sink.

I just printed up a new heatsink for the V6 that uses a standard 40mm fan.
I can't find the cooler that came with it. eh, it was 30 mm anyways.
painting that now.

yes i know these prints looks awful finish wise.
nobody will ever get that close to the hotend so i don't need a show piece.

now i need to create a part to mount my junction PCB to the carriage.

[BadMouth]

yes i know these prints looks awful finish wise.
nobody will ever get that close to the hotend so i don't need a show piece.

Looks as good as what I've been getting recently.   

I've been tweaking the Voron 2.4 trying to get more speed out of it, but am coming to the conclusion that the materials I print with (ABS & ASA mostly) just don't like to be printed fast. 
Even if I manage to print a layer fast, the print goes to crap if the minimum layer time isn't long enough.  This pretty much cancels out any increases in print speed when printing a single object that isn't huge.  Part fan cooling may make the parts look better, but will result in weak parts.  60mm/s walls and 150mm/s infill seems to be as fast as it wants to go and still produce decent looking prints.  Acceleration max is only set to 3500 because when I did the initial setup, whatever test results dictated that print quality decreased over that.  I've run the full setup including input shaping using an accelerometer. Someday I'll do all that testing again I guess. 

I got some Black Polymaker ASA and printed a bunch of tests to dial it in.  Everything looked great until I printed an actual useful part (fishing rod holder).  I dunno why the valleys look like crap.  Maybe having ironing turned on messed with the valleys.  The (supported)overhangs underneath look really bad, but they could be designed out of the part.  Seams always look bad.  Guess I need to spend days reprinting the same part to figure it out. 

[nitrogen_widget]

yes i know these prints looks awful finish wise.
nobody will ever get that close to the hotend so i don't need a show piece.

Looks as good as what I've been getting recently.   

I've been tweaking the Voron 2.4 trying to get more speed out of it, but am coming to the conclusion that the materials I print with (ABS & ASA mostly) just don't like to be printed fast. 
Even if I manage to print a layer fast, the print goes to crap if the minimum layer time isn't long enough.  This pretty much cancels out any increases in print speed when printing a single object that isn't huge.  Part fan cooling may make the parts look better, but will result in weak parts.  60mm/s walls and 150mm/s infill seems to be as fast as it wants to go and still produce decent looking prints.  Acceleration max is only set to 3500 because when I did the initial setup, whatever test results dictated that print quality decreased over that.  I've run the full setup including input shaping using an accelerometer. Someday I'll do all that testing again I guess. 

I got some Black Polymaker ASA and printed a bunch of tests to dial it in.  Everything looked great until I printed an actual useful part (fishing rod holder).  I dunno why the valleys look like crap.  Maybe having ironing turned on messed with the valleys.  The (supported)overhangs underneath look really bad, but they could be designed out of the part.  Seams always look bad.  Guess I need to spend days reprinting the same part to figure it out. 

I would flip that part on the long end to print.
but i'm guessing the velcro run inside it might cease to exist if you did that.
the fact you were able to get velcro to work with a print like that is amazing.
sometimes i can't even get nuts to sink properly.

I found a profile for pet-g on the vyper and it's printing amazingly. hardly any stringing and high quality.
printing loot studios swamp shack made for resin printers in pet-g and while annoying to print so many little parts to glue together it looks good.

I had to print some 30mm to 40mm extender adaptors in order to clear the mounting plate.
i'm doing a 35mm and 40 mm extender.
for the second time.
I didn't realize how thin they printed and broke both trying to remove supports so a reprint with different supports is happening now.
i'm building a franken hotend now.
i've been dragging my feet on the re-wire process though.

[BadMouth]

I would flip that part on the long end to print.
but i'm guessing the velcro run inside it might cease to exist if you did that.

That was my original plan, but I didn't like the way Cura was handling the bridge over the open velcro channel.
It chooses to print long lines over empty space rather than bridge over the narrow horizontal channel. 
I will still give it a shot.

I dislike a lot of the decisions Cura makes about the toolpath.
I try SuperSlicer every few months, but end up getting annoyed with it. Why can't I just add a friggin' filament? Instead it requires you to make changes to an existing one & save it as something else. 
The last time around there was something in the g-code for the included calibration prints that the Voron didn't like.  There is always something to make me give up and go back to Cura.

EDIT: Trying SuperSlicer again with this Voron profile: https://github.com/AndrewEllis93/Ellis-PIF-Profile .  The printer was trying to extrude my purge line before heating, so I had to add heat and wait lines before print_start in the superslicer start g-code.  This is probably what was happening with the built in calibration tests last time I tried SuperSlicer.  Not crazy about 40% part cooling fan for ABS/ASA, but I want to try the guy's settings without changing anything.  Even if the part comes out nice, I will probably break it just to see how much strength was lost.   

[BadMouth]

Surface finish and corners look sooo much better with the borrowed SuperSlicer profile.
Had a couple issues with the supports though.  The ceiling popped right off and left a decent finish.
The floor that was touching the part however, is adhered to the layer below it.  No way to get it separated, but sand it off.
Superslicer also added supports with brims around the bottom of the part all the way to the bed.  The brims need trimmed off my part now.
This could probably be improved by tweaking the support settings.  In Cura, using tree supports that only touch the build plate would solve it.

Biggest issue I have is that the part broke fairly easily when I squeezed both hooks together.  That 40% fan probably both made the surface finish look good and killed layer adhesion.
I could not break the old part that was printed without a fan (except for bridges).

