Locating Voltage Regulator

[RJDog]

I am designing a board which accepts a variety of daughter boards, but I just figured out the voltage regulator is underneath where some larger daughter boards would be; being that I expect up to 1.5A draw from the daughter board, the heat sink on the 5V linear voltage regulator is not small and therefore in the way.  I figured if I move a couple things around I can make room for the voltage regulator elsewhere on the board to be totally out of the way of any sized daughter board, but that would now make the Vcc traces to the daughter board connector about 8 inches long.  Is this a problem?  Do I need to make the Vcc supply traces larger to compensate for voltage drop?  Does it really matter where on the board the regulator lives?

[jeroen79]

Short answer: It depends.

What will the voltage drop be at 1.5A?
How thick are the traces? What is their resistance?
What range of supply voltage is acceptable for your daughterboard?
How would a drop in voltage effect it's functioning?

[RJDog]

Thanks for the quick responses!

I have capacitive decoupling throught the board, but I've never done inductive filtering... is there a general guideline for this?

The PCB will be made with 1oz./sq.in copper, and I previously had the Vcc supply lines using 50mil traces, but that was when they were ~1in. long.  I've seen a couple of trace width calculators but was a little confused by the temperature rise... what is acceptable?

The brick that I plan on using is 9VDC, so ~4V drop.  On the line of thinking of a DC-DC converter, would something like this be worth looking at?

http://power.murata.com/data/power/oki-78sr.pdf

It is a little more expensive than a linear regulator, but definitely fits within my space restrictions for the original location of the linear regulator.  Is the 5V output reliable enough to not need another linear regulator after it?  The daughterboard and other components will accept 4.9-5.1V... not a huge tolerance but enough I suppose.

[RJDog]

These responses are much appreciated, evb149.

I'm starting to think though based on your responses that locating the voltage regulator ~9 inch trace distance away from the load is not a great idea; calculators say i'd need around 85~90mil trace... not that the trace width is the end of the world I suppose, but still, it kind of seems like a pain in the butt.  The switching DC-DC converter I found sounds like the easiest way to go, but your observation about the tolerance being wider than the regular linear voltage regulator tolerance is noteworthy.  One of the IC's I am particularly concerned about is spec'd for 4.9V on the low side, and I know for sure it is unstable at 4.7V... not sure about 4.8V.  The other IC's I might possibly be worried about are 74HCT series, which are fine down to 4.5V so no worries there.

I'm wondering at this point if it's just as well to say "sorry guys, larger daughter boards don't fit". 

[RJDog]

Or, here's a question; the DC jack is panel mount, so I wasn't worried about where the voltage regulator was in relation to the DC jack, as the wires are just free-air anyway, so where it attaches to the board could be anywhere.

Could I do the same to get from the linear regulator to the high-current-draw daughterboard connector; that is, a free-air wire of some gauge?  Is there a convention for this sort of thing?  Is it a no-no, or am I in the exact same situation with the 85mil trace?  Everything else other than the daughterboard connector could be done on the PCB of course; we're just talking a few mA for ATtiny and some 74HCT.

[RJDog]

http://www.semicon.sanken-ele.co.jp/sk_content/mpm80_ds_en.pdf

Spec'd 4.9V to 5.1V, same as LM7805.  It's a little more expensive than the above chip (CAD$7 on DigiKey), but we might have a winner here...

[RJDog]

I just watched a video where 8-bit guy tore down a Roland CM-300... inside, it looked like its voltage regulator itself was off-board, connected with some short wire leads (maybe 2" long) to the PCB and was mounted to the metal chassis with the chassis acting as the heat sink for the voltage regulator.  Genius!   This would work perfectly in my situation; the leads to the voltage regulator could be very short (1-2" long), and I just mount the regulator to the bottom of the chassis.

Now I just have to think of how to make or procure a metal chassis...