[Re: thermal tips re expo converters:]

[Harry Bissell]

My rule of thumb would be small is about .010-.015 (10-15 mil) medium is about
.020-.025 (20-25 mil) and large would be .040-.050 (40-50 mil) any larger than
that is xxl. I wouldn't try to make a trace using home etching  tecqnuques any
smaller that the .020-.025 range. It will still be better than .050. Also, use
1oz copper, you will get less undercut than with 2oz. An aquarium pump makes a
great "bubble" etcher to keep the solution moving evenly. Or... you might try
putting pads on the board and running fine magnet (insulated with varnish)
wire to span the gap. Good luck. These thermal techniques are a must in
multilayer board design, if too much heat can be absorbed by the inner layers,
the board can't be soldered. So, tiny pads are used with etched-out "lakes"
around them, and tiny foil bridges jump the lakes to prevent heat losses. 
Good Luck Harry ;-)

owner-synth-diy at mailhost.bpa.nl wrote:
How small is small for the traces? Since I make my own PCBs from scratch, I
usually have somewhat large traces.
They don't exactly look like those micro-miniature commercial PCBs. I also tin
all my traces, so I guess I
wouldn't do this on the 3046 traces. Right?

SV


Harry Bissell wrote:

> Been thinking about the heated expo converters... here are some thoughts to
> share:
> 1) RE: How chips lose heat see National Semi LM-34/35 precision temperature
> sensors... The plastic package is actually a thermal insulator, most of the
> heat reaches the chip through the leads. This has been documented in some
app
> notes for this part. So, when you think about putting thermal insulation
> (styrofoam) over the chip, also consider it Under the chip.
> 2) Following the same logic... rout a slot in the PCB around the chip, and
> leave just enough room for the runs... This will make heat entering or
leaving
> the chip radiate through the air, which is a hell of a lot less effective
that
> conduction through the PCB.
> 3) And... Use tiny traces in foil, just enough to carry the current you
need.
> If you need more current (Gnd, power) use two tiny traces in parallel. This
> will conduct heat less heat than one trace of 2X the size. Again, you want
to
> be less efficient in heat transfer.
> 4)All of the above would give you the smallest thermal mass, (hence faster
> thermal response. This is a good thing for a feedback loop. It might be hard
> to layout the PCB like this, but probably worth the effort.
> 5) a last thought... When you glue styrofoam around the chip try not to
touch
> the circuit with the glue. Make the insulated volume as small as you can.
The
> glue, or any moisture that could possibly collect (when cold) might cause
> leakage. Digi-Key carries some little plastic dip plugs, that have some
little
> plastic header covers available. These would be perfect for covering a
single
> (heated) chip. Look for em. (sorry no catalog at home... but they are in
> there)
> Whaddya think?  ;-) Harry Bissell
>
> ____________________________________________________________________
> More than just email--Get your FREE Netscape WebMail account today at
http://home.netscape.com/netcenter/mail




____________________________________________________________________
More than just email--Get your FREE Netscape WebMail account today at http://home.netscape.com/netcenter/mail