Sponge + Ferric Chloride Method -- Etch PCBs in One Minute!

Step 5 Finished Product, and Your Results

Here's the finished product. Not too bad for 5 minutes from start to end!

I hope you will agree that this method is faster, cheaper, and more exciting to watch than other etching methods you may have used in the past. I'll bet you'll never use the ferric chloride tub or tank immersion method again.

I'm not sure if this low-volume wiping method will work with the muriatic acid and hydrogen peroxide etchant, but it is worth a try.

Some information I read indicated that adding a small amount of citric acid to the ferric chloride will make it an even more effective etchant for use with the sponge or immersion methods. You can find citric acid powder at beer and wine making shops, and even on eBay.

Go ahead and try this sponge method, and let me know in the Comments section if this will become your new method for etching circuits like it did for me.

Comments

[triodo]

alguien sabe si este producto de radioshark lo venden en ecuador sur america ?

[sudgeist]

me parece que hay steren en ecuador...ahi lo consigues...

[SniperProf]

Very nice :)

[Rock D]

i cant find fecl3 solution here by around my city. they offers me a crystals of fecl3. so help me to make solution of that????

[TechShopJim]

Yes, that will work. You just mix it with water to activate it.

Remember the safety rule: "Add acid to water, just like you ought'er."

That means you put the water in the container, and slowly add the FeCl3 powder to the water. NEVER do it the other way around...for example, NEVER pour water into a container of acid or other chemical because it could heat the water and burst into steam and kick the chemical up into your face.

Enjoy!

[rykonen]

Great job

[RES_Q]

Just a brief note regarding above mentioned UV Resist boards, it doesn't work using Sodium Persulphate or Ferric Chloride etching agents either. UV Resist wasn't designed to withstand the slightly abrasive nature of this method. The toner from a printer is a little stronger when laminated to a board properly and can often hold up to even harsher treatment.

However I have found that higher density/quality boards take a lot longer using this process, exposing the resist (any type) to a longer period of scrubbing, and is generally less reliable for fine traces than traditional heat-dip-agitate methods. It works a treat on the thinner copper laminates though, and I myself use this with great results for quick 'n' dirty simple prototypes. Not discouraging this method, just serving up a warning from 15 years experience to help avoid costly mistakes.

Great Instructable, just get the materials right and you'll have sharp looking boards, dressed to impress when following along.

[PCBgroove]

I use the Pulsar method of etching boards in a plastic seal-able baggie with Ferric Chloride. I tested Pulsar's data with using a laser printer and their foils to make boards verses the old UV photo resist method. Both methods used the same sponge on a baggie method. What I found was that the laser method pits and I can only get down to 10 mill reliably and even then it is dicey. Together with the pitting some traces break very easily.

Compared that to the UV photo resist method from MG chemicals. The traces show no signs of pitting, the density and resolution increased down to 3 mill reliably, and I beg to differ the photo resist holds up quite well to the sponge. I use negative method and maybe the positive method is not very strong. I can say this much, I am still very impressed with the UV photo method.

Timewise it takes longer for the photo resist. I have a quick single sided board I can in under an hour design to etch a board using the toner transfer method. I also have large .050" traces and ground plane distance of .025".

If I have a component that is a TSSOP with a lot of pins, I go with the photo resist method. I have a large format camera and a copy board I use with 2:1 masters. I get perfect dark black and clear film to make the photo resist with. I get perfect registration down to 4 mils, and my 10 mill traces show no signs of losing integrity after etching.

Here is a couple of picture wiht such a device using toner transfer method and the photo resist method. Again both were etched in a baggie of one oz of Ferric Chloride with a sponge to wipe the boards with. It generally take my 5 minutes to etch a double sided board. This same sponge has gently etched over 10 boards already with the same one ounce of etchant. I am going to treat the sponge with citric acid (it is safer to use) and I am sure I will still be able to continue to use this same sponge for some time to come. No hazardous waste.

