# Ultra high precision and accuracy printing

I work in a teaching hospital and we have a research project we're interested in pursuing. We'd like to 3D Print tubes we'd implant into rats to help with nerve regeneration. We're interested in the shape of the tubes right now, more so than what material it is or whether it's biocompatible etc..

So this question isn't necessarily about what type of plastic or whatever we should print in. My question is more so:

We'd like to print a tube that's 1 mm in diameter, about 1 cm long and has as many micro "tubes" crammed through it as possible, something like this:

I currently have a Stratasys j750 in my lab, a UPrint Se and a Prusa i3 Mk3s. They all work well but for the detail I'm looking for, come up a bit short. They have advertised accuracies of 14 microns (well, the j750) but thats just in the z direction, x and y are more like 200. If I went to get PRECISE, what technology should I look into?

• If you want to squeeze maximum accuracy out of your J750, you should talk to Stratasys' or your reseller's support channel. There are some non-obvious tweaks you can try out to get better accuracy, and they may even be able to give you access to settings that aren't on general release yet. Apr 17 '19 at 21:15
• Should that say "if I want to be precise" rather than went to get? I can't work out if PRECISE is a brand of printer [part] - although there doesn't seem to be such as brand... Apr 18 '19 at 9:20
• Also, whilst the question is interesting, I'm not sure if it borders on a recommendation which, strictly speaking, aren't really allowed on SE sites, unfortunately. Apr 18 '19 at 9:23
• Yes, it should read "want to be precise". Not a brand as far as I know, just bolded the word to emphasize it. Apr 18 '19 at 16:41
• Neat idea! "As many as possible" may be overkill though. Depending where you're looking, peripheral nerves have a surprisingly low number of fascicles. Those fascicles are often different sizes and shapes too. Do you have a cross-section of the nerve you want to use? "How do I make this?" might get better results. Apr 21 '19 at 1:49

This is an interesting question. A good thing to note when we start talking about SLA and other jewelry grade 3d printing, that you will have to factor in the materials toxicity when we start talking about medical applications. You can also look into DLP 3d printers but they will not have as good quality.

What can help you right now is these SLA printers I see that form 3 has 25 microns, with a laser dot of 85. Well within your tolerances. Just make sure to get dental grade or medical grade resin.

Note that if there are bends, then you will have issues with SLA printing, depending how steep the angles are.

There are other technologies, such as metal 3d printing or SLS printing, that will likely be out of your price range such as the HP Metal Jet that can do 21 -xy by 50 -z microns. Or binder based 3d printing which will have the best internal geometry, as powder will act as support, and is easily removed.

There are many specifically in the bio printing area that might be what you are looking for as well, this is more an alternative. You can look into what the researchers at Penn has been doing, where they have been using Sugar to create vascular networks. As far as I know they do not sell it as a product at this time. Here is more information about scaling vascular networks

• I've spoken with representatives at Form about the form 3, and apparently the 85 micron laser doesn't equate to 85 micron thickness walls. Their best guess was 125-150, which to be honest isn't much better than the 200 we currently have with the j750. An expensive printer wouldn't be an impossiblity (again, we have a quarter million dollar j750 at the moment) but yes, it'd be nice if I didn't have to go out and find funding for a giant printer for these tiny tubes. Do you know of any websites/service bureaus that can print in this detail as a service? Apr 17 '19 at 20:35
• yes shapeways.com Thats interesting that they say that. I bet that has to do with the fact you have elongation of the dot from the mirrors angle. Do note that post processing (drilling it out) might save you some $. I would make a cast of the thing you showed with no holes hit it with a CNC or some sort of precision drill. Apr 17 '19 at 20:36 • I'd suggest contacting studiofathom.com . I got a tour of their facilities and was very impressed. They showed us some 3d printed chips with really small cavities for micro fluidics applications that they prototyped. Apr 17 '19 at 21:34 • Update: Contacted shapeways and studiofathom, also had an extended conversation with formlabs. Shapeways and studiofathom were both very polite but passed on the project, studiofathom said "we reviewed the conversation and those requirements will not be possible on our equipment" and shapeways response (via phone) was similar. Maybe its a weird scale in between microfluidics/nano and normal printing Form however changed their tune (compared to what previous sales rep) told me. Their solutions specialist on the phone seems to think the holes I need could be down to about ~50 microns Apr 18 '19 at 16:20 The typical low-end consumer printers that are so common now in the$100-800 range (yes, I've seen small ones sell that cheap) cannot achieve this, even with special nozzles.

But your machine is a bit nicer than that. Looking online at the manufacturer's page for the machine, we find these specifications:

LAYER THICKNESS
55 microns

BUILD RESOLUTION
100 +/- micron accuracy


That's not gonna cut it. 100 microns is .1mm. +/- 100 microns means it can error that much in either direction from where things should be. To illustrate why this is too much, I'll use the image of the tube in the question, with the 1mm diameter as a scale, and layer the 100micron error potential as a grid over the image:

This grid doesn't show the size of the filament: just margin for error. It's getting close to what we need, but it's not there yet. To reliably create your tubes, you want the error to no more than 1/4 the diameter of a tube itself. Based on the image, the tube size is about 180 microns, so you want to look for something with an error in the range of +/-45 microns. _Maybe you could do this with a tube — more of a rounded rectangle — lying on it's side.

• The online manufacturer page you listed links to the wrong printer. j750 vs j700, resolutions are a bit different. And I've mentioned the j750 is not going to cut it. The question is not: what accuracy do I need, the question is how do I get there? Apr 18 '19 at 16:32
• @JosephCrozier what is your smallest wall feature? If it is less than the smallest nozzle you can get or smaller than 0.15 mm, no 3D printing type can do this. Apr 18 '19 at 16:46

Update on this:

There are a few 3d printing technologies beyond your typical FDM/SLA/Polyjet that can get this small.

• Melt Microextrusion
• two photon polymerization
• Something called MEW
• continuous liquid interface production

I found various articles where someone "printed" that small, but it was often kind of a misnomer where technically something was made additively, but it wasn't a "printer" that you could go buy. I think two photon polymerization may be the "best" actual printing method for what I want, though the price tag associated with that style of printers may be out of my range. But it can definitely get that small, this technology can apparently get down into the nanometer range.

FDM/FF printers can#t achieve those resolutions due to several constraints. The biggest problems are wall thickness and resolution:

The print needs to be made from walls that are at least one nozzle wide. Commonly available nozzles go down to the 0.15 to 0.2 mm area, so the thinnest wall has to be at least this wide or be ignored.

The resolution is probably an even bigger problem: common consumer grade and industrial machines can get dialed in to have errors down to the 0.1 mm. Which would be, with the aforementioned smallest nozzles, up to half a wall width shift! Shifting the print by that would destroy the functionality of the print surely.

## But, is there a solution?!

The only solution that comes to my mind is using a resin based printer like SLA or DLP. SLA printers can get down to 0.025 mm layer height, usually demand a minimum of 0.14 mm wall thickness, but newer machines can cut that almost by two.

The error on the SLA Form 2 machines is, compared to its minimum wall width, pretty much nonexistent for features of about 10% larger than its spot size (0.144 mm), so for 0.15 mm and up it prints pretty much spot on. Extrapolating this means that you'll need a SLS machine such as the Form 3, for minimum reliable wall lines of roundabout 0.09 mm. However, one wall thickness is usually bad, but it might achieve walls of 2 perimeters, coming down to about 0.16 mm.