Inkjet PCB printing.

[moonshadow]

Just to recap, I am thinking about an A5 size printer that can print resist, solder mask, artwork, flux etc, direct to a blank PCB. The printer would basically be flat bed printer. The print head moves on one axis and a sliding table, on which the blank PCB is fixed, moves on the other axis.
One of the problems some people had with constructing a milling machine was either a lack of skill or a lack of machine tools. So any design has to take this into consideration. Since we are talking about printing, it would make sense to use the innards of an off the shelf printer, which most people could pick up at a bootsale.
Today I did a post mortum on two different domestic quality printers, a Cannon BJC 1000 and an Epson 760.

Cannon BJC 1000.
The case is snap together and all the snap points are marked with a small arrow. The case came off with no problems at all. The Cannon mechanism was ALMOST exactly what I was looking for. The printhead slide assembly mounts vertically and the single control PCB mounts neatly behind the slde. The power supply is in a neat plastic box that is ideal for this application.
I completely removed all of the paper feed components, again with no problems. The resulting assembly looks similar to a gantry crane and the idea is to run the table between the 'legs'.
There are two problems with this printer. First it has a three colour cartridge but no black. Second, the designer has put the traverse stepper on the right hand side and the paper feed stepper on the left side. However he then runs a shaft from the left side back to the right side to operate the Head cleaner. This shaft passes through the space where the table needs to be. I think there are at least two solution for this problem. There is about one millimeter of fine head adjustment.

Epson 760.
The 760 does not use steppers but a DC servo system, with the motor on the right hand side and the optical encoder system on the left hand side. The optical encoder is coupled back to the right hand side by a shaft. This shaft is above the table / board path and no problem. However the chassis is awkward, as part of it is in the board path. This could be overcome, but it would also mean removing the rear part of the box chassis which houses the control PCB. There is about two milli-meters of fine head adjustment.

Conclusions.
Other printers need to be examined to find the most suitable. Using an off the shelf printer does elminate a lot of problems, such as the original electronic control board and power supply can be utilised. Also the standard printer driver software alread exists. All that has to be constructed is the moving table. At first glance one might think that coverting the rotary movement of the paper feed stepper to a lateral table movement might be difficult considering the accuracy required. However there is a fairly simple way of doing this.

John K.

[C H Whitehead]

Hi John,

Since you appear to have started 'examining' printer mechanisms for your A5 printer/plotter, have you considered using a modified version of the original flat bed plotters? I'm sure you should be able to find an old HP or the like for that.

Regards,

Clayton.

[moonshadow]

Rogers remarks about UV interested me. I have been consider two types of inks, solvent and UV cured. The UV cured ink seems to offer a lot advantages for high speed printing of PCB's and would be a lot more durable during processing than a solvent ink.
I have located some suitable UV cured inks, that cure in a fraction of a second. My question is .... exactly how much UV is require to do the cure ?. My only previous experience was 50 watt halogen lamp, followed by a very expensive 405nM filter and then piped that through glass fibre optic cable to the target. 405 nm output of the lamp is down to about 9%. It might be one way of doing it, to get a low mass illuminator mounted on the print head. I am also wondering if a UV Led (or cluster) might do the trick, mounted just behind the print head ?.

John K.

[moonshadow]

Hi Clayton,
As PCB production seems to be a general problem I was looking for a general solution that anyone could apply. If the idea works then a specific printer would have to be bought so that a matching kit of parts could be made available to do the conversion, ie to make the table etc.. One of the reasons I find the idea so appealling is because a printer would be so much faster than the plotter/milling approach. Thanks for the thought though. John

[roger]

Hi John,

If you have any details of the fast cure UV inks I will look into the question of using UV led's or a UV laser diode for the exposure. The simplest solution is to use a UV laser diode with a short UV fibre optic cable for collimation and a final output lens for focussing, (final spot size). However, I have been looking at high power UV led's (MUCH cheaper) and it may be possible to use them depending on the exposure power required. The disadvantage ?? of this is that we are back to single point scanning on a line by line basis, and some software would be required to raterise the pcb design for LED on/off switching. this is not something I would be able to write easily, but maybe somebody else could. However, this would be no real disadvantage timewise over mechanical milling, but would have the advantage of considerably higher resolution for smt work etc. but no real problem for most board sizes.

