First build - large Hex framed Delta - 3DR Mega style

Firstly as this is my first post on RepRap a big hello. I have read a lot of posts researching Delta style printers with the aim of building my own.

I bought a Flashforge Creator Pro in Jan 2015 and have around 2800 print hours and over 50Kgs using exclusively ABS. I can print dimensionally accurate parts so will be using this to print parts I need for the Delta and hence this post.

Delta's look cool and give me some serious build height that I would like occasionally, will admit some of the vases look stunning.

I have been significantly influenced by the 3DR Mega but I'm looking to encompass improvements to increase the rigidity and hopefully print quality.

Design Goals

Min 400mm diameter print area.
Quality is of higher priority to speed
Rigidity is key
Speed must be at least equal to the FFP as a minimum, I comfortable get 65mm/s with a 0.4mm nozzle and ABS.
Heated bed.
Full enclosure
Hexagonal design to make enclosing easier.
Whilst cost is a consideration it is not a priority over key goals above.


Current decisions:

20x40 T-Slot for the external framing 1600mm long
20mm case hardened rod 1600mm long aka like the 3DR Mega.
10mm GT2 steel reinforced belt.
8x6 mm Carbon fibre rods length to be finalised
Traxxas style rod ends, however not traxxas joints themselves.
E6D hot end.
Base and top 6 sided with printed internal and external corners and 20x40 t-slot sides 250mm long each.
Will not be Bowden feed. This leaves a flying extruder or effector mounted direct drive.
6mm MIC6 ALU plate for the bed covered in kapton tape for ABS.
Duet controller board, hoping the new version post 085 will be out by the time I'm ready.
No internally located cooling fans, all cooling air will be ducted in from outside the enclosure.
24V 350W or 500W SMPS unit, have a few lying around from another hobby.

Currently Undecided:

Nema 23 or 17, leaning heavily towards the 23's to handle the potential mass of the carriages and effector setup.
Type of heater to use... Leaning to 240v version (Australian based).
Extruder and Extruder drive.....

Thoughts:

Reading everything about Bowden tubes and how they get worse when they get longer, then for a printer of the size I'm planning it's not sensible. A flying extruder seems popular but it seems to me that you still have to move the entire mass around and despite any counterweighting you can't negate the actual mass you are moving around. In watching videos it seems that the issue is after the move is complete you are fighting to stop both the effector and the flying extruder with the extruder potentially being pulled out of place.
I learnt with the flashforge that an extremely important point is not only machine rigidity but what it sits on must be rock solid. Originally I had it on my flat solid router bench and all was good, moved it to the general workbench as part of a garage reorg and quality went down, till I noticed the workbench moved fractional as the printer worked, fixed by securely attaching workbeanch to garage frame. Ultimately I will bolt this new one down so it can't move.

I think having the extruder directly mounted to the effector gives a solid coupling of the mass with the negative of lower acceleration, fairly certain I can live with that.

This brings me back to the stepper motor choice and just what can the Duet drivers handle?

Mind you everything is up for suggestion and comment, please remember that this is not about saving money it's about building quality.
Comments, thoughts and suggestions welcome, with the becoming a build thread with pictures and hopefully a working video

Oh everyone seems to come up with cool names for designs....... personally I have no idea for this

Here is a screen grab from the Tinkercad model of the base design... the plane is the top will be identical. the round plate intersecting the round rod is just there to give me a sizing dimension at the moment but it is planned to be the base for the build plate. Not shown are the internal corners the reinforce each joint.

That's going to be a nice machine! A few suggestions:

1. Do not use printed corners, they are not rigid enough for a large printer.

2. Yes use a 240V bed heater. That in turn allows you to use a smaller 24V power supply to power everything else.

3. Nema 17 vs Nema 23: I guess it depends on the mass of your flying extruder. Without a flying extruder, Nema 17 motors should more than adequate, unless you think the greater belt mass is going to be a problem.