So now I guess I compare the superslicer settings to my Cura settings and decide which one to move forward on.
Both slicers are within a minute of each other, so the speed settings I arrived at myself in Cura must not be too far off.
The surface finish is much better on Superslicer though.

[BadMouth]

Last one I post.  I swear.

This one was done in Cura, but oriented on its end like the Superslicer one.  Surface finish is much better.  This one was printed with 0% fan (40% on bridges).  It is STRONG.
Tree supports were used which kept the left side perfect.  The inside of the right hook which printed on top the support leaves a bit to be desired, but could probably be improved.
The holes and recesses are kind of messed up from where the supports connected.  They probably weren't necessary and I should have blocked them.
Anyways, the left half of this print is up to my standards now.   

Probably sticking with Cura.  I did notice that Superslicer has the option to add a shrinkage amount to the filament properties, which is awesome.
I've been redesigning parts to fit through trial and error when I should have just been scaling them in the slicer.  I thought shrinking while cooling was too complicated to be a set percentage.  After some research, it seems I may be wrong about that.

[BadMouth]

I did a lot of testing yesterday trying to get nice overhangs with Polymaker ASA and no fan. This stuff is crazy strong when printed with no fan.
Conclusion: Without using the part cooling fan, the only solution is to go slow.....stupid slow....5mm/sec slow.   
I would have never thought to go that slow, but that is where the testing led me.

There is an experimental setting in Cura for overhang angle and overhang speed % (% of wall speed).
I set that to 10% and then look at the "speed" color scheme in the preview to see how fast the overhangs are printed.
I printed a Benchy and the sides of the boat further up were still crap, so I reduced the overhang wall angle to 20% so that would print slow as well.

Also did some experimenting with minimum layer time on the benchy pipe up top.  Lifting head resulted in a worse print.  Pausing to wait at all resulted in a worse print.  Filament oozes while printing has stopped and makes a mess.  It's better to keep the nozzle moving, but get the longer layer time by going.....stupid slow again.  Under cooling, I set the minimum speed to 5mm/s then adjusted the layer time until the pipe printed around that speed.  (so the nozzle doesn't stop moving)  Ended up with an 18 second layer time and the pipe comes out perfect every time.  This is actually faster than the 25 second layer time I've seen some people use for ABS.  The one I printed with slow sides and pipe was just over 2 hours.  It still has issues on the posts of the cab at the level of the windows.  I am pretty sure that could be fixed by printing the entire thing slower, but I don't feel like doing a 3 hr Benchy.

Thought I would share since I've never seen anyone anywhere printing this slow before.

[BadMouth]

Voron 1.8 widebody edition, with double the toolheads. 

I was a bit apprehensive of going down this path, but it wasn't as hard as expected.
It took an evening and a morning to get the config file squared away.

Nowhere near printing yet, but I can home and control the toolheads independently.  Everything moves in the direction it supposed to.
I'm not using anything that wasn't already in klipper (hybrid_corexy kinematics, dual_carriage).  Getting all the directions correct so the X motors don't fight the Y motors took some trial and error, but wasn't that bad.
The only issue is that when homing Z, Klipper doesn't have an option to automatically switch to the required toolhead if the wrong one is currently selected.
It can be worked around, but is more complicated than it should be.

I started printing the parts eddietheengineer has on github https://github.com/FrankenVoron/Tridex, but didn't want to buy NEMA 14 motors and didn't like how thin some parts of the front idler assemblies were.  I also didn't want to spend the time & plastic printing parts that will most certainly be revised later.  So for the Y axis, I used a pair of NEMA 17s from the dead printer junkyard and mounted them the quickest/least effort way with parts I had on hand.  He had an issue with the backs of the x-axis belts rubbing against the linear rail blocks.  My setup also has this problem.  Either the idlers on the x/y gantry need moved forward or the extrusion between them needs moved back.  For now I ordered smaller idlers which aren't ideal, but will hopefully move the belts forward a tiny amount.  EDIT:  Also wondering if using toothed instead of smooth idlers would give any clearance.

eddietheengineer's BOM calls for a mcu that can handle 7 motors.  This is only the case if you are using CANBUS to control the two extruder motors.  If not, 9 motor drivers are needed.  The CANBUS setup is another layer of expense & complication that I am not ready for.   I can get away with 8 motor drivers(Octopus) since the 1.8 only needs two for the Z axis instead of 3 like the Trident.  Right now I am running both Y motors off the same driver, so could add a third Z for a Trident conversion.  I need to see if the linked Y motors cause any issues first.  I already had a 3 driver expansion board on the way from aliexpress.  When it arrives I will see if it works with the Octopus.

This is a test rig and pretty much everything on it is temporary.
I wanted to make sure that I could get the kinematics working before sinking more time and money into details.
Now that the concept works (I think), it can be refined.

[nitrogen_widget]

is this for water soluble supports or just to be an absolute madman?

[BadMouth]

is this for water soluble supports or just to be an absolute madman?

Soluble supports.  It might be HIPS with limonene though, I haven't worked out the math.  HIPS is cheap, but the limonene solvent is expensive.  PVA filament is expensive, but water is cheap.  I also think HIPS is fairy brittle, so it might break off easy then just require a spot clean with the solvent?

