0
\$\begingroup\$

I'm trying to make PCBs for the first time, but I hit a roadblock and don't know how to fix the problem.

I etched my 3rd developing attempt, but it looks like only the border around the traces was etched, not the open areas of copper between traces.

Here's a picture of the transparency I used and the resulting PCB:

Looks better in real life

enter image description here

enter image description here

These are the materials and methods I used:

  • Transparency printed by FedEx - I combined 4 layers
  • GC Chemicals positive per-sensitized board (from Amazon - product link)
  • Exposed with sunlight for 20 minutes (clear day, had board perpendicular to sun, sandwiched copper board and transparency between non-UV proof plexiglass which I've used for other photo projects successfully)
  • Developed board in Lye (sodium hydroxide) solution (1 tsp. lye in 16 oz water, diluted slightly more after first attempt seemed to develop too quickly). The resist looked solid after developing, like expected and similar to examples from the web.
  • Rinsed in water for a while (minute or two?)
  • Etched in hydrogen peroxide, vinegar, and salt solution (1/4 cup each peroxide and vinegar) following this guys directions (peroxide was a new 3% strength bottle, vinegar was previously opened old gallon jug, salt was just normal salt: no iodine, but ingredients are "salt, sodium silicoaluminate," whatever that is, and it was kind of old and chunky). Had solution and board in a plastic container in a warm water bath.
  • I left it in the etch for several hours (about 4). At first I tried swirling the solution every couple of minutes, adding more salt and vinegar in small amounts (maybe 2 tblsp total vinegar) to keep fizzing. I tried brushing the board to remove copper but this started to remove traces, so I didn't do that for the last few hours.

At the end I took it out thinking it didn't work at all, but when I held it up to light, I could see through around the traces, so it looks like the borders etched, but not completely. I tried using a dart point to see if the copper that should have etched would come off, but it didn't. The area that should have etched has a different appearance than the protected traces.

What could have prevented complete etching? I'm wondering if under-developing could have left resist (but then again the borders etched), or if I should spend the cash on ferric chloride etchant. Could the boards have just been bad? There are some kind of textured-looking areas (3rd pic, lower portion).

Edit - problem solved
After trying this again and agitating nearly continuously, as suggested below, I got good results. For the sake of completeness, this is what I did to get a decent result:

  • Exposed with 4-layer transparency for 20 min in direct, mid-day sunlight (no clouds)
  • Developed about 2 min in lye-based developer
  • Developed in 1:1 hydrogen peroxide and vinegar solution (same as above) with about 2 teaspoons salt. Heated vinegar first (not quite to steaming). Nearly continuously swirled/sloshed gently the dish. Brushed gently with foam brush every couple of minutes. I noticed the copper would appear tarnished if the solution was still. Brushing would completely remove this (and copper, I assume), and swirling/sloshing would also, but not as thoroughly, and some spots would remain "tarnished." Finished in about 45 minutes.

Observations:

  • After about 15-20 minutes, the solution got pretty volatile: lots of bubbles, off-gassing (looked like steam but solution was barely lukewarm)
  • After about 35 minutes, the solution got very still, tarnishing stopped. Etching was about 90% done. Tried warming in a hot-bath, didn't change much after a few minutes. Added more vinegar and salt, no significant difference after a few minutes. Added more peroxide, solution got more active (not as active as before), tarnishing resumed, and etching finished.
  • Resist was starting to come off at the end, especially with brushing.
\$\endgroup\$
4
  • 2
    \$\begingroup\$ This is why I just buy boards from circuit board aggregators. They do a much better job than I do, and it's less muss and fuss. \$\endgroup\$
    – TimWescott
    Mar 2 at 16:05
  • \$\begingroup\$ 20 minutes seems like a long time...I use a really ugly "blue" CFL bulb for UV exposure, and it takes no more than six minutes. \$\endgroup\$
    – user156429
    Mar 2 at 17:46
  • \$\begingroup\$ @TylerStone It seemed like a really long time to me, too. The first time I tried I did 10 minutes, but all the traces disappeared in the developer after briefly appearing. Any idea if over exposing would cause this etching problem? \$\endgroup\$
    – Evan
    Mar 2 at 21:54
  • \$\begingroup\$ That sounds like either overexposure or a devoloping solution that is too aggressive. I had really mixed results before using the CFL bulb for exposure and mixing the developer really precisely using a CC-graduated syringe. \$\endgroup\$
    – user156429
    Mar 2 at 23:53
1
\$\begingroup\$

I'm suspicious of your etching solution, but I don't think that's the problem in this case.

This sounds like you weren't agitating your solution enough, and the solution over the large copper areas got saturated and couldn't accept more copper. If I'm right, then the reason you got a deeper etch around the traces is because where there was resist, there was a reservoir of fresh solution.

You need to agitate your solution continuously. Like, if the motion stops, you're not doing it right. The technique that I've always used for hand-etching is to rock the etchant tray back and forth, so the etchant gently swishes across the board. You want the etchant to be moving with respect to the board, and you want it to mix, at least gently.

I'm also concerned at your reported etching time. Did you control for temperature? Maybe it's correct for that etchant, but with the ferric chloride that I've used, if you etch at too low a temperature it takes longer (if I remember correctly 30 minutes is way too long) and the etchant will undercut the resist, making the traces rough around the edges and sometimes cut all the way across.

Get an old ARRL Handbook (I know the process is in the 1976 book) and look in the construction section. You'll see a motorized etchant tray under a heat lamp. Take that as a clue.

\$\endgroup\$
2
  • \$\begingroup\$ Basically, once I found dead-bug construction I stopped even trying to etch my own boards. If you're making a bunch of boards, and if you live in a 1st-world country, it's easier to just get them made. If you're making one, it's easier to dead-bug the circuit than to design a board and etch just one. Note -- if you're living in an underdeveloped country, then yes -- etching your own boards may make sense. \$\endgroup\$
    – TimWescott
    Mar 2 at 16:16
  • \$\begingroup\$ I hadn't thought of agitation - thanks for the tip, I'll try that when I get a chance. I let it go that long just to see if something would work since it didn't look promising after 30 min or so (based on what I'd read about ferric chloride). My motivation is mostly just to learn the process, and I'm make individual guitar effects pedals, so I really only need one-off pcbs. I'll check out dead-bugging, hadn't heard of that. Thanks! \$\endgroup\$
    – Evan
    Mar 2 at 16:24

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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