4. The existing Duet should in theory be able to drive 1.8deg/step Nema 23 motors with a low enough current rating (below 2A) using a 24V supply, however you might not achieve such high speeds as with Nema 17 motors. Don't try to drive 0.9deg/step Nema 23 motors from the Duet, because I expect they would need more than 30V to reach reasonable speeds. The next-generation Duet will probably have drivers rated at a higher current than the existing Duet, but it will still be limited to about 30V.

Or you can use external DSP-based drivers attached to the expansion connector of the Duet.

5. Another possibility: instead of a flying extruder, as you have 3 spare vertical towers, put another stepper motor, belt and carriage on one of them, and attach the extruder to that. The firmware would need to be changed to keep the extruder at a sensible height depending on the effector XYZ position.

---

Large delta printer [miscsolutions.wordpress.com], E3D tool changer, Robotdigg SCARA printer, Crane Quad and Ormerod

Disclosure: I design Duet electronics and work on RepRapFirmware, [duet3d.com].

Plastic corners cannot give you bending stiffness, but if you stabilize the printer either by bolting panels or making a cross-bracing, they will be sufficient. The hexagon design is quite practical for that.
I am for panels because I prefer enclosed printers, especially if you print ABS. Also, heavy wood panels dampen vibrations and noise.
In principle, the height of a flying extruder don't vary much because as an average, the height of the effector vary very slowly.
So, their inertia is somewhat negligible for loads, you only have to live with their dead load. You can probably get an extruder for 250/300g all included, which could be easily lifted by Nema17, because one column will rarely carry more than 40% of the effector weight.
With standard boards, I am for Nema17, because you cannot improve with Nema23, and lose microstepping precision (the microstepping intervals are said being more accurate on Nema17 than Nema23). If you get in trouble, limit the travel movement speed.
Also, I am for larger minimim arm angle (say 23~25°) instead of usual 20°, which limit the maximum carriage speed at the cost of slightly longer arms. That is also better for printing accuracy at large diameter.

If using traxxas style links, you shall find a mean to increase friction, as this dampen vibration and limit banding effects. spring between arms is not elegant, but that helps.

You may use ball and socket from Igus, they have way too much friction for a normal delta, but that may do the trick for a large unit and being all plastic, they are very light.

You may have a look at the italian Wasp printers, they exist in large size, use flying extruders and Igus components. They are near of what you want to do.

Also, some have developed light extruder for delta with small steppers, but that may limit the retract speed.

If really at the limit, you may slightly increase the voltage near the board maximum (30V) , but I may be afraid of back EMF from steppers.
For a 24V setup, use 12V fans, that is easier to handle. A DC/DC converter cost nothing.

Edited 1 time(s). Last edit at 03/13/2016 10:30PM by PRZ.

Also, quite large printer (usable diam 500x height 710/750) are the printers from nederland, at [www.opiliones.nl]
They have made interesting mechanical choices (the carriage are maintained by magnets on the tubes). No data about steppers nor board.
They have prototyped stuff with machinekit, but I don't know for the serial units.

Thanks for the quick response...

Quote
dc42
That's going to be a nice machine! A few suggestions:

1. Do not use printed corners, they are not rigid enough for a large printer.

Finding metal corners for 120 degree joints is impossible in Aus and I haven't seen them anywhere honestly, I think i've engineered the printed versions with more than enough, photos to follow in separate post. Testing of the outside corner shows no deflection or flex with significant leveraged force. I'm more concerned about flex in the 20x40 and 20mm rod at this time.

Quote
dc42
2. Yes use a 240V bed heater. That in turn allows you to use a smaller 24V power supply to power everything else.

OK 240v heater it is, I better add a DC/AC SSR into the parts list then....I do Christmas lighting for a main hobby, and never run "smaller" supplies, the smallest I have for general duties is 350W

Quote
dc42
3. Nema 17 vs Nema 23: I guess it depends on the mass of your flying extruder. Without a flying extruder, Nema 17 motors should more than adequate, unless you think the greater belt mass is going to be a problem.

Concerned about overall mass, especially if I go with effector mounted extruder.