This thing is a long term project with an indefinite time frame.
Today I reworked the STLs for the ends of the gantry to add clearance behind the X belts and secure the Y belts. 
Those will print tonight.

I flipped the Y motors to higher up inside the frame and am working out a pulley and tensioner system.

After the motion control is worked out, there are still many issues that I am choosing not to think about until I have to.

[BadMouth]

Dunno if anyone here cares about my 3D printer tinkering, but I don't belong to any other forums to post on. 

Settled on this boring belt path for the Y axis.  I planned to put a tensioning adjustment in the back, but there isn't room so need to rework the front.  This is just a plate for testing.  The final version will have the bolts holding the idlers supported on both ends.  I flipped the X motors over so they are above the back crossbar instead of below it.  That with this Y motor mounting puts all four X and Y motors in the same area which will eventually be isolated from the build chamber and cooled.  I think the Octopus control board can fit in there as well which would make every run of wire in the machine much shorter.

I installed the new X/Y joints which pushed all the belts out 2mm.  This looked like too much and the back idlers pushed the belts out way too far.  So I did what I should have done the first time around and measured the distance between the belts and the linear rail at both ends.  It turns out that leaving the rear idlers where they were and only pushing the front ones out 1mm would have been the correct thing to do.  So another overnight print and I have X/Y joints that give the belts 1mm clearance over the blocks.  The top and bottom belts are in perfect alignment with each other which gives me a warm and fuzzy feeling.

Since the belts are still pressed against the blocks where they are attached, they can still rub when both X axis blocks are near each other.  So next up is to rework the x-carriage to incorporate a 1mm shim behind the belt being attached.

[Howard_Casto]

I've been following along... I just don't have anything useful to comment on.   The documentation of your experiments are very much appreciated. 

[bobbyb13]

I am merely a voyeur here because the stuff is fascinating but I still know almost nothing about it all !

[BadMouth]

Progress:

The belts for each X carriage must pass through each other. 
In the last installment, I moved idlers forward 1mm to provide 1mm of clearance behind them.  (pic1)

The other worry is that the belts will rub the mounting screws.  I verified they did not by attaching the belt to only one carriage and installing only bolts in the other so that everything was visible.  (pic2)
I may have thrown in the towel or restarted from scratch if this hadn't worked out.

The next issue was to raise the secured belt 1mm off the block so it doesn't rub the other carriage when they are close.  I started out with a shim, but really hated the idea of having to hold the shim, belt, and carriage all in place while installing screws.  Finally came up with a different way.  (pic3)

Final clearances.  (pic4)  Top is secured, bottom is pass-through.  Other carriage will be opposite.  Passthrough tolerance looks super tight, but I pushed on the belt in all directions and it moves (meaning it has room to move).  Did the same test with the carriages at the ends and close to each other.  An official design would probably have more clearance toward the front to account for variation in parts used.  I am running with this for my build though.  The parts pictured are test pieces.  I will print the full carriage parts tonight.

[BadMouth]

Complete dual toolheads now mounted.  Waiting on Y drive parts to print, then it will be alive again.  It turns out that the old induction sensor will not work with the newer Voron carriages (it touches down before the nozzle!).  I had a pair of the newly specc'd ones on the way from Ali-Express, but shipping was delayed.  So I ordered a knock-off from amazon along with some HIPS filament.  Hopefully I will be pushing plastic next week.

The BTT EXP-MOT expansion board came in.  It allows you to drive 3 more motors.  It works fine with the Octopus.  This means that I can run the Y motors independently and also upgrade the to the Trident Z axis.

If anyone has a need to add a few more motor drivers to a BTT board, here is what you need to know:

It connects to EXP1 & EXP2 that the screen would normally use.  If you have a setup where the display connected to the mcu is required, this is not an option.
The included cables are only about an inch long.  This is because the board was designed to work with SKR 1.3/1.4 and sit right up against it.
I used a set of cables that came with a display and had no issues.

All pin assignments just use the EXP1 & EXP2 pins on the Octopus.  There are no separate addressable pins.
Wipe out your display pin assignments before connecting the expansion board.
It took me longer than it should have to work out the pins because I tried to translate an SKR 1.4 example to the Octopus.  Something was either wrong on the example, or I transposed a couple things. 
The pinout included with the expansion board documentation is from the silkscreen on the back of the board, so it is a mirror image.  It is not a guide as if looking down on the board like the documentation for other BTT products.

Anyways, worked through it and pic of pin assignments is attached if anyone is looking for them on a diy arcade forum.  I only tested Motor 1.  Now using all of them and they are correct, but always check every pin assignments yourself.  Jumpers were set for 3.3v logic & UART mode.  Other jumpers next to EXP2 removed.

This expansion board is a neat thing to have on hand.  I haven't gone through everything required, but with one of these someone could probably run a Voron 2.4 with a single old SKR1.3 or 1.4 that they already have on hand instead of buying an Octopus or second MCU.  A Trident could definitely be run off of it.

EDIT:  I also should have mentioned that I cut the sensorless homing pin off the stepper driver per Voron instructions, as there is no jumper to disable it.

 

[yotsuya]

Your progress is inspiring!


Sent from my iPhone using Tapatalk

[BadMouth]

The Y drive is back together.  Everything fits inside the frame with absolutely no travel lost.  The Y limit switches needed a home so I added a place for them last minute before printing.
The back pulleys are a couple mm too high.  I thought I had fixed this, but might have started with the wrong file when I added the limit switches. 
It will need at least one more revision, but I'm rolling with these parts for now to see if any other changes need made.

