Anyone photo etch PCBs? Help!

[madlabs]

Hi All,

Decided to try my hand at photo etching PCBs. My buddy with the CNC router has been flaking and I want to get my new alt charge controller on a baord.

I am using the Datak system. The procedure is to spray the boards with photoresist (in a darkroom) then dry 'em in the oven (keeping 'em dark in a tin) then expose 'em to UV light (~ 3 mins) with the artwork on top. Then develop them, 30-40 seconds in one tray, dry, then another 10 secs in another tray.

I follow the instructions to the letter, but I am getting only the faintest of patterns. My last one I set the bulb closer and got no pattern at all, so maybe I'm over exposing. Going to try to reduce the UV on the next try.

Any ideas what it means to get no traces or almost no traces? And does the developing need to be done in the dark? My instructions don't say, but I have been assuming yes and keeping it dark.

Thanks!

Jonathan

[Wolvenar]

I just use the old trick of a  heavy laser print on glossy catalog paper..
Iron that image to the board, take the paper off under running water, then etch.

Works pretty well most of the time.

[madlabs]

I have tried toner transfer for years, the thing is that it is erratic, sometimes great and sometimes not. And the more complex and finer traces the harder it gets. While phot etching is PITA I'm hoping for more consistent boards.

But right now I'm not getting any boards.

[Bryan1]

Hi Madlabs,
                 I too use the glossy paper for my pcb boards and just use an iron to transfer the image, now the trick is only use thick glossy paper like the heavy grade glossy paper business's use for flyers. Avoid any images and ensure the copper board is scrubbed totally clean with a scourer and don't have the iron too hot. I usyally iron away until the image shows thru the paper. I have even ironed on some smd pads etc with 0.4mm traces with great success and one can also convert a laminator over to use for the transfer method too.

              I found printing a pcb image onto a transparency with my laser printer I do need to do do 2 copies and I get mixed results even then so now I just expose UV boards fully so I can then just do the iron on method now.

Regards Bryan

[oztules]

I use the Kinsten stuff... cheap from Hongkong ebay.
I use the 15w desk flourescent.
I use a single transparency.
20 min exposure
then NaOH and then ferrric chloride or ammonium persulhpate.

perfect results every time... very fine detail if you want.

I don't like spraying my own resist boards from copper, as it never covers perfectly evenly, and the results are flaky.

Iron on sounds ok, but the kinston is the easiest to get perfect results every time.

Also note from the last PCB in the CDI story, I try to use up all the copper space I can.... saves on etchant. I used to make spider web looking boards.... now as chunky as I can.


....................oztules

[rossw]

Any ideas what it means to get no traces or almost no traces?

Last time I did a "prepare it yourself" photo board, it was available as either positive or negative. Which have you got?

The most common commercial boards are negative photo resist - meaning you print black everywhere you DON'T want a track.
The UV getting to the resist hardens it, the developer removes all the un-hardened resist.

So when you say you can't see any sort of traces - is the resist comming off?? Or it is all going hard?
If it's all comming off (suggesting massive under-exposure), are you sure your light source has enough UV?

Another old trick is to use a stepped greyscale from completely transparent to completely opaque. Run a test exposure on a small strip of board. Check/develop. Use that to see where your exposure is to get optimum results.

Good luck!

[madlabs]

I'm using a negative.

I am making some progress. First thing is although I can barely see the pattern, it does resist. I got a usable board done today, a simple relay board for a pump. I think the opaque areas aren't opaque enough and thus I have to etch longer and eat in to the traces. Im getting small holes and imperfections of the traces. I need to experiment with developing anf exposure times. Looks like it might take some practice, but I hope to get it down.

The chemicals are nasty though.

Thanks for the help folks!

Jonathan

[frackers]

I started with spray on resist about 40 years ago but had to always touch up with a resist pen and moved onto toner transfer about 20 years ago.

That worked well but was slow as you do have to be careful about iron temperature, paper type and soak off time.

Nowadays I'm spoilt because I can buy the precoated Kinsten board and use the etching facilities at work - UV exposure box, heated and agitated tanks but still my own Dremel in a stand for drilling.

On drilling, I'm gradually moving to surface mount but the nice thing about leaded components is that they make the layout easier (can jump over 2 tracks with a resistor!!) and I still prefer to knock out a few single sided than try one double sided!!

[WooferHound]

Another old trick is to use a stepped greyscale from completely transparent to completely opaque. Run a test exposure on a small strip of board. Check/develop. Use that to see where your exposure is to get optimum results.

I had  a photo darkroom set up in my bathroom for 10 years and worked in darkrooms all my life. We used a trick to determine exposure times and development times. You just need an opaque cover, like a piece of cardboard.

Setup to make an exposure, but cover the PCBoard with the cardboard. Begin to make your exposure and uncover an edge, every 20 seconds uncover another area. When you are finished you will have 10 or 15 exposure times on the same PCBoard. Develop normally and you can choose the timing that makes the best result.

I looked at the Kensten website and found this information . . .
--------------------------
Overhead transparency sheets usually produce very poor results, as they cannot aggregate the ink or toner density enough to produce high contrast. Some materials are not UV transparent, and are also unsuitable for using as PCB artwork. Using a suitable base media with an inkjet or laser printer, good results can be obtained.
--------------------------

[Janne]

I used to do a lot of DIY PCB's with the photo-resist method, but lately I've been getting lazy and just ordered boards ready made. For hobby projects I've been using ITeadstudio(google it if interested), because they're cheap and do adequate job, especially for the price.. 10x of these boards cost me 25$(+shipping). Of course, if you're in a hurry it's bad since china post takes ages to deliver. These boards have the smallest traces 8 mils and smallest spacing also 8 mils.. no problems in the boards.





With the photo-etch method I initially had lots of trouble. What woofer suggested regarding doing a test PCB is a good idea to get the UV exposure right. Once you find a good setting, write the numbers up. Not much good if you've forgotten the good setting next time you do it. Also document the developing solution.. Too little or too much NaOH and it's not going to work well. Also, in time the solution will concentrate, if water is let to evaporate (been bitten by that a few times as well..) Lastly, stick to the same manufacturer of PC board, I've found that the optimum settings differ quite a lot from one PCB manufacturer to the next.

Most of the time nowdays I tend to build quick one-off projects on perf board, for the more layout critical stuff / SMD boards just order the boards professionally made.. Cost is about the same for having them made for me to buy the raw materials anyways =)

edit. I always used the pre-sensitized boards, I found the spray coating of UV sensitive layer way too error prone, at least for me. Also, I remember a good advice in mixing the developer.. Add NaOH to water until the solution feels "slippery" to the fingers

[madlabs]

Was away for my 48hr shift. Back at home and at work!

So, my buddy with a CNC router offered  to lend it to me to play with. So instead of zeroing in on the photo etch, I may switch over to the  CNC. However, there is a steep learning curve there too, and some hardware and software to buy. Anyone here make PCB's with a CNC?

Jonathan

[Bryan1]

G'day Jonathan,
                               I've made my own cnc and I use Kcam for the interface software but I will be switching over to EMC2 in the near future. Now one has to first use a pcb based software to design the pcb image then export a 'gerber file' for the CNC interface then for the drilling export the 'excelleron drill file'.

                             Personally I use Sprint Layout for the pcb design and it does have the export function. To date I haven't milled a pcb board but I have done the simulations in KCam and everything looked OK. Now for the cutter bits I just use my broken carbide drill shanks then grind up a D-bit to different widths. Also one will need 20,000 + rpm for using a D-bit tool.

       Regards Bryan