Top board is the TTF method notice the pitting the shorted 10 mil pads on the IC.
Bottom board is the UV Photo resist notice the edge detail even close to the edges of the board. Now agreed these both do not pass muster. The point of this exercise was to compare to methods in a timely manner. With more time the UV photo method would even be sharper and I would have sued new boards. Both boards are 5 years old at least.

The weird bottom pads are for a surface mount crystal and two 1206 capacitors. That shoud give the idea of size. The bottom picture the resist was left and you can clearly see no damage to a 5 years old resist using the sponge method with ferric chloride.

[RES_Q]

Although I spoke somewhat against rubbing the UV Resist, it was only to point out that it was not intended for this purpose and will be rubbed off if not careful, but it will withstand some abuse before failure. Positive vs Negative in abrasion resistance, I can't comment on. Possible failure happens to be the case if you try this method with 1 and 2 ounce copper boards. The 1/2 OZ boards work great with this. For myself, I find that I do a lot of fine pitch layouts and this method proves too unreliable, using either TTF or UV. If it's a small thru-hole board, then I'll use this method with any boards, just after soaking them for 5 minutes in etchant first. This reduces the chances of rubbing off the resist before the copper, although patience and a gentle hand are still needed for thicker Clad. I agree on the pitting issue with Toner, as again, it was not purposely designed with PCB re-transfer in mind. I do not use toner on traces less than 20mil, and even then, my lightbox has been seeing a lot more use lately. My particular combination and environment allow me to produce almost manufacturer quality double sided boards in less than 2 hours. The extra 15 min in UV Photo prep and dev. are quite worth it. I run 3 mil lines all the time and have yet to have one broken after 'old fashioned' etching. I'm not trying to argue, but rather point out some of the inadequacies of this particular method in an attempt to help save others costly mistakes.

It really boils down to patience, materials, board requirements and working environment. Personal preference plays a role too I think. Wish this had come up a bit ago, I would have posted pictures of some 10"x8" DS boards with 3 - 25 mills running everywhere (processor and logic board) to show just how nice UV can turn out, unfortunately they have been populated, sold and put into service. 

Keep up the great instructables and comments, because although it may not be the 'best' way, and may cause contrary comments, it is in fact a working method, and a wealth of collective knowledge for novices and professionals alike...

[PCBgroove]

Thinking about what you said about careful has great merit. I do these with the sponge method by hand because I can control the feel of how hard the process is working. That being said is probably why I get good results. It is not meant for  a machine to do unless hand feel and complete feedback is seen. I tend to do center first and the work to the sides rather then the whole board at once. This gives the edges a break where rubbing to hard would allow for the edges to wear fast. It is like photography and developing ones own film. There is a right way and a wrong way to deal with the agitation.

BTW, I added 10% in citric acid to my ferric chloride (again only about an ounce) and lo and behold the etching time came back with the only thing is the sponges got softer. Probably because the iron oxide was less.

My limit on my laser printer seems to be about 12 mils. I get about 4 mils of accuracy with the photo system although it takes much longer to produce. I am hoping as I get my photo method dialed in I will get to the 3 mil current PCB spec. Anything lower I think will require better photo lith films. My pad limitations are with the copper eyelets I use which translate into a .050" pad. For .015" vias, I probably will use a silver loaded epoxy under vacuum and a laser to drill the holes. Now the trick is to get this complex via technology to work. Getting to that point is exactly where Taiwan's PCB industry is at with 64 layers limit. That level will be my 12 month goal.

Sorry my big fingers and my little keyboard my email address is misspelled [e-mail]

[PCBgroove]

If you still use a laser printer you will never be able to print small traces because the grin size is simply too big and the micro vibration that laser printers use to gently shake the toner while it heats makes the traces join. I have done everything with every laser printer imaginable that uses the right old fashioned toner. All the problems of shorted traces when down to 0.65 mm trace pitch it simply fails. Printing onto plain paper is better but that will not transfer well. Using the toner transfer paper form Pulsar works great but the fine pitches do not fair well. I equate the output to printing on a transparency. The two methods: transparency printing and Pulsar printing, yield the same results in pitch variations.

