PCB Etching

[duttyrok69]

Hey everyone,

I'm a noob here and was looking for info on etching techniques for pcb's. i have layouts to many pcb's and was looking to further my engineering skills by doing my first boards. has Dave done a vid on it yet ( I'm still going through all the blogs, there so great. ) Well any info would be appreciated thanks.

[mianchen]

I guess Dave hates etching PCB at home, you won't see a video on this. Search youtube there are plenty of videos..

[IanB]

Home etching is less popular than it was. Double sided, plated through holes and SMT technology are the reasons. However, a good overview can be found here:

http://www.electricstuff.co.uk/pcbs.html

[Chasm]

Making PCBs is not really hard.
The thing to remember is to keep it simple, if I need complex stuff I'd rather order them from China.

Some random thoughts
- Keep single sided, double sided is a lot more work
- Use large structures! You can do finer stuff once you know what you doing.
- Try to use SMD, way less drilling. Just saying.

Mikes article is good, here is some stuff I do differently.

I use the toner transfer method.
While I did learn to tape out and have all the gear I do have lots of blank copper PCBs but no more photoresist type.
Toner transfer obviously not suited to make multiple PCBs of the same layout, in that case just order them.

If you use photoresist but have problems with your artwork medium. There are still a few companies around which do Reprofilm for you, usually for a small amount of money. Listed under Typesetters in Mikes article, these are really good and durable too. Actually, try to find prices for this service before buying a box of OHP transparencies for your laser printer. Good transparencies are expensive, and using the wrong ones can ruin your printer. (Esp if you use a color laser! Those have much higher fusing temperatures. Copiers or high volume printers are usually even higher.)


I like to use sodium persulfate instead of ferric chloride etchant. Less of a mess and much easier to see what is happening.

[Psi]

If you get a nail or screw and a hammer then dent the center point of each through hole pad it makes easier to drill.
If instead you try and drill without a little dent in the center the drill will wander all over the place and once it starts you cant move it.

Another way is to transfer the design to the pcb with the drill holes present.
That way when you etch it the pads end up with a little hole in the center where the copper is missing.
It can be a little tricky to get right though, if they're too small they don't transfer correctly.

[6502nop]

I really like Quinn's guide (but I also prefer toner transfer), as it's informative, easy, and funny as all get-out:

http://quinndunki.com/blondihacks/?p=835

(she's also a member here, but don't tell anyone - even her!)

Anyhoo...

I also prefer the H2O2 + vinegar + NaCl formula. It's cheap (about a buck for 16 fl. oz. of each, and salt is practically free), it's easy (equal parts liquids, add salt), and Quinn also addresses the whole "disposal problem" that others don't seem to mention.

Now, when you've got kids running around the place, you do NOT want toxic chemicals like FeCl3 or persulphates anywhere. Using Quinn's guide, you only mix up what you need, etch, and dispose. Far fewer problems.

Yes, you can do double-sided work. Just use fiducials! "Overprint" some marks outside the actual PCB artwork, and line those up when sandwiching the PCB between toner sheets (or folds). If you're familiar with DTP, these would be called "registration marks" in the "bleed" area.

nop

[free_electron]

I occasionally etch a board at home. I use the pulsarpro kit toner transfer technique. Works fine.

As for etching : cut a 2 by 2 inch square of those blue household spunges. ( you know the yellow with one green sode scotchbrite ? There is a blue variation that is softer ).

Take a sturdy freezer bag with ziploc. Pour two tablespoons of fecl on the spunge. Put pcb and spung in ziploc bag, close bag.
Grab the sponge and dab it on the pcb ( do NOT use the scrubbing side ! Use the soft side ).
Dab hard enough to create foam.

Every time you pull the spunge of the board it reabsorbs all the etchant. Simply work your way over the board. No rubbing , dabbing is the key. You are essentially aerating the etchant and this speeds up the process. You can etch a 5 by 3 inch board in a few minutes like this. You dont come into contact with chemicals as everything stays inside the bag . You use a minute amount , two tablespoons. No spills. And it etches fast.

[mianchen]

I occasionally etch a board at home. I use the pulsarpro kit toner transfer technique. Works fine.

As for etching : cut a 2 by 2 inch square of those blue household spunges. ( you know the yellow with one green sode scotchbrite ? There is a blue variation that is softer ).

Take a sturdy freezer bag with ziploc. Pour two tablespoons of fecl on the spunge. Put pcb and spung in ziploc bag, close bag.
Grab the sponge and dab it on the pcb ( do NOT use the scrubbing side ! Use the soft side ).
Dab hard enough to create foam.

