Moving blog hosting

With our revamped website, we are moving this blog to be hosted at SquareSpace.  This move will allow better integration with our eCommerce solution, as well as making our soon-to-be released Kossel documentation easier to reach.

Going forward, please follow the blog at:  Our latest Kickstarter update can be found there as well.

Shipping update – Kossel Pro

First, the bad news. As some of you may already suspect, the delivery of Kossel Pro kits will be delayed, due to a few minor technical difficulties. We have faith in our engineering and design but we will not be able to make the original Christmas deadline.

As someone who had delivered the last kickstarter ahead of schedule (and had padded this kickstarter schedule too to hope to achieve the same thing), I am obviously disappointed at having to make this announcement and take full responsibility for the delays. We understand that some of you may have purchased this as a Christmas present; we are going to create a set of downloadable engineering prints that people can print themselves as a token gift for Christmas, with the understanding that the real McCoy will follow shortly. We will also be looking into some form of compensation, likely in the form of a discount coupon to the OpenBeam store, to ship with our kits. We are currently projecting a ship date of early March, 2014, based on schedules outlined by our injection molding, extrusion, machining vendors as well as our lead times for electronics.  At the end of the day, given the scope of the project, and how tight funds can get, we did not feel it was prudent to launch the capital expenditure in tooling without sufficient testing and until the rest of the printer, including the electronics board, were fully developed and at least protototyped.

The Kossel Pro features very high degree of intergration; we specifically laid out our control board with delta geometry in mind and paid a lot of attention to connector placement and compatibility across the various sizes in the Kossel Family.  To give you an idea; the original OpenBeam project raised over $100,000, and less than $10,000 went into tooling. The rest of the expenses went into spinning up the company and support infrastructure, as well as merchandise for the initial system launch. We had 5 molded parts that didn’t really have to interact with each other, an extrusion and 2 stamped parts. We could afford to be more aggressive on the launch schedule for all the tooling; in fact, we purchased all the tooling “at risk” before we had money in hand from the Kickstarter to get a jump on shipping.

As a comparison, the OpenBeam Kossel Pro currently has over 20 unique injection molded parts, with suppliers in the EU, Asia/Pacific Rim, and the US. The budget for our tooling is well over $40,000.00. The rest of it had to be earmarked for the BOM of the printer. We simply cannot afford to make a mistake – a few thousand dollars here and there in a mold change will quickly make this project run into the red. As the great Shigeru Miyamoto once said: “A delayed game is eventually good, but a rushed game is forever bad”. We’d rather take our time to make sure things are done right and do the right amount of testing, before releasing the printer.

Now, for the good news:

We have received the first engineering prototypes of the new Brainwave IIs and although there were a few hiccups with initial testing, we believe that we are on a good path forward.

We have signed the final round of die approval prints and paid for our extrusion die. We hope to have first articles in about 4 weeks, and we were able to keep this part of the project on US soil by finding what appears to be a good, competent US based aluminum extruder.

We have started building the test fleet of printers to stress test the new Brainwave II boards. Part of this printer fleet will be deployed at Metrix Create Space in Seattle to print – you guessed it – more reprap Kossel parts.  The original Mini Kossel Prototype 2, featured in the 2014 Make Ultimate Guide to 3D printing, have now logged hundreds of print hours. We are looking at the different failure modes that have occured and making sure that these design issues are adequately addressed in the Kossel Pro.

We have received all the extrusions for kit fulfillment, and the quality is top notch. We are in the process of sourcing validation quantities of all the rest of the printer’s hardware – the power supply, bearings, screws, nuts, bolts, etc.

We have already received into inventory some of this material, such as the timing belt pulleys and ball bearings. The longest “tent pole” in the schedule is our injection molding.

We have released all parts, except for the auto-levelling probe, to injection molding. We will be launching the tooling this coming week on all parts except the touch probe; we hope to launch the touch probe tooling the week of Dec 2nd when I return to Seattle. Currently injection molding is sitting at 8 weeks lead time for T0 (first shots off the mold). These first sample parts have to be qualified – basically measured against the engineering prints and signed off, before mass production can begin.

There’s still plenty of work to do, even during this time when tooling is built; packaging needs to be defined, documentation needs to be written, machines needs to be tested, CNC machines (performing the second-operation machining after extrusion and cutting for the metal vertices) needs to be programmed, videos needs to be shot and edited. Our team will have a busy season ahead of us, but we will rise to the challenge. In fact, while I am writing this update in Germany on my honeymoon, the crew at Metrix is already busy putting together the printer test fleet, and Matthew Wilson is diligently working on the bootloader and patching Marlin for our new Brainwave II.

Finally, for a glimpse of the Kossel and how it stacked up against other machines, consider buying a copy of the Make 3D Printer Review. We are very proud of how the prototype did and what lies ahead for the project.

Thanks, The Kossel Development Team


And now, a guest post from Mrs. OpenBeam

Bride and groom.