Personally, I have no objection to etching pcb's, but it would be nice to print direct to the pcb without having to produce transparencies, lightboxes, developer etc.

The advantage of your system is that a small addon drilling head for pcb holes could be added next to the exposure unit.for drilling after exposure thus saving remounting the pcb.

I have a laser/led power meter and some suitable optics so would be willing to run a few exposure tests on coated pcb or the UV inks that you have data on with UV leds. If the tests were successful I would be able to produce the UV units easily, maybe with a holographic focussing lens, which makes using leds a doddle.

Sounds an interesting project, let me know your thoughts

Roger

[moonshadow]

Hi Roger,
I know very little about PCB's and am busy learning at the moment. I feel that this kind of project can really be split into stages. The first is using acid resistant solvent inks which would possibly be OK for the occasional user. I will have a go and knock up a 'proof of concept' model to play with. The only problem with this is finding some electronic way to return the head to zero after a print. To make this easier I may remove the head cleaning drive rod and put the table in it's place, for a trial. I obviously don't know what other problems I may encounter.
A possible next stage is to move over to UV cured inks, which I feel might be more suitable for this application. For anyone else interested I will say that solvent also evaporates inside the head and can cause clogging. UV cured inks are actually a resin and more similar to the toner idea. I did wonder if one could mount a page wide UV source on the other side of the Print asssembly gantry but this takes us back to the subject of exposure times and power required. I will get back to the ink people and get some more data on the UV inks. Exactly how available and practical they are in terms of price, I have no idea at this time. I would think it would be wise to sheild the head from the UV light to stop cure at the nozzles.
Another possibility is to forget etching, copper and boards completely and use metallic inks if they can be had at a reasonable price. I am in contact with one company but it involves a ND, so I cannot print it here. However from the results they are obtaining, it looks as if it could be the way to the future.
For the moment I want to knock up something that works so that I can have a play with it and get onto the learning curve. Would you happen to know what the typical distance is between paper and head for cheap inkjets ?. To simplify a prototype, I was thinking of making the table height fixed, so that it would only handle 1.6 and 1.0mm board and use the paper thickness lever as the fine height control. That would be enough for me at this stage. I'll send the links in a separate posting. Cheers ... John

[moonshadow]

Whoops ... I mean't return the TABLE to zero !. jk

[Maverick]

http://www.new-system.com/products_n...t_moreinfo.htm

[roger]

John,

Acid resistant solvent inks in a piezo drop on demand head would be the simplest approach, presumably you know of suitable ink sources.

The head to roller platen distance on my cheap Canon BJ10 is approx 5 mm, but it is a bubble jet, not a piezo head.

I believe a Cambridge company, Xaar, do suitable heads for various inks but thats all I know. presumably you have got a suitable head in mind.

Cheers Roger

[moonshadow]

"The advantage of your system is that a small addon drilling head for pcb holes could be added next to the exposure unit.for drilling after exposure thus saving remounting the pcb".
It would be nice to have the drill on the same assembly at the head, but the Cannon carcass is a bit on the flimsy side for that. So I think I decided to make up a drilling jig to drill two index holes in each blank PCB and use the same jig to put two index pins on the table. The index holes would allow accurate transfer of the board between the Printing process and the drilling process. Ay the moment I am thinking about a small chain conveyor belt. Boards from the board hopper, would be gravity fed to index with pins on the 'belt'. So I did wonder if I ought to do a belt version first ... but eventually decided to take this in small steps and use a table on the first one.
Just to get it out of my system, I was also thinking of a plastic chain type conveyor to take the printed board through the various etching stages. I thought that since I might want to do a hundred boards at a time, I would need to include compensation for temperature and saturation of the etchant. A nice little PIC project. However that is for much later. The index holes are a quick and cheap solution for the time being. John.

[roger]

John,

You have moved way out of my league with your continuous flow production system.