Every time you pull the spunge of the board it reabsorbs all the etchant. Simply work your way over the board. No rubbing , dabbing is the key. You are essentially aerating the etchant and this speeds up the process. You can etch a 5 by 3 inch board in a few minutes like this. You dont come into contact with chemicals as everything stays inside the bag . You use a minute amount , two tablespoons. No spills. And it etches fast.

interesting technique, might try it this weekend. thanks

[Chasm]

I really like Quinn's guide (but I also prefer toner transfer), as it's informative, easy, and funny as all get-out:

http://quinndunki.com/blondihacks/?p=835

Not bad, note the large clearance for the ground plate even when usiging fine structures. - Makes life much easier.

I also prefer the H2O2 + vinegar + NaCl formula. It's cheap (about a buck for 16 fl. oz. of each, and salt is practically free), it's easy (equal parts liquids, add salt), and Quinn also addresses the whole "disposal problem" that others don't seem to mention.

That is basically hydrochloric acid and peroxide.
Which emits chlorine and hydrogen chloride, which is toxic, which will also damage any nearby metal surfaces including non reactive ones like gold... [willitblend]Do not breath this![/willitblend] Use outside or under a fume hood (no, not the one in the kitchen...).
And, as always with chemicals, know WTF you are doing - preferably before you are doing it.

[hlavac]

I'm making single sided boards at home, It's fastest, can make the board in 30 minutes.

I'm using the photosensitive resist method, the kind you can buy ready made with blue opaque foil on the photoresist side. Easy to work with & cut before exposure and fast.

I'm printing the masks on a laser printer. I was experimenting with printing on a clear transparent PET foil, kind of works but larger toner areas tend to not stick properly and there were tiny cracks in the traces.
Got much better results with printing on tracing paper so that's what I'm using at the moment.
Also tried the Transparent spray and normal paper, it kind of works but the greasy goo interferes with the developing solution so it has to be cleaned off the undeveloped board without exposing it to light which is too impractical for me.

I have made a small fixture to hold the pattern flat to the PCB during exposure out of two transparent plastic plates, four screws and a bit of flat rubber with adhesive on one side (used to silence PC cases).

Mask any unused areas on the borders of the PCB when exposing to save on the etchant.

For UV exposure I have used a black light bulb at first but the wavelength was too short.
So I have built a flat UV floodlight with UV LED matrix. Kind of overdid it with 160 UV LEDs arranged in 16x10 matrix 1cm apart ;-) It takes 12V/1.45A or so and exposition time is in seconds... currently I use 30 seconds. It took several minutes with the black light lamp. It can make a fluorescent marker on the other end of room shine brightly, probably not very healthy to look into so I have made a black cover around it that can be conveniently placed over the PCB holder fixture (like a box with missing bottom that has the UV LED matrix on top inside).

For developing the photoresist I use solution of 10g of NaOH in 1L of water (bought big box of it as a drain cleaner). Easy to make when I run out.

For etching I use FeCl3. It's a bit slow when cold but that can be an advantage if you don't want to overetch. Better to fiddle with it for 10 minutes than have the board dissolve because I left it in for 5 seconds longer as with the acid/peroxide etching solutions...

I have also experimented a bit with soldermask, the type you laminate onto the board and expose with UV using the methods and equipment above. Exposition went OK but I had not properly cured it because I was impatient. I'm sure next time it will be better

[6502nop]

Chasm
That is basically hydrochloric acid and peroxide.
Which emits chlorine and hydrogen chloride

Yeah, but the points I am making are:

- It's no better or worse than Ferric Chloride, when etching.
- Before mixing, they're harmless Vinegar, H2O2, and salt. Unlike FeCl or HCl.
- A trip to Walmart and about $2.50 is all it takes to start etching at 2 A.M.
- The byproduct is mostly Copper Chloride, which is used as a tree root killer here in the States. Most water treatment plants can easily handle it going down the drain.
- It stays in the separate states until mixed, and it's ridiculously easy to mix: I use an old measuring cap from a jug of liquid laundry detergent and use only what I need - no more, no less.

To add, I've also found some very useful items at local dollar stores:
- Plastic ice tongs (check wedding/party supply aisle) for handling boards.
- Storage containers for the etch tank (like Quinn uses on her blog).
- Gloves, just in case.
- Actual ingredients (H2O2@3%, Distilled White Vinegar, and salt).
- Foam brushes for wiping boards as they etch.

So, for less than the price of the RS kit, you're good to go.

And, as always with chemicals, know WTF you are doing - preferably before you are doing it.