I need to make a decision on whether or not to move the octopus board up between the motors before I can start rewiring everything.

[BadMouth]

So I'm printing parts on the Voron 2.4 and they start looking weird.  The back is perfectly fine, the front is missing starting at a certain layer.
Filaments spool preload might be a little tight, so I loosen it up and start the next print.  Again, printer stops extruding filament toward the front.
Diagnosis = extruder motor wires are probably broken inside the drag chains.   

I already went through this with the inductive sensor a month ago, but was hoping it as an isolated case.
I guess the silicone ribbon cable I used just wasn't up to the task. 
So now a complete rewire of the toolhead is required when I'd rather be working on other things.


Again, this is a horrible hobby.   

EDIT: Fixed after three hours of work.  I had just enough of various color silicone wire to make a new run for the extruder motor.  The only other run from the toolhead using the bad wire was for the induction sensor which already failed and was already replaced a month ago.  Now my "by the book" black and red V2 has a rainbow of wiring colors.

[bobbyb13]

The fortitude you guys have on getting/keeping these things running is admirable.
And I thought learning CRTs was the ultimate time-suck.
What a freakin rabbithole printing is.

Like my neighbor the programmer (whose computer I used to play Zork on back in the mid 80's!) used to say about IBM.
It actually stood for-
Interesting
But
Maddening

[Mike A]

If you spend the money and don't buy a kit, then a 3D printer is a tool that just works with minor maintenance. Same with a CNC machine.

These guys are buying kits and they enjoy the tinkering.

[BadMouth]

If you spend the money and don't buy a kit, then a 3D printer is a tool that just works with minor maintenance. Same with a CNC machine.

These guys are buying kits and they enjoy the tinkering.

Kit?  What kit?  I did it the hard expensive way and ordered every bolt and piece of wire separately.   
For the one I'm building now, I cut, drilled and tapped the extrusion myself.

[nitrogen_widget]

So I'm printing parts on the Voron 2.4 and they start looking weird.  The back is perfectly fine, the front is missing starting at a certain layer.
Filaments spool preload might be a little tight, so I loosen it up and start the next print.  Again, printer stops extruding filament toward the front.
Diagnosis = extruder motor wires are probably broken inside the drag chains.   

I already went through this with the inductive sensor a month ago, but was hoping it as an isolated case.
I guess the silicone ribbon cable I used just wasn't up to the task. 
So now a complete rewire of the toolhead is required when I'd rather be working on other things.


Again, this is a horrible hobby.   

EDIT: Fixed after three hours of work.  I had just enough of various color silicone wire to make a new run for the extruder motor.  The only other run from the toolhead using the bad wire was for the induction sensor which already failed and was already replaced a month ago.  Now my "by the book" black and red V2 has a rainbow of wiring colors.

this is what i bought for my chiron hot end.

https://www.amazon.com/gp/product/B07T4SYVYG/ref=ppx_yo_dt_b_search_asin_title?ie=UTF8&psc=1


also just for the heater.
https://www.amazon.com/gp/product/B0746H2K6G/ref=ppx_yo_dt_b_search_asin_title?ie=UTF8&th=1


have yet to put it in.
too many other things going on now that it's spring.

also that feeling when you are 48 hrs into a print and realize stupid fing cura decided to not put supports on one small 90 degree part.
it's just the mast mount for a ship piece i'm printing vertically.
i can cut the part out of the model, reprint then glue it in but it's going to be a pain to dril it out of the deck

[Howard_Casto]

If you spend the money and don't buy a kit, then a 3D printer is a tool that just works with minor maintenance. Same with a CNC machine.

These guys are buying kits and they enjoy the tinkering.

I wish that were that case Mike, but it really isn't.  Even with a top-of-the-line printer the success rate is probably 80% on prints at best.   Filament reacts to moisture, bed adhesion fluctuates, hot ends overheat ect.   Remember we are essentially making a cake by stacking row after row of squirts of icing on top of each other and hoping that it doesn't fall over..... that ---steaming pile of meadow muffin--- shouldn't even work so it's a miracle that it even works half of the time exactly as you intend it to. 

[talkgeek]

...honestly I think 80% is a bit low, and that will scare off people getting into this  - its prob closer to 90+% - and my experience is most issues are filament (temp / moisture / bed adhesion) or slicer config - once you have tuned in your printer, ..and yes, (like anything including building arcades) you have to invest time to get the most out of it - the more prints (experience you gain) the less issues.
I have 3 filament printers (and 2 resin printers)
1x Ender3 Pro (basic mod's - BLTouch, SKR-Mini3 board, Dual drive Bowden - my first printer)
2x Ender 3 standards
- one heavily modified including V4.2.7 silent board with direct drive dual extruder, strong bed springs, dual Z axis drive, BLTouch, glassbed etc (tuned to print mostly Ninjaflex TPU)
- the other E3 standard (last purchase) has had very little done to it (bar PFE Capricorn tube, strong bed springs and metal dual gear bowden drive) - and (probably because) I run same brand PLA only on that machine, almost never have issues with it (that aren't caused by me forgetting to add (or remove) supports when slicing)

[BadMouth]

Thanks for the links nitrogen_widget.  Those are cheaper than what I remember coming up in my searches.  I am wondering if silicone cable is a good choice for inside cable chains though.  It's flexible, but also grabs everything it comes into contact with.  When the first wires to the inductor broke (four conductor silicone ribbon cable), I left them in the cable chain because it was easier than removing them.  This time around I removed them and the other broken wires.  The first set which was no longer connected at the toolhead end had worked itself backwards through the cable chain five or six inches.  I am wondering if the sticky silicone ribbon cables weren't "walking" themselves backwards against each other as the cable chain rolled and unrolled.  If so it would be an argument for $PTFE$ cable.  I already ordered a larger gauge silicone ribbon cable on the current build, but it won't be inside a cable chain.

