[Arm-netbook] laptop 15in pcb1 progress

Luke Kenneth Casson Leighton lkcl at lkcl.net
Tue Sep 29 10:57:49 BST 2015

On Tue, Sep 29, 2015 at 10:47 AM, joem <joem at martindale-electric.co.uk> wrote:
>>  yes i got the T862A, the one with an 800W pre-heater and a 200W
>> overhead lamp, i can actually see what's going on, keep an eye on the
>> solder as it melts and poke things with a stick if necessary.
> I don't dispute the benefits of a T862A, but get oven, destroy
> a few boards until you got it timed to working 100%, and then its a
> doddle to plant components, press go, and get working board with
> near 100% reliability.

 for this (first) board i need a slightly different approach, which
the T862A suits better.  i need to go carefully and slowly, populating
only a few components at a time, testing them, inspecting the board,
making sure there's no short-circuits (some of the components are
0402), verifying that the circuit is functional by applying power, and
correcting it there and then if necessary.

 if i were to populate the entire board with components, there would
be absolutely no way that i could test individual circuits or
components.  in some places for example resistors don't measure their
actual value because of other components on the board.

 i made one mistake by putting in the 3.3v regulator *and* the MOSFETs
for the LCD digital circuit.  the 3.3v regulator circuit was fine, but
the LCD digital 3.3v supply circuit wasn't.  so i dismantled pretty
much the entire 3.3v regulator circuit, had to cut PCB power line
tracks, and generally messed up the board, taking over 2 hours to
track down what was wrong.

 if i had put down only the 3.3v regulator and then tested that, i
wouldn't have had to remove any of its components.

 now multiply that scenario up for an entire board and it's easy to
work out why i'm doing this a bit at a time.

 so i'm using the T862A as a PCB development tool, not a PCB
*production* tool.  if i was doing small runs of PCB production, where
i was absolutely 100% confident that the circuit was proven, correct,
and functional, *then* yes absolutely i would get one of the T962


More information about the arm-netbook mailing list