SUNY New Paltz was the first college in the world to install a large array of 3D printers, at the time marketed

Our Array of Replicator 2 printers

by Makerbot as the Makerbot Innovation Center.  Thank to my friend Wallace Patterson for first bringing this idea to us!   We started up in February of 2014 with 25 Replicator 2 printers and 5 of the Replicator 2X.  And some scanners that never worked, but the less said about that the better. The Replicator 2 has been an outstanding printer and we have put thousands of hours on ours.  They are used constantly to support a wide variety of academic programs as well as external customers.  All of them have broken down, but thanks to my colleague Aaron Nelson’s ability to fix just everything and the work of our talented team of undergraduate interns, they are pretty much all still functional.  We have made some modifications.  Aaron and Mitch Wagner developed a control system using BotQueue and Octoprint so we could control them remotely, we have fully enclosed the printers to address health and safety concerns and we have replaced the build plates with a flexplate system.  We have been considering replacement options and thought that it would be interested to share these options and invite comments.  Note that an important component of whatever we do is our Stratasys Continuous Build Stack.  This does give us the ability to produce a lot of parts

Stratasys Continuous Build Stack

very quickly and the recently upgraded software has significantly improved its functionality.  It’s main drawback is that it only prints ABS and, from an environmental perspective, we would greatly prefer to use PLA for the thousands of student printers we do in a year.  And, frankly, the quality of a Stratasys printer is kind of overkill for beginning designers.

So here are the options we are considering:

#1 – Continuing to fix and upgrade the Replicator 2’s. This has been are favorite approach for a while because these printers have a really robust frame and are mechanically solid.  We would probably do that except we would need to replace all of the belts and bearings and we would really like to upgrade the electronics to take advantage of some newer features like using a touch probe and 256 microstepping.  Aaron looked at replacing the Replicator 2 board with something open source, but it turns out that would require also rewiring all of the end stop and stepper connections.  Next!

#2 – Purchasing a lower end open-source-type printer.  The one we considered the most seriously is the Prusa i3 Mk3  which runs around $1000 assembled.  We borrowed one from our CS Department last year in the faceshield frenzy and worked it pretty hard.  It’s a nice printer.  Our only issue is that it’s a completely open build platform and we have run into issues with our EHS folks about the potential nasty effects of nanoparticles being released from the hot end.  Enclosing the printer seems to mitigate that and also does improve print quality, especially for materials other than PLA.  because of the Prusa’s configuration and our space limitations, we haven’t figured out a good enclosure solution.  Next!

#3 – Purchase a lower end, enclosed commercial printer.  We dismissed this option in about a nanosecond.  From everything we have seen, these printers are not built with parts that are better (or even as good as) the Replicator 2.  We suspected within a year we would be right back where we started from and the printers would probably be even harder to repair than the Replicator 2’s.  Next!

#4 – Purchase higher end “desktop” printers.  These range from something like the Makergears (we own several of them and have been pretty happy with them) to the Makerbot Method to the enclosed versions of the Ultimaker.  These printers range from $3-7k or so.  We need a minimum of 15 printers in the array to meet our needs and buying 15 printers at this price range would be pushing our budget.  Since no matter what printer we bought, we work them hard enough that they will break regularly.  Generally, the more expensive the printer, the less open source it will be and the greater the maintenance cost.   Next!

#5 – High throughput printers.  We never really considered this very seriously, but it’s a neat idea.  There are printers that will run multiple jobs in a row without human intervention so each printer can run 24/7.  Since our Replicators probably only run 50% of the time even where we are busy, this would significantly improve efficiency.  The two main options are continuous belt printers and printers that swap build plates.  We investigated a few options, but reviews and our innate suspicion of relatively untested technologies suggested the time wasn’t right..  Next!

#6 – Build them ourselves.  This is the option we are currently pursuing using the VORON platform.  This is a open source printer that uses components that are as high end as can be reasonably purchased.  Both my colleague Aaron and I have built printers from a kit or from scratch and it’s surprisingly easy.  Our goal is to build a printer with similar quality mechanical components to what we could purchase in option #4, but only spend about $1500 per printer plus a couple of thousand dollars in intern time to do the assembly.  We also have the advantage in being able to print the plastic components out of PC/ABS on our Stratasys printers which will give very high quality parts.

Since this would give us an entirely open source printer we could upgrade easily as new options become available.  This not only seems to be the economical option, but fits our philosophy of pushing the envelope in practical applications of 3D printing.  We are getting ready to assemble our first printer and we will provide updates as they become available.


Keep on Printing!


5 thoughts on “Post #8 – Time to Replace the Bot Farm

  1. You may want to take a look at automating whatever printer you choose with Quinly from 3DQue. Continuous printing like a belt printer but on whatever printer you like. 99% operator-free for 24/7 printing who VGH would double your throughput.

  2. As I noted in the post, we are very interested in experimenting with continuous belt printers, but they are a pretty new variation on 3D printing technology that seems to introduce a fair number of new variables in terms of bed adhesion, print quality, print orientation, etc., etc.. With the thousands of part per year we print, we need a solution that we know will work for the next 5-10 years. If you would like send us a demo, we would be happy to try it out!

  3. Rep2’s are TANKS! I still have two at home and still run great. Fargo3D can hook you up with parts. Check the y-axis secondary’s as they wear out really easily with the stock plastic ones.

    I’d look at a Raise3D if you want to head towards the higher end of the spectrum. Their E2CF is coming out in Q4 it looks like. If you want continuous belt and want to experiment the Creality 3DPrintMill might be a low-entry one to try.

  4. We’ve been ordering parts from Fargo3D for a while and they are great!

    The Raise3D printers look good as do a lot of the desktop printers in the $3k-$4k range. I guess our thought process is that we believe we can make a printer just as good or better from $1,500-$2,000 that will be completely open source and easily upgradeable. One way to think about this is that the reason there are so many companies selling desktop-level printers is that they are kind of easy to make. The challenging part is to keep the running over an extended period of time.

Leave a Reply

Your email address will not be published. Required fields are marked *