Go to homepage
3D Printing
3D Printing Service
FDM SLA/DLP SLS MJF DMLS/SLM Material Jetting
CNC Machining
CNC Machining Service
CNC Milling CNC Turning Surface Finishing Part Marking
Injection Molding
Sheet Metal Fabrication
Resources
3D Printing Guide CNC Machining Guide Injection Molding Guide Metal 3D Printing Guide Knowledge Base Webinars Help Center
Get instant quote
Jun 2017
21 / 24
Jun 2017
Jun 2017
bardiir
Jun '17

I’ve looked through the suggestions for improvements and features yesterday and I’ve seen this request. While I feel like it would be a great addition to have a complexity modifier or runtime as a pricing variable I’m already doing this and thought I’ll just share my pricing setup as an idea for everyone who wants to include complexity into their prices.

Basically what makes an FDM printer slow is the perimeters. The more details you got the more perimeters you have.

So my basic idea was. The more details there are the higher the surface area is. I put the raw material cost into the ccm pricing and then added a cost calculation point for surface area, quite highly priced, uploaded a cube and tuned that price to my target for a low complexity object. Then uploaded a model of just thin walls and checked if the cost is where it should be. If the complex part was too highly priced I would shift some cost from surface to ccm for the cube baseline and if it was too low I would shift from ccm to surface

Then it was just a matter of relative pricing. If the filament was priced differently duplicate and correct ccm of the print speed was different for example flexible filament duplicate and adjust surface cost.

Prefect prices every time now. Adjusting the surface area this way and having a little overshoot also calculates quite nicely into support material cost.

bardiir
Jun '17

-> This is a point where 3DHubs could really improve a lot, offering you your own model uploads for pricing calculations/estimates in the preview not just different sized Marvin.

edservices
Jun '17

Really like the idea of uploading own models to estimate/preview pricing instead of Marvin method.

I had this issue before where a relatively complex part taking ages to print but quoted cheaply due to sliced volume pricing model. Later adjusted the sliced volume price but ended up simple part costing too much! Really appreciate your suggestion to add surface area costing method, will give it a go.

Odense3DPrint
Jun '17

i will give that a go also when i bring my hub back online after machine upgrade

a screenshot could have been nice

Odense3DPrint
Jun '17

it was easy enough to set up.

i used a combo of surface area and objecct volume. The latter because it makes it dead easy to set filament price… sliced volume is an object with only 25% infill. So going for 100% infill makes it more easy, just calculate the raw filament cost and enter it

Odense3DPrint
Jun '17

hmm…

i choosed one of those hairy models as the complex one and got the price where it would say it was ok

then i picked a 10cm marvin as the simple model… now the price is way off

there is an hour in difference for the print time according to cura engine in repetier host. but the price at 3d hubs is 50% lower

so what could be done wrong on my part?

Odense3DPrint
Jun '17

used this one as the complex: Rocktopus 2 ( 2 hands and hair ) by joeostrander - Thingiverse 1

it takes about 6½ hours to print and the price on github reflects that ok

then i took the 10cm marvin as used in the preview… its a hughe one and takes about an hour less… its a simple one with not so much surface, but it still takes time to print, pricing is screwed on that one

but if i take the standard marvin it hits the price ok again

is there anything that can be done about this? so close to perfect prices and still not at the finish line

bardiir
Jun '17

Simple part too cheap -> Set ccm higher, Surface lower

Simple part too expensive -> Set ccm lower, Surface higher

bardiir
Jun '17

I would count small marvin as complex, there is very little infill compared to the shells. So you have set the surface price too high and need to set higher ccm and instead lower the surface a bit. On the other Hand I would not select something like Hairy prints or other stuff that has some “special” properties as a base pricing. You’re more likely to encouter someone without a 3D Printer beeing interested in this: https://www.thingiverse.com/thing:2230899 1 rather than a hairy print which is sort of a 3D printer novelty and mostly interesting to people who deeply know the limitations and thus think the hairy stuff is cool :smiley:

Odense3DPrint
Jun '17

hehe

i just picked the first thing on thingiverse that would have a lot of surface area

will give deathwing a try

so you would recommend using deathwing as the “complex” and then a simple cube as the “simple” model to set prices?

bardiir
Jun '17

sliced volume, but actually, thinking about it object volume might make more sense as it doesn’t involve the perimeters. In the end using sliced volume and surface does count the perimeters double. And what you’re doing with using object volume and surface area is sort of replicating the sliced volume but with your custom weighting for material/complexity.

edservices
Jun '17

I am really very very shocked! Is this hairy thing really printable with FDM printers?seriously? Most likely I’ll reject such job if ever come across it, very low confident to deliver.

bardiir
Jun '17

Well Deathwing is as complex as it gets for me, lots of support, very fine details, lots of it. And a cube is as low as it gets so i guess these are nice extremes yes. Everything else like hairy prints or other special stuff would be stuff to discuss with the customer directly.

Odense3DPrint
Jun '17

i picked object volume since the explanation for it says that its 100% infill and that is what my raw price is once i have converted from kg to cm3 and figured the price

i assume that 3d hubs does some internal calculation to figure out how much material is used and then calculating the cost based on that

i tried to slice deatwing in repetier at 0.1 layer heigth, it looks like a dragon with the wing skin taken off

http://imgur.com/a/cCL29

any special settings to use to make the skin show up?

bardiir
Jun '17

Yeah it’s totally easy but it does screw up pricing a lot because the support is rarther special for that. Basically what it boils down to is a print where there are hairs standing out and you comb it into shape with a hair dryer to soften the plastic stands up again, this is a special technique in handling a special support structure, it’s not printed like it looks directly.

bardiir
Jun '17

The skin on the wings is veeery thin. Your slicer is probably ignoring it due to the minimal wall thickness. You can choose something different with a lot of wallt too of course, on the other hand, even without wings deathwing is quite complex, just not printable in that case :smiley:

Slic3r kind of has a hard time with it too, Simplify3D does handle it better if you choose very thin layers.

Odense3DPrint
Jun '17

for the purpose of prices i can set it as thin as needed to make the skin work

can you tell me how thin walls would be needed to make it work?

what i do here is to compare the time in my slicer with the final price in 3d hubs to check that i have set the price for surface right

bardiir
Jun '17

I’d advise you not to do that, because that would make print times unrealistically high. Be honest, you wouldn’t accept deathwing to be printed on the hub, would you? You should choose the most complex objekt you could think of that you would accept as an order, otherwise you’re skewing the scale potentially a lot.

25µm Layers will do the trick for me with my 0.125mm Nozzle size.

Odense3DPrint
Jun '17

no no—

what i’m doing here is using deathwing to figure what price i should input

hmmm call it a guide line for what my prices should be if i accepted it… which i would not do

you could also see it as a way to reach the very extremes

Odense3DPrint
Jun '17

funny

i told cura to use 0.125 nozzle and 0.025mm layers and even then it could not do it

what would an approx print time be with those settings? just for the fun of it

bardiir
Jun '17

As I can only print veeeeery slowly with the fine nozzle due to the pressure buildup Deathwing takes a whopping 470h to print, and that’s just the estimate, I guess after Acceleration kicks in it’s closer to 550h :smiley:

Odense3DPrint
Jun '17

thats crazy

i ended up with a hollow cube and the small marvin and the biggest one

once i get the extruder i will put the printer online and i guess i will just have to adjust surface as i get orders