through hole via plating?

[ftransform]

So I am interested in how through hole via plating works. Finding the proper industrial solution is not happening on google.
I found this http://www.frezycnc.eu/pcb-processing-manual/through-holes-and-vias/holewall-activation/ honeywall process but it states
"", most of the procedures encountered in the production of commercial printed circuits are fairly complex, requiring multiple process tanks, each with it's own control and maintenance requirements. "

this to me sounds like bullshit. I don't want to deal with some easy peasy ink bullshit. Too complicated? challenge accepted

What is the "complicated" solution to plating vias?

i am not scared of chemistry. At all. Seriously.

basically i am looking for a textbook that deals with the creation of vias and not a easy how to website oriented for hobbyists.

[8086]

As far as I'm aware, additive electroplating is one of the ways to do this

I found this after a quick search but I didn't watch it - looks like what you're after though?

[Rufus]

As far as I'm aware, additive electroplating is one of the ways to do this

I found this after a quick search but I didn't watch it - looks like what you're after though?

Yes it is a good video and his other tin plating one showing the rest of the process. Normally you would selectively plate the copper and tin so you don't put copper on the board to be etched off again. Don't know why he didn't do that.

I think overall it demonstrates what a good idea paying someone else to do it all is.

[free_electron]

What he is showing is the old black-ink process. this has long been put in disuse in the industry.

It's not so much the chemistry that is difficult , it is the alignment and subsequent etching....

Here is how the real process works :

- take a core doublesided board with solid copper.
- drill the holes using a CNC ... <- problem 1 ... you need a robot drill
- immerse the board in a cleaning solution that makes the surface of the isolation material slightly porous
- perform a microetch of the copper so any smudges are gone. the top 1 micron or so is etched off. they use Sodium persulfate to etch. NOT FECL !!!
- immerse the board in an electroless plating tank. this solution grows a few angstrom of copper EVERYWHERE. anything that comes in contact with this stuff gets coated in copper.
- take the board out and put in the electroplating tank. Coppersulfate and a current source is used. a donor copper plate supplies fresh copper in the solution.

After about 17 microns of copper is grown the process is stopped. For fine holes you will need agitation (NOT air bubbles ! the last you want is for airbubbles to be stuck in a hole preventing copper deposit... you need mechanical agitation) to get fresh liquid to flush through the holes... one problem is the sulphur content of the liquid . This may 'poison' the donor plate (sulfur deposits on the copper donor plate.. this reduces surface conductivity and has an impact on the process..... to prevent this the current needs to alternate once in a while. ( polarity reversal for like a 5% duty cycle .. )
- clean the board in fresh water. dry board.

Now comes the hard bit ... and this is where the home process fails. up till now anything i described you can easily do ....

Apply NEGATIVE dry-film photoresist ..... <- you need a laminator and negative resist... and a room with yellowlights ...
Once this is done : expose to artwork and develop.
Anything that needs to REMAIN is UNCOVERED. This is opposite to classical etching where you want the traces covered in resist. You want the traces and other features that need toremain in blank copper !

Perform a tinflash. This is an electrolytic tin deposit. Just like you grow copper you grow pure tin. This pure tin covers only thee xposed copper. It also grows inside the holes..

- remove fotoresist.

You now have a bare board. Anything that needs to remain after etching has been coated in tin. Anything that needs eating away is blank copper.

Etch in sodium persulfate... sodium persulfate is very selective. it only eats copper , not tin ...

So, we actually use tin as the etch-resist ! The reason is that you cannot trust the photoresist to form a perfect seal above the holes. If etchant gets in the holes you would eat away the plating in there, yielding a broken connection. Using the negative photoprocess and tin plating step you eliminate this problem.

So, in short the chemistry is not the problem... it is the dry film photoresist application , exposure and the automated drilling process upfront that is the problem..

There are other solution. There is a polymer compound you can squirt in the holes and cure. this gives a connection resistance in the order of 0.5 ohm and can carry about 20mA ... this is frequently used in cheap doublesided boards like remote controls, pocket calculators and such. Curing is in an oven. The drawback is that you cannot solder to this compound. it is purely for via's...

I got pictures of the inside of a PCB fab. I visit my suppliers once a year. Whenever i have a new employee starting that will design boards i take them down there. Having an understanding of how boards are made helps you avoid boneheaded mistakes in the layout that would yield no-fab boards ...