Quote
dc42
4. The existing Duet should in theory be able to drive 1.8deg/step Nema 23 motors with a low enough current rating (below 2A) using a 24V supply, however you might not achieve such high speeds as with Nema 17 motors. Don't try to drive 0.9deg/step Nema 23 motors from the Duet, because I expect they would need more than 30V to reach reasonable speeds. The next-generation Duet will probably have drivers rated at a higher current than the existing Duet, but it will still be limited to about 30V.

Or you can use external DSP-based drivers attached to the expansion connector of the Duet.

So would a "long" Nema 17 be better? Can I go to 28V to the board? (have plenty of 27V supplies then can easy go to 29V)
I guess the real question is..... What are the stepper motor specs that are recommended? Duet085? DuetNG?

Quote
dc42
5. Another possibility: instead of a flying extruder, as you have 3 spare vertical towers, put another stepper motor, belt and carriage on one of them, and attach the extruder to that. The firmware would need to be changed to keep the extruder at a sensible height depending on the effector XYZ position.

Interesting idea and I like the concept but that heads back to the Bowden feed and it's (imo) various issues

Quote
PRZ
Plastic corners cannot give you bending stiffness, but if you stabilize the printer either by bolting panels or making a cross-bracing, they will be sufficient. The hexagon design is quite practical for that.
I am for panels because I prefer enclosed printers, especially if you print ABS. Also, heavy wood panels dampen vibrations and noise.
In principle, the height of a flying extruder don't vary much because as an average, the height of the effector vary very slowly.
So, their inertia is somewhat negligible for loads, you only have to live with their dead load. You can probably get an extruder for 250/300g all included, which could be easily lifted by Nema17, because one column will rarely carry more than 40% of the effector weight.
With standard boards, I am for Nema17, because you cannot improve with Nema23, and lose microstepping precision (the microstepping intervals are said being more accurate on Nema17 than Nema23). If you get in trouble, limit the travel movement speed.
Also, I am for larger minimim arm angle (say 23~25°) instead of usual 20°, which limit the maximum carriage speed at the cost of slightly longer arms. That is also better for printing accuracy at large diameter.

If using traxxas style links, you shall find a mean to increase friction, as this dampen vibration and limit banding effects. spring between arms is not elegant, but that helps.

You may use ball and socket from Igus, they have way too much friction for a normal delta, but that may do the trick for a large unit and being all plastic, they are very light.

You may have a look at the italian Wasp printers, they exist in large size, use flying extruders and Igus components. They are near of what you want to do.

Also, some have developed light extruder for delta with small steppers, but that may limit the retract speed.

If really at the limit, you may slightly increase the voltage near the board maximum (30V) , but I may be afraid of back EMF from steppers.
For a 24V setup, use 12V fans, that is easier to handle. A DC/DC converter cost nothing.

The printer will definitely have bolted panels on 4 sides with the front two sides having a swing door.

Looks like the 17/23 research will go on for a while

Thanks for the WASP printers info I'll go and read what's available.
.
Interestingly I have a bunch of 24V fans in both axial and radial from both the FFP that is 24v based and cooling enclosures for Christmas lights. Though I also have on hand some 5v-24v output fully adjustable 5A DC-DC converters if needed

Flying extruders are excellent, especially if you want quality prints. The extruder itself doesn't really move when the effector does, it just sort of stays in the middle and tilts. It gives you the benefit of a short bowden cable, without the moving mass of an effector-mounted extruder. About the only downside is that they look a bit weird.

Another option is the flex3drive, but that is quite an expensive solution that I haven't tried.

Test prints of the 120 degree corners, printed in ABS on the Flashforge.
The large one is printed at 0.2mm layer, 4 perimeters and 50% full honeycomb infill sliced with S3D
The inside bracket is 0.3mm layer, same infill/perimeter (first thing i've printed at higher than 0.2mm)

The SD card adapter is to provide scale, all sides that will bolt to the extrusion are 10mm thick to match with 16mm bolts



Here is three inside brackets arranged to check angle accuracy... the first batch was out by 1 degree


And this is how they will be arranged on the inside/outside of the extrusion


I haven't designed in motor mounts as the steppers will be mounted above and independent of the corners

Quote
nebbian
Flying extruders are excellent, especially if you want quality prints. The extruder itself doesn't really move when the effector does, it just sort of stays in the middle and tilts. It gives you the benefit of a short bowden cable, without the moving mass of an effector-mounted extruder. About the only downside is that they look a bit weird.