After years of researching all of this I came back to film based negatives and using the photo method recommended 15 years ago where one prints at 200% and reduce to 100% in the camera. Film grain is about 1000 % smaller than toner grain and can handle the fine pitches. I also can grow trace and pad widths with exposure and developing techniques thereby mitigating any issues with undercut.

The sponge method I like because it concentrates the liquid enchant and move it around. It is reasonable in speed and the undercut is minimized. I do not use anything but 1/2 oz board. Post drilling and eyeletting, I cover the board with liquid flux and with hardly any solder and a medium heat, I coat the board thus building up traces to well over 2 oz board specs. That is an old technique alive and well in my lab.. This enables all the components to tack onto the board and with the fine pitch devices that is more than enough to solder each pin. That means I merely touch the pins with a hot iron and it reflows the solder that is on there, making a great solder joint.

This process works great now but I only do small boards. My camera is 4X5 and is the limit in size my copy stand can handle. To do anything larger I use a service in Silicon Valley that does laser plots. However my PDF's post Illustrator are good enough for my tiniest of vias and tiniest of IC's, I do not use Gerber and if I did the laser plotter would be fine. It costs extra money to convert from a PDF to a Gerber file. Hardly worth it. I get very good quality negatives in minutes where the laser plotting service is a 2 week window.

It is nice that the US could make 3-25 mills running all over a logic board. We are rather lost with this as Taiwan has closed the market gap such that this 29b dollar industry, the US has all but failed to compete now globally. When I have a board made here it costs me over $50. When it is made in Taiwan, or China it costs me $5. The problem is pollution and productivity that I think the sponge method can mitigate both issues. If anyone is interested in developing this further I think it has merit. If the abuse can be mitigated my 1 minute times the addition of an acid will accommodate that. I like citric acid it works just the same as HCL and is far safer to use.

Please contact me at [e-mail] if you would like to know more and are interested in pursuing putting the US back in the game competing with Taiwan so that the US sees exports rather than the import imbalance we currently see. I think this is one big reason our US economy is in tatters.

[RES_Q]

PCBgroove: I'm in Canada, lol. Vancouver Island to be exact (yup, had to go see the games!). But I understand where you are coming from on the issue of national competition and import/export imbalances. And wild ideas are what have made for some of the greatest advances in modern technology, so I agree with your wanting to further develop this method. Just as it stands it is not overly useful to me, but despite the near perfection of my boards with my process, I, like any other person, do not like waiting. ;-) I have not tried adding citric acid to the etchant before, but given a reasonable knowledge of chemistry, it can't hurt to try... you never know, I might just convert...

Anyhow, as of this moment, my process entails my Brother Laser printer at 1200DPI on HQ laser transparencies, and processed using an intense exposure, a weaker than suggested developer and agitated FC dip. The strong exposure reduces UV leakage to printed area edges by reducing the time it has to seep in and do it's evil (toner isn't 100% opaque) and the weaker develop helps ensure this integrity when I'm finished. This helps reduce undercut during etch and gives very, very crisp lines under the scope during QC down to 3 mil. I take my time in prep as that is the real secret in my recipe. If I'm completely organized before hand, I'm free to monitor and tweak where necessary. No 2 boards ever process 'identically', but I've got it nailed down to about 98%. It's all been a big, long learning experience, but I'm happy to keep pushing envelopes within reason, so I think this method here, as you have said, deserves some further experimentation and advancement.

Time for a bite to eat and to start my next project...