I am wrapping up a book on surface mounted technology and one of the chapters deals with the board. I got pictures of the process line from one of my suppliers. and the process is described in detail.

I will see if i can get permission to publish the images here on the forum as well... don't want any trouble...

[digsys]

Free_electron is detailed and accurate in his comments. I was involved in helping set up a professional
PCB manufacturing business many years ago, and it was a VERY complex involved process.
The worst part is - there are DOZENS of stages involved, and if ANY stage plays up and you don't pick
it up immediately, you can end of with lots of place mats VERY fast ... and that's using 10's 1000's $$s
equipment. The plating current profile eg is a bugger to keep perfect ALL the time.
GOOD LUCK :-)

[free_electron]

ah , yes i forgot that aspect... get any sort of contamination in the electroless copper plating bath , or any kind of smudge prior to plating or air bubble or other thing wrong and all bets are off. all you have are coasters....

and the plating current needs tight control.. too much current and you 'burn' the electroless copper off... failed metaliisation. a little bit too much current and you can grow the top shut but the bottom still open ... now no liquid can move through the hole ... problem.

it is very hard to get a uniform wall covering. also the copper donor plate needs to be as pure as can be. just any old sheet of copper will not work due to the impurities in it. In the industry they use solid copper balls that go in a kind of microporous  'sock'. this acts as a filter to keep stuff from growing on the copper balls...

What is shown in that youtube video may work for 1 mm holes... but once yuo get to via's below 0.5mm ... forget it. You'll never get it right in a reasonable amount of time , and for one-off boards it just is too much of a hassle and way too expensive. by the time you want to make a second board a few weeks later your chemicals have gone bad ...

And when you are done : where do you leave all that stuff... local recycling places will not accept them becasue they don't know what it is. And some of that stuff is very nasty ...

it's simply not worth it to muck about with this kind of stuff. the chemicals are nasty , they don't keep  ( tinflash goes bad exposed to air.. ) and you will never approach commercial quality. need a souble sided board ? -> china is your friend ( sad to say )

By the way .. don't even think about gold plating at home All those electrical pens with a solution that deposits 'gold' so you can make gold plated faucets ...... here's a secret : it ain't gold... its rhodium ...

Gold plating involves goldcyanide... and that is not something you want to be around... In the industry this is done in a separate area of the building where only one operator is allowed.. If you see him keel over gasping for air .. DO NOT APPROACH ! the process has gone off and hydrogencyanide is being released in the air ... get out of there.

[Rufus]

I am wrapping up a book on surface mounted technology and one of the chapters deals with the board. I got pictures of the process line from one of my suppliers. and the process is described in detail.

Then you should know the difference between panel and pattern plating.

[free_electron]

Most of the suppliers i know use panel plating. Pattern plating is a technique that pops up from time to time but has its own problems.

Especially thickness is hard to control.. a large open space will attract more copper ions than a small sliver somewhere else... so if you have microvia's ... trouble brewing ...
Pattenr plating gives you an uneven crosscut edge. If you make a crosscut you will see a clear ridge where the copper sheet and the plated area meet. this is not the case with panel plating. And if you have uneven thickness across the board you will get varying degrees of underetch...
The etching process with panel plating is also easier afterwards since the panel is uniform in thickness. For the copper cost it is irellevant as pcb fabs regenerate the etchant and extract the copper out of solution... and throw it back in the plating bath.
Panel plating is superior when it comes to impedance controlled boards. You just can't do that with pattern plating.

Patternplating is really only useful if you are working with very thin copper foils so the aspect ratio between starting copper thickness and what you are gorwing is large. for a 50/50% it is not doable.  Patternplating starts getting interesting again when you do plugged microvia's and is done when you have laservia's for fanout. I have one supplier that switches from panelplating to pattenrplating the patternplating is only done on the outer layer where the laservia's are situated.

Most machines for patternplating use a horizontal process . I know that we use it to make VQFN packages (SAP process).
Another problem is resist entrapment...

Pattern plating has its use , but not for mainstream boards. Flex , thin prepreg and via in pad / blind via ( plugged or grown shut ) can benefit from pattern plating. most PCB shops are not set up for it.

[ftransform]

Ohhhh airless chemistry fun. 

what is this micropourus sleeve that you speak off? Can a unglazed ceramic pot (for planting) be used?

And how do they measure the thickness of the coatings? Electron microscope? Or can it be mathematically determined based on current measurements and chemical concentrations?
Why is the CNC machine so important? What is the difference between that and a higher quality drill press?