Printed some brackets to hold the electronics where I would like them to live.  The goal is to have the electronics and motors in a compartment insulated from both the frame and chamber with a couple big fans moving air across them.  Still don't know exactly how that will work, but this lets me visualize how things fit.   A lot less wire will be required.  Still haven't lost any Y axis travel.  Gonna start rewiring & printing parts for the Trident Z axis conversion.  After the Trident conversion is done, I am going to try to incorporate the filament spools inside the frame under where the toolheads park on the sides. 

 

[RandyT]

If you spend the money and don't buy a kit, then a 3D printer is a tool that just works with minor maintenance. Same with a CNC machine.

These guys are buying kits and they enjoy the tinkering.

I wish that were that case Mike, but it really isn't.  Even with a top-of-the-line printer the success rate is probably 80% on prints at best.   Filament reacts to moisture, bed adhesion fluctuates, hot ends overheat ect.   Remember we are essentially making a cake by stacking row after row of squirts of icing on top of each other and hoping that it doesn't fall over..... that ---steaming pile of meadow muffin--- shouldn't even work so it's a miracle that it even works half of the time exactly as you intend it to.

Agreed.  These two responses demonstrate the difference between marketing-influenced perception, coupled with seeing the results from someone who knows their machine(s) really, really well and someone who has actually tried to make good prints with any sort of regularity on a machine they own.  A chisel is also a tool which just works with minor maintenance, but if you don't know how to properly sharpen one, the results will show it.

...honestly I think 80% is a bit low, and that will scare off people getting into this  - its prob closer to 90+% - and my experience is most issues are filament (temp / moisture / bed adhesion) or slicer config - once you have tuned in your printer, ..and yes, (like anything including building arcades) you have to invest time to get the most out of it - the more prints (experience you gain) the less issues.

The Ender series are probably the best of the low-cost bed-slingers, but not out of the box unless you get very lucky.  They still really require the upgrades you mentioned to take away a lot of the variables which are just part of the process. And those aren't even worth anything if you don't REALLY understand the slicer software and can make proper adjustments to get rid of the problems.  Even that won't fix all of them, forcing you to go back to your CAD software to make changes to the part to be more conducive to the process and how you want your workflow to go.  Experience is everything with these machines and it transcends the hardware.

[nitrogen_widget]

Thanks for the links nitrogen_widget.  Those are cheaper than what I remember coming up in my searches.  I am wondering if silicone cable is a good choice for inside cable chains though.  It's flexible, but also grabs everything it comes into contact with.  When the first wires to the inductor broke (four conductor silicone ribbon cable), I left them in the cable chain because it was easier than removing them.  This time around I removed them and the other broken wires.  The first set which was no longer connected at the toolhead end had worked itself backwards through the cable chain five or six inches.  I am wondering if the sticky silicone ribbon cables weren't "walking" themselves backwards against each other as the cable chain rolled and unrolled.  If so it would be an argument for $PTFE$ cable.  I already ordered a larger gauge silicone ribbon cable on the current build, but it won't be inside a cable chain.

Printed some brackets to hold the electronics where I would like them to live.  The goal is to have the electronics and motors in a compartment insulated from both the frame and chamber with a couple big fans moving air across them.  Still don't know exactly how that will work, but this lets me visualize how things fit.   A lot less wire will be required.  Still haven't lost any Y axis travel.  Gonna start rewiring & printing parts for the Trident Z axis conversion.  After the Trident conversion is done, I am going to try to incorporate the filament spools inside the frame under where the toolheads park on the sides.

as for silicone grabbing
I also have this stuff for my cnc and it's slickery.
i'll either use original wire loom inside the chain or use this stuff.

https://www.amazon.com/100ft-Expandable-Braided-Sleeving-Sleeve/dp/B074GM1PK1/ref=sr_1_1_sspa?crid=1O6NPW0S4REVA&keywords=wire+loom&qid=1652046194&sprefix=wire+loom%2Caps%2C107&sr=8-1-spons&psc=1&spLa=ZW5jcnlwdGVkUXVhbGlmaWVyPUExQVRVV0hCVTNBN0hJJmVuY3J5cHRlZElkPUEwMTUwODQzM1QzVDZQSERKOEgwQiZlbmNyeXB0ZWRBZElkPUEwMTg4ODcxMTg2WTlPUjhHTDA1MiZ3aWRnZXROYW1lPXNwX2F0ZiZhY3Rpb249Y2xpY2tSZWRpcmVjdCZkb05vdExvZ0NsaWNrPXRydWU=

[nitrogen_widget]

If you spend the money and don't buy a kit, then a 3D printer is a tool that just works with minor maintenance. Same with a CNC machine.

These guys are buying kits and they enjoy the tinkering.

