--> ap/xxxxx

* __

Second set of scrying board prototypes from: (2008.03.11:2)


Working succesfully.

Sidenote: Now resolved on 6min50s for Artoz pastel foggy paper and SMD components (6min30 for general purpose).

First scry prototype back from batchpcb: (2007.10.18:1 pcb#15 scrying_tech_notes#33)

A few (of my) mistakes.

KiCad: (2007.09.24:3 pcb#14 schematics#11 tech_notes2#12)

1] Plot -> PDF route sometimes chops board - simply reposition on grid

2] Remember to mark tracks/pads which should be attached to the (GROUND) fill before filling the zone

Sc/krying boards PCB designs: (2007.09.20:2 scrying_tech_notes#31 pcb#13)




Schematics to follow shortly...

Export Eagle -> KiCad (2007.09.20:1 scrying_tech_notes#30 pcb#12)

Unknown library with CR2450_KEYSTONE battery holder part for CR2450 battery. Exported from design at:


1] In Eagle, Open Board then - File->Run->exp-project-lbr.ulp

2] Exports board as .lbr file and opens library editor

3] From here File->Run->exp-kicad-lib.ulp to convert to KiCad .mod

4] This last script is available from:


KiCad drilling (2007.09.19:2 scrying_tech_notes#29 pcb#11)

file misaligned:

1] Plot settings - Gerber, plot origine: absolute, spot min: 0.015, Lines Width: 0.0010, all other settings though greyed out must be as defaults

2] Drill export settings - suppress leading zeroes, 2:4, drill origine: absolute, via drill 0.025, drill sheet: none

3] Check alignment using gerbv: http://gerbv.sourceforge.net/

free Gerber view package (KiCad's gerbview does not yet support examination of .drl drill files)

4] Check also by way of:


For SMD PCBs overhead folie Folie 3562... (2007.07.04:1)

is preferred

Timing from 2min30s to 2min45s. Rest of process as usual. Reduced scrying test board also adjusted for more distant traces.

increase track size, clearances for home PCB/SMD: (2007.06.22:1)

1] Tracks too thin and also etch poorly (7 minutes maximum UV)

2] Already increased connector pad size to 1.8mm

expose 7 minutes for fine-traces/SMD/T (2007.05.23:1)

PCB photo etch measurements/guide: (2007.05.10:1 pcb#7 tech_notes2#6)

1] Using foggy/tracing paper (Artoz Perga Pastel) gives better results than Folie 3562 Overhead transparency when printed with laser printer (600 dpi) though latter can be used if fill in gaps. (SMD/T fine with six and a half minutes below).

2] Warm-up UV lights (see pcb below) for two minutes. Switch off briefly and line up with ink facing PCB (in dim daylight is fine). Expose tracing paper for five minutes/ overhead transparency for two and a half to three minutes (heavy books on scanner/UV box lid). Develop for three to five in the Natriumhydroxid (Sodium Hydroxide I think (10g in 1000ml)). Wash in cold water. (above in dull light).

3] Now etching in Natriumpersulphat (Sodium Persulphate) which is faster, cleaner and more accurate. Dissolve 120g in 600ml water. Heat up in bath to 40 degrees or more. Etch board for around 15 minutes until copper vanishes - agitating as much as possible. Wash and dry boards and save solvent removal of UV paint until just before soldering.

KiCad/GIMP printing: (2007.05.09:1 pcb#6 tech_notes2#5 schematics#7)

1] When we export by way of plot option and then ps2pdf and then into GIMP the image is too small. We need to scale X and Y in the plot option to 1.08.

2] To test boards with three options:

a) latest foggy transparency paper

b) clear OHP film

c) foggy as diffuser and then with clear OHP film

... for SMT, double-sided, new VLF design, spectral boards (inc. TV transmitter)

Printing for PCB machen - foto resist using the Gimp: (2007.04.23:2 pcb#5 tech_notes2#3)

1) Import the PDF into the Gimp (open and set resolution to 600). Manipulate and copy as required as a layer.

2) Print settings: Set Image to line and output to Black and White

Etching PCBs using photo-resist method: (2007.04.23:1 pcb#4 tech_notes2#1)

Light documentation from workshop 21.04.07: workshop

1) UV apparatus constructed using gutted scanner with three 8W UV lights/lampe (Ersatz-UV-Roehren) from reichelt.de, three starters (from junked fluorescent tubes) and three 8W transformers from ebay.de.

useful link (see also below):


2) Design laser photocopied from software-created original in this case onto diffuse transparent paper (milky). Completely transparent OHP material to be tested: (Avery Folie 3562 Reichelt again)

3) Design is placed with ink in contact with photo-sensitive (fotopositiv) PCB material from Conrad. Plastic protection is first peeled off - all of this process in near darkness. PCB is exposed for three minutes to UV light using a darkroom timer to time this.

4) Pre-mixed developer/entwickler (Natriumhydroxid from Conrad - 10g to 1 litre of 20 degree C water. Development time varies from three to five minutes as developer ages.

5) After development rinse in cold water and develop in etch solution (see: simple_pcb_guide for this part of the process.

6) Scrub off the yellow photo-resist with water and a brillo-pad and clean a bit with alcohol. The results:

photoresist also: (2007.01.16:2 tech_notes#11 pcb#3)


Printing for PCB machen from PCBNEW/KiCad: (2007.01.16:1 tech_notes#10 pcb#2)

1] Straight print from PCBNEW menu under a range of options is unusable (tracks joined, offset and pads connected randomly).

2] Choose plot with Postscript option from file menu and write (in this case) xxxxx-Component.ps. Printing from this using gv is better but unusable also.

3] Run ps2pdf on this .ps and open resulting .pdf with xpdf. Print from this.

(further notes: print command for our CUPS shared printer is lpr -P printer, resolution was set by way of web interface to 600 dpi)

next stage exposure (UV light) and photoresist:


Using gschem (as part of geda) part one (2007.01.08:3 pcb#1 tech_notes#335 schematics#1)

to edit a schematic and generate a netlist before we make a PCB design in pcb (all GNU free software) - pcb is installed under Gentoo with seperate emerge command.

gschem (also gattrib for editing attributes of .sch files)


and the warmup:


perhaps try CVS version

Progress so far:

What are the differences here between Default - reference component, embed component and component as individual objects (latter option seperates out and can thus be edited)?

Use reference I think...

ATmega8 component lacks VCC and GND and pin 1 is in an awkward place. First two remedied by modal operation/keystroke driven approach ... where we want pin ap keys and pull line with left mouse into body of chip. middle mouse to stop behaviour. an to add net/wire with similar results. select object with left mouse squaring over and add aa to attach/add attribute such as a pin number. ea to edit attributes of selected piece/element.

er to rotater elements, fs to save, z and Z to zoom, w to zoom next selected square, ve to view extents (?)

But how do we edit the ATmega8 stuff as it stands to switch pin 1?

(only problem is crashes on update)

see Wiki above:

Update Component (ep)

warmup above is a good place to start for simple operations.

note also:

it only makes sense to create pins while creating or editing symbol files...

note also url for gattrib: http://www.brorson.com/gEDA/gattrib/