Several years ago our own inimitable Max published a blog “Want a Voltera Desktop PCB Printer?” that described a PCB printer that was being developed in Canada and funded through Kickstarter. The product itself was interesting enough, but that it was being done here in Canada got me really intrigued. I got myself on the mailing list, and when they recently sent a postcard advertising the finished product I showed it to my boss and he decided that we should get a demo. I contacted them and rather than them coming to us, we were invited to go to Voltera. They are only an hour away and so a road trip beckoned.
Recently I have noticed a lot of discussion about a significant uptick in tech innovation in Canada. Somehow I think this may be the unexpected upside to Blackberry’s malaise. Since the jobs at Blackberry were just not there, I reckon it inspired a lot of engineers to start their own developments. You heard this silver lining theory here first! Anyway, Blackberry is based in a town called Waterloo in Ontario. Waterloo is in fact part of a larger metropolitan arrangement consisting of 3 cities: Kitchener, Waterloo and Cambridge. An interesting aside is that Kitchener, established by German immigrants, was originally called Berlin. In 1916, for obvious reasons the authorities renamed it for the British Secretary of State for War, whose image we are all familiar with. This is a very long winded way of telling you that Voltera is in Kitchener.
Lord Kitchener in a well-known pose. (Source: Wikipedia)
The University of Waterloo is an institution of some repute in the engineering and scientific world and has one of the few co-op engineering programs in Canada. No doubt it is one of the sources of these up and coming innovators. To spur this development the university has created the “Velocity Garage” the “largest free startup incubator in the world”. Voltera is part of the endeavor and located in this building.
The space is rather like a large warehouse turned into a student union establishment- communal meeting rooms and kitchen facilities and each startup in an area with minimal partitioning. Voltera also has a workshop right next to the work area, obviously with doors that can be shut and locked. I began to understand the reason for the new work sharing paradigm I wrote about in “How Well Can You Work With Others?” Funnily enough, the Upverter package I included in the blog is also part of the Canadian innovation wave. As I looked at all these keen faces concentrating on screens and diagrams, I wondered and then dismissed the idea that I would like to work in this environment, but then I was lucky to even be allowed into the building. After all, the ink on my birth certificate is dry!
But you all would like answers to some of the questions Max (and others in the comments) posed back in 2014, so let’s see if I managed absorb the presentation correctly. There are several videos and pictures on Voltera’s web site, so if you want to see it in action head over to it. There are also several videos on YouTube.
The whole assembly, and indeed aspects of the operation, unsurprisingly reminded me of an old X-Y plotter. The mechanism consists of an X-Y controlled carriage with a Z dimension since the carriage also has an adjustable height. The bed of the device is thermally controlled. It appears to me that the mechanism and the heated bed are operationally mutually exclusive.
Figure 2: The V-One. The heating pad is flat the black horizontal rectangular area and the carriage with a conductive ink cartridge installed is visible at the top right. (Source: Voltera)
The carriage is capable of carrying different heads and this approach contributes to the versatility of the device. There is a probe, the trace deposition “head” (made up of a conductive ink cartridge and a nozzle), a solder paste deposition “head” (made up of a solder paste cartridge and nozzle) and a few more that I will get to as we proceed. The V-One is aimed at the high end of the hobbyist market and low end of the professional one- hence some compromises have been made to contain costs. There are some manual operations as part of this economy and changing these heads is one of them- surprising given the accuracy in placement they achieve.
Obviously before you start you have got to have your PCB design in Gerber format.
After clamping the blank PCB to the bed, the first step is inserting the probe in the carriage and the system goes through X-Y calibration. For trace deposition the nozzle must be held within very tight tolerances above the PCB and so the probe is then used to sense the undulations of the PCB, mapping more than a few points to record the landscape. On a blank PCB the exact location probably doesn’t really matter, but in the case of certain boards (patience, we’ll get there) there need to be fiducials and so the probe is lined up with two of these by eye.
With ink-jet spray the ink is atomized and accelerated with high voltage. On 3D printer the filament is normally heated till it flows. It appears to me that the conductive ink is extruded out of the nozzle under pressure created by the mechanism rotating a toothed gear at the top of the cartridge. Deposition of the traces is quite quick, taking several minutes for a relatively complex side. There is a flow quantity adjustment which I guess you learn to fine tune for your needs. At this point the traces are wet. You can simply wipe them off – so you need to be careful not to touch with your finger. They have to be cured and that’s the primary reason for the controlled heating pad. Apparently curing is improved with the wet traces facing down towards the heater, so the board is flipped and mounted on rails. It says a lot for the ink’s viscous properties that it can be manhandled like that and not be deformed. Curing takes about 30 minutes.
For double sided boards, the process is repeated on the blank side (remembering the fiducials). One of the shortcomings is the inability to drill through-holes. Voltera is in the last stages of producing a drill head that will address this deficit and of course it will be backwardly compatible since it mounts on the carriage. The conductive ink will flow through existing holes and plate the via, although I am not sure of the minimum size of the though-hole. Somebody put a lot of research into that ink!