However, while doodling over a coffee or two, I came up with the thought that, for the hobby user the biggest problem always seems to be to get the pcb design transferred to the pcb. Why not then use a standard inkjet printer modified to accept a moving table as per your idea, but print the pcb design using the standard inkjet ink cartridge directly onto photosensitive board. (I don't know yet whether the ink will take to the photoresist surface). Then expose, develop and etch in the normal manner. This means that the question of acid resistant inks is avoided, but eliminates the need to produce master transparencies.

Are you going to use the canon chassis for your first trial machine?

Roger

[moonshadow]

http://www.smh.com.au/articles/2003/...?oneclick=true Article on conductive inks and their applications.

http://homepage.tinet.ie/~ei9gq/pcb.html How to make PCB's using toner transfer, using UK available paper.

http://americanprinter.com/mag/print...ntroducing_uv/ Article on UV inks.

http://www.capv.com/home/CAPStats/CS...3/3.30.04.html Article on UC cure Inks giving 0.2 second cure time and advantages over solvent inks.

http://www.signindustry.com/outdoor/...-JL_UVpt2.php3 Article on UV cured inks.

http://members.whattheythink.com/all...e.cfm?id=17518 Article on commercial UV cured inks.

http://www.oce.com/en/News/Newsbulle...zonaT220UV.htm Commercial article on UV cured ink.

http://www.oce.com/en/News/Newsbulle...zonaT220UV.htm UV inks and machine development

http://www.screenweb.com/inks/cont/UVCuring.html Mysteries and myths of UV cured inks.

http://my.ecplaza.net/cnkcho/2.asp Small commercial thick material flatbed printer showing table height mechanism.

http://www.circuitree.com/CDA/Articl...121582,00.html Article on the application of Jetink printing of PCB's and limitations.

http://www.xaar.co.uk/ One of the leaders in this field. Printheads, Ink etc. A company to watch.

John K.

[moonshadow]

Hi Roger,
My first step is to get a table and table drive done. I think I will use the Cannon chassis as it is to hand. Until I get that far using normal solvent ink I will not know if any of this is practical. I suspect there may be problems with solvent ink in the acid bath. I tend to be a bit of a pessimist and look ahead for problems, to try and work around them. Standard, off the shelf solvent ink cartridges would be great if they worked ... or could be made to work. Whatever is in the Cannon cartridge is what I will be playing with first. I have got flushing fluid so I can try alternatives if I need to.
I popped into town (Andover) this morning to see what I could do about the table. For DIY projects I make it a rule only to use off the shelf components that would be readily available to all. So I plundered my local DIY shop and came back with an assortment of brackets etc .... and a pair of steel, ball bearing linear draw slides £3.50 a pair). Personally I only want to do small boards but it would probably be easier to aim at an A4 print area machine for those that need it and these slides look long enough to do that as well as A5. So I am going to sit here and knock up a design for a simple A5 table using these stamped out components ... if they are accurate enough for the job. Then I will assemble one and do some run out checks with a DTI to see how they perform. This may take a little time because I have other things on at the moment. We have a possible offer of a collaboration job with a UK conductive ink manufacturer, any ideas on how best we can approach this ?. They will want a NDA. It would be one way of getting samples etc.

John.

[moonshadow]

Both printers seemed to use the same printer mechanisum, so I decided to 'examine' the HP600. I still have little idea of how to get the case off 'properly'. Although these printers have a pressed steel bottom cover the actual printer chassis is all plastic. The traverse slide assembly blocks the intended table path as does the electronic card. The card could be moved, but part of the plastic chassis would have to be cut away weakening the whole assembly. I felt these two printers were definitely not going to be suitable for the project.

John K.

[moonshadow]

http://whttp://www.smartshopper.com....ts/default.htm Supplier of inks and solder resist materials for inkjets.
Refill Station PERMANENT MARKER 488 50 for simple and clean refilling of PERMANENT MARKER 350/352
Prevents 'overfilled' markers
Polypropylene tank
Content 30 ml
Available in 4 colors: black, blue, red & green
Product code: STD488 50
(Maplins sell Lumoclor 318 as an etch resistant pen.)

http://www.pulsar.gs/1-tts/c_pcb/pcb...ing_store.html
Solderable silver ink pen. Solder mask pens etc.

http://eeshop.unl.edu/pen.txt
Private study on the effectiveness of various etch resistant pens. Staedtler
Model: Lumocolor 313 AV, Ink color: Red wins.

http://www.everythingpcb.com/p10666.htm
PCB supplies including conductive ink.

