I built the cheapest 3D printer available online so that you don’t have to: iNSTONE Desktop DIY (review)

How cheap can a 3D printer be and still function? Although they seemed plucked out of science fiction, there’s not really that much to these machines. A few stepper motors, some switches, a control board, a heating element, and a nozzle are really all you need. It’s the software, and the expiration of a bunch of patents, that kicked the 3D printing revolution into high gear.

Is it possible to assemble the right collection of components to make a functional 3D printer for less than $100? iNSTONE thinks you can, and they are not wrong.

Behold, the iNSTONE Desktop DIY.

iNSTONE Desktop DIY. Photo by Author.

This is the best printer you could build for $99. It’s terrible. I love it. You absolutely should not buy it.

This is a cantilever printer with a small build space, no heated bed, a fairly standard OS (that connects with a type-B USB port, which I haven’t seen on a piece of consumer electronics since 2015), and absolutely no frills. I never plugged in the provided power supply; it will not spark joy.

If the design philosophy behind the Anet A6 is “how do we make the most 3D printer for $200?”, the philosophy behind iNSTONE is “how do we make the best 3D printer for less than $100?” Honestly, I think they pulled it off. At a quarter the build volume, there’s more metal in this machine than in the Anet. Its main structural elements are extruded aluminum. The steppers are decent quality. The hot end is a very cheap E3D clone. The mechanism on the Bowden extruder is dense injection-molded plastic. You can pick it up and move it around without feeling like everything needs to be re-calibrated.

It has a rotary encoder on the control board, something the Monoprice Mini Delta lacked.

The kit also came with more tools than any of the other printer kits, including everything you need, this awesome little snipper, and several wrenches that don’t actually fit any of the bolts on the printer.

If you were to send me to a market in Shenzhen with $100 in Yuan and ask me to assemble the best 3D printer I could, I really doubt I could do better than this machine. There are a ton of clever little design details that demonstrate the shear amount of thoughtful engineering that went into making this thing work. This printer is going to occupy a place of pride in my workshop.

Unfortunately, the iNSTONE Desktop DIY is objectively terrible. Unless you’re a 3D printer enthusiast looking for a curiosity to add to you collection (and a really fun build), you absolute should not buy this.

For an explanation of our testing protocols, please see: We’re gonna beat the heck out of these machines: The search for the best dirt-cheap 3D printer for fieldwork.

The Burn In.

Building this printer was a whole heap of fun. The instructions were clear enough if you’re familiar with 3D printing but probably too vague if you’ve never built one before. Everything is mounted to the extruded aluminum frame with t-nuts, creating a very strong, stable structure. The kit was missing some parts, but fortunately I had replacements in the shop.

There are simple mechanisms to adjust each axis. The z-axis has the strangest coupler I’ve ever seen. Rather than attach the threaded rod to a piece of machined aluminum or even a 3D-printed part, the iNSTONE Desktop DIY uses a small piece of PVC tubing (more on that, later).

There’s not much else too this. The printer has no heated bed or blower fan to cool the print. The fiberglass build plate is held in place by two OEM binder clips.

Getting a good print out of this machine is an exercise in patience. The build plate is adjusted by three thumbscrews, one of which is nearly impossible to reach. The three point design is not solid. The build plate loses alignment constantly, sometimes even in the middle of a print. It took almost a week of daily tweaking and tightening and calibrating to get this machine to the point where it would reliably lay down a solid layer and get to printing.

Most of the demo prints failed because of z-axis wobble (more on that, later). Finally, after a herculean effort, the first passable print emerged from the provided g-code: this vaguely spooky mask.

Vaguely spooky. Photo by author.

Burn-in Score: C

The Benchmark Test. 

Oh Benchy, I’m so sorry. Photo by author.

There’s nothing glaringly bad about this Benchy other than the obvious wobble everywhere. It printed. Over-extrusion and a wobbly build plate means that you can’t read any of the bottom print and there’s some slight elephant footing. There’s a big vertical jump near the top from the z-axis slipping (more on that, later). Nothing is awful, but everything is just slightly sloppy.

Benchmark Score: C

The Replication Test.

The purpose of the replication test is not to make 3 pretty ok Cute Octos, it’s to make 3 consistent Cute Octos. The lack of cooling, unstable build plate, and absence of a heated bed means that each Octo warped and weirded in different ways.

No. Photo by author.

This test was a failure.

Replication Score: F

The Functional Parts Test.

When it comes to the threads of a screw, it’s all about z-axis control. And this printer does not have it (more on that, later). The threads were sloppy and loose. There was a ton of wobble between the two parts and they did not hold firm. They work, but just barely and not in any way that would be useful.

It’s… not as out of focus as you think it is. Photo by author.

Functional Part Score: D

The Complex System Test.

It’s time. We need to talk about the z-axis. I thought the worst thing I’d seen on a 3D printer is those terrible flexible couplers on the Anet A6. I was wrong. This is how the z-axis connects to the motor on the iNSTONE Desktop DIY:

What?! Photo by author.

It’s just a piece of PVC tubing held in with friction. For a printer that ticked so many geek boxes for me, that’s so disappointing. This is why there’s no control over the z-axis, why there’s so many gaps in the vertical shells, and why, if you try to print out something at 200 micron resolution, it looks like this:

Hello darkness my old friend. Photo by author.

Simply put, this machine just doesn’t have the range.

At 100 microns, the Niskin receiver took 11 hours to print (2 to 4 is the average on the other tested printers). It was flimsy, the servo didn’t fit snugly, and the parts all had way too much play. As a testament to the brilliance of the Niskin3D design team, the water sampler still worked, but just barely. It feels like it’s ready to snap apart at any moment, so I didn’t even attempt the in-water test.

Complex System Score: D

The Precision Instrument Test. 

Can a $99 3D printer print a precision scale suitable for a navigational instrument at 50 micron resolution?

Printrbot in silver, iNSTONE in orange. Photo by author.

No. It cannot.

Precision score: F

The Educator’s Test.

I was really hopeful going in to this one. It was the final print of the review, the printer settings were as dialed in as humanly possible, and the Isopod model was intentionally designed to be easy to print on most low-end printers.

It might be an Isopod. Photo by author.

Z-axis wobble strikes again (more on that, earlier). The vertical layers are inconsistent with big gaps, much of the detail is gone, the base looks awful, and there is warping on the telson. You can tell it’s an isopod, but that’s about it.

Educator’s Score: D

The Tear-down.

Like the Anet A6, pretty much the entire printer is exposed, so there’s not much to tear down. I spent a lot of time with this printer over the last few weeks and am intimately familiar with its inner workings. I’m not exaggerating when I said this is the best printer you could build for $99. It’s well designed and well made. The Bowden extruder shows signs of stress from filament being pulled in and out, but otherwise there’s no obvious signs of wear, no weird rubbing, and nothing that would raise a red flag.

It’s great for what it is, just “what it is” isn’t a great idea.

Note: In a lot of ways the Burn In and Tear Down tests for these reviews are redundant. Kit printers get exhaustive Burn In sections and fairly sparse Tear Downs. Pre-builts get the reverse. It all evens out.

Tear-down Score: A

The Trash Test.

This printer produced an order of magnitude more waste than any other printer tested, and that doesn’t even include all of the failed prints during burn-in.

Total mass of waste filament: 27.71g

Reviewer’s Discretion.

There’s a ton of things you could do to improve this printer. Adding an all-metal (or even 3D printed) z-axis coupler would go a long way to improving it. Redesigning the build plate and adding a heated bed would work wonders. It desperately needs a cooling fan directed at the build area.

But that would spoil the spirit of this machine. The iNSTONE Desktop DIY is a $99 3D printer, not a $99 3D printer + a week of post-purchase engineering + 5 years experience in 3D printing.

You don’t build a hundred-dollar 3D printer because you should, you build it because you can.

Final Assessment.

Honestly, I can’t even give this printer a score. You should not buy it.

Final Score: Stop. Don’t.

The rankings so far:

We’ve only got one printer left, the highly recommended $194 Creality Ender3, so come back soon for the final results!


Depending on how successful this project is, I may expand to include printers in the $200 to $400 price range. If you want to help make that happen, you can either use the Amazon Affiliate links in the post to buy printers, consumables, anything else or you can sign up for my Patreon and help support Southern Fried Science.

Please review the comment policy before you reply.

This site uses Akismet to reduce spam. Learn how your comment data is processed.