Home PCB Production

Comments from the Nutters

Some friends comment on making printed circuits at home.

Jenny:

This is all about Printed Circuit Board production at home, so those not interested, leave now!

Bernie has been making PCBs for the both of us using the following method:

EasyPc to produce schematic, library symbols, and using the derived netlist, the final PCB. I am getting better at this process. My standard PCB has been double sided, one side mostly earth plane..

Use an Inkjet and transparencies to produce artwork.

Expose pre-coated board to UV

Develop and Etch using developing trays.

My most recent PCBs have been more challenging. 10 thou tracks, tracks between 1206 resistors, double sided SM. The yield is not good.

Bernie and I are considering investing in a 3 tank etching unit. Develop (temp controlled ), wash and bubble etch.

So the question is....

Will 400 for this improve our yield ? We just don't know if this is good value. We are looking to Tris and Nigel for opinions, or anyone else who is interested.

J2 will be after at least 4 more prototypes over the next months. I know that Bernie is doing something similar. Is anyone else interested in the 'Nutters' prototyping service ? I plan to kidnap Bernie's light box and have a go myself (if Jennifer will let me boil etching fluid in her kitchen).

Professional prototyping should cost about 150 for a panel. This is another option.

Thoughts ?

...and here is a URL:

http://www.megauk.com/

 

Dave:

What's your main problem with getting a good yield? 
I notice 'How to make really really good homemade PCBs' says that the main problem with inkjets is getting a really good, solid black at a high res.

Have you looked at the Epson 2100p or the SP950 inkjet printer? The 2100p is an A3 inkjet with the ability to take thick, flat media - it can be used to print directly onto a CD.

Would it be possible to print directly onto the PCB? 
The print res is higher than a laser - 2880 x 1440. 
Also, it has two types of black ink cartridge in it (photo black or matt black), ensuring a more solid black coverage. 

http://www.photo-i.co.uk/Reviews/printers/Epson_2100p.htm 

Just a wild thought, 

 

Tris:

A few questions spring to mind.

At the UV exposure stage it is very important not to over expose the board. Even a little or thin tracks may lift while etching. 

Do you know that the laser printer/inkjet ink is sufficiently UV opaque? How old is the resist coated board stock?

Buying an expensive processing unit will, of course, improve yield only if you satisfy the few conditions above.
Are you heating the ferric chloride to about 50 degrees and stirring the solution regularly while etching?

Rapid Electronics have a triple tank unit for about 250 (I think).
Is the 400 unit the one in RS/Farnell? Rapid also offer photo-resist coated copper-clad board at good prices. They have CEM1 for non critical stuff and FR4.

Prepreg and Kapton based copper-clad flexi is available.

How about just a simple tank like a fish tank, with a fish tank heater to get the etchant up to temp? That's seems to work very well (take the fish out before adding the etchant).

 

Nigel:

I too have some questions:

  • Where do you think your process is deficient?
  • What process are you using, Positive or Negative, contact or not?
  • What chemicals are you using?

I built my system some time ago from odd bits, fish tank pumps and the like. I can only tell you of the results of my experiments.

It is clear that heating and bubbling the etch improves the consistency of the results. Heating speeds up the reaction. Bubbling ensures that the concentration of the active ingredients in the etch is uniform. Without bubbling, large areas of exposed copper result in local dilution of the acid in contact with the PCB, this leads to uneven etch times, with some sections being over-etched.

If you are using pre-coated resist boards, the age of the coating is important. Old boards take longer to develop and may need stronger developer concentrations. Again it is more important that the pre-coated boards are stored in suitable environment. The colder, darker and dryer the better. I store in the garage, in a black plastic sack.

Changes in storage temperature may result in micro cracks in the resist. Heat acts the same as light in hardening the resist. I use positive resist because the process is simple. The professionals use a negative resist process which means that tracks tend to get thicker rather than be undercut during the etch process.

I have had good results using pre coated overhead film in an Epson Stylus photo printer. This printer has software that allows very dark over printing. I had intended to see if any of the other colours were opaque to UV light. The transmission of visible light is not a good indicator of performance at UV wavelengths, where the photo reaction takes place. The reason for looking at other colours is that some ink colours flow better than others and fill in the gaps between the ink dots. I now have a Cannon S800 printer but have not tried it in this application yet.

In this application contrast ratio is key to good results, hence the uses of overhead film. Note, bending the film after printing can lead to micro cracks in the ink tracks, resulting in breaks in the final PCB track. It is also very important to keep the process clean, avoiding grease, dust and hair on the master film and unexposed PCB. I use an air extraction system in the exposure box, and wear a lint free hood and gloves when performing the photographic bits of the process. 

To obtain good contact between the master film and the PCB during exposure, I print the film so that it is the ink side of the film in contact with the resist of the PCB at the point of exposure. The film is held in place by a thin sheet of flexible polythene plastic. The plastic is held down by a vacuum generated by a fan mounted under the exposure table in a sealed box. The table being perforated for this purpose. This avoids the need to use a glass sheet to hold the film in place. Glass, other than quartz, being a poor transmitter of UV light. 