[mnementh]

Hi guys,
Sorry if I seem to be hijacking this thread but as an electronic designer and prototyper, you really seem to be taking the difficult path to making a PCB here.

It all seems to be too much work, to be honest.

My method is very simple, by comparison, i.e. laser printer, melinex sheet, light box, develop, etch. JOB DONE!!

If your project is to produce a milling option for PCB's, then I once again apologise for hijacking the thread but if all you're trying to do is make a PCB, then make life simple and use the standard approach.

Apart from anything else, it's a damn sight more accurate and less likely to produce short circuits, due to swarf.

1) laser printer. Produces a dense black track
2) melinex sheet (drafting sheet). Transparent to UV
3) UV light box. Available from Maplin, etc.
4) developinging solution. Sodium Hydroxide (caustic soda, available from most hardware outlets)
5) etching solution. Ferric chloride (available from Maplin, etc.)

HTH,
Sandy

[david]

Hi Sandy,
Could you please elaborate a bit more on your method?
Is melinex the same as glace paper?

Cheers,
David

[moonshadow]

Hi,
Please feel free to comment on this theme, I posted it because I wanted information and feedback. So I cannot complain if I get it Hi!.
Not all users of this site are experts and many are beginners. Perhaps posting an article on the User Site, on your conventional approach to making PCB's may help them ?. Toner transfer seems to be quite popular too at the moment.
Am I doing this to mill PCBs ?. The answer is no. I decided that making PCB's and drilling holes in them is only a small part of making a complete project or product. Apertures have to be cut into cases, front panels engraved, etc.. So I decided to have a re-think on the whole subject. It is simply my own opinion but I think there has to be a better way of doing things, certainly in my case. The problem interests me and I have both the time and interest to play with it.
Printing PCB layouts direct onto a substrata has enornous advanatges and can also print layout artwork, solder resist, flux etc. as well as pads and tracks. It can also do it at the same speed that an inkjet prints on paper. Using metalic inks one can make a PCB as fast as printing a page on paper, without light boxes, photo sensitive boards, developers and the acid bath !.
I think it is worth the effort of trying to find out what can be done ... and how. John

[mnementh]

Hi David & John,

At the place I work, we use a milling system to produce PCB's in small quantity as well as cutting thin metal sheet material in intricate shapes and as far as I'm concerned, it's pretty much a "magical" system but for the average Joe, that doesn't have access to such intricate machinery, it's easier, if a bit messy, to do it the "old fashioned way".

I can certainly appreciate the convenience of the electronic method of transferring data to a milling machine and watching the board generating itself . I also very much like the idea of metallic ink printing. That's someting I would certainly like more info on. If you have any links, etc. It'd be appreciated.

David, to answer your question, Melinex is a draughting medium, used before computers and Autocad made drawing relatively simple, used by draughtsmen to draw technical drawings.

It's a tough plastic sheet, usually translucent rather than clear, that has very little "stretch". It will "usually" go through a laser printer and survive the heater roller. Unfortunately, due to computer generated drawings being the "norm", it's getting quite difficult to get hold of.

The main advantage with the old fashioned method is that very fine tracks can be produced fairly easily and repeatedly, down to a couple of thou wide, if needed.

HTH,
Sandy

[See_Mos]

I buy me laser film from Rapid Electronics here in the UK, it is not cheap at £5.50 for ten sheets plus a handling charge of £3.45 for orders under £25 (all plus vat) but the results are excellent.

I see on the Crownhill website that they can produce PCBs and there is a picture of a small milling machine but there is little detail of what they can do, turnround times and costs so perhaps they could give us a little more general information.

