3dhubs.com Service Bureaus

[C855B]

Does anybody here use a 3D printing service they found on 3dhubs.com?

I ask this since it seems that nearly all or at least most of the discussion around here has been about Shapeways. In my case, I've been shopping for alternatives especially since the end cost of producing a moderately large project I have almost ready to go - the UP Kelso station featured in MR, but in N - is closing in on $500 at Shapeways. Obviously I'm balking at this.

This project appears to be workable on a larger-format hi-res printer, such as Shapeways' ProJet 3000 in UHD (FUD). Somebody with a middle-range Envisiontech printer (Vector Hi-Res 3SP, maybe?) could probably handle it, too.

What hits me as strange is the estimates from 3dhubs.com put the cost from 1/2 to almost 10X the Shapeways pricing. Yes, that's right - there were estimates north of $4000. WTF? Maybe that's assuming a "fully-finshed" model? I guess my biggest concerns about 3dhubs.com experience are, "Were you surprised when you got the bill?" and "How was the quality?"

TIA!

[wcfn100]

Do you have a rendering of the 3d drawing you can show to see how you're trying to print it?


Jason

[C855B]

I'm sure with work I could break it down into components, such as collapsing the walls of the core and even doing the walls on a Silhouette cutter instead. That's been discussed over the layout thread. I have also uploaded a test with just the roof and decorative parapets, and it saved $100 or so. However, and frankly, the more hands-on effort on my part to make it work, the less likely it's going to happen. I'd like to simply hand off the STL files and push a button, more or less, trading dollars for time, within reason. One-armed paperhanger situation is going on at the moment.

[wcfn100]

Mike, on the Shapeways page that shows all the materials and pricing for the depot part, can you hover your mouse over the price for the FXD and capture the relevant numbers?  Not trying to convince you of anything, but would like to see how much of the depot cost is the support material.


Jason

[wcfn100]

Just as an example (and why I'm so interested), I started drawing this tank car a few days ago.



Here's the price break down.



So for a $29 model, $11 is just support material (the red).

Jason

[C855B]

Looks like nearly half the cost is support material. Building core and colonnade, respectively:

[ Guests cannot view attachments ]

[ Guests cannot view attachments ]

[wcfn100]

Thanks Mike.

I would think just separating the roof from the walls could save a bit because you wouldn't need the support throughout the inside of the building.  But I do understand the want to keep it as simple as possible.  I struggle with that every time I start drawing.

Jason

[Lemosteam]

Mike there is no real reason to keep it simple if this is primarily for you (or even if it is not).  Use the same techniques that model kits did when you were a kid.

Also, the machine space is a huge chunk of the cost.  If you had separate walls, each would be $1 apiece, but the machine space would be vastly minimized (height also comes into play hear in the machine space too).

Each section of the colonnade could be designed to fit into the next, for example with the ends a separate glue on piece.  I have developed a rule of thumb for this: .075 gap all the way around fitting pieces.   So if you have a pin or rectangular pin the mating hole would be .075mm gap all the way around. 

This gap is so tiny yet printable it is barely noticeable and you won't believe how accurately they will assemble.

The roof could also be done with lap joints along the mating ridges with each gable end having pins to align the roof sections like a gauge.  The tiles along the ridge would hide any gaps. Just take your current model and copy it 20 times, and chunk away what you want in each piece, and add tabs and pins where you need them.

You would have to continue to play around with the orientation of each piece, then arrange them in a single multi-model and the only additive cost would be the per-part fee.

To avoid this fee you could sprue all of the parts together with a 1mm square or round sprue

Just thinking out loud here.

[C855B]

One underlying problem here is the source files are merged into single objects. It will be an immense amount of surgery to cut things up for output economy. If you read the MR article, you'd discover that at nearly every design step he merged everything, albeit keeping the intermediates (something like 200 of them) as undo files. I only have the resulting merged products of the two sub-structures.

What's worse is he did the files in Autocad 123D Design, freeware which is no longer supported - thankfully I found a downloadable that somebody had rescued. Conversion to another format was tried, we had two weeks of back and forth trying to find a convertible common format, but as I discovered once I had the source files in-hand, one of the big problems with the Autocad freebie was its conversion (in)capabilities. Something simple, no problem, something complex, forget it. I think the merged objects played a role here, too.

So... yes, under normal circumstances we could marquee-select object groups and prepare the output in pieces. With the CAD files I have in-hand, not so much. That's why I'm saying while cost is a consideration, it's not the only consideration, I don't have the time resources to manage all the reverse-engineering necessary to control output costs. If I can find a service bureau that could output the files as they are with decent quality for, say, $250 total, I'd be fine. That's my pain threshold, and is why my original query was about alternative service bureaus.

[Scottl]

I printed my rock sheds via a provider I found on 3Dhubs  that was local.  It worked out very well and economically for me (1 day turnaround, low cost, no shipping) but he was limited to PLA which is not suitable for finely detailed models.  Factoring in all costs, it worked out to about 1/4 of the cost Shapeways quoted, and their turnaround is slow by comparison.

[tom mann]

Is the building symmetrical? Could you just print 1/2 of it 2 times?

[railnerd]

Might want to also check out https://fictiv.com/

[Lemosteam]

@C855B Surfaces ain's so bad to work with... but I love this stuff so...

[C855B]

@C855B Surfaces ain's so bad to work with... but I love this stuff so...

Clearly you do. With your relative success with the converted vector file, I ought to take your advice and take another whack at separating the walls in the original CAD file to print as flats. Still need to come to some conclusion about the windows, especially determining whether available Tichy windows can be made to work or if I have to draw those too.

That will have to wait for a bit; I'm making headway again on the C855 trio, so the scarce modeling time this month is consumed by that project.

[bbussey]

I would separate the building into flats as much as possible and orientate them as flat, which would save tremendously on the cost as well as eliminate any visible stepping.