Yes, when mixed, it's nasty stuff. It eats metal for Pete's sake. We get that. The problems I have with FeCl/HCl is availability, price, handling, and safety of a couple of very nosy kids. So far, the only problem with the H2O2/Vinegar has been the long etching time - but that can actually be an advantage!

I endorse Quinn's tutorial, and suggest everyone try it at least once, as they probably have the chemicals lying around anyway. See for yourself. You can just hand-draw something on some scrap board, mix about 2-3 fl. oz. each of the ingredients, add a pinch of salt, and have a go. Afterwards, you can throw in a spoon of baking soda to make sure things are neutralized, if you're paranoid.

nop

[dadler]

I occasionally etch a board at home. I use the pulsarpro kit toner transfer technique. Works fine.

As for etching : cut a 2 by 2 inch square of those blue household spunges. ( you know the yellow with one green sode scotchbrite ? There is a blue variation that is softer ).

Take a sturdy freezer bag with ziploc. Pour two tablespoons of fecl on the spunge. Put pcb and spung in ziploc bag, close bag.
Grab the sponge and dab it on the pcb ( do NOT use the scrubbing side ! Use the soft side ).
Dab hard enough to create foam.

Every time you pull the spunge of the board it reabsorbs all the etchant. Simply work your way over the board. No rubbing , dabbing is the key. You are essentially aerating the etchant and this speeds up the process. You can etch a 5 by 3 inch board in a few minutes like this. You dont come into contact with chemicals as everything stays inside the bag . You use a minute amount , two tablespoons. No spills. And it etches fast.

Reviving an old thread-

I have the Pulsar kit (originally bought it for the Decal system, now using it for etching PCBs).

Today, I tried etching via the above "sponge" method for the first time. This method is also described on the Pulsar website.

Etching a small 2"X1.8" PCB, double sided (back side protected with contact paper), 1oz copper.

Using brand new ferric chloride, I applied 1oz of FeCl3 to my 2x2 sponge with a transfer pipette.

I tried dabbing the board, but it wouldn't leave much FeCl on the surface. The sponge sucked it all back up.

After about 5 minutes of this, with no apparent progress, I added another 1oz of ferric to the sponge (attempting to saturate the little sponge). This helped, but still wasn't really working. I had to start squeezing the liquid out of the sponge to let it cover the surface (with a meniscus), leaving it for a minute or two. I would then vigorously dab over the board in sections using the corner of the sponge (using the whole sponge just sucked all the FeCl3 back into the sponge).

I could tell it was working, but this was taking forever. I had to start scrubbing-more-than-dabbing to finish the job. There were some stubborn spots around the edges, where I had to really apply force and scrub to get it to etch. The Pulsar TRF is tough stuff and put up with the abuse without issue.

Anyways, it took me nearly 25 minutes to fully etch one side of the board using this method. I must be doing something wrong? I thought this method was supposed to be fast?

Is it perhaps because I am using 1oz copper? I have 1/2oz copper clad, but prefer thicker copper.

[tautech]

I thought this method was supposed to be fast?

Heat will help.

[dadler]

I thought this method was supposed to be fast?

Heat will help.

Ah. Good idea. I was hoping that wouldn't be necessary with this method but I will give it a shot.

I'm about to etch the back side now, I'll try my mini heat gun to warm it up during etching.

[KL27x]

I do it like the pro's. Vertical etching tank with cupric chloride and insane amounts of aeration/agitaton from compressed air. This does both sides at once, and there's very little under-cutting. Also, you don't have to dump the etchant when it gets old, ever. Add water and acid, and you just make more of the same etchant. Siphon it off for a backup supply, and when you have too much, leave it outside in a container with the top off, so it will evaporate into crystals. No rush, because you do not create much excess, unlike with solutions that wear out and must be tossed.

With no heater, it takes maybe 45 minutes for a full 1 oz board. 20 minutes for a half oz board. But while it's etching, I'm doing something else. The very low under-cutting also means if I accidentally leave the board in for hours, it's fine. Unless I actually go to sleep for the night with a board in the tank, I don't generally screw up a board, despite gross indifference/negligence. When you've done hundreds, the magic up mixing up chemicals and hands-on manipulation of a board while watching it etch loses its charm. Time is the least important factor to me. Amount of work, attention, and mess is what I rather avoid, and my setup is super low in these factors. The ethant stays in the tank. Put the board in, pull it out.

