Hardware

h1. Design rules

h2. PCB facturing

|Base material                  |  	FR4 TG 140°             |

|                               |Technolam UV-Block FR-4-86 681 |

|PCB Thickness                  |  	1.6mm (0.063inches)     |

|Base copper thickness          | 	18µm                    |

|Min. processed copper thickness| 	35µm                    |

A 15cm trace with minimum width (0.15mm) has an resistance of ca. 1 Ohm. A shortcut would limit to 3 Amps.

A 2cm trace has an resistance of 0,12Ohm; 25 Amps.

| 15cm | 0.15mm | 1 Ohm    |

| 15cm | 0.2mm  | 0.7 Ohm  |

| 15cm | 0.25mm | 0.56 Ohm |

An QFN pad has an width of 0,25mm;

h2. Traces

min. drill diameter: >=0,3mm

min. trace width:    0,150mm

pad diameter DuKo:   0,6mm (spec. 0,7mm?) !!

hole to hole:        0,35mm (not specificated. Via was missing. Wert empirisch. Does not happen when following other design rules)

Distance to cuts     0,35mm (cut-through to easily by grinding)

* Power: 0,2 distance; 0,2 width

* Powerline (240V): 1,5 distance; 2,5 width (CANSwitch)

2,5mm: ca. 1.5A bei 10°C rise

!design_rules.png!

h2. Text varnish

Lack:

Gut, fuer längeren Text:

0,7x0,7; Staerke 0,1mm (NUC64Debug V0.2.x)

(zu klein fuer Schrift in Copper)

Zu klein:

0,4x0,4; Staerke: 0,05 (CAN/IO)

Geht noch, nicht kleiner:

*0,6x0,6; Staerke: 0,05* (CANDisplay::Eval; Miniminutnik)

Geht noch, eher nicht kleiner, manche Labels nicht mehr eindeutig:

0,55x0,55; Staerke: 0,15 (CAN/IO v0.6.x, einzelne Referenzen nur erratbar)

h2. Text copper

Copper:

Gut unverzinnt:

0,6x0,6; Starke 0,08mm

schlecht verzinnt:

0,7x0,7; Staerke 0,1mm (NUC64Debug)

nearly ok:

0,75x0,75; Staerke 0,15mm (Candisplay Eval)

Still dots not ok:

0,8x0,8; Staerke 0,2mm (CANIO 0.6.x)

next test:

0,85x0,85; Staerke 0,25mm (EEPROM)

h2. Standard LEDs; 2V @ 3V

120 Ohm: very bright

240: bright, ok for blinking, still to bright for permanent on

480: should be the usual case

1k Ohm: very dark. eg. ok for power LED