I wish that were that case Mike, but it really isn't.  Even with a top-of-the-line printer the success rate is probably 80% on prints at best.   Filament reacts to moisture, bed adhesion fluctuates, hot ends overheat ect.   Remember we are essentially making a cake by stacking row after row of squirts of icing on top of each other and hoping that it doesn't fall over..... that ---steaming pile of meadow muffin--- shouldn't even work so it's a miracle that it even works half of the time exactly as you intend it to.

Agreed.  These two responses demonstrate the difference between marketing-influenced perception, coupled with seeing the results from someone who knows their machine(s) really, really well and someone who has actually tried to make good prints with any sort of regularity on a machine they own.  A chisel is also a tool which just works with minor maintenance, but if you don't know how to properly sharpen one, the results will show it.

...honestly I think 80% is a bit low, and that will scare off people getting into this  - its prob closer to 90+% - and my experience is most issues are filament (temp / moisture / bed adhesion) or slicer config - once you have tuned in your printer, ..and yes, (like anything including building arcades) you have to invest time to get the most out of it - the more prints (experience you gain) the less issues.

The Ender series are probably the best of the low-cost bed-slingers, but not out of the box unless you get very lucky.  They still really require the upgrades you mentioned to take away a lot of the variables which are just part of the process. And those aren't even worth anything if you don't REALLY understand the slicer software and can make proper adjustments to get rid of the problems.  Even that won't fix all of them, forcing you to go back to your CAD software to make changes to the part to be more conducive to the process and how you want your workflow to go.  Experience is everything with these machines and it transcends the hardware.

I pissed away a roll of pet-g on multiple 40-50hr prints (same model) that failed half way through (layer shift) because there was slop in the bearing roller for the bed.
since i replaced it i'm back to back 60+hr prints with zero issue.

[RandyT]

I pissed away a roll of pet-g on multiple 40-50hr prints (same model) that failed half way through (layer shift) because there was slop in the bearing roller for the bed.
since i replaced it i'm back to back 60+hr prints with zero issue.

Sounds about right.  That's where, I think, most of the waste and difficulty occurs.  Not knowing where the problem is and whether it is an isolated "glitch" somewhere or a genuine machine issue which needs to be addressed.

I had a similar situation on an ET4 where I would print a bed full of small parts and would inevitably get a layer shift or the head would stall and deposit a pile of plastic and then back the filament all the way out of the tube.  Fewer parts on the table and it would usually print fine.  Thinking the firmware was crap, I upgraded it and still had the same issues.  Then after replacing the mainboard after smoking it by shorting the heater with some tweezers due to the aforementioned blobs, I put the latest Anet firmware on it, thinking it might solve the issue (same base machine and MB.) But still the exact same problems.  Wasn't until I found a Marlin replacement firmware that the issues went away and the machine started behaving.  If I didn't have the sense that the firmware was the root of the problems, the "rabbit-hole" would have gone much deeper and been more costly to remedy.  I don't know how the Chinese engineers manage to bork Marlin so badly when they skin the software for touch screens, but somehow they do.  If I were to guess, knowing how they tend to operate, one guy does a shoddy job and it gets copied/modified by everyone, with existing bugs plus potentially new ones, which are never fixed because it almost works well enough to sell the printer.

So there's another bullet point for my list of machines to avoid; anything which doesn't have a reasonably easy path to make it swallow a stock Marlin firmware or a community modded one which has been well-tested.  In the low-cost market, that means sticking with the most popular brands and models, even if they tend to cost a little more than other similar machines.

[nitrogen_widget]

I pissed away a roll of pet-g on multiple 40-50hr prints (same model) that failed half way through (layer shift) because there was slop in the bearing roller for the bed.
since i replaced it i'm back to back 60+hr prints with zero issue.

Sounds about right.  That's where, I think, most of the waste and difficulty occurs.  Not knowing where the problem is and whether it is an isolated "glitch" somewhere or a genuine machine issue which needs to be addressed.

I had a similar situation on an ET4 where I would print a bed full of small parts and would inevitably get a layer shift or the head would stall and deposit a pile of plastic and then back the filament all the way out of the tube.  Fewer parts on the table and it would usually print fine.  Thinking the firmware was crap, I upgraded it and still had the same issues.  Then after replacing the mainboard after smoking it by shorting the heater with some tweezers due to the aforementioned blobs, I put the latest Anet firmware on it, thinking it might solve the issue (same base machine and MB.) But still the exact same problems.  Wasn't until I found a Marlin replacement firmware that the issues went away and the machine started behaving.  If I didn't have the sense that the firmware was the root of the problems, the "rabbit-hole" would have gone much deeper and been more costly to remedy.  I don't know how the Chinese engineers manage to bork Marlin so badly when they skin the software for touch screens, but somehow they do.  If I were to guess, knowing how they tend to operate, one guy does a shoddy job and it gets copied/modified by everyone, with existing bugs plus potentially new ones, which are never fixed because it almost works well enough to sell the printer.

So there's another bullet point for my list of machines to avoid; anything which doesn't have a reasonably easy path to make it swallow a stock Marlin firmware or a community modded one which has been well-tested.  In the low-cost market, that means sticking with the most popular brands and models, even if they tend to cost a little more than other similar machines.

yeah.
the only way i found it was by pressing on stuff and tightening fasteners.
again.
I found 4 on the underside for the rail the rollers run on were loose.
then found the bad bearing shift but it only happened when one corner was pushed. i don't push on my printer beds so....

it is amazing how some dude who is a hobbyist can make marlin work better than the manufacturer of the printer aint it?

