[Arm-netbook] Use of 2N3904

[luke.leighton]

On Sat, Nov 10, 2012 at 11:39 AM, jm <joem at martindale-electric.co.uk> wrote:
> On Fri, 2012-11-09 at 20:49 +0000, luke.leighton wrote:
>> On Thu, Nov 8, 2012 at 11:08 PM, jm <joem at martindale-electric.co.uk> wrote:
>> > Hi,
>> >
>> > I'm helping out with design checking
>> > http://rhombus-tech.net/community_ideas/kde_tablet/news/
>> >
>> > Can anyone tell me the price advantage of 2N3904 over a 2N7002 (Far East
>> > pricing).
>> >
>> > The changes as example to section with Q3:
>> > 1. replace 2N3904 with 2N7002
>> > 2. remove C160
>> > 3. remove R58
>> > 4. Change R56 to 100K
>> > 5. Change R61 to 10K
>>
>>  ahh, you're a star.  ok i take it that's for the NRESET transistor as well?
>
> Yes it does for NRESET.
> If the lines are rarely used (as is normal for reset signals),
> then they can be 33k with no adverse effects.

 ack.  err... which ones - both of the resistors?

>> > would make the circuit work equivalently and less prone
>> > to effects of touching and hand grease finding itself
>> > on the PCB and adversely affecting circuit.
>>
>>  superb.  thanks joe.  wouldn't occur to me, that :)
>>
>>  ... would you be willing to look at the audio (transconductance amp)
>> and the mosfet h-bridge one as well?
>
> To be quick with circuit + PCB analysis, the following can help you and
> me:

 ack! that works for me

> 1. Leave links to datasheets to important chips below diagram.

 oo that's a good idea - i wondered how to do that!  i could put it as
a field into the database of parts, then, hmmm, that goes into the BOM
if you ask it to... yeah, i'll do that.

>    Keep those datasheets in one directory so that once that
>    place is found, we can find all the related datasheets quickly.
>    STM32F103 for example. The inductor. S6M1, LM13700, RT8020,
> LC5730,etc etc
>
> 2. When drawing circuits, draw them as you might find them in a text
> book. When orientations and positions of components change, its hard
> work locating mistakes quickly.

 i started out doing cut/paste of app notes and other peoples'
circuits.  am veeery slowly progressing in a random direction beyond
that point :)

> 3. In complex systems like computer boards, use drawings of chips that
> most resemble physical pin layout. It helps plan final layout, and with
> debugging.

 ahh, ok.  right.  oh well.  i changed the STM32F one to be a split
part because i found an eagle jobbie with that, but yeah i get it.

> U34 for example is prone to pick up so its good to know where
> to place the mic in components and you may as plan from the schematic by
> drawing the shape of U34 and its pins accurately.

 ah.... i think...  you mean the .... unf?  where's U34?  oh
rightright, ok, yes i'd deleted one of those, so yes, the LM17000 -
yes, that came from allegro's own database, yes it's a bit shit
layout: parts that i've created (except the STM32F where i did
actually originally copy the pinouts oh well) have followed that
rule.. mostly :)

> Its very important
> that all the chips and numerous wires and positions of components are
> subtly planned from schematic so there is less work in the final PCB
> layout. Another example, the inductors for U28 pins 2 and 8 can't sit
> next to each other in case they
> interfere with each other. Knowing the pin positions accurately
> to the real chip would allow us to plan where the inductor is positioned
> in the schematic to hint at where it should be in the final layout of
> PCB.

 ok sah.