The 2/3 peroxide 1/3 HCl mix in particular gives off a lot of corrosive fumes and is very aggressive at undercutting the resist. You have to be on hand to watch this etch or else you will ruin some boards, and your pitch may in fact be limited a bit by your choice of etchant, here. The data is out there. HCl and peroxide will fail miserably at etching a very fine pitch or a very thick copper pour, for instance. Not sure about the vinegar variant, but I would assume it's very similar, just less concentrated. And these peroxide mixes have to be mixed fresh each time then tossed, the container/tanks cleaned. Biggest work/mess factor. Ferric is actually quite reusable, with aeration, because it just turns into cupric chloride. Just add acid and water when the salts/oxides start to precipitate out of solution. But the Fe stains are something quite horrible to deal with.

If anything, SMD makes DIY boards more attractive. Drilling holes is the biggest bother. Even with a CNC machine, setting it up and changing drill bits is a bother.

Kids? Well, this stuff isn't cyanide or warfarin. It's not poison. It's toxic, as is any household cleaner, any solvent, paint, oil, or the like. It's mildly acidic and it kills sea life and snails. I wouldn't stay up at night worrying that my kids are going to go in the garage and start drinking and splashing motor oil around the house, either. Trace copper ingested over a long period can kill you. Drinking a bunch at one time will make you sick and vomit.

[con-f-use]

If anything, SMD makes DIY boards more attractive. Drilling holes is the biggest bother. Even with a CNC machine, setting it up and changing drill bits is a bother.

Agreed. Solder mask and vias are annoying, too.

[ez24]

Home etching is less popular than it was. Double sided, plated through holes and SMT technology are the reasons. However, a good overview can be found here:

http://www.electricstuff.co.uk/pcbs.html

Added link to
https://www.eevblog.com/forum/beginners/electronics-primers-course-material-and-books/

thanks

[cdev]

I am pretty sure that the part about needing a different etchant to get a fine pitch has been disproven. In that regard, the methods all are about the same. I use peroxide/vinegar/salt and I can even put very fine text on there. The trick is basically taking it out at the right time. Some people who like informality may have a tendency to be a little more sloppy time wise with the cheap ingredients, which is fine, too. Just make your traces a bit thicker. But if you are willing to babysit the board and remove it when its done just so, then you can get very high precision traces.


Use acetone to remove the black stuff at the end.

[KL27x]

I am pretty sure that the part about needing a different etchant to get a fine pitch has been disproven.

This isn't a myth. This is very observable. Ferric and cupric can be used to etch even deep engravings with good accuracy and predictable and even undercutting. I imagine it has to do with the number and size of the ions, involved. Ferric is supposed to be a little better than cupric. edit: And the alkaline etchants are even better in this regard. HCl + peroxide is pretty close to the bottom of the barrel.

An etchant that uses smaller ions and less of them, once a small "hole" or "tunnel" starts in any given place (such as under the resist), it will get in and etch there easier than an etchant that needs more/larger physical ions to be in proximity to that Cu to etch it. I think of it as HCl is a chisel, and ferric chloride is a planer.

PCB manufacturers titrate the concentration of HCl in their cupric chloride, specifically to avoid undercutting. A higher HCl concentration is undesirable, even though (up to a certain level) it etches considerably faster. 

This is observable in practice, even if you're not going for 6 mil traces. With cupric chloride etchant, if there's a "cold spot" in the bubble tank that doesn't etch as fast, it's no big deal. You can leave the whole board in there until that final spot finishes etching. You don't really have to worry about moving the board around. Just drop the board in and wait extra long until it's completely done. So even though I know from experience that a given board is probably done at X minutes, I don't bother to even look at it. I just wait until X minutes * 1.5 to 2 and take out my finished board. The traces that were in a hot spot and finished 20 minutes earlier don't thin out much. Of course, if you're going for tiny traces, you will have to pay more attention to this. But given the same amount of attention, you will have a better etch with the cupric or ferric vs HCl and peroxide, with less ragged edges and less chance of lifted resist and broken traces.

I have never used vinegar. It may be different than HCl and peroxide. The anion in acetic acid is much larger than a chloride ion, so it could conceivably be a more "accurate" etchant than HCl + peroxide. Not sure how that works out, since you're adding Na+ and Cl- to the mix.

Curpric isn't the fastest. It doesn't have the least undercutting. But it's quite good and it has very little cost and maintenance. Overall, considering maintenance, yield, and cost of reagents, time and effort are often saved over the alkaline etchants and/or ferric.

http://www.chemcut.net/wp-content/uploads/2015/03/Chemcut_Bulletin_8_Cupri_Chloride_Proces_-Parameters.pdf

[dadler]

I ended up etching the other side, and it went a lot faster. 1oz copper done in about 5 minutes.