Whilst on the subject of the ink, the resistance of a trace often comes up in discussions. Sheet resistance is given in the bizarre units of ohms per square so you have to do some calculations (see the link) to get to actual resistance. Obviously the track width is the one used on your Gerber output, but the thought occurs to me that it would be nice if there were an “economy” option (either in the CAD package or in the V-One software) where you thin the traces for this prototype stage and save some ink. For similar reasons Voltera recommends that earth planes be cross hatched rather than solid as you can see in Figure 3.
Figure 3: Printing with the conductive ink (Source: Voltera)
You are going to have to make room in the beer fridge- the ink has to be refrigerated until use, and then left at room temperature for ~15 minutes before attempting to plot. This brings some questions to mind- how do they ship the ink? Is refrigeration only required after first being used?
There is something interesting to consider. The PCB printer will print not only on FR4 PCB material but on many other surfaces. We also saw several examples of different substrates including flexible PCBs and even a ceramic base with the track being used as a heater element. It will also print over silkscreen and existing copper tracks, and make a connection to those tracks. That means you can have templates of boards with connector patterns already placed and even a number of vias which you can use if you so choose. Voltera offers some of these templates for Arduino and Raspberry Pi and some others. We intend to make our own for the weird PCB profiles that we use. Figure 4 shows one of our weird profiles (please ignore the components and semi-straight edge on the left in that figure) and our intended approach to providing a rectangular shape to mount on the V-One. Of course there would be only the 6 connector plated holes that you see at the step edges at a minimum. Perhaps we could add the voltage regulator (since there is always one) and some vias distributed around.
Figure 4: One of our weird profiles. (Source: Author/ Emphatec)
The next step of the process would be solder paste, which is apparently fairly simple to do obviating the need for a stencil. According to Voltera, some of their customers use this feature for non-conductive ink prototypes simply because of this. But that brings us to the biggest stumbling block – pick-and-place. A pick-and-place head is in the works, but I was unable to gauge when that might become available. At this stage you are left either placing the components by hand (as we all know solder paste does have a centering effect) or acquiring some pick-and-place equipment. Some of these like the Neoden TM220A or the SMT Place 2000 can cost significantly more than the V-One itself, and there is also the ezPick which is much more affordable.
Soldering is simply a matter of initiating the heating plate which has the thermal profiles needed for the conductive ink. You can use it for non-conductive ink solders as well by changing the heating profile as needed. Another question I have- what happens with components on both sides of the board? I will have to ask.
As a prelude to the subject of reworking let me describe another plus in the additive approach. In products with a higher current requirement, we often “plate up” to thicken the trace. The V-One allows you to print on top of the existing traces to thicken them up, and it can even be done selectively. In addition, patch-up of damaged boards can be done in the same way. When it comes to reworking a board, it seems to me that this printed PCB is a little less robust than what I am normally used to. You need a special solder and a temperature controlled soldering iron limited to 200°C. But at least in theory, you can repair the damage as described above.
I was given to understand that a cartridge of ink (at $99) could complete up to 15 Arduino Uno sized PCBs of moderate complexity. I assume that to be single sided. So approximating to 8 double sided Arduino boards, it would cost $12 a board excluding the cost of the substrate. Unquantified is the time saving, especially if a second spin of the board is needed.
I asked if there were any plans for the ability to place text on the board as with a silkscreen. Of course text in conductive ink is not a problem, but although this question appears to have been asked before, it is not high on Voltera’s list of priorities.
No one is perfect and besides the lack of the pick-and-place option and drill (for the moment) there are some other drawbacks. The maximum board size is relatively small (5.3” (135mm) x 4.4” (113.5mm)) and it can only produce double sided boards. The V-One cannot handle very fine pitch components; I frequently use a TQFP package micro (0.5mm pin-pin) and this is outside the 0.8mm spec on the printer – this will definitely cramp my style. A custom board shape cannot be milled. I also believe there is some restriction on the upper frequency range, but it was in the GHz range and I am sure many other effects come into play as well.
It occurs to me that in the above discussion I have focused on creating standard boards as an end in itself. Although I do quite a bit of simple analog design, I lack confidence in my abilities so I try to make sure that I have a feasible approach before committing to a PCB. It is possible to simulate certain approaches (stand back before the Pease lightning strikes), but it is nice to be able to try out just a small part of the circuit to see if something works. Today I use solderless breadboard or vectorboard, but with SMD components it becomes harder requiring breakout boards and wires going everywhere. A printed solution would allow several alternative approaches to be produced on the same simple substrate to compare and contrast.
According to this article there are 5 companies in the additive PCB printer business and no doubt quite a few more that use the milling machine approach. Is there a PCB printer in your future? Even though the ink on my birth certificate is dry and definitely not printed in conductive ink, there is one in mine.