[PCBgroove]

Out of Edinburgh Etch comes an acid and ferric chloride based mordant method of etching metals. Kieth Howard is the one that did the original research (as far as I can tell) of a method that goes back to the Middle Ages of metal etching. As far as the "chemistry" it is well known. Adding citric acid in the right molar strength makes the iron oxide go back to iron chloride. That is all. Howard says they increases their mordant mixture's effectiveness to one year of viability, probably because of the high levels of molar quantity of copper chloride which is an ineffective slow enchant. Others have posted on this and Ferric Chloride is safe and effective to use. Citric acid while hazardous if used improperly (for that matter all etchants are very hazardous if used improperly), in the right concentrations is very safe to use in a ferric chloride.

To read more on this fascinating topic see:

www.polymetaal.nl/beguin/mape/edinburgh_etch.htm

Here one can get the right molar strengths that according to Howard are safe to use.

[PCBgroove]

I am very happy to see you are having success with a Brother laser printer. I have heard they have gone to a epoxy based system which probably makes for a nice transparency.

The only other alternative is a laser plotter but it uses the same film I use without the steep cost. The nice thing about black and white orthographic films is if they are processed right (and that is a no longer a challenge for me) they will last 1000 years without degradation. That means in the future, say in 20 years, I can take that saved negative and reuse it. I am not sure digital prints outside of a laser plotter have that same integrity. But laser plotter's are also 2400 DPI and higher.

The problem is under a scope are jaggies seen in the angled traces and are the edges fuzzy or razor edge sharp on round pads. It is real easy to print vertical or horizontal lines it still takes an interpolation to make any other angle. It is this raster interpolation that gets these systems locked into medium resolution (10 to 20 mils) and possibly to one off high resolution (6 to 10 mils) and no further. A laser plotter uses postcript to plot lines as lines and not a raster image. I get that same level of integrity by a 2X blow up of the master and reducing it 50% in my camera where the moleculer cloud is still well below the 10 mil traces. Anything lower than that is a real challenge even for the high end board makers. Most board makers I know have a spec of +- .004" which is a 8 mil accuracy. That means a 3 mil traces has a very good chance of not being there. It also means a 10 mil traces could reduce to 6 mils. A six mil trace is dicey.

As far as using transparency and a laser printer. Do you get bleed through? Can you put your transparency up to the light and see no light coming through? Do your fine line traces hold up to no breaks? The surface tension of a liquid does have its mechanical limits. If your clear areas are truly clear and your black areas really black even with your hairline traces, you have something remarkable.

What I have done is to use a foil to overlay the toner image and that seems to intensify the contrast of the image where it can act as a negative properly. The issue there for me is the foil tends to stick to the transparency too thereby creating unintended interconnections. I am pretty fed up with the toner transfer method for anything complex as the foils are persnickety to humidity to the point if we have a storm the foil refuses to stick to toner even after I fully dry the PCB and heat it up to over 150°C.

So over the years I have accumulated the photo equipment to make real negatives that black areas allow no light to shine through and I take my dense PC board masters at 2X and make perfect 1X negatives with the emulsion side up (I have to get the emulsion next to the copper otherwise a lens effect takes place and the print appears fuzzy). It is not easy to do this, but I get very clear white areas, and very dense black areas with no jaggies. Home Depot has a nice daylight flourenscent bulb in a great package that fits into a defunct Brother printer that I use as a transparency box. The short distance gets me to 6 minute exposure. That keeps the light creep to a minimum and the light is not harmful to look at just very bright. They also sell a florescent fixture that has the right kind of connector and transformer thereby hacking the fixture and adding in a power switch and cable, one has a compete system to run one bulb. Doing a PCB this way I get accurate double sided boards. That means in my 12 month time line as I do have access to a laser cutting system, I will get to micro-vias, and multiple layers. After that it will be to get to the 64 layer spec Fujitsu is pursuing.

I would like to see some of your fine work. Can you or have you posted any?

I will post a recent board of a rotary encoder that uses a tiny 16 pin SSOP device: www.austriamicrosystems.com/eng/Products/Magnetic-Encoders/Rotary-Encoders/AS5040

It is these kinds of devices that I run into problems with where I get razor edge negatives. The toner transfer method completely fails with this chip.

