3-D printers

[Sclass]

Nice subtle look at where you went to college

I’m really proud of that casting. It has been a tradition since the 1940s for Stanford engineers to make these castings in the engineering school foundry. Old guys bring their old medallions in for reunions. Many alums have this on display around their businesses or offices. Because the original positives are reused every year the degree of detail indicates when it was casted. It’s almost like a diploma on the wall. There is also this little brass magnifying glass we have been forced to make for decades.

I’ve seen a similar tradition with a Tau Beta Pi key with Caltech engineers.

We are basically forced to make that thing our first year. For many, including myself, it is an unpleasant experience. Sand casting and foundry work is harsh. For a lot of the kids it’s the first time they get really dirty. Like coal mine dirty.

My experience went poorly. I forgot to use the release powder on the mold and I opened my mold after pounding sand (a no no) and it got slightly deformed. It came out ugly. I almost didn’t claim mine out of the pile the next day because it looked worse than the others. I wanted to switch mine for an unclaimed one but I swallowed my pride and took my special one home. I remember seeing the pristine one donated to the engineering library by an old graduate and I thought maybe I should keep it. It sat in my garage for years. I used it as a trivet for heated items in my shop. I almost left it in a scrap metal bin during a move a few years back. But I kept it...for 25 years.

Now I’m kind of proud of it. More so than my diploma. I keep it at my workbench to remind myself that I’m kind of a screwup with a knack for digging out. I have that magnifying glass too. I think they get special meaning over time because it isn’t just given to you.

https://twitter.com/StanfordEng/status/ ... 77984?s=20

https://engineering.stanford.edu/get-involved/alumni

Young student making her sand casting.

https://images.app.goo.gl/VMEc2hTtoHiv1dAX6

[Tyler9000]

That medallion is badass. I did something similar in a sculpture class*, and there's nothing that compares to seeing something in metal made by your own two hands. You should be proud of your imperfect result, as I can guarantee it was better then my attempt!

(*) I do find it kinda funny that my undergrad ME education never did anything like this, but my random community college art class did. Stanford has the right idea.

And speaking of castings, if you want to make some custom car parts that look a little less 3D printed, you might also consider using your printer to make master patterns for a silicone mold. With a little hand sanding and a coat of paint you can make the master pattern nice and smooth, and the finish should replicate with the molding process.

[7Wannabe5]

Super cool.

[unemployable]

(*) I do find it kinda funny that my undergrad ME education never did anything like this

Neither did mine, and it's a school that likes to think it's almost as good as Stanford. Maybe I would have been more inclined to actually go into engineering as a career if they had.

Interesting stuff.

[Sclass]

First print. The vent came out well. 1.5 hours. Not much filament used. The solid parts are hollow and honeycombed inside. I made a mistake with the CAD tool that I didn't see on the screen. I also measured the bores on the AC unit poorly. So there are some fitment issues. The great thing is this is like word processing. You just go back to the cad, make some edits, then reprint. The quality of the print is good. It seems pretty strong and the fibers are fused together well.

I learned the skill by watching roughly five hours of YouTube videos. I don't think this is that tough but there are a lot of details. It seems to be the hobby of data driven nerds. The good news is they post up their parameters (there are dozens of settings) online and you can download their configuration files and just hit print.

The vent in printing.


Finished. An hour and a half later.


Installed on the AC unit. Doesn't fit. Tomorrow is another day. Looks a heckuva lot like the originals which were also white plastic.

[SavingWithBabies]

It takes a little getting used to but if you bust out the calipers (I'm sure you already did) and then start tinkering, you can get the fit down right. One trick is to print just a couple layer thin ring that is the size you think you want, check it for fit and tweak with more prints until you get the exact fit. Then go do the whole part in that size. Much faster iteration cycle.

Eventually, you get a feel for what the fudge factor is and parts just start fitting.

Nice choice on the printer. I've got the Creality CR-10S and I don't need this huge print area at all. Wish I'd gone smaller as big prints take so long that I rarely do them. Handy once in a blue moon I guess.