As some of you know, Terence and I got married a few weeks ago (October 26, to be exact!) The wedding was fantastic and we loved getting to see all our family and friends from all over the world. We figured this would be a good time to show off a little bit of our geeky side, and we integrated some laser cutting projects, some robotics, a 3D printer, and of course, some touches of OpenBeam.

The table numbers were something very simple I designed, cut out of 3 mm Baltic birch, and mounted in OpenBeam.

Lasercut wood and OpenBeam table numbers.

The card box was a simple box utilizing Baltic birch and OpenBeam again.

The OpenBeam card box. Yes, we packed a 2mm hex driver. 🙂

Our “ring bear wrangler,” the daughter of one of Terence’s friends/colleagues, is a budding geek herself, and an active Hackerscout, so we knew she would be up to the task of driving the ring bear robot down the aisle. It was controlled via Bluetooth on a Nexus 7, building upon the work done by Keith Baston, who won the OpenBeam Build Off a few months back.

The ring bear robot is powered by a Sparkfun Redboard, a Bluetooth Mate Gold Bluetooth modem, and 2 Hub-ee wheels.  Terence put the robot together in about 3 hours.  During rehearsal and testing though, one of the wheel’s internals melted and caused the robot to swerve as our ring bear wrangler tried to navigate it down the aisle, adding to the challenge.  Fortunately Terence had bought a spare wheel, though he left it at home so we didn’t get a change for a full practice run with the working robot.

3D printing was also featured at the wedding.  We 3D printed a few LED tealight holders and small vases and gave them away at the end of the night:

We were glad to have a number of Terence’s 3D printer and geek buddies in attendance:

Back row: Pierce Nichols (Logos Electromechanical), Catherine Nichols (Hackerscouts Guide 004), Matthew Wilson (Brainwave/ Brainwave II), Johann Rocholl (Father of Deltabots). Front row: Richard DeLeon (Metrix staff), Plamena Milusheva (Metrix staff), Choong Ng, Marc Goodner (Deltabot Alpha tester) and Marta Johnson. Metrix must have been quiet that evening!

Last but not least, the laser cut guest book. The hinge is based on the living hinge used in Snijlab’s folding wood booklet and I used binding posts from the scrapbook section of the local craft store to fasten about 30 pages into the book. The pages were printed with photos of us from the last 3 years, including many of our engagement photos, with plenty of room left for guest signatures.

Guestbook (Better photo to come)

I think that about covers the relevant bits of the wedding! Thanks for being patient with us as this has taken over a lot of our life for the last couple months. I think we are both looking forward to getting back to “normal,” whatever that might mean!


Rachel, Terence and furry monster puppy

Rachel, Terence and the furry monster puppy 🙂

Kossel / Kossel Pro – In progress pictures


First, a personal note.  I’m getting married this Saturday!  I’ve been completely swamped.  The future Mrs. OpenBeam had commissioned me for a few OpenBeam related wedding decoration.   On top of all this is my day job, plus the usual logistics challenge of coordinating a typical wedding – Relatives from around the globe, transportation logistics, etc.  Come think about it, it’s not very far off from supply chain management…

I just want to share a few pictures of the work-in-progress Kossel Pro and Kossel redesign:

Our ball bearing ball joints.  Each ball joint contains 4 bearings, for butter smooth, low friction, zero backlash motion.

We’ve managed to retain about the same dimensions as the popular Traxxis 5347 ball joints!  This means that carbon fiber tubes cut to length for the Traxxis ball end rods can also be used with the OpenBeam ball ends to make a rod with the same center-to-center distance.





A set of these babies on carbon fiber rods.

Here’s the carriage, with our belt end clamp.  The belt loops back onto itself, you pull it to tighten it around the peg.  We’ve ported this geometry back into the RepRap branch as well, and we give you two options:  Build the unit with captive M3 nylock, or build it with a captive M3 x 37mm spacer:

The use of a spacer greatly improves rigidity, as it couples both ends of the ball joint together.


And here’s a sneak peek at the new end effector, with integral cooling ducts.  The ring around the hot end is fed by two top mounted fans that are PWM controlled, while the hot end gets its own dedicated, always on ducted fan cooling system to prevent thermal jamming issues.  We are VERY proud of the work we’ve done on the end effector; as you can see it packs a LOT of technology into a very small package without growing the overall size by too much.

You can see more “in progress” pictures on the OpenBeam Flickr page.

I’ll update more after my wedding!


-=- Terence





[In Depth] – The Kossel linear rails

Terence’s note:  This blog entry is a work in progress.  We are releasing it ahead of our usual editorial process as our backers are now receiving their linear rail sets and the linear rails contains an instruction sheet that links to this article.  Photos and video will be forthcoming.  Thanks.

Hello Kickstarter backers!

By now, all the “rails only” Kickstarter rewards have shipped via Fedex to our backers. You should have received a Fedex shipping notification email with your tracking number and information.

