3D-print at low costs again - Mallory, how to realize it

To sum it up.

Basic investments:
printer 445 Euro (WANHAO Duplicator i3 with minor customizing)
2017-08-19 edit: Finance office refunded the VAT to me (19%). I am a freelancer and run a mighty one-man-enterprise. So I paid 360 Euro only.
CAD-software 850 Euro (Rhino 3D)
same edit: means 689 Euro.

1300 Euro to start at the lower consumer area is okay for me. The results are not overwhelming but satisfying - as expected.

The conclusion from nearly five years testing phase is:
You should know what you want and how to realize it. Hard- and software doesn't matter that much. Knowledge and experiance count!

The best way to check out the use of 3d-printing is to set up a meaningful goal. Once I saw a picture:

And I thought by myself: "Wow yeah, that's pretty cool! I would like to have this babe sitting at the corner of my window." That was the basic idea. I had a real need to realize it. I own a 3d-printer and some software. Can I model this woman? No. I am a mechanical engineer. I can make rectangular and circular geometries. But I will fail with natural bodies. And here comes the tricky part: How to get a good 3d-model? I ordered it from an experianced modeler (Eugene Petrov, ZBrush). Costs: US$ 400. I called her Mallory according Mallory Kane in the movie Haywire.

The goal is to get a photo of my printed and painted figure wich looks nearly like the original photo. I've made a first comparison in 3DS Max:

Not perfect yet but acceptable for me so far.

Eugene sent a splitted model to me:

The reason for the splitting is simply a manufacturing issue. It's easier to paint single parts. Old Eugene made a very poor version of the Desert Eagle. So I decided to make my own version. I searched for an existing good free 3d-model, used it as base to build a new one wich was waterproof and merged it to model.


Next important question: How big shall Mallory be? Half a meter seemed comfortable to me:


Well, my printer's volume is 200x200x180 as shown in red in the picture:

Not that big challenge, I think. I can split the model into smaller parts. But there are other challenges. How can I print the parts without or with less suspension? How can I fit the parts together? How can I stick the parts together? How can I make the figure stable?

My printer uses PLA. There is no glue that helps with PLA. In difference to ABS wich can be sticked easy by Aceton. Vice versa: I'm afraid you can make a tank for Aceton from PLA. So what to do to connect parts. The idea is to print a holow part with locking device, thin wall (0,8 mm) and fill it half with nearly fluid gypsum. Then add the next part and fill it half again with gypsum. So the print is the outer hull only. It can be thin - fast to print. The gypsum will deliver connection, stability and weight. That's the theory and what I'm trying atm. Printing the upper body atm:

The pink parts are the "connectors" wich will be printed seperately. They define the accurate positition. This is an important part of the story. The printer isn't very accurate. It squeezes all dimensions by 0,1 to 0,2 mm. Holes are smaller than the original model and shafts are bigger. Using seperate connectors offer the ability to change the connectors only if they don't work correctly and print them again quickly.

 

Well, first part done. Looks pretty good so far:


6:40 h, 58 g PLA

 

Haha, well the goal is to check out the possibilities of the low cost consumer area. Since the hard- and software were bought for several projects the costs per project are nearly insignificant. The control unit of my printer tells me that it was working for 18 days and 14 hours already. That means 446 hours. So I am at 1 Euro per hour until now. But I hope it'll work a bit longer. Rhino was bought for other purposes. I work ocassionally as freelancer and earn money. Meanwhile I got thousands of Euro because I own and use this software - several times the sum I once paid. Only the time I spend with the project counts. 4 hours so far. But it's a hobby. Fun even. Other folks pay for their hobbies. For me it's for free. What I really have to pay are the material costs: 0,02 Euro/g. Total costs for me so far: 1,16 Euro PLA and 6,40 Euro printer. I call this cheap.

But that's my personal situation only. The reason for this thread was to let other people know, how cheap 3d-printing can be. And how hard it could be to get good results if one can't do all this work by himself. I could ask shapeways or other services. But I'm afraid the model is too big for shapeways and if not it would cost hundreds or thousands of Euro.

 

Uhhh ohhh!
A quick test run at Shapeways:



Holy crap! That's just one small part of the model!

Additional I checked Shapeways' PLA print parameters compared to my printer wich costed me nearly exact the same sum as they offer for this small part - ROFL!