To obtain double sided boards, I buy double sided resist coated copper board. I cut the board a bit bigger than is required. I expose one side, develop and drill reference holes. I then remove the plastic cover from the unexposed side, and use the reference holes to align the second side film. A red light in my exposure box allows me to align the holes with the film. This work well, but I am sure this is not the best method. I would like to expose both sides at the same time, and leave the drilling till later. I think this would require the film for both sides to be mounted in some form of jig. The jig might need to be constructed of quartz glass sheets to act as support for the film, and allow per alignment. 

I use Potassium Hydroxide solution as the developer for the resist. It is important that this solution is kept out of contact with the air as much as possible as it reacts with CO2. So I use a sealed tank and agitate the developer by placing the tank on a mechanical rocker table, driven by an electric motor. Again this ensures no local dilution of the hydroxide.

I use a home brew of acid etch, made up of a mix of Ferric Chloride, Hydrochloric acid, and a few drops liquid detergent in pure water. The detergent lowers the surface tension of the acid solution allowing it to wet the PCB. (Warning, do not add too much detergent or you will have a foaming mass of liquid, bubbling out of the tank, eating all in its path. A Calcium Hydroxide solution is handy for such events)

I make up my own Ferric Chloride solution. Which I filter before adding to the etch mixture. The Hydrochloric acid stops the solution from going turbid with solid copper salts. The solution starts clear brown and turns a clear green in colour. The reason for all this is to stop solids from being deposited on the PCB blocking the etch action. I use distilled water to make up the etch. This avoids putting lime scale and other salts in the tank. Lime changes the PH, and over time leads to floating solids in the solution. 

My acid tank is heated by a 100W light bulb, mounted in a black tin box on which the plastic acid tank stands. A diode contained in a length of sealed fish tank air tube measures the acid temperature. Low and High voltages circuits being isolated by opto and transformer circuits.

I never liked the idea that the 240 volt heater should be immersed in the acid. A fish tank pump provides the bubbles.

My PCBs float horizontal in the tank, but vertical is better.

I hope this helps. 

I tin plate using a home brew electrolysis tin solution. The only problems with this are that once mixed the chemicals have a very limited life, are expensive and very poisonous.

PVC gloves, and face mask are required at all times. 

 

Bernie:

We got the triple tank from Mega today (they are actually local - at Linton, next to Eric Johnsons).

The bubble etch has made a big improvement to etch consistency across the board judging by Jenny's trial run today/tonight. This is always where I thought my production deficiency was, given that I have been producing quite good boards consistently over the last 6 months.

Apparently one of the coated boards supplied by Farnell and RS are processed at Linton. They do not supply better dielectric board than FR4, but they will coat any that you obtain yourself (Rogers do a 4001 [I think it is] - its a material that works like FR4 but is still good to 20GHz. It's only marginally more lossy than PTFE, so that is worth knowing.)

We do need to keep cleanliness and safety at the top of the list, I agree about that. l thought the triple tank is much more sanitary. It does end up exiting much more developer and etchant per run than old method.

We ought to sort out a neutraliser for the ferric chloride - any suggestion?

 

Jenny:

The bubble etch tank / developer tank seems to make the etch more reliable. We made a couple of perfectly aligned double sided boards by enclosing the PCB inside an 'envelope' of transparency, held in one place with double sided sticky tape.

 

Nigel:

I would like to see the new tanks, but I would get lost on the way. You will have to wait until I get a new car with GPS.

Re my last email: Ferric Chloride can be neutralised with calcium or sodium hydroxide. I use bicarbonate of soda, it froths a bit, but you
know when the reaction is done, when the fizzing stops. The reaction give table salt and ferric hydroxide which quickly turns to
ferric oxide (rust) as it dries. Much less toxic than the ferric chloride. You could use garden lime or washing soda, but this can burn
the skin almost as bad as the ferric chloride. 

If you are using sodium or potassium hydroxide as the developer, then just mix the two. Warning the reaction will get very hot.
That is why you should wash the board between developer and etch.

Warning, Do not neutralise in your tanks or you will have a very solid mess to clean out.

Jenny's method of making double sided boards works, but I have found that the film tends to slip out of alignment some of the time. This tends to be the case if the PCB is thick.

Please, Please wear some protective clothing. A face guard is vital. I have a number of full face guards, some designed for my more dangerous experiments. RS can supply.

A single drop of developer splashed in the eye WILL PERMANENTLY BLIND YOU. 

The Gloves need to be PVC, not rubber. The last thing you want is strong acid or alkali inside your glove, skin burns are not very nice. The gloves also help to keep the board free of skin grease. 

I am afraid that someone at work has borrowed one of my books (Electroplating and electroforming for artists and craftsmen) and failed
to return it. I have reordered it, but it is currently out of print.

If you see a copy, I strongly suggest you buy it.