[BadMouth]

It doesn't surprise me when it comes to most Chinese manufacturers.  R&D is not their strong suite, hence reliance on hobbyists for design, software, & support.
Cost control is their strong suite and I do appreciate that to an extent.

The youtube channel Clough42 recently did a review of the Quidi i-Fast where the conclusion is that it's a great printer with half assed software.  It has a camera installed inside, but you have to separate camera software to view it.  It is not even integrated into the 3D printer control software.

It's a shame because it seems to have everything I'm trying to accomplish with my current build.  Dual hotends and an 80 degree heated chamber.  He shows huge chunky ABS parts that printed with absolutely no warping.  If I didn't already spend so much building a Voron (and another), I'd be tempted.  Not sure if I'd pull the trigger on it or not though based on the software and concern about future parts availability.

EDIT:  watched the video again and he says the specs claimed a chamber temp lf 65, but it went to 80.  I wonder if the parts it's built from are even rated for continuous duty at 80.  I've beem looking at ceramic heaters.  Some of them come with pc style fans and say they are limited to 65 because of the fans.

[BadMouth]

Did some more tinkering....

A few weeks ago it was unseasonably cold and I had a few failed ASA prints.  It was cold in the basement and chamber temps were only getting to the mid 30's (celsius).
So I finally ordered a pair of 50mm fans to tinker with a PTC heater that I had purchased long ago.  Also orderd a 500w personal heater from amazon warehouse for $15, figuring that it might be a simpler and cheaper route.

Supposedly the aluminum fins on these heaters are carrying 110v when they are active.  I couldn't get anything with the meter, but am treating them as live just in case.  There is a cover over the ceramic portions, which may be where the angry pixies are.  PTC is labelled 500w.  Harbor Freight clamp meter says 3 amp, so probably more like 360w.  Fans are 12.6 cfm each and seemed to provide enough airflow.  It took 12 minutes to heat the build chamber of the 350mm Voron V2.4 from 27 degrees to 50 (No other heat source.  Bed wasn't on.)  It took forever to reach 51 degrees, so I think any further increases would have been minimal.  I am not looking to get it hotter than 60-65.  All the small fans in the printer are only rated for those temperatures, and the linear bearings (which have plastic races) aren't rated for a whole lot more. 

There are a lot of considerations to work through before I would feel safe leaving it unattended.  The fan must be on when the heater is, so it would be bad practice to rely on software to control the fan.  The fans I am currently using are 24v.  They could be connected directly to the power supply which would make them run whenever the machine's 110v power switch is on.  Not sure if that is better or worse than using a 110v fan.  If something caused the fans to stop, the area surrounding the PTC must be able to tolerate the heat until a thermal fuse or relay is tripped.  Relay probably isn't the way to go as it would reset itself when it cooled down and then the PTC would just heat up again.  There are probably other things I haven't even thought of yet.

Not sure how it could be incorporated into the 2.4.  Best thing I've come up with is a blower and ductwork attached to the outside of the back panel with the PTC mounted about where the exhaust outlet currently is.

I worked out an idea for the new IDEX build where it could be mounted underneath the chamber in the center with ductwork pulling air from the sides.  I am thinking heat rises in the center.  Cold air falls, but is pushed to the outer edges by the heat rising off the buildplate.  The only problem I could think of with this setup (in addition to crap falling in the ducts), is that on really tall prints the bed will get close to the PTC heater and that might trigger a "bed heater not behaving as expected" automatic shutdown.  Dunno, but I've arranged things so the possibility is there.  The ductwork would be a pain to print.  The picture of it under the IDEX is just of it sitting in place.  Nothing has been done as far as adding it. 

[BadMouth]

IDEX build has been half apart and back together a couple times.

I had to flip both x steppers back over so they hang under instead of sitting on top.  I do not understand, but I could not get the kinematics to work correctly with them on top.  I reversed motor direction in software, homing direction, physically reversed wiring, etc.  All combinations either lead to having to home toward the front or move 0,0 coordinates to the back.  I gave up and redid the back assembly and belts, etc.     It is still irking me.

Trident Z axis conversion is complete and everything is working correctly.  It really opens up visual space (and real space) on the sides.  Original plan was to incorporate spool holders on the sides, but there was enough room in the back.  They are far enough back that oozing filament can't drip on them.  Only problem is that the bed has to be all the way up or down to load filament.  Swapping rolls in the middle of a print may not be possible.  The spool holders are mounted with removable t-nuts, so nothing is set in stone.  There is just enough room on the side for a 5lb roll from Keene Village Plastics, but it would be super close to the bed.  I had a piece of textured ABS that was the right size for the bottom panel, so I went ahead and added it.  I will probably change it to aluminum composite later.

Still haven't wired the toolheads.  Not looking forward to it.  The silicone wire I planned to use droops instead of arching like an umbilical.  It will take a lot of wirres and I have not found the perfect toolhead connector yet.  I happened to come across a new BigTreeTech canbus pcb for "introductory pricing" which was not much more than simple "Voron" connector PCBs go for.  So I ordered a pair.  I will be in over my head with the canbus stuff, but it would mean only having to run 4 or 6 wires to the toolhead instead of 17.