Another option is the flex3drive, but that is quite an expensive solution that I haven't tried.

Yeah, forgot to mention the flex3drive in post one..... it certainly should be on the short list based on what I've read about it and I don't consider it to be overly expensive, just not convinced.

Quote
nebbian
The extruder itself doesn't really move when the effector does, it just sort of stays in the middle and tilts.

What nags me on every video i've watched is either the tilt that must impart some pull and potential tilt on the effector. I just can't get my head around how if the bowden tube is short enough in the 4 to 6 inch range that the mass of the extruder then becomes more than loosely coupled to the effector. This of course is a non issue for short infill work or continuous outline work but surely becomes an issue for fast moves around the bed?
As i plan to do a lot of multiple part printing this is a concern for me

Question to the forum:
I have 8 x 6mm carbon fibre tube specified.
I see that roll wrapped tube is overall better but the smallest size I can find is 10mm x 8mm, though it weighs only 25% more at @ 40gm per metre

is 10x8 to big for arms that will be around 470-480mm long?


Second question whilst i'm posting:
Back to the stepper motor sizing.

each rod carriage will weigh in around 350 - 380 grams including LM20UU bearings, printed carrier, nuts bolts and the arms.

If i was to drop a Bulldog XL and E3D v6 hotend directly on the effector we add some significant weight, bulldog is listed at 560gm + e3d v6 + effector (100gm) = odds and ends so say 750gm.
Now right next to a tower any one extruder may need to move around 1000gms of mass.

This could be offset by counterweighting (and happy to do this) the entire weight of the rod carrier. Leaving only the static weight of the effector and the full active mass for acceleration.... Again I'll say happy to trade off acceleration if this negates any bowden setup.

Edited 1 time(s). Last edit at 03/14/2016 02:32AM by aussiephil.

That is fairly heavy.
You may use Nema17 , but you will have to limit accelerations. BTW, for stepper choice, I have written this page :
[reprap.org]
But with such weight, you may began to have problems with belts and you enter in the range of 'very large printer' with lot of specific solutions. That may not be justified for a 400mm diameter printer. Increasing the weight will lower the vibration frequency, but reducing the vibration may became quite difficult.

An example of relatively Heavy effector on a delta is found on the spiderbot (double hotend with full metal construction), but there was some reports of 'flying effectors' and they had to improve their magnet system to cope with that.

I advise against counterweight, as this double the weight so the inertia, already high. I have used counterweight on CNC for safety reasons, but high accelerations are more important on a printer than on a CNC (I use exclusively honeycomb infill, which is very demanding mechanically - with 8000 mm/sec acceleration on the D-bot).

Rods are not a good solution for large delta, even for medium size, because they are not sufficiently stiff without increasing significantly diameter, which drive to heavy bearings.
With an extruder on the effector, you want relatively stiff carriage.
Off the shelf, one of the best solution is V-slot with polycarbonate small wheels. You may have difficulties of supply in australia
Another choice (mine for the D-box) is steel wheel on steel profiles. I had good experience of it on CNC, so I used them for printer. My return of experience is not huge, but yet, I am happy with that. This is much lighter than any rod bearing.
This solution is also used on the Delta-pi with SS square tubes, and scaled up delta pi carriages might be a light and stiff solution.

Bulldog are full aluminium and not a good option for a delta.
I am not fond of super small steppers, but have a look at Ultibot extruder.

Think of delta moving parts as an airplane, you chase gram after gram.

For large printer, you will find some info in this post : [forums.reprap.org]

Edited 1 time(s). Last edit at 03/14/2016 07:58AM by PRZ.

