The Cheap Taobao knock-off achieves an astounding 200mm/sec print speed; Offers have come in to evaluate helping with the libre laptop batch-printing.

Evaluating 3D Printers


This simply should not have worked as well as it has. For a budget of around $USD 250 all-in so far, a printer which, as-is, is barely struggling to be able to print its own parts, has been successfully transformed into a speed demon. By the metric set for this project (millimetres per second per dollar) this 3D printer is by far and above the highest achieved to date.

  • The Mendel90 has a confirmed metric of around 0.33 (200 mm/sec for a $600 budget).
  • The Ultimaker-2 is around 0.125 (150 mm/sec for a $1200 budget).
  • The MendelFlex is around 0.35 (350 mm/sec for a $1000 budget)
  • This cheap-and-cheerful clone is a whopping 0.8 - four times higher than the Mendel90 or the Ultimaker-2.

That's still with potential improvements to be made and evaluated. It also sets an extraordinarily high bar for the planned new design of 3D printer to exceed. For example: if the Bill of Materials for the planned design comes to say $USD 300, then the target speed has to be in excess of 240 mm / sec, just to reach parity. And that's without having tested the cheap-and-cheerful printer at faster speeds, nor making any additional improvements.

Since operating the Mendel90 over 4 years ago now for at least 2 years, a lot has been learned. The trick about using insanely-high acceleration figures simply wasn't known at the time, for example. It is now suspected that there's a subtle flaw somewhere in the Marlin firmware, where it is better to trust the belts and the stepper drivers to be able to handle abrupt transitions than it is to have Marlin firmware slow down the extruder. 200 mm / sec seems to be what the 1.3 amp A4982 steppers can handle, as can the low-cost 1.6 A NEMA-17s. However the speed demon in me just can't leave it there: an effort has to be made to take it to the next level... all whilst keeping an eye on the goal of bringing mass-volume eco-conscious computing to a wider world and fulfilling my promises to the backers of this current campaign at the same time.

Now, it has to be said that the quality is not perfect, and it is still early days. Several more tests are going to need to be carried out: the laptop's parts are quite sophisticated, and challenge even the best design of 3D printer. One of the things then that I am doing is investigating skeinforge. Chris Palmer (aka nophead) exclusively uses skeinforge, but he edits the configuration files directly: this was simply too much when I first began.

Now, however, there is a feature of skeinforge that may prove critical: arcing on corners. One of the problems with high-acceleration on corners is that if it is a straight right-angle the print-head and the belts simply can't handle it: the print-head tries to keep going and you get a "ringing" effect in the plastic due to the printhead bouncing back and forth. Skeinforge has a critical setting that places an "arc" on corners, thus preventing the printhead from being subjected to abrupt changes in direction beyond the capacity of the 3D printer's capabilities to cope. Cura does not have this "arc" capability; Slic3r is too damaged to produce quality slicing; repsnapper does not compile properly, and that leaves skeinforge.

Offers of assistance

Thank you to everyone who contacted me after the last update. Peter for your kind consideration to test out printing of the laptop's parts: if we had 10 to 20 more people like you this would be an easy task. Neil in particular for your experiences in batch 3D-printing in China: I loved hearing about the hilarious things you got up to, but also to hear that you also managed to ramp up a set of cheap-and-cheerful China clone reprap-style 3D printers. I look forward to seeing what you get up to on the list.

But in particular there is one piece of news that is very exciting yet still much too early to get properly excited about: an evaluation - note the careful qualification there - is underway to see if it's practical to use a private collection of over 150 networked 3D printers in the U.S. Much as I would like to be able to tell you who they are, that is down to them to decide when (or if) it is appropriate. As I always love an opportunity to tell people about failures (so that they don't have to make the same mistakes, but can also see what insights drove me down that path in the first place) as well as the successes, I am always puzzled slightly when other people prefer to only announce good news, but it is understandable particularly when there are many people involved in making a decision that they should act in their best interests, when it is very early days. In this case, in particular, we have to see if their 0.5mm nozzles would work with these parts when they were specifically designed for a 0.4mm nozzle: I simply cannot expect them to swap over machines that are being used round the clock to a different mechanical setup, so 0.5mm has to be tested.

What I can say is that two people very kindly spent an hour on the phone with me to provide some very valuable insights into the puzzling "ghost town" feelings that I had when returning recently to the reprap forums: corrections which are quite important to communicate. It turns out that Adrian - whose inspiring innovation basically started the whole 3D printing movement - is close to retirement age, and that naturally he has been, for some time, slowing down on the pace of innovation. Consequently, with the reduced income that resulted indirectly from that natural decision he decided to call it quits. Some rather direct comments can be found here

As can be seen from, for example, the reddit 3d printing subforum the rest of the Western 3D printing world however is steaming ahead and doing rather well... just not on the reprap forums, which the remaining innovators would dearly like to be able to continue (and revamp) but, due to Adrian's retirement and resultant resistance, are giving serious consideration to just cloning and forking the forum and the wiki, instead. This places the 3D printing community in something of a conundrum, particularly given that the 3D printing community as it stands would not exist without Adrian's pioneering work.

Revision 2.7.4 EOMA68-A20 Card