Screen Printing Problem

Does the resistance variation follow the same trend for all parts (e.g. always higher in certain regions), or is it more-or-less random?

The variation seems to be random although, as we increase the number of items tested, a pattern might appear.

Check again your screen, how is it good prepared for your printing process? What is a screen resolution?

It would be useful to compare your hand made production with automate printing machine. If you meet the same problem with deviation of resistance. Problem is rather in ink "chemistry".

I would have to ask our supplier about the screen mesh size but the medium being printed is high in solids content so I suspect that a fine mesh would not work. The pattern is a simple 'pin cushion' shape approximately 50mm diagonal width with no fine detail.

We only have hand printed items to look at for now, automation is for the future when we need these in higher volume.

So resistance might be depend on:

pattern design density; printed pattern dot gain factor; state of screen (how it "obstructed" by ink after several passes); state of screen (how it's tight); thickness of ink on item, how printing operator makes his work.

Hmm.. isn't it too much to be tracked?

I worked in printing industry quite long time but for producing general things.

You didn;t answer what deviation of resistance you were faced and what deviation is permitted.

Maybe tampon printing with foil ink? I am not sure has your industry to be used such printing technology?

The most extreme variation we have seen is a 50% difference in resistance. We believe that we would need to achieve 10%.

I don't know if the supplier has tampon printers but they use screen print for other heavy metal content printing although this would be for metallising in preparation for brazing where the uniformity probably isn't critical.

Excuse me if I couldn't help you as it so hard to figure out your processing troubles remotely. There is my last entry in hope to help:

So well, first. I'm assuming you're preparing your screen through photo pre-processing. Maybe you would change the design of your pattern to find out more stable?

Second. Can you place on your screen a quite "fine" image to be printed on the placed aside sample in order to control stability of printing process for your disks.

Sorry I can't invent more for a while.

regards, caramba

The pattern is fixed, it has to have specific electrical properties.

We have 2 small printed rectangles alongside the main pattern at 90° to each other. These were intended to check the main pattern resistance, I think this is what you are suggesting.

Thanks for your comments, you have given me a few things to think about.

In supposition you have pattern something like as shown in fig

I wonder again how can vary a dot gain factor depends upon a lot of reasons mentioned before plus how accurately and uniformly can be carried out photo exposure process when screen prepared. It's likely that you've got instability measured in decades of percentage for such handmade multi-chain procedure.

I do not know whether or not it's available for technology but why wouldn't you have applied further adjusting of your production by means of simple cutting off some exceeded printed areas due to make its resistance maximum closely to desired. I did something like when adjusted "fresh baked" gauges.

regards, caramba

The pattern is a solid 'colour', like a square with concave edges.

I know what you mean by adjusting after manufacture but I think we need to get the thickness over the whole pattern area uniform, we can't trim the edges.

I did some silk screen resistors for thin film electronics during my apprenticeship, a pneumatically operated screen printer could achieve maybe 20% accuracy at a best case. The ink load on top of the screen being one of the major factors.

to obtain any sort of accuracy these where all screened onto a ceramic substrate and designed to be around 50% below the required value .. then while probing the resistor we would drive a laser to cut across the resistor till about 5% from the value and then cut down the length of the resistor to finish to 0.1% tolerance or so.

i figure that since you are doing anodes then it is the surface resistance you are trying to achieve.

maybe some sort of process after the printing (eg baking in an oven or dipping in acid) could alter the resistance ? It depends on the tolerance you require i guess.

You may be right about some sort of post processing. Our supplier can't measure the surface resistivity but they are going to try to judge the uniformity by back lighting the substrate & looking for uniform light transmission.

OK. A couple of questions...

- What type of material are you using for your screen?

- Have you fooled around with adding a thinner to your resist?

- Have you considered processing the finished resist covered part through some sort of fine, wet-grind (or diatome) process to equalize the surface thickness?

- Have you considered using a photo-resist type film applied directly to the substrate?

The screen issues, I would have to talk to the supplier.

It may be that we have to consider skimming the surface of the part to level it but I think that would be a last resort.

We've looked at various different resists in the past but these will be used in a hard vacuum so must have no outgassing & must survive a 350°C extended bakeout.

How about applying several layers of thinned ink?

There is a good possibility that your losing you properties of the ink via air drying.

I would suggest that you go with a UV cured ink as it dries instantly. The properties would possibly be more stable upon completion as there is no loss of ink mass and it is more easily controlled.

