3d printed stainless steel

The automated blind I described in my last post has been reliable, operating every day since. The curtain puller design I also mentioned proved not to be reliable due to the cord piling up on the pulleys randomly, causing its tension to be variable. I redesigned it to use T5 timing belt that I had left over from my early use in 3D printers. I didn't have enough belt to span the windows twice, so I made up the rest with the original cord.

So there is a pulley for the cord at the idler end and a printed T5 pulley for the belt. I couldn't get enough tension to stop the belt slipping over the pulley teeth, so I printed a semicircular guard to press it against the pulley and prevent it skipping.

 

This design worked reliably for more than a year, but then the spare bedroom curtain's printed parts broke. This is the smaller of the two curtains that I automated, but they are thicker and heavier material than the lounge. These are the PLA parts that failed.

In hindsight, the pocket to hold the bearing left very little material in the middle of the clamp. Possibly the wooden pole swelled a bit in damp weather and broke it, or maybe I just tightened it too much. Once the clamp broke, I think it had the knock on effect of breaking the belt guard. I think that would be strong enough if the clamp had held.

A fix would have been to make the clamp wider and thicker, but then I would have needed to replace the shaft and beef up the other half of the clamp to match. But since the screws into the case would then be further apart, I would need to reprint the case and then strip the whole device down and rebuild it.

To avoid rebuilding the unit, I started thinking about stronger materials to keep the clamp the same size. I think I could have milled one out of aluminium, but as I haven't done that before it would have involved a lot of work. It also has features on both sides, so it would be two operations, one possibly on a lathe.

Then I remembered that PCBway also do 3D printed metal parts. I uploaded the STL of the part to see how much it would be. In aluminium, it would have been about \$28, which seemed good value, so then I tried stainless steel and that was only about \$2 more, \$30.73, so I went with that. I am sure aluminium would have been strong enough, but it might tarnish. Stainless shouldn't tarnish and is over the top strength wise, but it makes more sense to 3D print it because, unlike aluminium, it is notoriously hard to machine.

The STL has polyholes, a hanging hole and a teardrop_plus cutout for improved FDM printing, but it doesn't matter if these artefacts end up in the SLM print. I could have changed the model, but life is too short.

 My understanding of the SLM process (Selective Laser Melting) is that a layer of fine metal powder is laid down, and then a laser melts the areas that are part of the model and the bed drops and the next layer of powder is laid down and melted. It results in a fully dense part, unlike laser sintering.

The total cost was \$59.06 with shipping and bank charges, but PCBway have said they will reimburse me for writing this blog article. It would be more economical to order more parts and share the shipping cost.

The build time at PCBway is 9 to 10 days, so twice as long as PCBs, but I think the actual process is probably faster as there are a lot more steps to making a PCB. I assume they just have much more capacity for PCBs.

I placed the order on Sunday 29th of June. The part was shipped on 7th of July, and it arrived here in the UK on the 10th of July. They actually sent two, so like PCBs, they seem to do one extra in case there is a reject. There was a slight difference is quality between the two though. This is the best one:

 

  

 

 

 

 

 

 

The other one has some strange artifacts that look like z seams you get with FDM. I have no idea how they occur with SLM. Here they are under a cheap microscope:

 

 

Odd, but they wouldn't affect its function in this case. I will use it on the lounge curtain if that ever breaks.

 

Here is the other one fitted to the guest bedroom curtains.

 

 

I reprinted the belt guard in "silver" PLA to somewhat match it.

 

Before fitting it, I made some measurements to compare the accuracy with my PLA original.

 

	NetFabb		PLA		Error		316L		Error
Width	20.33	mm	20.29	mm	-0.04	mm	20.24	mm	-0.09	mm
Length	43.65	mm	43.83	mm	0.18	mm	43.52	mm	-0.13	mm
Height	20.75	mm	20.67	mm	-0.08	mm	20.68	mm	-0.07	mm
Bearing Socket	16.08		16.12	mm	0.04	mm	16.11	mm	0.03	mm
Volume	7.79	cm3
Weight			6.785	g			57.28	g
Density			0.87	g/cm3			7.35	g/cm3

The NetFabb column is the STL dimensions as reported in NetFabb Studio. The PLA column is my FDM part printed on a Mendel90 and the 316L column is the stainless steel part. 316L is the grade of stainless steel. Both measured with Mitutoyo callipers. The accuracy seems broadly similar.

The density values are the weight divided by the volume of the STL. 316L is reported to have a density of 8g/cm³, but as the part is a little smaller in each dimension it is probably not far from that. 

The parts are grey, presumably because the outside faces have the same granularity as the metal power before it was melted. I lapped two of the faces on 240 grit emery paper, and they then look metallic, but there are still a few low spots that would need a lot more sanding to remove.

 So I think parts could be sanded and polished to make them food safe, for example.

All in all, a great fix for my design problem that didn't need much effort from me. I will definitely use this service again if I need a complex part that needs to be super strong or corrosion resistant. It would be great for use outside, for example.