---

Pierre

- Safety [reprap.org]
- Embedded help system for Duet and RepRap Firmware [forums.reprap.org]
- Enclosed delta printers Lily [rouzeau.net] and Lily Big [rouzeau.net]
- OpenScad delta printer simulator [github.com]
- 3D printing on my site [www.rouzeau.net]

PRZ, thank you, more reading and thinking material.
I take the point about the counter weights. The reverse thought then as this.... The counter weighting is essentially in the vertical direction whilst we have increased the total mass iT actually becomes easier to move back and forward. The see saw analogy comes to mind, with equal weight on each end it requires small force to rock it.

Take these thoughts with large grains of salt as they likely don't translate to real physics, forgotten nearly all my high school physics.

The goal is 400mm actual print size, bed will be closer to 500mm

Yeah v-slot is either impossible to source or stupidly expensive by the time it arrives at the door.

Whilst the 20mm rod pairs will 1600mm long, the longest unsupported length should be 1200mm or a bit less.

Ok, I made some exercise to see viability of system
first , one of the best online calculator for beam is here :
[skyciv.com]

two rods of 20mm have a bending inertia of 15700 mm4
an angle of 40x40x3, rotated 45, in its weakest orientation, does have an inertia of 14500 mm4, nearly the same of your two rods, at a fraction of the cost and weight.

But thats a bit too small for your machine, so I made calculation for 50x50x3 , which gives me 28800 mm4.
A bit surprised, I checked those number on old paper tables, and they are ok.
I have no data for rotated square tube, but you may have similar or better bending stiffness with a square tube 30x30x2 (delta pi is 20x20x2)

Note that it is imperative to use cold rolled angle, as rolling on hot rolled angle simply don't work, even with thorough cleaning and brushing.

A try to scale up the D-Box carriage for this 50x50x3 (which will be published for 40x40x2, while prototype is 35.5.35.x1.5) gives that:



Total weight including 13mm bearings may be somewhat around 120~130g
arm space is 84mm, same as D-Box effector, that I will published unmodified.

---

Pierre

- Safety [reprap.org]
- Embedded help system for Duet and RepRap Firmware [forums.reprap.org]
- Enclosed delta printers Lily [rouzeau.net] and Lily Big [rouzeau.net]
- OpenScad delta printer simulator [github.com]
- 3D printing on my site [www.rouzeau.net]

I have some original vslot and carriage and they really nice, but since you close to china .. I know robotdigg just get some vslot carriage, wheel and extrusion, I dont know about they quality but they are the exact copy of the american made one and so far robotdigg is a good china supplier with decent quality, you may want to try those, a least a sample to check the quality. Plus you can get your pulley and belts there and more (they have steel renforced or carbon gt2).

Quote
PRZ
Ok, I made some exercise to see viability of system
first , one of the best online calculator for beam is here :
[skyciv.com]
two rods of 20mm have a bending inertia of 15700 mm4
an angle of 40x40x3, rotated 45, in its weakest orientation, does have an inertia of 14500 mm4, nearly the same of your two rods, at a fraction of the cost and weight.
But thats a bit too small for your machine, so I made calculation for 50x50x3 , which gives me 28800 mm4.
A bit surprised, I checked those number on old paper tables, and they are ok.
I have no data for rotated square tube, but you may have similar or better bending stiffness with a square tube 30x30x2 (delta pi is 20x20x2)
.

PRZ
Again thanks for investigating and running numbers, very much food for thought, neither the rods nor t-slot is ordered so will pause and digest everything so far

Quote
GroupB
I have some original vslot and carriage and they really nice, but since you close to china .. I know robotdigg just get some vslot carriage, wheel and extrusion, I dont know about they quality but they are the exact copy of the american made one and so far robotdigg is a good china supplier with decent quality, you may want to try those, a least a sample to check the quality. Plus you can get your pulley and belts there and more (they have steel renforced or carbon gt2).

My only concern about ordering lengths 1600 to 1800mm long direct from China is I've seen how parcels generally arrive from China (I buy a fair bit of other stuff) and I would be worried about bending and dents.