I now have a heated bubble tank for etching. This speeds up the process quite a bit. The results are much sharper edges to the tracks and I don't have much cleaning up to do when I have finished. Before I got the tank I made my own using a plastic food container intended for breakfast cereals, being tall and narrow meant it needed less ferric chloride to fill it. I put the board to be etched next to the container and filled it with etchant to a little more than the height of the board. The container was then lowered into a bucket which was then filled to a suitable level with hot water to heat the etchant. For the bubbles I use a cheap fishtank aerator pump and pipe. The pipe was sealed at the end and had 6 holes drilled at half inch intervals form the end, using a small pcb drill bit. The pipe is lowered carefully into the etchant before the board is placed into the tank. When the etching is done the aerator pipe is lifted out of the etchant and lowered into the bucket then removed from the bucket before the pump is switched off, this stops etchant being drawn back into the pipe. The whole process can get a little messy but it works. A word of warning though, don't try this on top of your new stainless steel kitchen sink !!! a splash of etchant will remove the surface finish long before you can mop it up (guess how I discovered that ?).

regards, Trevor

[Mark Rodgers]

Just a quick comment, I used for many years an HP5 laser jet for making PCB's by the bucket load(I sold hundreds of pic programmers at Radio Rallys) they all had tracks down to 0.5mm and were very reliable, I still use the prototype now some 10 years on and it was made with humble Tracing Paper from WH Smiths.
I did try it with an ink jet with ok results about 6 years ago, I would expect that the improvement in technology would make it much better now.

[FirePIC]

Hi,

I use a standard inkjet printer with 3Com or HP transparencies, the trick is to set the ink transfer amount to the highest setting the printer will give (most recent HP printers have this setting in the driver). Doing this I get very good film to then use.

As a sidenote, I also use double-sided padded adhesive tape to do two-layer PCBs, you print both films, lay them on top of another (some crosshair marks help), then stick the tape on two connected sides of the PCB, instert the PCB, develop and process.

I second the idea that a direct PCB printer at hobby scale is taking things too far, it sounds like a really nice engineering challenge, but I personally wouldn't have the time to experiment.

Best regards,

Mike

[Calum]

Just a quick comment on a couple of points..

1) Conductive Inks. There are lots available, but the ones I know are designed to be screen printed so are fairly thick. I'm not sure how they would work with an ink jet printer. The ones I've played with have also been very soft and exceptionally easy to damage when hand soldering, so it's worth checking that out before spending too much time on the delivery system.

2) You asked about cleaning copper to improve resist adhesion. If you do it chemically, you need a combination of degreasing and microetching (the copper is likely to be ixidated, or to have an anti-tarnish layer, either way, you want to remove it). In a home set-up, you are best with abrasion - pumice works well.

Cheers

C.

[moonshadow]

Hi Calum, your points about the screen print inks are noted. There are several companies involved in conductive injet printing and I noticed a release from Epson today on the home brew list, concerning their 20 layer micro-thin PCB's.

http://www.epson.co.jp/e/newsroom/news_2004_11_01.htm


A UK company is also using the same technique of spraying on each insulating layer. Some are printing resistances as well and printing semiconductor devices is also a reality.


John.

[Tim]

Again some info on Epsons pcb printing http://www.newscientist.com/news/news.jsp?id=ns99996618

[JonW]

Very interesting. A few weeks ago I visited a PIRA conference on printed electronics. There are many companies investing in printing not just tracking but plastic (polymer based) transistors and a like. Most conductive inkjet printing uses a nano sized conductive particulate in the solution, some are using a UV curable ink to speed up the curing of the tracking. There are still issues regarding the conductivity and the thickness of the conductive tracking. This will obviously effect the power handling capability but also limits the RF performance of the circuit. For many applications this may not be a problem, but for masss volume markets and multi-disaplinary designs, RF, high power, high speed, these printed technologies seem limited in comparison to conventional techniques for PCB manufacture.

Good luck john

regards

[Calum]

It is interesting. I am fairly sceptical about the range of applications for the reasons you mentioned - bulk conductivity, repeatability, and geometry - all of which would tend to steer you towards simple digital circuits and not much else.

There is also the question of how you connect the components - as I said before, soldering to these can be a real problem (at least in my experience). When we move to lead free soldering in a couple of years, the replacements will require more heat, exacerbating the problem.

So my gut feel would be to stick to copper and concentrate on the best way to create an etch resist pattern.

C.

[Richard]

Check out the Cutronic Foil -

http://www.techsoftuk.co.uk/MATERIALS.htm