I was also tempted by this Biqu hotend kit hat uses a proprietary "USB-C" cable for all the hotend connections.
https://www.aliexpress.com/item/1005001995220414.html?spm=a2g0s.8937460.0.0.6b772e0eARS7X4
I would have to order a bunch backup cables though, because you probably won't be able to get them in the future.


Not much else is going to happen until I wire the toolheads so the actual IDEX testing can begin.
I am fiddling with the camera mount and lights to avoid working on the wiring.
Was going to start on the insulated enclosure, but all the local big box stores are out of 1/2" polyiso.  Might break down and use 3/4....

[pbj]

So I need basically the simplest thing in the world printed.... more or less this design:

https://www.thingiverse.com/thing:3223106

I had an outlet installed in my gameroom, but it's mounted to a brick wall that was covered with sheetrock.  (God bless the former owners)  So that means about 9/16" of electrical box is exposed between sheetrock and back of the outlet face plate.  (I was made aware of this before the electrician installed it and I told him to do it anyway)

All I need is a rectangular frame that lines up with the edges of my wall plate cover.  The edges have to be narrow because it needs to clear the mounting tabs of the outlet itself.  I made one out of wood but it was too thick and looked dumb.  Right now I've got it covered with a few wraps of black velcro tape.  Your eye passes over it, and it's better than exposed metal, but I'd like something more finished. 

That's a company that makes these rings, but I'd need three of them:

https://www.kyleswitchplates.com/depth-ring-wall-plate-extenders-1-gang/

And the pricing seems absurd.

Anyone?

[HaRuMaN]

I can do that...  just get me some dimensions.   

[pbj]

Outside dimensions should be 4.5" tall by 2.75" wide.  I wouldn't make it more than 1/8" thick from outside to inside edge.  And 5/16" tall.  I'll try to get some more precise measurements.

 

[HaRuMaN]

Outside dimensions should be 4.5" tall by 2.75" wide.  I wouldn't make it more than 1/8" thick from outside to inside edge.  And 5/16" tall.  I'll try to get some more precise measurements.

 

Ok cool

[BadMouth]

Made some progress, but still not pushing plastic yet.

The wiring to the toolheads that I had planned to use was too droopy to work as an umbilical.  After weighing options and wanting to try something not seen before, I ordered a length of 20 conductor motion rated cable from Igus (part no. CF9-02-20).  24awg covers everything but the hotend heater, but there are enough conductors to double up two pairs for the hotend.  I knew the thickness ahead of time ( around 9mm), but it's not quite as flexible as I'd hoped.  It should still be fine with a decent size arch.  I wouldn't use it on anything less rigid than the Voron setup for fear of it making the toolhead flex out of alignment.  In retrospect, multiple runs of simple stranded CAT6 is probably the better way to go.  This is top quality wire though.  Lots of strands and strong. 

I looked at PCBs for toolhead connections, but couldn't find exactly what I was looking for.  Started to mock some things up with generic prototyping boards and realized that what I had in mind was possible without any pcb.  Soldering all those pins took a while, but wasn't too bad.  There isn't much plastic where the male pin passes through, so they would go all wonky when soldering.  Plugging in the connector while soldering kept them in place.   There wasn't nearly as much room for excess wire as I had pictured in my mind, but it fits.  I am still making some design decisions about the cap.

Also made some progress on the chamber heater.  I decided to violate my own rules and have the fans be software controlled.  Otherwise they will either run constantly or only run while the heater is on, meaning they would shut on and off as the heater cycles on and off.  To mitigate the danger of the heater getting out of control, it is held in place by metal parts and has a 120 degree thermal fuse attached to the outside fin.  I may switch to a lower temp fuse, but that is what I had on hand.  Basic idea is that the fans pull air from the edges of the chamber and push it through the heater in the center.  I still need to cut the deck to accommodate the intakes and heater.

Next time around I'll get the toolheads wired so IDEX tinkering will be back on the menu....which is the whole point of this project anyway. 

[BadMouth]

I was hoping to have a cooler print to demonstrate with, but I'm short on time and won't be able to work on this again until next week.
This is just a test to see if the printer will work the way I intend to use it; using the second toolhead exclusively for support material (currently HIPS filament).


Getting to this point was straightforward enough that I didn't expect the start routines and tool changes to be so much of a PITA.  Cura automatically inserts gcode to switch toolheads before my start macro, which is annoying because Klipper will refuse to switch toolheads if the printer hasn't been homed.  If Cura would put it after the start gcode where it belongs, several problems would go away.  (only the first toolhead has a sensor for bed leveling).   I added some stuff to the start macro to temporarily switch to the first toolhead if the second one is active, but for some reason it doesn't behave correctly in all situations.
I could manually change the gcode for every print, but don't consider that or postprocessing scripts to be acceptable solutions.

There are other problems like Klipper only allows you to specify one part cooling fan, despite having two toolheads.  I can specify the second one as a generic fan, but am not sure the commands from Cura will register correctly.  I am sure there is a complex workaround, but for now I am using the multi_pin option in Klipper to run both fans as one.  I'd prefer to get it working as simply as possible before adding layers of complexity.  That's also why I've stayed away from forked builds of Klipper.  I'd rather get it working as good as possible with the standard build first.

[RandyT]

Very nice!   But man, those extruder swaps definitely need to be moving in tandem.  I would think that two motors on the same axis, with the same feed rate and direction with differing destination points, should never result in a conflict.  Or is this just a safe test?