Any further questions? LOL!
Especially the accuracy at Shapeways is really a no-go! The worst case means: the hole is 0,5 mm to small and the shaft is 0,5 mm to big. ONE millimeter overlapping at a maximum length of 285 mm? How ever, this is the worst case. But what to do when this tolerances are very important for my model? I made a few tests to learn more about the tolerances of my printer.
This one was with an offset of 0,5 mm and it was to much.

Costs: 0,90 Euro
It lead to 0,4 mm wich is okay. It envelopes a tolerance of +/- 0,2 mm produced by my printer. This test would cost at Shapeways:



I think I will stick with my better accuracy and 7,56 Euro instead of 430,00 Euro. Nonetheless I will check all the other parts at Shapeways as well when I'm done with them. I'm afraid I will end up at something like 4000 Euro there. Seriously: Who would pay so much to bring a photo to the corner of his window?

 

Okay, Friday a new roll of grey PLA-filament arrived. Time to continue!

Now it becomes really tricky. Next part shall be the lower body of Mallory:

While the upper body was completely printed without any suspension or filling, this one will need many suspensions. Without them the printer has to put material into the air at to much places. I tested another position:

But it's not really helpful. So I went back to the first version and decided to accept all this suspensions - outside and inside.

I think this is the point to name the software I use for this process. The main helper is Repetier-Host V2.01 for computing the printer-data. It's for free and in fact an enviroment for position, scale, multiply and place parts at the one hand and set up all parameters for the printer and slicer at the other hand. It offers different slicers:

I checked Slic3r in the past but wasn't very satisfied with it. Usually I use CuraEngine. Probably if Mallory's arse becomes too tricky I will check Slic3r again.

Cutted polygon nets exported from Rhino often tend to be not waterproof. Repetier-Host gives a warning about it but offers no help for repair. That's the reason I use netfab Basic. It is for free as well and recommended by Shapeways btw. Usually the repairs from this Basic version worked very well for me - and shapeways (see above).

The start is an important part. The print must stick at the print platform:

Later it's important to check the layers for air-printing. You can see a part that would be printed in the air without infill:

An infill of 3% helps a bit. The print won't be very good. But since this detail dosn't matter much (it's inside) I accepted it.

I set the infill to a very low value because I have to remove it later. I want to fill the model with gypsum. The outer suspension is minimal. This is important for finishing the model.

Good feature in Repetier-Host is, you can inspect single layers:

I hope it'll work. We will see. I'll start tomorrow. Meanwhile I will ask Shapeways for costs

 

It took 40 mins to upload the STL (12,7 MB/100k cable connection) and calcutate it. That's okay.



What's not okay, is this:


And the connector:



Three parts only:

Costs at Shapeways so far ~852 Euro. Estimatet 21 Euro for me. And I have no clue how the quality would be from Shapways. Not to speak about the incredible tolerances.

 

The printed connector fits well to the upper body. Exactly wat I wanted! So I continued with all other connectors I designed so far for upper and lower body. It's a good idea to keep all the small parts until the current roll of filament is nearly at the end and use the new one for the bigger parts.

Here is, what Shapeways offers for the following parts:


All the connectors so far:

Mallory's arse at 8 mm hight atm:

Looks good so far. But all in all it'll take more than 11 hours. I think I will watch the original Mallory Kane (Haywire) meanwhile, sleep a bit and inspect the printed part tomorrow. I made a sheet with the data:

It's really impressive, I think.

btw. that's my little workshop:

 

And here is what I found in the morning:

The printer decided to become a lunatic. Very sad because it was nearly ready. Only ~20 mm of hight before the end. And I don't know why. The good part is: I noticed that the holes for the connectors in the legs are twice as deep as they should be. So I have the chance to correct the model before I try again.

 

Well, the scheduled change with the depth is to difficault. Rhino don't accept my boolean operations for this. And I'm too lazy to struggle with Rhino at this unimportant detail. I made another decision: I splitted the part nearly at half hight and added a further connector.

11 - 15 hours for another try with uncertain result is too much. Positiv side effect: I can place each of the parts on the print table with a wide flat area and avoid all external suspension - hopefully. Presumably I will have to increase the infill a bit.

7,5 hours

8 hours

 

Ahhhhrg! I found the reason for the breakdown when I prepaired the printer for the next try:

The filament left the drum and was stuck at the mount. So simple!
Okay, I will go back to the original model and start the next print tomorrow noon. So I can observe the process. Maaaan, I hate such banal shit.

Well, experiance even ...

