9
\$\begingroup\$

I just tried to make my first PCB at home, but it did not work. I don't know what went wrong, so I'd appreciate some advice.

I was using the photoresist method, so I printed my circuit on a transparency using a laser printer and made this template:

enter image description here

Next I exposed my presensitized PCB and developed it to get the following photoresist mask:

enter image description here

The mask looks quite functional at this point. All traces are well defined.

Next, I submerged the PCB in Ferric Chloride to remove the unmasked copper. This is where things went wrong. The copper between the traces on the top side of the PCB just wouldn't come off. I ended up over-etching the board, as can be seen from the fact that the photoresist mask started to break down, yet the unmasked copper between the traces at the top still wouldn't come off! Here are the pictures:

enter image description here

And after removing the photoresist mask:

enter image description here

It seems that something was different with the top of the board, but what? Was it a defective board? Did I do something wrong? I'd really appreciate some advice before I try again.

\$\endgroup\$
  • \$\begingroup\$ What kind of etchant are you using? Ferric Chloride or Cupric Chrloride? \$\endgroup\$ – angelatlarge Apr 1 '13 at 6:11
  • \$\begingroup\$ I am using Ferric Chloride \$\endgroup\$ – Val Blant Apr 1 '13 at 6:11
  • \$\begingroup\$ Have you used this batch of etchant before? Have you used it a lot already? I found that after a while you need to separate the grime collected on the bottom from the liquid parts: not doing that seems to cause exactly the kinds of problems you are seeing. But if the etchant is new, than this wouldn't apply. \$\endgroup\$ – angelatlarge Apr 1 '13 at 6:12
  • 2
    \$\begingroup\$ I'm thinking, either that PCB was defective (it wasn't clean when they coated it) or maybe it was insufficiently developed in that spot: a very thin layer of resist remained there, blocking the etching. \$\endgroup\$ – Kaz Apr 1 '13 at 6:51
  • 1
    \$\begingroup\$ At what temperature did you do the etching, and did you move (agitate) the pcb and/or the etchant? \$\endgroup\$ – Wouter van Ooijen Apr 1 '13 at 7:47
5
\$\begingroup\$

I've had the same problem in the past, here's what I did to fix it:

  1. You have some fuzzy lines, probably because the mask isn't close the board in some places. Wedge your board and mask between two glass plates and clip them together with office clips.
  2. Your board looks a bit under exposed, in my case I had to double the exposure time.
  3. The longer the acid takes the more distortion your tracks incur. Try using the Hydrochloric & Peroxide mix. A 50/50 mix gave me an etching time of a minute.
\$\endgroup\$
  • \$\begingroup\$ I think this probably correct. In my experience, pitting, uneven etching, and general etching weirdness resulted from etchant being too weak: too used up in case of FeCl etchant, and not enough Hydrogen Peroxide for CuCl etchant. \$\endgroup\$ – angelatlarge Apr 1 '13 at 15:51
  • \$\begingroup\$ @s3c: I just tried your suggestion of using a 50/50 Hydrochloric & Peroxide mix. I am very surprised that it worked for you so fast! Mine was not able to remove all copper even after waiting for 20 minutes. Here is a picture of the result: i296.photobucket.com/albums/mm185/vace117/HCl_20min_2.jpg \$\endgroup\$ – Val Blant Apr 2 '13 at 3:18
  • \$\begingroup\$ The concentration for Hydrochloric acid is fairly constant, Peroxide on the other hand can be purchased in various strengths. I can't recall which one I used but it was fairly high. If the problem persist increase the exposure time for your boards as well as the development time. It really is that quick if you get it right. \$\endgroup\$ – s3c Apr 2 '13 at 12:21
  • \$\begingroup\$ +1, especially for #2 (underexposure). There seem to remain some green shades in the areas that need to be blank. My experience is that even a little bit of remaining photo resist takes all the fun out of the adventure and gives exactly the result pictured above. \$\endgroup\$ – zebonaut Oct 1 '13 at 19:03
3
\$\begingroup\$

I have seen this problem before. I etched many PCBS, but occasionally one would turn out like this. I can only imagine that it was a poor quality board. However, some advice on etching your boards correctly:

After you cut the PCB to size, sand off the cut sides, and bevel the cut edges on the side you're going to etch. This is to make sure that there are no burrs sticking up which will prevent the mask making perfect contact with the PCB.

Make sure you print out the mask mirrored, and expose the PCB with the ink side of the mask touching the pcb. This is critically important. If you have the mask the other way around, with the ink away from the PCB, then the light will leak into the gap between the ink and the PCB, causing fuzzy edges, just like the ones you have in your image.

Make sure everything is clean and dust free. Be anal about this. Get an air duster, and use it to remove dust firstly from yourself, then from the UV box, mask and PCB. Place the mask on the glass (ink side up), and slide it around a little. You will be able to feel and particles scratching around.

Don't remove the protective layer from the PCB until the last moment. This layer not only keeps out the light, but also prevents the PCB photoresist layer getting horribly scratched like yours.