[DemonSpawn]

 So i was wondering how you got the black lines off from the transfer paper and also is this method accurate as my pcb is very colose together lines and also very small?

[0x0539]

how did you make your etched pcb black?

[TechShopJim]

Hmmm...I'm not sure what you mean. It isn't black. It is copper on a semi-transparent epoxy board.

[0x0539]

usually the insulating layer is this ugly green color: http://www.instructables.com/id/S3YRO5HFYTD4Y8Y/ . so it's just the pcb you bought?

[TechShopJim]

Yes, the board I used was translucent, so you are just seeing the black background through the board a little bit.

You can see I am holding the PCB up to the light in this image:

http://www.instructables.com/files/deriv/F4E/Y7UR/FX23V78R/F4EY7URFX23V78R.MEDIUM.jpg

Thanks for the questions!

[profpat]

Great Idea will try this method the next time i make some pcb's for my projects.thanks for sharing!

[vmspionage]

Hmmm, I wonder if I can use the sponge method with UV resist and ammonium persulfate... I'll give it a shot and let you know!

[vmspionage]

Just a quick update, this does not work with ammonium persulfate etchant. It reacted with my sponge and made some stinky gasses, and the sponge seemed to accelerate the breakdown of my resist layer leaving me with some of my finer traces over etched. Oh well!

[dgdimick]

Thanks for trying it out - I was wondering the same thing.

[lcohen]

Jim Great Instructable! Congrats! Really solves a problem I am dealing with right now! Thanks!

[aplnx]

Tacunada... Borrokera Ducarai...

[valeen1959]

Check out their toner reactive film. I use the green for sealing the toner. As toner does not cover 100% (large areas will show some pitting) The TRF product makes a much better board. Frank at Pulsar talked me through modifying a laminator to work with the products and I have been making boards for a few years now. Note that the etch time will vary on the thickness of the copper used, all I can get here is 1 oz and it takes about 15 min to etch even with the sponge. Great product and thanks again Frank for making life easier.

[doylep]

Valeen, Any chance of getting some instructions for moding a laminator I have the standard GBC model that PCB in a box use but it is no good for 1.6mm pcb

[Wieniec]

You can use hot (180-200C) iron against laminator. And of course "chalk" paper as base (somethimes i use paper from color magiznes). It's well know and using method in Poland :)

[TechShopJim]

Yes you can certainly use a clothes iron to remelt the toner and perform the transfer. I had problems with that, because I would either get the toner too cold (so it failed to stick to the copper), or too hot (in which case it would squish and smudge the toner and ruin the traces). The $70 laminator that Pulsar sells is TOTALLY worth the money! I get PERFECT transfers every single time now. And hey, if you don't want to buy your own laminator, this one will be available for members to use at TechShop Menlo Park...when I get done playing with it. ;) It is clear that I am going to have to do an Instructable on the exact process I use to do toner transfer. It really is perfected. This method is so cheap, easy and consistent that it would probably benefit folks here to at least see the process.

[[TheDarkTiger]]

I will test it as soon as possible ! It's looks great !

[freinds323]

Thanx..That`s really good..
i used to put the board in the acid and "vibrate" it up and down...it doesn`t take more than 3-5 minutes...
i`ll try this method next time!

[jimyxt]

hey man thanks for it I had made PCB 20 years ago, I've tried a lot of methods but never this, that's really amazing, fast and safe for your cooper tracks (some slow methods can damaged the resistant and you will lose the tracks, congratulations for your discover!!.

[ivaneduardo747]

I will give it a try (if i find the ferric chloride thing in Dominican Republic).

[luispic]

that's something you'll probably find in any electronics shop, if there are any in DR it's called "cloruro férrico" here in Guatemala and it's what every electronic engineer is thought in school.