I think you can print in PETG with your printer -- it might be worth picking up a spool to have something more temperature resistant. Oh, and another trick with PLA: if something is slightly too big, drop it in boiling water for a bit and then take it out and it'll shrink a small amount. Supposedly, you just "annealed it" (put in quotes as not sure that is technically the correct words to use) which has made it stronger/more temperature resistant. I played with that a little bit but don't know how truthful the strong part is but it seemed to be.

[Sclass]

Eventually, you get a feel for what the fudge factor is and parts just start fitting.

Wait, are you saying the finished parts have a wide tolerance? I never bothered to check if I got the same diameter I plugged into the CAD model. I though I measured the hole pretty well. When it’s light out I’ll check the fit again.

[SavingWithBabies]

Well I tried making a nut and bolt and I found if I was 100% on the mark (exact dimensions), it didn't work as it was too tight. I had to add some fudge to the values so the fit wasn't ideal. Which is kind of obvious in hindsight I suppose. Apparently, the standards for nuts and bolts have some fudge in them for metal parts but you need more for plastic.

The other thing is you can calibrate the extruder as you might find it over or under extrudes and that impacts +/- in sizing over your dimensions. I found mine to be fairly close though. All kinds of calibration cubes and doodads on Thingiverse.com.

Also some filament is more variable in dimensions as in one role won't be consistent in size with the previous. I haven't really run into this but I've read about it. So you can take measurements for the actual size in a couple of spots and average it and put that in as the actual size instead of 1.75mm. I think these days filament is better but I mention it as it's something to be aware of.

In any case, if it is off, for me it has been repeatedly off so it's consistent at least . Nice choice on the printer! I have the older brother (CR-10S).

[SavingWithBabies]

I realized maybe I'm not being clear. So what I would do if I was making a part like that to fit is I would measure the diameter of the hole. Then I would subtract maybe 0.5mm from that diameter and print test circle (just couple layers high*, nothing fancy -- quick to print). Test that for fit and if good, print the item (scaled to the needed diameter if needed -- assuming just a tiny scale adjustment needed). I'd scale it in the slicer (I'm partial to Cura). If I hadn't designed it yet, I'd design it to the determined diameter.

* Might need to be higher to not be too sloppy if a precision fit is required.

[Sclass]

@SavingWithBabies

Hey thanks. 0.5mm is quite a bit. I haven’t spec’d turned threads in a long time. My recollection was the tolerances were unforgiving. Unidirectional tolerances? Printing threads makes these machines attractive over other CAM tech.

I thought about getting the bigger printer too, but I’m running out of space on my bench. Also I couldn’t resist the Ender 3 price. I’m still shocked at the quality I got for my $200. It is stunning they can deliver this precision device for so little.

I originally got attracted to 3d printing when I needed a cam timing gear for my Elna 62. It’s a plastic helical gear. They commonly lose a tooth and they skip zig zags. Luckily the gear is available on eBay for $20 but it would be a possible candidate for printing. Your Elna has one so it is something to look out for as it ages. A 3d print is probably strong enough to work.

Thanks for the scaling tip. I was about to adjust the dimensions in Tinkercad and redo slicing. Your trick will work.

Just researched PLA annealing. Looks like it’s worth a try given the interior temperature in the truck.

[SavingWithBabies]

Yeah, 0.5mm was a complete guess. I'd probably go with something between 0.1-0.5mm which is quite a range -- most things I'm making are small so I'm towards the lower end on that range. I edited a couple times before posting last night trying to remember what actual values I use.

The annealing might work on that Elna gear too although I found PLA surprisingly strong/robust. I think the shrinkage was something like 2 or 3%? But if I had a part that fit, I guess I would scale it up by the shrinkage amount and then anneal it. It was fun playing with the annealing. I printed a ring to glue onto the radiator on our SUV. Basically, the plastic hose "barb" (mount?) for the upper radiator hose had cracked. So I printed, annealed and then glued it on with two part epoxy. So far, it's holding but I'm going to replace the whole radiator shortly (and will take a look to see how it held up for the past 3-4 months). I would have used PETG instead of PLA but I didn't have any. I want to try PETG next.

[Sclass]

Unfortunately the annealing didn't work for me. Due to the shape of the grill vent it shrunk in a really uncontrolled way. It was highly distorted after twenty seconds in the boiling water.