Then drop the PCB onto the mask, it should make that lovely sound of two clean, flat surfaces coming into contact. Slide the PCB around feeling for particles.

Once the PCB is exposed, develop it in fresh developer solution. Once it's developed, take it immediately to the sink, and run it under water, rubbing it with a good quality paper towel, like Bounty. Some more photoresist should come off, leaving the image on the PCB crisp and with good contrast.

Now put the board in your etch tank. If it's a bubble etcher, make sure the bubbles are all going up the side of the PCB being etched, not up the other side. When the board is half-way etched, take it out, and rotate it 180º. This will help the board to etch evenly across its surface.

\$\endgroup\$
2
\$\begingroup\$

I see several possible issues.

  1. You have an uneven density of copper on the board. The right side copper is completely etched away while the other side has more copper. It takes more time to etch away the 'empty' areas. So the areas that need less etching will be exposed longer than needed resulting in less copper there. Try to even out the copper density by adding back the copper in your CAD files -- just don't connect them to anything for electrical reasons ( or you can connect it to GND for better noise immunity).
  2. The solution may be weak as others have mentioned.
  3. Have you heated up the solution? It needs to be warm to work better.
  4. Are you moving the board around enough in solution. The echant may not be circulating enough to get into the voids.
  5. Your photoresist registration looks bad. As mentioned above, check your board cleanliness, and exposure times. Also fill in any scratches or voids with a sharpie. If your mask is bad, then the etched board will be no better.
\$\endgroup\$
2
\$\begingroup\$

Usually printers are not that good to provide a solid toner layer, there are small holes / imperfections left.

When doing photoresist I usually print two copies of artwork and align them together by putting a drop of water between - water surface tension acts as glue, letting you to perfectly align traces. Then I use a drop of water on the board to stick the artwork to it and use pieces of taper towel to squeze the excess water out by working it from the middle to the edges.

If you align the artwork perfectly, you'll notice that your printer has some "dead pixels" - spots where there is no toner printed ever. Taking out and cleaning the drum with alcohol might help. Or using two different printers to print two copies.

Unfortunately the quality of photosensitive PCB is not the factor you control, that's why I had to abandon photo method in favor of toner transfer.

\$\endgroup\$
0
\$\begingroup\$

Thank you for all the great suggestions!

I have tried 3 different ideas recommended here: 1) Using 50/50 mix of hydrochloric acid and hydrogen peroxide. This didn't work, b/c the solution could not remove all the copper even in 20 minutes, so the resist started dissolving. It seems that I am having drastically different results from those reported by many others. There must be something wrong with my board... enter image description here

2) I have exposed the board longer (8.5 minutes instead of 7). I have also developed the board longer (5 minutes, instead of 3) to make absolutely sure that I am not ending up with a thin layer of resist remaining between the tracks.

3) I have heated up the Ferric Chloride. I don't have a thermometer for liquids, so I heated the echant until it just started visibly steaming.

4) I have used another PCB board, although it was purchased from the same batch.

Combining 2,3 and 4, and somewhat over-etching, the board came out like this: enter image description here

enter image description here

Definitely not ideal, but at least I have something usable now.

\$\endgroup\$
  • 2
    \$\begingroup\$ it really looks like your boards are old; maybe these are bargain boards where the photoresist has started to go? fresh chemicals, fresh photoresist, heat and agitation are all important, as mentioned by others. clean is also super important, but again you're using presensitized boards so I would start to suspect that the boards have been sitting too long. This doesn't apply to presensitized boards, but clean copper is pinkish, not copper-coloured. \$\endgroup\$ – akohlsmith Apr 2 '13 at 12:35
  • 1
    \$\begingroup\$ With only one guess I'd say your exposure time is too short, a thin resist layer can still be present even if it's not visible to the naked eye. Increasing my exposure time from 4 minutes to 8 fixed this problem for me. The boards looked exactly the same after exposure but the one etched really fast. \$\endgroup\$ – s3c Apr 2 '13 at 18:31
  • \$\begingroup\$ Problem is, that laser printed mask doesn't have a lot of contrast. \$\endgroup\$ – Kaz Apr 2 '13 at 21:47
  • \$\begingroup\$ @Kaz: What would you use to print the mask, if not a laser printer? I used three images overlayed over each other to increase contrast. It was opaque enough that I could not see any light from the sun when looking through the black parts. \$\endgroup\$ – Val Blant Apr 2 '13 at 22:39
  • 1
    \$\begingroup\$ @Val This depends on your setup. The UV lights used for my first attempt were fairly weak and took a while. My current lights work much faster, afraid this depends on your particular setup. The ideal way to test this is to expose one board with varying time intervals, this can be accomplished by placed a thick black piece of paper between the board and the light source and pulling it out a cm at a time every minute or so for 10 minutes. The resulting strip that has the best after etch results is your optimal exposure time. \$\endgroup\$ – s3c Apr 3 '13 at 9:54

Not the answer you're looking for? Browse other questions tagged or ask your own question.