What would you build if you were starting again now?

bot

I would make a few small iterative changes to my current design. Actually, I will be making printer number two soon with a few changes:

Drive the Y axis only by one side instead of both. Currently I use a ridiculously long primary belt to drive both sides of the Y axis. This causes unnecessary backlash because the primary is as long as the secondaries (lol. I know now.) I will just use a short primary like on X. Hopefully this immediately solves the backlash on Y (it's in the order of <100 microns of backlash)

ajdtreyd

@dc42

I have to say I agree, though IDEX and core-xy are really tempting.

I have a SeeMeCNC rostock max v2 that I built from kit in late 2014 as my 1st (and still only) printer. The early days were constantly alternating between joy and intense frustration.

I have updated most every aspect of it over the years and now have a printer that I really enjoy. When I get tempted by Voron videos and the like, I just think about how easy my delta is to work on. I can't see any other platform being so simple, accessible and quick.

So like you, I would build another delta, only bigger with an all metal frame!

ajdtreyd

@dc42 does that new SE plan also include an integrated accelerometer? Surely there's room!

dc42

@ajdtreyd said in What would you build if you were starting again now?:

@dc42 does that new SE plan also include an integrated accelerometer? Surely there's room!

Yes, of course! Although my delta exhibits very little ringing. Basically it will be a combined Smart Effector and Tool Board, although we will probably leave out some of the additional IO pins.

PaulHew

@zapta I hate the gantry. 'U' shaped. Personally I think there is tooo much happening with it. I think I should of built the 1.8 where X/Y is static and the bed moves...
And Klipper. I updated it and Klipper threw a its rattle out of the pram.
IMO with Klipper and SKRs, everything is a workaround. Did try Klipper on a Duet2/Duex5 on the RailCore and the Bed level hardly worked. Back to RRF and worked everytime...

PaulHew

@phaedrux Friend of mine has his own toolchanger on a Voron 1.8
https://www.youtube.com/watch?v=ZICOevBpO0A

fcwilt

@paulhew said in What would you build if you were starting again now?:

@zapta I hate the gantry. 'U' shaped. Personally I think there is tooo much happening with it.

Agreed.

zapta

@paulhew said in What would you build if you were starting again now?:

And Klipper. I updated it and Klipper threw a its rattle out of the pram.

I am not sure if V2.4 requires Klipper or any specific hardware. For example, Youtube Tom built recently a V2.4 with a duet.

o_lampe

@droftarts said in What would you build if you were starting again now?:

@o_lampe said in What would you build if you were starting again now?:

a # hash-style printer

I guess you mean https://youtu.be/tzf8903FvCs
Looks interesting!

Ian

Exactly,
it's a bit of 'hen or egg' problem regarding slicers and firmware, but RRF kinematic model is already working. Until then it's a single object quad-color printer with very quick toolchanging. My macro for ooze compensation based on standby time has already prooven to work.
I'd build it with 'anti-racking constraints' instead of the classic 'shaft & dual belt' approach shown in the video.

PaulHew

@zapta Voron2.1 you had 2 options, RAMPs on Arduino boards and a Pi or Duet2/Duex5, from the BOM and there were baseline configs for both.

The BOM now states, SKRs and a Pi and there is a baseline only for Klipper.
Within the Voron Discord, there is a Duet_RRF group, still.

My V0.1 is running a Mini5+ with RRF.

opentoideas

oh boy you have all given me lots to think about...

one of my biggest regrets was when I bought my CR-10 I think the chap was emigrating and was also selling his Delta with Duet. I offered him more than it eventually sold for but he wanted to let it run on Ebay and I missed the end of the auction.

probably would have still been cheaper than what I ended up spending to get my printer where it is now and I suspect it was a very nice machine but I was more than a bit daunted by it.

that hash design is interesting and perhaps a reason to go with a croXY style as it would potentially be adaptable.

lots to think about and I am not looking to build just yet so off to read all I can on DC42's Delta

o_lampe

@opentoideas said in What would you build if you were starting again now?:

that hash design is interesting and perhaps a reason to go with a croXY style as it would potentially be adaptable.

Just started to play with some ideas to build mine

dc42

@opentoideas said in What would you build if you were starting again now?:

lots to think about and I am not looking to build just yet so off to read all I can on DC42's Delta

The details are at https://miscsolutions.wordpress.com/dc42s-large-kossel-build/. These days I would choose a Duet 3 Mini to control it.

fcwilt

@dc42 said in What would you build if you were starting again now?:

@opentoideas said in What would you build if you were starting again now?:

lots to think about and I am not looking to build just yet so off to read all I can on DC42's Delta

The details are at https://miscsolutions.wordpress.com/dc42s-large-kossel-build/. These days I would choose a Duet 3 Mini to control it.

Are large delta's easier to get working well than the mini's?

I built three different mini's and was never happy with the print quality after I built my first Cartesian.

Frederick

dc42

All deltas need to be constructed with precise geometry. Cheap delta kits fail to achieve that.

zapta

@paulhew said in What would you build if you were starting again now?:

The BOM now states, SKRs and a Pi and there is a baseline only for Klipper.

I think that the Voron V2.4 crowd is now moving toward a Pi + a single controller with 7+ driver such Spider or Octopus.

The Spider for example is said to be developed with input from the Voron dev team https://wiki.fysetc.com/Spider/ and I think it's supposed to come also with a more capable CPU that is will RRF compatible (more RAM?).

I wish that Duet would have a reasonably price stand alone board with 7+ drivers and WiFi. That would be my first choice for my V2.4 (a work in progress).

o_lampe

@zapta said in What would you build if you were starting again now?:

I wish that Duet would have a reasonably price stand alone board with 7+ drivers and WiFi.

...like the Duet Maestro with 2-channel expansionboard?

zapta

@opentoideas said in What would you build if you were starting again now?:

With the knowledge you now have if you were building a new machine what geometry and machine type you build?

I am actually building a new machine, a black and red Voron V2.4 250mm with Clockworks (direct) extruder.

Sources the parts individually per the sourcing guide, which took a while, and the frame and x/y/z movements are already assembled.

I picked the Voron because it has a strong supporting community (I can get answers within seconds on Discord, or watch detailed videos by Nero 3DP) and it's a proven design with 1000+ people that built it. The design is on the overkill side, but I am taking my time and am happy with the results so far, especially with how stiff the x/y carriage is compared to my existing HEVO.

zapta

@o_lampe said in What would you build if you were starting again now?:

...like the Duet Maestro with 2-channel expansionboard?

I meant a single board, no expansions, for simplicity, and for avoiding CAN bus based issues.

As for the Maestro, IIRC it has no Wifi and limited MCU that is marginal for the upcoming input shaping.

The Chinese competition, with capable low cost non-clone new boards is becoming strong, and Duet's lead is shifting to better customer support and warranty rather than hardware feature set.

droftarts

@zapta I think @o_lampe maybe meant Duet 3 Mini 5+ WiFi with Mini 2+ expansion board. Also pretty cost-effective (when you factor in documentation, support, community as well) against cheap Chinese boards.

While there is a market for standalone boards with 7+ drivers, it's not a big one. Most machines still only use 4 or 5 drivers. There's also more to go wrong on a larger, integrated PCB, and brick a whole board. There was also more demand for having tool boards rather than monolithic, single wiring point electronics, hence CAN is used now.

Ian