GT2 10mm Steel reinforced belt has been ordered and shipped just so I can get hands on it

.

Edited 1 time(s). Last edit at 03/14/2016 06:13PM by chris33.

---

Check my rubbish blog for my prusa i3

up and running
[3dimetech.blogspot.co.uk]

Chris33..... is that a cryptic pointer to your blog?

Quote
aussiephil

Quote
GroupB
I have some original vslot and carriage and they really nice, but since you close to china .. I know robotdigg just get some vslot carriage, wheel and extrusion, I dont know about they quality but they are the exact copy of the american made one and so far robotdigg is a good china supplier with decent quality, you may want to try those, a least a sample to check the quality. Plus you can get your pulley and belts there and more (they have steel renforced or carbon gt2).

My only concern about ordering lengths 1600 to 1800mm long direct from China is I've seen how parcels generally arrive from China (I buy a fair bit of other stuff) and I would be worried about bending and dents.

GT2 10mm Steel reinforced belt has been ordered and shipped just so I can get hands on it

They dont use china post, they use DHL, Fedex and similar service. You can also ask robotdigg to put the extrusion in a solid shipping tube I bet they could do that

AussiePhil, where did you get the GT2 10mm width (and pulleys) ?

Pierre

Quote
GroupB

Quote
aussiephil

Quote
GroupB
I have some original vslot and carriage and they really nice, but since you close to china .. I know robotdigg just get some vslot carriage, wheel and extrusion, I dont know about they quality but they are the exact copy of the american made one and so far robotdigg is a good china supplier with decent quality, you may want to try those, a least a sample to check the quality. Plus you can get your pulley and belts there and more (they have steel renforced or carbon gt2).

My only concern about ordering lengths 1600 to 1800mm long direct from China is I've seen how parcels generally arrive from China (I buy a fair bit of other stuff) and I would be worried about bending and dents.

GT2 10mm Steel reinforced belt has been ordered and shipped just so I can get hands on it

They dont use china post, they use DHL, Fedex and similar service. You can also ask robotdigg to put the extrusion in a solid shipping tube I bet they could do that

The Fedex and DHL drivers I know on a first name basis, very little do I ship via china post, even so boxes will arrive dented, it's just a fact of international air freight handling. I have however browsed RobotDigg site a few times now
Cheers
Phil

Quote
PRZ
AussiePhil, where did you get the GT2 10mm width (and pulleys) ?

Pierre

Hi Pierre
got belt and idler pulleys from here -> [www.aliexpress.com]

I didn't order drive pulleys as the Nema motor sizing is still up for grabs and I didn't want to order two different shaft sizes

Cheers

PS really happy how fast they shipped the item, though I did leave it on the slow china post route as I'm not in a rush.

Edited 1 time(s). Last edit at 03/14/2016 07:06PM by aussiephil.

Despite have Traxxas rod ends on order I've decided in the last few days to head down the ball/socket set up for the effector/rods/carriage. Not magnetic but using spring/tension retention. I've read people having big issues getting them connected up but they seem to work well and give full movement.

So the Easter weekend has been a design exercise for the effector and carriage to enable mounting 10mm M4 thread balls into each and to design the actual rod ends for the 10mmx8mm carbon fibre rods.

Here is a screen grab of the current effector design.

90mm between the ball centers for 90mm rod spacing
the middle hole between the balls will hold a small eye bolt to attach the line for spring retention.
The centre hole is 44mm diameter to hold custom removable inserts for hotends or alternative extruders.
The additional holes allow for wiring or air feeds to pass through to keep things neat.
Can't see in this grab but each edge has 4 x 3mm spaced 20mm apart for custom fan brackets if needed.

Effector plate is around 60 grams and seems to be flex free.

We will be going with a flying extruder setup now

First print of the two part effector design has been done and everything is a nice snug fit, the inner ring is basic to check for fit and will be customised for the E3Dv6 mount and cooling ducting to allow for externally ducted air to cool the hotend.

the two pieces apart


and together with one M4 machine screw to check it can be screwed down,

I use this system with two wires per side on the D-Box [rouzeau.net] [rouzeau.net] and Lily [rouzeau.net].