 

As mentioned already: Do that project and show it to the world. The best way to prepair children for all the future struggles: Build your own solutions. Just to teach them that things are worthy because someone spent time and material for it. Such education and experiance is unpayable.

Today I had another issue and coudn't start the 11- 13 hours observed job. But I will try it again during the next days.

 

Okay, lower body again. With increased walls this time (from 0,8 to 1,2 mm). Estimated print time 16 hours. Wish me luck!

Oh btw. if anyone wants help with all the software I use (Rhino3D, 3DS Max, Netfab, Repetier-Host) and prepairing geometry for printing, just send me a PM. I offer help via TeamViewer for free. Rhino is for free for 90 days and again for 90 days at a second computer (= half a year), Netfab Basic and Repetier-Host are completely for free, 3DS Max ... hmm, not really for free. I can use a full student version until November 2018. Probably I can help with printer issues as well. Feel free to ask.

 

Thank you!
Looks like it helped:

 

I learned again from this one. I will never print such a big part as one single piece again. The main reason: the printer was completely confused at the very last Millimeter:

It didn't finish the last five layers but started to print in the air somewhere like a lunatic. After more than 16 hours of doing a good job. I switched it off and it's all good since this detail is really marginal. But imagine that it had happened after 15,5 hours, ~ 10 mm before the end! The printed part would be scrap again. The risk is too high. Even at a 6 or 8 hours-part it would be a tragedy.

There was a mysterious phenomenon at my first try. A deformation wich is neither in the original 3d-model nor in the print-model:

Good to see: the inner crossed structure. My conclusion was: It's a result of the very thin wall (0.8 mm). One reason for me to increase the wall thickness to 1,2 mm. Now look what I got this time:

Even in the print-file it looks completely smooth:

I have really no clue where this bullshit comes from. Usually I would make some tests now. Not this time. It's at Mallory's back, no one will notice it and I don't want to print the part again.

All in all: You should print each single part in an adapted position and with an adapted setup. Probably several times with different arrangements. Again: Experiance counts.

 

Next parts will be the legs. Hard to estimate how much effort it'll take and how to handle them.

Four or five parts I think.

At this point I think about replacing the old sheet of adhesive plane for the printing bed. It did an incredible good job as far as I adjusted the printer correctly. It had to suck terrible treatment by to close Z-adjustment (vertical axis) wich burned the material into its surface, rough procedures to free printed parts or hardcore glue to fit a basic structure. You can see deep wounds at SE of it. I still don't believe this fucking adhesive plane served me that good way. Thumbs up BuildTak. Costs: 27 € for three of them.

Tricky part is: while printing gum-PLA you don't need any heating of the bed. Crazy!
Well, it looks terrible but still works. I will keep it for a while.

Time for a short reflection. I paid 445 Euro for the printer and $ 400 for sculpting the 3d-model. 25 Euro for material and 9 Euro for the plane. And Shapeways wants ~900 Euro for some small parts of the figure I want? Really?

My costs 30 Euro.
Shapeways 900 Euro?

Really? Can someone explain to me what causes the 870 Euro difference? At lesser accuracy and smaller print volume. Not to mention additional delivering costs.

 

I don't want miss to mention the modeler. It is Eugene Petrov (Евгений Петров). I don't know much about him. But he did several good jobs for me so far.
https://www.artstation.com/eugenep
During our cooperation he became a very good designer of 3d-models for 3d-print. Because there are a few tricky parts in it. Some significant differences between models for games/animations/pictures and for 3d-prints at the other hand. I advocate him to everyone who wants good - and printable! - 3d-models. Even from very poor basic material. Only one single photo in this case.

 

Meanwhile I think about a small model wich fits to my printer as one single part. I never saw Mallory in real life completely so far. Would be a nice motivation. But oha! Netfab went down to its knees:

This was from an already reduced model! Rhino can handle it easy including turning and moving it at the screen. Okay, seems I have to reduce it again. Or wait ... "Ignore" is an option!
After hitting "Ignore" again I got this:

Okay, was a try. Here you can see what I got in Repetier-Host from it:

LOL!
Holy crap!

Rhino can't do the job to reduce the number of polygons from 3.000.000 to 1.500.000? Why not? I'm really confused about this issue. How ever, I think it's the end of this short interlude. As mentioned, I'm too lazy to fight with Rhino and Netfab. Too sad. Would have been a nice gig to show the complete Mallory. I can't remember such a lot of error messages in a row. No worries, I will continue as scheduled with the legs next days.