Xaar joins the Si-MEMS printhead club

It’s been a few years since Xaar announced the development of a silicon-MEMS thin film piezo printhead to form the basis of their P4 (platform 4) product range. So has it been worth the wait? On 2 June, 2016 we got first details at the product launch and it certainly looks both impressive and competitive. Here I’ll explain why.

But first let me just explain what silicon MEMS (abbreviated to Si-MEMS) actually means and why it’s being increasingly seen as the future for piezo printheads. The use of silicon as a base for developing powerful computer chips, memory and other electronic devices has revolutionised industry and our lives. These devices are made by the diffusion and coating of materials in layers at extremely high resolutions. MEMS technology is making structures in and on silicon using similar techniques. This involves etching as well as depositing materials. Over the past decade or so etching techniques have become much better, faster and cheaper, making the mass production of MEMS devices viable. The technology is used for a very wide range of applications, and one of those is in ink jet printheads. Thermal or bubble jet heads can be considered MEMS devices in that physical processing is often required in the manufacturing process, and Memjet takes the process much further by building CMOS circuitry with the actuator structure on top.

But the other term often heard in conjunction with Si-MEMS is thin film piezo, and they go hand in hand. Si-MEMS allows a printhead manufacturer to make actuator chambers and ink passageways at a finer scale than previous techniques. The material is much stiffer too, so that the walls between chambers can be made thinner. Unfortunately as the chamber dimensions are reduced, you need to make a thinner and more flexible piezo actuator layer to flex into the chamber to displace a drop from the nozzle. Until recently ink jet devices used bulk piezo, made from wafers of the material. Thick film deposition of piezo has also been used, forming layers around 20 microns thick. For the use in Si-MEMS devices much thinner coatings are used, with piezo layers 2-5 microns thick on top of flexible diaphragm layers only half that.

For Xaar, moving to Si-MEMS offers many benefits. Perhaps the biggest one is it will finally allow them to work with aqueous inks which are becoming increasingly important for commercial and packaging printing, and are of course essential for textile printing. It will also operate at much higher linear speeds than possible with Xaar’s previous designs.

Xaar 5601 GS3p0 printhead
The new printhead consists of 4 chips mounted in a Z-shaped module. Each module can print over roughly a 4.5 inch (115 mm) width at 1,200 dpi and at 120 metres per minute. Alternately the printhead can be supplied with 2 colour inks and print at 600 dpi. Inks can be either aqueous or solvent-based and the printhead can be operated in binary or greyscale mode with drop sizes ranging from 3-21 pl.

The actuators are thin film piezo configured in roof mode, with nozzles at a pitch of 300 per inch. That equates to a pitch between nozzles of 85 microns. Each row of nozzles is offset, with each pair at 1/600th of an inch and the pairs offset by 1,200 of an inch. This allow single or 2 colour printing.

Xaar 5601Although later to the game than some of their competitors, Xaar has taken the opportunity to use the latest technology internally. For instance the connection technology used is state of the art. They have also focussed on producing a design that is easy to implement. The head has alignment features to allow easy replacement, with the final nozzle alignment taking place in software. As the head covers a wide width there are fewer ink and electrical connections reducing costs and complexity. The Z shape enables printheads to be assembled in wide arrays with stitching also taken care of by software. The heads are also reasonably narrow in width, so stacking heads together for a full colour system will be compact – Xaar claims only 200 mm for 4 colours.

The printhead has been designed with a high drop velocity – a figure of 9 metres per second was quoted at the press conference but I understand higher velocities are possible. This should allow wider spacing between printhead and substrates while maintaining print quality.

Drop placement accuracy will also be high – Ramon Borrell, CTO of Xaar, stated they were achieving 3 sigma within 5 micron at 1 mm spacing.

One of the design considerations with a high density piezo printhead is thermal management. The heat from the actuators heats the ink, and if the temperature of the ink varies then so does its viscosity and therefore the drop volume. So Xaar has incorporated its high flow rate technology which will equalise and stabilise the temperature across the printhead, and which will also keep ink fresh at the nozzles and rapidly remove bubbles by passing ink right behind the nozzles.

AcuDrp Technology
A new feature incorporated into the printhead design compensates for manufacturing variations between nozzles. It allows the printhead to be calibrated to equalise drop volume and velocity across each printhead and all of the printheads in an array. It claims to do this dynamically, so perhaps can also be used to reduce crosstalk and maximise printhead life.

Partnership with GIS
Xaar also announced a partnership with Global Inkjet Systems who will supply drive electronics and ink supply systems for the 5601. Xaar’s overall aim is to minimise the total cost of ownership by reducing development and integration costs and time to market.

Product introduction
Xaar aims to be manufacturing the printhead at the end of 2016 with the first product – a textile press – appearing mid-2017. A version capable of handling UV-curable inks is due in September this year, and some manufacturers will be adopting the printhead for 3D printing with a major vendor launch in 2017. Xaar is out-sourcing the silicon MEMS manufacturing, focussing on final module assembly and testing, so won’t be investing in expensive silicon fab facilities.