As for recycling that sounds like a job for a rotovap to reduce the liquids into solids in order to make disposal easier. I feel that reclaiming the metals would not be too difficult and the organic crap could be burned off or destroyed in a kiln under high pressure confinement.

[free_electron]

it looks like a fiberglass sock ...

[ftransform]

How come small rivets are not used instead ? I would imagine a crimped rivet to be a more durable solution then any kind of plating...

[free_electron]

that exists and was used in the 70 for cheap consumer junk.
but good luck finding a rivet for a 0.3 mm hole ...
and there is the problem of inserting and crimping them... and the potential for oxidation. you will need to use a copper rivet and then solder it on both sides..

[ftransform]

well all the information you provided in this thread is a great overview but do you know of a good resource(s) where I could educate myself about the nitty gritty details of the process(s)?

[nctnico]

I have a book called 'desiging and creating printing circuits' by Sikonowiz (ISBN 0-8104-0964-X). It describes all kinds of etching processes in great detail. But the bottom line is: you can't do thru-hole plating at home as free_electron already explained in great detail.

If you want to do double sided boards use wires to connect vias between top and bottom or use components which can be soldered on both sides. For surface mounted circuits I simply use the 'solder side' as a ground plane and use wires to connect to ground. This works pretty well.

[ftransform]

challenge accepted!!!!!

[nctnico]

Its gonna be legen --wait for it-- dary 

[free_electron]

challenge accepted!!!!!

The chemistry is the least pf your worries...
Its the dry film application and drilling that os a problem... And then tin
Lating and tin stripping...

[hlavac]

so could i:

• drill holes
• make inside of holes conductive with graphite ink or something
• apply photoresist on everything i want to go away
• electroplate copper on everything not covered including hole walls
• put chemical tin on everything not covered including hole walls
• remove photoresist, leaving exposed copper i want to go away and tin covered what i want to stay
• etch in something that eats copper but not tin
• put more chemical tin to cover sides of etched traces

And be done? Why remove the tin?

[free_electron]

correct. that is the 'old skool' black ink process. use ammonium persulfate or sodium persulfate. that eats only copper but not tin.
your main concern is drilling holes without visually knowing where they need to be., so you needa kind of cnc drill...
and then applying photoresist and aligning the photomask to the hole pattern

[nctnico]

your main concern is drilling holes without visually knowing where they need to be., so you needa kind of cnc drill...
and then applying photoresist and aligning the photomask to the hole pattern

One could print the hole pattern on paper, stick it to the PCB and drill the holes manually.

[Christopher]

Just a side thought: when making my homemade double sided boards, (Min track width of like 1.5-2mm), for vias I put wire through the hole and solder both pads. (I know this is a sloppy method and unrelated to the topic, but cheapest for hobbyists short of getting the board professionally done)

[free_electron]

your main concern is drilling holes without visually knowing where they need to be., so you needa kind of cnc drill...
and then applying photoresist and aligning the photomask to the hole pattern

One could print the hole pattern on paper, stick it to the PCB and drill the holes manually.

that is a solution. but, good luck on a board with a few hundred holes if you need via's it's typically because you do SMD work. For through-hole stuff you can use the component pin as 'via and simply solder on top and bottom.

there is another caveat though ... the drill must run very fast and stable. you need to have a smooth edge without lliquifying the resin in the hole. if you run too fast or wobbly the plating will not tick to the hole wall very well and it may peel off.
anotherproblem is the 'strike' of the drill. if you punch through too fast you will create 'dimples' : your baord will no longer be flat the copper on the backside will be pushed somewhat outward. this can lead to cracs in the plating and , it may pierce the dry film photoresist ... with all kinds of other misery.

it ain't easy...

[hlavac]

What they do to fix the drilling problems is they put a thin aluminum sheet on top and a wooden one on the bottom and drill thru the whole stack

[free_electron]

yes they do... and they also have a vacuum table and tensioning mechanism to keep everything perfectly flat.. they also use spindles running at 100K rpm and not a dremel at 5K rpm.. they also use tungsten carbide drills and not hss (high speed steel... i call it high- speed shit ) drill bits. they also change bits every 10k strikes.
they also control the downforce , strike speed and pull out speed...

They cool the drills ... and much much more. a cnc pcb drill is not a simply drill in a press stand...

[nctnico]

Nobody with a right mind uses hss drills in FR4 board. They are completely dull after 4 holes.