I used a tiny piece of sponge, and about 40ml of ferric chloride. This time, I was more vigorous with my agitation of the board. I left the board covered in fluid, waited until I could tell it had etched a bit (color change), then scrubbed vigorously with the soft side of the sponge. Repeat. Board was just sitting flat in an open-air gallon ziploc bag.

The Pulsar system really seals the surface, so there isn't really any risk of removing the protect layer even with vigorous scrubbing.

I never ended up using heat, as I feared it might soften the toner that bonds the GreenTRF to the board.

All is well. I used mainly smd components on this board, so I still need to drill the vias. Considering the plated-through-via method using windshield defroster repair solution from the auto parts store, but I think I will just solder wire in place like usual and leave that for next time.

[KL27x]

I started out using ferric in a shallow glass tray, and when my solution started to get old, I would shoot it with a heat gun. The solution was shallow enough that the air would expose the bare board as it heated/mixed the solution. It was easy to hit the slow spots and watch the copper stream away.

Adding acid and peroxide will also keep your ferric working a lot longer, but as some point you would need to add aeration or dump it and start over. (My cupric started life as ferric. I just never stopped using it. It's been through so many dilutions, it's pretty much pure cupric, now).

+1 on Pulsar. It feels like a crime to pay 1.50 for a sheet of paper, but the time and effort it saves is great.

[dadler]

I started out using ferric in a shallow glass tray, and when my solution started to get old, I would shoot it with a heat gun. The solution was shallow enough that the air would expose the bare board as it heated/mixed the solution. It was easy to hit the slow spots and watch the copper stream away.

Adding acid and peroxide will also keep your ferric working a lot longer, but as some point you would need to add aeration or dump it and start over. (My cupric started life as ferric. I just never stopped using it. It's been through so many dilutions, it's pretty much pure cupric, now).

+1 on Pulsar. It feels like a crime to pay 1.50 for a sheet of paper, but the time and effort it saves is great.

Ahh maybe I will try that method next (super shallow tray, heat gun to agitate). Since I have been using such small quantities of ferric chloride, I have just been neutralizing it with baking soda, filtering out the precipitate, and dumping the liquid down the sink. Probably not super eco-friendly, but I've only used about 100ml of ferric chloride so far. It starts looking nasty after it is used, seems to stain worse.

On the expensive paper: I received two packages of the dextrin paper with the DecalPro system and I still have a lot left.

I also print the transfer first on heavy stock paper (dead center), then cut out the center section around the transfer area and toss it. Next I cut a piece of dextrin paper the exact size of the hole, and then place a thick mailing label on the back to attach the dextrin paper to the larger sheet. I guess this is why I've been able to make the transfer paper last. Granted, I've only been using the system for decals up until this current PCB attempt. The decal system works great and yields very professional-looking decals, but it has many more steps than the PCB workflow. Once you learn how to do it, it is easy, though.

I have the Apache laminator. Now they apparently recommend a different laminator, looks higher quality, although I don't know if it gets as hot.

[KL27x]

I have just been neutralizing it with baking soda, filtering out the precipitate, and dumping the liquid down the sink. Probably not super eco-friendly, but I've only used about 100ml of ferric chloride so far.

I tried to do that once, but the precipitate was bulky and wasn't all solid. There was a high volume of gelatinous matrix that made the filtering problematic/inefficient.

If you let pure ferric evaporate without any buffers, it turns back to ferric chloride, which itself is pretty icky... very hygroscopic. But at least it's not a huge volume of ick. Used ferric that has been maintained with sufficient HCl will evaporate out into a mix of sticky brown crystals and green copper crystals. If you don't maintain the pH and chloride content, a lot of white precipitate comes out of solution.

I also print the transfer first on heavy stock paper (dead center), then cut out the center section around the transfer area and toss it. Next I cut a piece of dextrin paper the exact size of the hole, and then place a thick mailing label on the back to attach the dextrin paper to the larger sheet.

What I do is select just the dimension layer and print as many copies as I need to make on regular paper. Then cut the dextrin paper to size and tape it to each page to line up over the dimension outline. Then put the stack back in the printer, and off to the races.

I have the Apache laminator. Now they apparently recommend a different laminator, looks higher quality, although I don't know if it gets as hot.

I use a GBC Creative. I egged out the holes that hold the top roller so that full thickness boards will fit. And I leave the cover off, because I shoot the board with a heat gun while it's going through on the second pass. It will work without the heat gun, but the heat gun ensures even a large and thick board with a 2 oz pour will work on just two passes. It also speeds things up if you don't want to wait for the laminator to heat up all the way. The heat gun is also handy for drying the board after the pre-etch and rinse.