We use a UV cured ink here and we can maintain our ink thickness to within 4-5 microns. That is using a 305 mesh with metalized inks.

Just a thought, may be worth a try. Consult Sericol for you ink mixes they may have other solutions as they are the best in the business

These posts are certainly making me look at the problem in different ways.

These inks (more like pastes really) have to have very specific electrical resistances. Sericol don't appear to do them (I did my apprenticeship with Sericol) & the current supplier does not seem to do a UV version. The ink is fired in air at 900-1000°C, I believe it has to be in air as oxidation is part of the process.

I've just measured a dud anode & we have a coating thickness of about 10µm but a 50% variation which ties in with the 50% resistance variation I mentioned in an earlier post so it would point to the problem being one of coating thickness rather than ink consistency.

First: I have no experience or knowledge in this area, so, please ignore me if this is rubbish. Also It may already be done like this.

Could you incorporate a net or mesh in the screen to fix the thickness of the ink. When you remove the screen the gaps should close up:-

That's essentially how the silk screen works, you choose a thickness of mesh to suit the job & squeeze the ink through. You then let the printed pattern sit for a while so that the ink flows into the gaps left by the screen material.

By the way, your pattern diagram is exactly the right shape.

Nigh,

I have spent a lot of years in the printing industry and also in silk screen.

There is a lot of information you have not given us.

But, if you are talking about hand screening, i.e., a screen and a hand held squeegee, then as far as uniform lay-down on any one piece, much less piece to piece, I don't think so.

Much as some folks would like it to be otherwise, a man is not a machine.

Pressure on the squeegee is bound to vary across the screen and as well from piece to piece.

Depending on the solvents the viscosity of the ink will change from minute to minute and hour to hour.

Depending on the mesh of the screen it will be more or less plugged at any one time.

Screen tensions are another factor. Are you using metal or fabric screens? Do you have the instruments to measure screen tension. Heck, even in the bill board industry we kept control of our process in part by keeping control of screen tension.

Again, thinking in terms of ink solvents and viscosity, are you in an air conditioned work area. When I worked in the printing industry there in England many years ago, they were trying to run litho plate manufacturing in a plant, 3M incidentally, that was not air conditioned. Plate sensitivity depends on control of the emulsions viscosity.

I probably could go on for a while. A whole series of factors need to be considered and controlled and even then I am not sure you can get the kind of control you need with silk screen.

You say you are making anodes and printing a resist on ceramics. What is the rest of that process? I am not familiar with ceramics as anodes although I can imagine ways to do that.

Nonetheless, even with screen machines I don't think you are going to get close control with screen printing techniques.

I have years of experience in printing and industry in general. Describe completely what you are trying to do, start to finish, so we can apply our general knowledge to your problem. Obviously if you are printing a resist on ceramic material you are not etching, So why the resist?

I am fascinated, something I may not have run into before, which is rare.

Tell us all and then perhaps we can come up with a process to achieve your end. I just don't think screen is the way to get consistent and uniform lay-down although if we were talking about metal etching it would not have to be that carefully controlled, just meet a minimum.

j.

I haven't given a lot of information because I don't have it. We buy the finished anode from a supplier who is making it to our spec. but struggling with quality. I guess what I'm trying to do here is to define the sort of things I should be asking this supplier in order to work with them to get the quality we need. Your post & others have already provided me with a lot of useful information & pointers on the sort of questions we should be asking.

The parts we have had so far have been hand screened but they are about to make another batch will be made on an automated screen printer, we will see if they are any better.

I don't think that the resistive ink has any volatile solvents in it, that's why it needs a high temperature air bake to cure the ink.

I can't answer the questions about the screen but our supplier is a well organised professional outfit & I would expect them to have the appropriate process controls in place.

The anodes are circular alumina ceramic disks about ø50mm with a pin cushion printed shape in the resistive ink. Randall's sketch in post #19 got the shape just about right if you ignore the grid. When an electron hits the pattern you can determine where it hits by monitoring the signals arriving at the 4 corners. I don't understand the electronics involved but we have been using similar devices for a long time. The anode is used inside a vacuum tube & the electrons come from photons passing through a window & hitting a photocathode layer.

I don't think I've ever mentioned using a resist?

The ink used is not like the sort of stuff used for printing nameplates etc., it is much thicker with a lot of suspended solids. The supplier is currently having to blend different resistance value inks to achieve the properties we need but we are going to try using the 'off the shelf' inks to avoid this blending process. We are not sure what this will do to the electrical output.

As a completely different angle on it, could you perhaps accept the product as is (or as good as they can supply), and calibrate each anode with some 2D linearizing algorithm?

John, that's exactly what we've done in the past when we've been building the electronics as well. For this particular product, we are supplying it as a plug in module for our customer to use with their electronics also, we are the 2nd source for these units so we can't ask our customer to change their electronics to suit.

Running out of ideas. As you're trying to get a uniform coating over a fixed region (rather than an actual pattern of lines etc.) maybe you need to look at a different application technique - spraying, or some kind of vacuum deposition (?? out of my field).

Now I see.

When I mentioned resist I was think in terms of engraving because when we speak about screen printing we are usually speaking about a graphic product.

When I started to read about alumina and realized you were using it for it's photo-electric properties I immediately had the same thought John did, adjust the circuit as per each tube. But for your situation that is out.

Obviously the anode in this case is used in some sort of photo-sensing device because of the electronic solid state properties of alumina. The printed electrically resistive coating has to be kept within parameters because if not it cannot serve its purpose of location in regard to the light source hitting it.

The disk, now that I have made the conversion to inches (We are backward here in the states), is relatively small and I would bet your concern is not resistance changes over the surface of each disk but rather substantial changes from disk to disk.

Are you ganging, i.e., printing more than one disk at a time? My guess is that you are. It would kill production but you could print fewer disks. My bet is that would not eliminate variation from group to group.

Is screen printing the method previously used. The patent reference I have attached below speaks of screening through a mask, so probably not.

Is the accuracy required of these devices you are dealing with any different than previously? If so perhaps you need to change your method of depositing the resistance material to something that would provide better control.

There are methods of painting that essentially deposit paint as a charged particle.

As far as what you are doing now I would suggest, as a means of improving control of your printing process that you consider the stability of your screens. If not already using metal screens you might consider that.

I would also insure that there is absolute control of the resistive inks viscosity. That would require talking to the ink suppliers and ensuring not only that they control viscosity in manufacture but also that you understand what means must be used to ensure maintenance of that viscosity when you open the can and apply the ink.

Apparently the ink base is a photo-polymer. Do humidity and room temperature effect viscosity.

If you have used other printers to supply these disks in the past what is the difference between their printing environment and the present one.

My thinking is obvious. If you stay with screen printing as your method of deposition you need to ensure rigid control of all the possible variables in the screen printing process, i.e, screen tensions and materials, ink viscosities, uniform dispersion of the pigments in the ink base (Which apparently you have already considered), uniform contral of the squeegee, etc.

That pretty much exhausts my thinking for the moment. But do let us know how you resolve this.

j.

Title: Electrodes for solid state devices Document Type and Number: United States Patent 4388346 Abstract: Contact electrodes on a semiconductor device (50) such as a photovoltaic solar cell is formed by screening through a mask (38) onto a surface (40) a pattern of ink containing a dispersion of lower melting, sinterable metal (tin) coated base metal (copper) particles dispersed in a liquid vehicle including a vaporizable binder polymer and a fluorocarbon polymer. On firing the screen coated device (46) in an oven (48) the binder polymer is vaporized, the fluorocarbon vapors etch the surface and the coated metal particles sinter without being oxidized to form an adherent, coherent, contact electrode (53).

"I would bet your concern is not resistance changes over the surface of each disk but rather substantial changes from disk to disk."

Not so, we are seeing variation in resistance across individual discs.

"Is screen printing the method previously used."

Yes, screen printing was used before but, as far as I can tell, the company has only made around 4 or 5 of these devices per year over the past 10 years. The problem with the printing has always been known but turnover was low enough to simply select the good anodes. Now we have taken on a contract which requires us to make 50 devices per month & we have to address the yield of the anodes.

"I would also insure that there is absolute control of the resistive inks viscosity"

The anode supplier has been mixing these inks to achieve the values we specified but we are getting them to try some using the ink 'as supplied' to see if that helps.

"If you have used other printers to supply these disks in the past what is the difference between their printing environment and the present one."

The supplier is the same one we have used in the past but they have moved to a new factory. The conditions now are much better controlled than they would have been in the past.

Nigh,

My brother used to work for a couple of companies which made thick-film resistors, among other things. He knows a bunch about your problem. Email me, and I'll put you two in touch with each other.

Bill

Nigh,

One other thought that I must preface by saying no insult intended because in some areas we have the same problem here.

England is the oldest industrial country. When I worked there as a lithographer in the sixties there was resistance to the changes needed to do the new thing, lithography.

As a rule, there as well as here, when old techniques and industries, run into new materials or products or purposes, there may be resistance, perhaps unintended, to what is needed to make things work. That resistance is based on tradition.

You say the screen printing is being done by a sub-contractor. My bet is you are the assembler of the final product, e.g., the vacuum tube.

If I were in your shoes I would want to spend some time, days if necessary, at the screen printer's carefully observing the whole process implementation from start to finish. It could well be that there is where you will find your problem.

It could be something as simple as the screen operator adding solvent to the ink to make it work better.

When I worked here in the States running the prep department in a forms printing plant, a young lady that worked for me seemed to be unable to create a proper plate. No matter what it was upside down.

I took one of her plates to the pressman, one I had personally checked and so knew was correct, and watched the pressman install the plate upside down.

He didn't know that I was also a pressman.

His problem. The new thing at that time was women working in the printing sections of a printing company. He did not want her preparing his plates.

Sad, but a true story that could be affecting your process.

Go watch the screen printing process being implemented.

j.

I have not taken offence at any of the comments in this thread, quite the opposite, many of the comments have been useful & are pointing me towards the issues that must be addressed.

You are right in that we build the finished product, our sub-contractor supplies the anodes as a complete assembly to our design.

We are visiting the supplier to go through the process, I should also add that this supplier is being completely open & as helpful as possible in investigating the problem.

Nigh,

The last clears up any other questions I had or might of had. I had thought of making the production runs and selecting only the good ones.

No doubt you are aware that in the manufacture of solid state devices, the wafers that are the bottom line of those devices, do not, in production meet all the requirements. It is a given there that a certain percentage are discarded. Insofar as they are producing hundreds of thousands, at least for the main devices, they throw away a fairly large number.

Obviously your pricing in the past accommodated throwing away a few insofar as you were producing only a few. Now, and obviously this is a business decision which only you can make, with a larger production run you may have to change pricing or change your method of production.

As far as silk screening, if everything I have suggested does not pull production into the values you seek, I cannot think of any other components of that process to work with, and I must say that I have years of experience in the graphic reproduction industry involving lithography, silk screen, and lesser methods, i.e., photo engraving, letterpress.

Actually, I had thought of one other possible method.

If the disks will accept a fair amount of pressure I had thought of printing an ink on them with the litho process in a flatbed litho proof press. There are as well methods of making thin relief plates that are used in a process sometimes referred to as "dry litho." Again that might work on a flatbed litho proof press where you might get tighter control of the ink lay-down.

Because the process is dry, i.e., imaging the plate does not require water because of the raised image, you might be able to use the screen inks you are now using. The controlling factor will be how thick an ink layer you require.

There ought now to be in England litho printing trade houses that produce color separations and print them on the proof presses I have referenced above. I would talk to such a house and see if your problem might be solved that way.

As far as screen I have laid out all the controlling factors.

Do let us know how this problem works out for you. I am also interested in the end use purpose of the tubes you manufacture if that does not violate any body's trade secrets.

Technology, of any sort, always interests me.

j.

Jack,

We accept that we will have some wastage but, as it stands, we can only use around 40% of the items. If we can get that to say, 60 or 70% we can live with it, any better will be a bonus.

I'm not sure about the dry litho process, the ceramic disc is strong in compression but only if well supported, otherwise it will just shatter.

Our supplier will be producing another batch soon but using a semi-automated screen process. I intend to visit them to watch how this goes. It will take a while for us to receive & test the finished anodes but I will report back on the outcome.

The finished tube is part of a passive missile detection system. We can produce photon detectors that, even in strong sunlight, will pick up the UV signature of a shoulder launch missile from the sort of height that a cargo plane operates at. This alerts on-board systems which allow flares, chaff or other counter-measures to be deployed.

The printing, i.e., ink carrying, element on the proof press is a rubber blankets on a cylinder which is inked by the relief, in the case of dry litho, image plate.

Typically for printing on paper we employ a "squeeze" of .004 thousandths of an inch. We would normally set that squeeze with a .004 piece of film that should offer firm resistance as it is pulled out of the nip.

The bed of the press, where your wafer would rest is absolutely flat. If your wafers are flat and the same dimension there should be no problem.

But let's see how your new runs turn out.

An interesting use which makes sense but which I would not have guessed.

j.