Oh well, there is always HTPLA, ABS, PETG as you suggested. I'll see how hot these get in the truck. The truck is a light color and doesn't get hot inside. The AC is under the dash near the floor. I have hopes but I may have to learn how to print nylon. There are some really cheap all metal hot ends on ebay right now. They cost about $6. It may work for Nylon.

In the meantime I need to get better. I had a lot of starts and stops today because of mistakes in my CAD CAM files. The printer was well behaved and reliable.

This is where I am today. They look good. I'm off to the races.

[Tyler9000]

Well I tried making a nut and bolt and I found if I was 100% on the mark (exact dimensions), it didn't work as it was too tight. I had to add some fudge to the values so the fit wasn't ideal. Which is kind of obvious in hindsight I suppose. Apparently, the standards for nuts and bolts have some fudge in them for metal parts but you need more for plastic.

There are a couple factors that go into FDM tolerances. One is the print tolerance of the machine mechanism. A second is how the slicing algorithm accounts for the steps in the outside surface due to the layer thickness. Is it measuring to the outside corner, inside corner, or the midpoint? And finally, even if the dimensions are perfect with reasonable clearance, something like a thread will always be tricky in practice due to the increased friction of the stepped surface. But practically speaking, I think you have the right idea to figure out the "fudge factor" for your particular printer that just works as a good rule of thumb.

@Sclass -- looks great! From a design perspective, one trick I've learned in plastic part design is to the use the material to your advantage. Plastic is flexible. To make your vent part easily repeatable even with wide tolerances, you might try designing it with snap features on the outside perimeter that flex to different hole sizes and perhaps even hold the part in place. Figure that out, and they'll fit every time with the extra benefit that they won't rattle.

[Sclass]

a design perspective, one trick I've learned in plastic part design is to the use the material to your advantage. Plastic is flexible. To make your vent part easily repeatable even with wide tolerances, you might try designing it with snap features on the outside perimeter that flex to different hole sizes and perhaps even hold the part in place. Figure that out, and they'll fit every time with the extra benefit that they won't rattle.

Hey brilliant. I have everything fitting, but the vents are just held in by friction. They blow out and rotate around when the air is on high. I tried making some barbs in cad on the inner lip at the start but my skills with Tinkercad aren’t good enough yet. I’m still trying to manage a “snap to” in the software. 3d makes it more tricky to get the tabs where I want them. Since the vents took 1.5 hours to make, I think I’m just going to make some tabs with some molten drops of filament and a hot knife rather than reprint.

[SavingWithBabies]

@Tyler9000 It's really a fascinating topic. I started writing a reply earlier and deleted it but I agree. I also think about the variability in terms of "how you do it". For example, if you tend to level the bed slightly higher or lower, it'll impact the first layers more so than the higher layers. Most of the time getting it roughly right is good enough. But if different people are using the same printer and leveling it differently but printing the same files, they may get slightly different results. I think a lot of this comes down to the fact that the printer is completely dumb. The printer doesn't have a feedback loop for observing what the result is. If it did, it could potentially warn/stop if the first layer was not squished just right or it could adjust the feed rate if it determined the middle of the spool had slightly thinner filament than the outside or it could stop a print if an overhang attempt completely failed. Making the machine smarter and/or at least having a feedback loop (which is quite complicated -- do you use a camera with vision techniques or some sort of other sensor(s) approach) seems like the future but at the same time, as a 3d printer operator, once you get past the hurdles such things are less needed particularly considering the expense. It'll be interesting to see if things change there.

@Sclass You might consider using Fusion 360. I resisted for a long time and it does have a learning curve. I'm still figuring out some basic things. But there are fairly good Youtube videos out there for it. I do agree a snap fit would be good but it might be as simple as putting a lip around the inside of the outer vent cylindrical surface. I guess the tricky bit there is your tolerances still need to be good. Also Fusion 360 has some interesting STL to model import features. I haven't tried it but it might work to get your model in (but hopefully, there is a better path from Tinkercard -> Fusion 360 if you want to go that route). Or you get to learn 360 by redrawing it .