So, will the Xaar 5601 printhead family succeed? It does seem to tick all the right boxes for high quality ink jet printing at speed. In terms of linear speed and print resolution it is up there with the Fujifilm Dimatix Samba and Kyocera printheads, while being easier to integrate into a process. Landa Digital claimed 300 metres/minute with Samba printheads but like others this is achieved by doubling up the printheads, which Xaar could also do. HP is running web presses at 120 metres in ‘quality’ mode, but their claimed 2,400 nozzles per inch is actually 1,200 dpi with 2 nozzles in line to get 4 grey levels.

A modest yet significant product announcement at Drupa 2016

You expect at the Drupa trade show, currently running at Düsseldorf, a certain amount of showmanship. Demos with lots of loud music, Xerox has Ministry of Sound shows 3 times a day, and the show everyone wants to hear is of course from Benny Landa about Nanography.

imageBut tucked away at the end of Hall 8a is a prototype machine easily missed. Not just the machine but the significance of it too. The machine is the Canon Voyager, demonstrated every hour with the usual display of print samples under glass close by. Canon is not really keeping it quiet, it is a public display and they have a page advertisement for it in every issue of the Drupa Daily. But that’s pretty modest compared to others.

So what’s so special about it? Well, at Drupa 2012 Landa Digital showed their presses for the first time. Instead of ink jetting the image directly onto the substrate, they printed onto a flexible belt, dried the image, and then transferred the image to the substrate. The advantage of using intermediate transfer is that the process becomes almost substrate independent, one of the concerns of direct printing. Of course there is a major problem to be overcome, as you need the intermediate transfer surface to be non-wetting at the transfer stage, yet wetting when you jet the image.

I track all of the ink jet patent applications published, and can tell you that Canon probably has the most relating to intermediate transfer, with others with IP in this area such as Xerox, Ricoh, Seiko Epson and Fuji Xerox.

By now you should have worked out that the significance of the Canon Voyager is that it also uses liquid intermediate transfer, this time using a drum rather than a belt. It’s a B2+ press capable of 3,000 sheets per hour. New fixed array bubble jet printheads print 7 colours using pigment-based inks onto a drum covered with the intermediate transfer blanket. The image is then dried and transferred to the final substrate.

Little information has been given about Canon’s intentions for the press. It is planned for a launch in 2018 and will be aimed at high-end commercial printing applications demanding the highest print quality.

Canon already has experience of producing high quality from fixed ink jet arrays in the DreamLabo photofinishing machine, although in that case printing is directly onto coated photo paper.

So with a confident Landa Digital announcing the building of factories ready for volume production of the Nanographic presses, and now Canon also planning intermediate transfer, what will we see in 3 years time at Drupa 2019. My prediction is we will see ink jet presses, both direct and intermediate transfer, vying side by side for customers. If anyone thinks ink jet is close to the limits yet they are wrong.

Drupa 2016 – no surprises? What have the ink jet engineers been up to?

Mike Willis, Managing Director of Pivotal Resources believes this year’s show will be evolution not revolution


So far pretty much all of the pre-announcements of digital press technology at this year’s Drupa trade show are similar to those made in 2012. So what have the industry suppliers been doing for the past 4 years?

Back in 2012 many ‘launches’ were, in reality, technology demonstrations. There is nothing new about this; I’ve seen the same thing since the first Drupa I attended in the 1980’s. At the last show the big story was Landa Digital’s Nanography. A few print samples were shown, demonstrating potential for the process but far from saleable print quality. This year Landa is back with what should be machines close to being ready for sale.

The same goes for Xeikon, who demonstrated their liquid toner press in 2012. This year we are promised a full colour version. And there are many other examples of this: demonstration at one Drupa, commercialisation at the next.

So why does it take 4 years from showing a working prototype to having a machine ready for launch? What on earth do engineers get up to for 4 years? Well, for a start there may be many changes made to the machines over that time in response to 3 factors:

Feedback from customers
At the last Drupa, potential customers will have made comments on what they liked or didn’t like about a press. This feedback may provoke extensive revision of a machine. As an example, we know from Landa Digital that the large touch screen controls on the side of the presses were not liked, and the panel is now more conventional, and at the delivery end of the machines.

Process development
Building an ink jet press is relatively straightforward. You just source ink jet printheads and ink and most of the job is done! Well, there are still a few people who think that way, but most have learnt that the reality is very different. There are a wide range of areas requiring careful design and development, such as head mounting and alignment, nozzle maintenance, ink supply, data paths, drying or curing… the list is long. If the process involves fundamental new technology, like the Landa Nanoinks and intermediate transfer belts, they don’t develop themselves either, but require considerable work. There are many other issues that are special or critical for high-speed single pass printing too, but I’m saving that for a later blog.

Changes to the design
As demonstrations for a Drupa show come closer and closer corners are cut to get everything working. After the show a major ‘post mortem’ takes place. Are we using the best technology now? With hindsight what would we do differently? What changes should we make to meet the manufacturing cost budget? Now we are aware of the competition what changes should be made to remain competitive?

I’d like to point out that the engineers don’t get 4 straight years to do these redesigns and developments. Ideally it’s much better to get to the next Drupa having been through beta testing at selected customer sites and to have made further revisions in the light of operational experience. So allow perhaps a year for this and we are down to 3 years.

So a development engineer’s job is pretty tough. And it’s going to get worse. The next Drupa is 3 years away, not 4. So if you meet an ink jet development engineer at the upcoming Drupa be sympathetic!