One of the things that separate the Kossel Pro from the rest of the machines on the market and at this price point is the use of linear recirculating ball rails on the machine chassis. There are a few great Open Source hardware options when it comes to linear bearings: Bart Dring’s MakerSlide, Mark and Trish Carew’s OpenRail and V-Slot, and Steve Graber’s W wheels are three of the systems that makers are most likely to be familiar with. When designing the Kossel Pro, we also took into consideration patents that are existing in the field, namely this patent held by the 80/20 corporation. After all, we are a small start up challenging the status quo of expensive T-nuts with cheap extrusions in the T-slot framing space and we cannot afford any legal entanglements with a much larger corporation.

The decision to use linear rails become pretty apparent when you consider the following factors.  Being able to adjust the play between the carriage and the rail is good; not having to worry about that adjustment is even better.  Given that I speak both Cantonese and Mandarin, and given our investment in our trade and logistics capabilities, we are able to source these linear rails at a very competitive price.  In fact, our linear rail and ball bearing carriage subassembly BOM cost is actually lower than the retail cost of a comparable length of MakerSlide and its carriage and just slightly above the retail cost of 3 W-wheels from Mr. Graber.

The biggest advantage with the ball rails over other ball bearing on aluminum extrusion systems, however, is in the material selection. The rails are made from induction hardened tool steel.  In fact, the steel used in the rails is so hard, the vendor that we purchase the rails from uses a Wire EDM process to cut the rails to length. These rails are also precision ground to incredible levels of flatness. This, combined with our new extrusion vendor’s awesome aluminum profiles, meant that the linear rail assemblies on our machines will be straight and true, and likely will stay that way unless the machine is put through serious abuse. Remember, Aluminum Extrusion Council’s standard tolerances is about 1.1 mm per meter length and a competent vendor generally can be expected to deliver parts within half of commercial tolerances. That is still 0.25 to 0.5mm of deviation across the Kossel’s travel on the linear axis, and we would like to do better than this. Our rails, being a ground metal product, is flat to within 0.02mm per meter – order of magnitude flatter than normal aluminum extrusions. They also add an incredible amount of rigidity to the machine, and that is the secret to how we can build such a large, rigid machine chassis on such a thin aluminum profile.

The rails we source are Chinese made clones of Thompson / THK / HiWin linear recirculating linear ball rails. Contrary to popular belief, they are NOT from HiWin! Early on in the Kossel’s development, there was a cloner who would shamelessly come into Metrix, observe what people are working on, and proceed to stock his webstore based on his observations. As the cloner himself does not do any meaningful R&D work and we rely on future sales to recuperate our R&D cost and had already lost money due to his cloning of early legacy Kossel extrusion kits, we could not afford to lose the opportunity to sell the rails if the cloner beat us to the market. Therefore, we masked the true origin of the rails by calling it a (much more expensive) HiWin MGN-12C equivalent. In reality, our rails are the same generic ones that are sometimes sold on AliExpress; we just found a reliable supplier, qualified them, and verified that the rails that we are sourcing will work with our design.

There is a big cost difference between our rails and a genuine HiWin / THK rail. Although our rails are ground from hardened tool steel, the surface finish on our rails is rougher than a HiWin / THK rail, and as a result our rails run nosier. (Surface finish is measured in micro inches or microns. The rougher surface does not affect accuracy, just performance noise). The tolerance range on the grinding is also looser than what would be on a genuine HiWin rails and carts, and some carriages may be stiffer than others to move on the rails. None of these issues affect the accuracy or machine chassis rigidity of the rails and is not considered a defect, and a little bit of care and preparation can help improve the rail’s performance drastically.

The rails ships from the factory with a retained ball carriage. This means that there is a wire cage running down the length of the rail to make it harder for the ball bearings from falling out when you remove the carriage. Note that it is still possible for individual ball bearings to fall out, but it is also possible to push the ball bearings back in. There is a gap in the chain of balls in the recirculating ball race – in a higher end carriage, there would be a special plastic cage piece between individual balls that contains a reservoir of lubrication grease. For maximum life, it is recommended that you replace the protective oil that the rails shipped from the factory with a proper grease in the ball bearings.

I’ve found it easiest to slide the carriage off by laying it on its back and sliding the rail out from the top. This way any displaced balls hopefully won’t fall and get lost. A towel over the work area to capture the occasional lose ball is a good idea.

We are shipping all kickstarter rewards (and all Amazon rails) with a piece of lint free wipe to wipe off the preservation oil, as well as a light bearing grease for greasing the rails. The seals do a pretty good job at keeping debris and dirt out of the rails. Unfortunately, this also means that any grease applied to the outside of the carriage on the rail will likely be wiped off and will never make it to the ball race. We therefore recommend our users to remove the ball carriage carefully, and apply the grease directly to the line of ball bearings using the provided cocktail stick, moving the balls around and applying the grease. We have chosen a non-toxic grease that is rated for intermittent food contact that complies with the new regulations on VOCs in lubrication fluid.

i’ve taken a short video showing the cleaning process here. With a little bit of care, these rails can be a very high performance upgrade option to the 3D printer (or any other mechanical build). We hope you find this information useful for your project. Happy building!

-=- Terence, Rachel and the furry monster puppy

%d bloggers like this: