Stencil Printing for fine pitch componets ?

Community PCB Design
1

I have followed my insperation from spark fun and I am now soldering Lead Free boards in my modified oven toaster :smiley: . Picking and placing with an automated m/c I got on ebay for £300 (its a Japanease MDC ECM93). I am even solder paste printing with laser cut plastic film.

Where I do have problems is with manual stencil printing fine pitch components such as QFN’s. :frowning:

I have just done 130 boards which had two 40pin QFN’s. Although we succeeded in the end we probably printed over a 1000 boards to get our yield.

I started with a simple arrangemet with a laser cut .006" s.s stencil clamped down. It was possible to print with this but I found the stencil would “lift” in the middle causing a small gap between the stencil and the pcb (probably about 5-10 thou). This resulted in paste bleading out between adjacent pins and joining up. A good board should have neat little steep wall sided lumps on each indiviaual pad, rather than the range of sand dunes I often would get sprawling over all the pins.

To get round the “gap” problem I got the stencil mounted in a proper frame and got a cheep ebay printer from HK. This improved the process but results were never consistent. We often got the “sand dunes”, missed pads or poor alignment which meant we were cleaning a lot more boards than we were producing. Thus the bottle neck in our process was not being able to paste enough boards to keep the P&P busy.

The paste I am using which was recommended to me as good for fine pitch work is Kester EM918. We found that the more we “worked” the paste the “sloppier” it got and the worse the results. We found we got the best results when we took the paste straight from the fridge.

Are there some “must do” prerequisists I should be following ? Is a $2000 stencil printer going to solve my problems ? Is my experience a common problem ? :?:

2

You forgot one of the most important issues. What solids percentage is your solder paste? Did you get it in a syringe or “jar”?

I have found, if you get the solder dry enough to have individual, separated mounds on the PCB, the solder doesn’t have enough flux to make a good solder joint.

Are you using the solder at room temperature? How long did you let it sit and come to room temperature? What is that temperature? Is the temperature consistent?

I will almost guarantee a fancy stencil printer will not totally solve your problems. The sand dunes will be consistent, but still, total individual mounds will probably never happen unless you go with a nitrogen reflow oven.

The second issue is how much did you resize the apertures, and what is the board plating? Data sheets list aperture size based on a clean bare copper pad. If you are using a plated board (not including gold) with the apertures in the original size, you are putting too much paste.

We resize apertures for every plating other than gold. All others will mix with the solder, and cause an excess.

I also think your stencil is way too thick. I would go with a .004 thou or .003 thou stencil thickness.

J Long

3

I was doing this job during August here in the UK. The temperature in my workshop would have been approx 70deg F. As I said before we got best results using straight from the fridge rather than letting it warm up to room temperature.

The paste is Kester 918 in 500g Jars. http://www.kester.com/Data%20Sheets%5CS … May06).pdf

I collected a new jar from my local distributor and its been kept in the fridge since (approx 6degC).

The stencil apperatures were laser cut 20% smaller than the pads by the stencil maker http://www.photo-etch.co.uk/page1/page2/page2.html

The boards were HASL (RoSH).

I am not sure I understand the implications of your question

What solids percentage is your solder paste? Did you get it in a syringe or “jar”?

Did I buy the wrong paste or am I supposed to mix it with something else ?

When I mention “sand dunes” this is what the paste looked like on the pads before placeing and soldering. On the whole the soldering went well appart from a the occasional bridged pins on the QFN’s. I guess a thinner mask as you suggested would have help solve this.

:?: Is Kester 918 no good for my sort of application ? If not what should I be using ?

:slight_smile: I can use a thinner mask that seesm to make sense to me.

:?: Is my workshop to hot for good results, what is the ideal temp ?

4

With that quantity of boards I’d have had them assembled professionally. It isn’t expensive and the yield will be very good. Your setup is probably OK for the occasional prototype.

Leon

5

Not sure I would agree with this premise.

I had 100 or so boards done my usually ok local pcb maker/assembler. It took me over two weeks to go round all 28 pins of the QFN on each board under the mircoscope and hand check/fix many of the connections to get the boards to work. This is after giving him £1,500 for his trouble to use his supposed good kit, ie Auto P&P and Vapour Phase Reflow setup etc.

I vowed after this experience that if I was going to fix mistakes it might as well be my own rather pay for some one else to produce them for me.

I have invested a modest amount of money and have been able to get reasonable protoyping and small batch production capabillities.

I have “toaster” oven reflow with PIC based profile controller, works great and gives good results on Lead Free (cost ~= £100).

I produce D/S prototype PCB’s with 0.6m vias and 10thou track/space, no problem with bubble etch tank and home made UV box (now 15yrs old). (Cost £300)

I have old but servicable Auto P&P machine (cost £300). It does QFN and 0603 fine but “looses” about 30% of 0402’s if pressed.

I have new ERSA I-CON soldering iron, (cost £280), this has been an invalubale investment and worth every penny. For leadfree SMT soldering works great.

Bausch&Lomb stereo zoom boom microscope, (cost £120), can’t see how I would manage without one.

I have £120 IR rework station.

I have cheep manual stencil printer.

I have “home made” microplace station which can place chips to within a 2 or3thou using stero microscope vision and x,y,z & theta micrometer adjsutment table.

I have found that using mylar laser cut stencils I can easily produce boards with stuff down to SSOP chips and 0402’s.

I have to admit doing QFN’s has been a challange on the pasting front. Hence my enquires here as to why its going wrong for me.

At then end of the day it should be matter of getting the paste printing process under control, which I am not sure mine was. But other people manage this so why should I not be able to learn how ?

If its only possible to paste print QFN type boards in dust free air conditiondion workshops with auto screen printers costing $20,000 then perhaps someone will stick there neck out and say so.

My enquiry here is to ask if its reasonable possibly how have other people faired and where did I go wrong ? If the answer is its only possible if “your a proffessional” then I have to say my direct experience is even they can get it wrong :!: .

6

Who was your assembler? I’m in the UK, also, and want to avoid them!

Leon

7

You should be alright if your not placing orders with a certain company in Stratford-uon-avon.

8

I can recommend ASK Technology, Active-PCB Solutions, and Wilson Process (here in Hastings).

Leon

9

nurquhar:
I was doing this job during August here in the UK. The temperature in my workshop would have been approx 70deg F. As I said before we got best results using straight from the fridge rather than letting it warm up to room temperature.

The paste is Kester 918 in 500g Jars. http://www.kester.com/Data%20Sheets%5CS … May06).pdf

I collected a new jar from my local distributor and its been kept in the fridge since (approx 6degC).

The stencil apperatures were laser cut 20% smaller than the pads by the stencil maker http://www.photo-etch.co.uk/page1/page2/page2.html

The boards were HASL (RoSH).

I am not sure I understand the implications of your question

What solids percentage is your solder paste? Did you get it in a syringe or “jar”?

Did I buy the wrong paste or am I supposed to mix it with something else ?

When I mention “sand dunes” this is what the paste looked like on the pads before placeing and soldering. On the whole the soldering went well appart from a the occasional bridged pins on the QFN’s. I guess a thinner mask as you suggested would have help solve this.

:?: Is Kester 918 no good for my sort of application ? If not what should I be using ?

:slight_smile: I can use a thinner mask that seesm to make sense to me.

:?: Is my workshop to hot for good results, what is the ideal temp ?

So far, your problems seem to be inline with a stencil that is too thick.

The solder bought in a “Jar” is usually 88%-88.5% solids. Where syringe solder is around 86% solids. If purchased in a “jar” you are probably ok with the solids percentage.

Failures are going to happen, even with a professional assembler. Although the failure rate should be low when you are paying for it (10% or less).

You Workshop temp is ideal…maybe a few degrees cooler would help, but open air time (solder paste exposed to the air) will affect the solder. The more time in open air, the less effective the flux within the pate will be.

I personally would go with a thinner stencil and not worry about the sand dune effect. We have that happen all the time, and after reflow the solder will be correct.

Don’t think that you can not do this with your current equipment. It is just getting all the variables correct. You do not have to be a professional to assemble those parts. You just have to find the area that is giving you problems.

One issue that I have is using mylar stencils. I would check those stencils holding them flat in front of your face (looking at them on their thin edge) in the sunlight to see if flash exist. If it does, that will be a problem. Areas with many holes will tend to “seep” worse than others. The flash issue is a quantitative issue. The more holes, the higher chance the flash is preventing the stencil from laying flat.

We tried mylar in the beginning, only to find out the flash will cause problems every time (it lets too much solder “seep” under the stencil). We moved on to Kapton stencils. We still have problems, but not anything we can’t fix pretty easily.

You seem to be pretty close. I would continue on the path you are on, and fix the problems. Just the way I see it.

James L

10

If your design is fixed, you can also move to a stainless steel stencil, especially in the quantities you’re dealing with.

11

Dear propellanttech

Thanks for the directional advice. I think for the next circuit that uses QFN or similar I will try a .003" or .004" S.S stencil instead of the .006" S.S one I have just used.

Is see my paste does indeed have an 88.5% solids ratio.

The circuits I have done with mylar had low component density and nothing finer than an SSOP PIC chip. However it worked fine for me. Perhaps I was lucky not to get much “flash”. Andy at smtstencil.co.uk made my Mylar stencils.

If your design is fixed, you can also move to a stainless steel stencil, especially in the quantities you’re dealing with.

I did’nt every try to use a plastic stencil for “fine” work. I think it would be pushing your luck a bit far.

12

It definitely sounds like your stencil is way too thick. I have found a .003" stencil to be ideal for qfn parts. Additionally, you can individually modify certain pads (in this case I might have mad your qfn pads smaller - say 40% shrinkage).