Sunday, January 8, 2017

Home made PCB for a prototype board with TSSOP 28 tlv320aic and lessons learned from a simple uv led box


Printed Circuit Board (PCB) forms the foundation for many electronic projects and making a PCB at home is one of the useful skills to develop. If you think about the time and the headaches associated with home brewing a PCB, there are several cost effective solution which provides very nice boards at the cost of a long waiting time (for the postage). DirtyPCB and oshpark are some useful ones. But I decided to give a try and to learn some new things and a bit of chemistry. One of my old ugly attempt using salt water is shown is in this post on Si570 signal generator .

There are several detailed tutorials on making PCB by employing various techniques such as toner transfer, CNC router, laser engraving, using photoresist and an etchant. So this post is far from my usual programming experiments and is more of sharing some learning points and personal experience.

It was very tedious in the beginning with lots of disappointment and giving ups. But finally, it became better and better. The picture below shows several bits and pieces with various levels of etching and exposures

Failed boards placed on the uv exposure unit

The process used is based on pre-sensitised boards (Bungard) and ferric chloride etchant

UV led 

It was fun building the UV led unit. The initial plan was to use a usual 5mm led in 100 numbers but later shifted to using some high power led (3watt) and running them at a constant current of 400mA to get a stronger light from just 3 leds. ( the actually needed 700mA but I am afraid of my heat sink)

UV led pasted on a copper board (acts as heat sink)

I fitted it in a photo frame and later learned that the depth is too LOW resulting in uneven exposures

To drive the led I used a 12-volt adapter and then a two transistor based current controller (it basically cuts off if the draw is over 400mA)

After the failed result from photo frame , I kept the whole thing under a glass rack and then placed an A4 sheet of paper (dipped in olive oil and dried to make it translucent) over the glass. On top of the translucent paper, I kept the paper with the layout (again used oil to make it translucent, alternately you can print on a transparency). Used a drop of water to keep the print in close proximity to the copper board)

Results were improved by increasing the separation between the led and plate by 7-8cm

The picture below shows the prototyping of the led driver. The led were with 3.3volt forward drop and i used 3 of them in series.

SIMPLE DRIVER (MOSFET  doesn't need any heatsink in my case, and alternatively a transistor like MJE13005 also works and needs a higher bias current for the base)

LED driver for the UV exposure unit

Test patterns for PCB exposure

Final learning

  1. Keep the patience and use cut pieces of the board for testing
  2. Adjust the distance for uniform light
  3. Adequate exposure time ( a bit longer is better with low light outputs)
  4. Print side of the pattern should be close to the copper (use mirrored)
  5. Repeat all these several times! (be patient and keep sanity :) )

A few more pictures & tips

Always check the pad spacing for the components if you are doing it for the first time. I learned something here in a hard way

1) Inkjet printers are different from a laser printer in one more aspect!  YES, i choose to print the exact size as in pdf file or in eagle cad. But it appears that in my inkjet printer I need to scale the image by 106% to get the exact size. Laser printers are accurate. This may be due to the mechanical limitation in small inkjet printers which incorporate a scaling down somewhere in their drivers?

The simple solution is to set a scaling of 1.06 in eagle or in your cad software while creating a pdf or while printing.

checking the proper scaling e.s.p if you use inkjet printer 

After the UV exposure, it seems Bungard pre sensitized boards has a small extra layer. So while developing it is worth using a tissue paper folded into thick layer(hold with plastic forceps) and gently rubbing on the board to remove all of those resists in the exposed areas.
Post exposure and developing

checking the pad spacing

Pad alignment

Lessons learned with various exposures and options

Final points

It is fun to experience the challenges and make one PCB at home. It will be much cheaper to get it from services like dirty PCB esp if you want to use SMD parts and fine pitch.

making a breakout board for tssop28

A codec chip placed with a normal soldering iron and drag soldering, dirty but it works

Entire mess and successful i2c communication with codec chip for testing
There is no coding in this post. More time on trial and error and getting the PCB better and better!

The main idea was to learn things and finally it was very pleasing. Took several attempts to learn different faults as I numbered in the beginning

  • Adjust the distance for uniform light
  • Adequate exposure time ( a bit longer is better with low light outputs)
  • Print side of the pattern should be close to the copper (use mirrored)
  • Repeat all these several times
The main mistakes were uneven exposure and reduced exposure times which resulted in bad traces e.s.p in areas where the pads are too close. 

One of the final results is shown below ( again a blank run on a cut piece )

Better results after identifying the mistakes and correcting them. The lead spacing for TSSOP is  0.0256 inches (0.65mm)
This attempt also gave a quick start in using eagle cad to make a PCB layout from scratch. The ugly schematic and the usable board layout ( The final result is shown above) is attached below

Eagle quick start and board layout made out of it
The summary is: it is possible to make PCB at home for SMD use with a little bit of patience and time

Wishing you good luck and good patience in making your own PC boards


No comments:

Post a Comment