'Spectra' wire on the D-Box (60 kg ?) and 16kg fishing nylon on the Lily.

On the D-Box, which have a heated chamber, with my quite small bending moment (due to small space between wire and arm), I got effector bending, so I think a central wire is not a very good solution and might be unusable with a heated chamber. The bending is not quite visible on the photos, but my magnets raised by ~0.8mm. I am contemplating a reprint in nylon or polycarbonate (deflexion temp 150°C, so no fear of deformation).
It shall be noted that I somewhat overheated my chamber (~60°C) for the printer material (PETG), following a mistake (I made a short, and that made my chamber SSR permanently conductive).

The big advantage of this setup is you can modify the wire tension, which modify the ball/socket friction. This creates dampening which may allow to reduce/eliminate some banding.

I you print ball sockets, there shall be a snug fit which will reduce with time. It might be a good idea to make a high speed large 'air print' (to use full arm travel) of a few hours.

Edited 1 time(s). Last edit at 03/29/2016 02:01PM by PRZ.

Thanks PRZ, hadn't seen the dual line setup before and it seems to be logical and sensible....... I wonder if making it one run of spectre line looped at one end but still parallel to each arm and dual springs would allow for tension variation on each arm to naturally equalise?

Easy for a couple holes to be added to each side of my design though so thanks for pointing out the dual line configuration.

Cheers

Progress is somewhat slowed by parts not yet arrived from China, though steppersonline got the nema23's here pretty quick so whipped up a mounting plate for them. The plate is designed to mount to the planned flat base plate that will be on top of the bottom extrusion set and underneath the printbed area.
The 23's are 0.9 degree units and were not expensive so if no good then they can be used for something else.

The camera certainly highlights all the "flaws" on the print yet when I feel the prints they do not seem anywhere near that bad, the one mistake was not allowing easy access for the bottom nuts, lucky I could get to it via the hex holes



The rear has some M3 sized holes to attach a fan or cable management bits.

I don't know how you think to build your plate, but something like the image below may transfer belt load better and might be lighter and easier to assemble (note that it is for Nema 17, so proportions are not ok)

Bits keep dribbling in....
Arrived this week.
Various M4 and M5 bolts and nuts. more to arrive yet
The 2040 T-Slot
the 20mm steel rod

First thing to get test assembled was the steel balls to the effector and oh crap that is when I discover that the FFCP steps/mm is slightly out, just enough to give a 1% variance on the ball spacings on the three sides, had been unnoticed as all the other bits where printed in the same orientation.

one 100mm calibration cube later I got to reprint the effector


I'm down to 0.5mm variance in the ball spacings on the three sides.

Hand testing of the printed rod ends with the ball cups seems to confirm that they will work fine. The rod arms are now 10x8mm @500mm long


due to calibration issue above had to reprint the frame corner, test fitted the 2040 and 20mm rod to one... tight fit but a rubber mallet worked

Here is an image looking down from around 1200mm high, this was free standing on the floor with no top bracket, everything as square and solid. the yellow piece is the rod carrier... still waiting on the bearings to arrive from china.


Here is the corner bracket being printed, the slot for the 2040 is a nice tight fit.

I put the bracket into a vice and let 2 metres (6') of 2040 be suspended horizontally...... whilst it wasn't laser level, any pressure on the extrusion actually caused the vice to move on its mounting bolts to the bench rather than flex the bracket. I'm no longer concerned about flex on the printed bracket. Oh that was without any bolts in place

Why not use bigger hole in the effector for the ball so you can adjust them perfectly with something like a jig for spacing and a dial for the Up/Down? The jig then can also be use for the carrier. You can do better than 0.5mm , this is a important measure to get right, try to aim in the 0.05mm (0.002 inch), we talking about the tilt problem if those are not right. I know its hard and will take time to get that kind of precision but at the end you will avoid problem.