[Tyler9000]

Hey brilliant. I have everything fitting, but the vents are just held in by friction. They blow out and rotate around when the air is on high. I tried making some barbs in cad on the inner lip at the start but my skills with Tinkercad aren’t good enough yet. I’m still trying to manage a “snap to” in the software. 3d makes it more tricky to get the tabs where I want them. Since the vents took 1.5 hours to make, I think I’m just going to make some tabs with some molten drops of filament and a hot knife rather than reprint.

Working with the print you already have is always worth a shot. For me that little bit of hand work is part of the fun.

If you're brainstorming retention methods, acquaint yourself with my favorite FDM accessories: https://www.mcmaster.com/heat-set-inserts

You can either drill a hole in the part or (even better) design it into the print. Push the threaded insert in with a hot soldering iron and it easily melts into place. That will give you a nice secure thread for any type of screw that might come in handy.

[Sclass]

Got some new stuff to post up. I fixed my Black and Decker Workmate bench with printed parts but I'll put that in the fixit log.

What I have here is a tool that I made. I've been dreaming of having a very low clearance offset screwdriver that uses hex bits. I had been using a 1/4" gear wrench and hex bits but the bits would always fall out of the wrench while I was using it. So I decided I needed some kind of holder. I looked on Thingiverse and it turns out there was a guy who used a 15mm gear wrench to make such a low offset screwdriver. However a 15mm ratcheting boxwrench is huge. I decided I'd remix the design and make one for my smaller 10mm wrench.

I need to mate these two items together so they won't fall apart while I'm applying pressure to the phillips screw.


So I printed these little pieces. A 10mm hex or 16 pt head on one side and a female hex bit socket on the other. The little top hat brim allows you to put downward pressure on the screw without having the bit slip out of the wrench.











Look at the clearance on this thing compared to an off the shelf offset bit driver on the right. It can fit in much tighter spaces. These come in very handy when working on Mercedes cars that use a lot of phillips screws. If you can get in there with an offset you can often get away with taking less of the car apart to fix it. Therefore saving precious time.



The printer is turning out to be a very interesting tool. There are a lot of things online that can be printed. One thing I noticed is things that break get redesigned and uploaded. All you have to do is download and print, then fix. I'll show a couple I did in the fixit log. Okay, that is it for 3d printing. Make custom tools!

[Tyler9000]

That's awesome! I love smart & simple design ideas like that.

[Sclass]

I have rats around my house and in the attic. I tried catching them with snap traps and had mixed results. I decided to set up a webcam and I got some interesting data. The rats are afraid of the traps. They gingerly pick the food off the triggers. I went as far as riveting bulldog clips to the triggers and putting bait (raisins) in the clips and the rats gently chew off the bait leaving the remainder clamped in the clip. These guys know what they are dealing with. So I decided to make a new kind of trap that didn't look like a trap. I made something that looked like a sewer pipe.



I used some black abs pipe I had left over from a project. A big 3" nail. Some steel wire. Rubber bands from broccoli bunches. An o-ring. Some old #6 screws left over from a project. A bamboo skewer left over from kabobs. And four 3d printed items.

1) a nail holder. basically a big head to put on the nail so I could push it along with rubber bands
2) a nail guide. Kind of a tube to guide the nail along a constrained direction. keep it going in a straight line.
3) two saddles to keep the pipe upright.
4) a trigger paddle to get bumped by the rat's nose as it tries to go after the bait. No more nibbling. the trigger needs to be pushed out of the way and if it happens, blammo, the arrow fires down.

Right, it's basically a crossbow powered by rubber bands. I designed this thing while staring out the window yesterday. I was influenced by a video on something called the "Tilong" trap. Used a similar trigger. The trigger is an interesting concept. I never knew it but trap triggers are all about leverage. You hold back a really strong spring force with a tiny trigger frictional force using a lever or two. You can see the wood kabob skewer in the photo. It's burned and has kefta residue on it.

I'll report back if it works. It does severe damage to pickles.

[Sclass]

Made some improvements to the trap. I printed a back cover. I was going to buy a sewer pipe plug but why when I can make one.



Made some refinements to the trigger. Shortened things up a bit. The great thing about this tool is I can make engineering tweaks real time and try different setups.



Come on in! Free food! Lowered the nail a bit. Gives it a little more stability when cocked. Looks like I put that granola cluster too far forward. Oh well, got to give them hope.