Lpc2000

I thought of that, Tom (separate supplies), but I realized that I 
wouldn't get anything from it, because there is this single device 
that goes from <5 mA to 1.6A, and that is the radiomodem. I can't say 
when the RM will draw <5 mA and when will it draw 1.6A. This is for 
the GSM protocol to decide. Therefore, whichever power supply I'd set 
up for the radiomodem must be efficient for pretty much the same 
range as one supply that would feed the rest of the circuit as well 
(which by the way won't take more than 80-90 mA ever).

It is tempting to have a power supply specially tailored for our 
needs. However my schedule has become incredibly short, with the 
additional difficulty that we are based in Argentina, so it is hard 
to arrange such things. However, feel free to lend me your 
acquaintance's e-mail address through a private message.

Guille

--- In lpc2000@yahoogroups.com, Tom Walsh <tom@o...> wrote:
>
> Guillermo Prandi wrote:
> 
> >Hi! I am in trouble trying to find a voltage regulator suitable 
for my 
> >project; my project features an LPC2138, a GPS and a radiomodem. 
When 
> >idle, it draws around 20-25 mA from the 3.3V source, 500 mA when 
> >working hard and with short (50 mS) peaks of up to 1.8A. The 
supply is 
> >a lead-acid 12V battery (can't change that), and occasionally, 
when the 
> >battery is charging, the input voltage will rise up to ~20V. The 
> >problem is the regulator efficiency. We were going for the LM2676-
3.3V 
> >switching regulator, because it seemed simple and in range, but 
two 
> >things made it a poor choice: it draws about 4.3 mA from the 
battery 
> >just for its own supply, and its efficiency when the load is in 
the 20-
> >25 mA range is awful (less than 50%, not counting Iquiescent). I 
looked 
> >everywhere but I couldn't find any other switcher with better 
> >performance in these conditions (12V in, 3.3V out, 20-25 mA 
average 
> >load, but >1.5A capability). And there's a catch! The radiomodem 
is 
> >extremely sensitive to VCC getting lower than 3.3V especially in 
the 
> >current peaks! Also, it should be relatively easy to find in ~400 
> >quantities (i.e., Digikey, Mouser or your suggestion).
> >
> >  
> >
> Keep in mind that most off-the-shelf supplies are intended for 
general 
> purpose use / short run / prototypes.  Anyone doing significant 
> production quantities is not willing to accept the built-in premium 
cost 
> for the convenience of not designing your own.  I stopped looking 
at 
> off-the-shelf supplies for my boards / systems a long time ago.
> 
> My reaction from what you were saying was that you need multiple 
> supplies rather than a single supply.  Different regulator types 
for 
> each current load profile.  Most switching supplies require a 
minimum 
> load on them otherwise they can go into conduction (switching gate 
left 
> on), you hear a term called "discontinuous mode", that helps 
prevent 
> stalling.
> 
> What it sound like to me is that you need a modest supply to run 
the 
> processors circuitry, then a larger supply to drive the heavier 
loads.  
> The larger supply could be one that can be shutdown, thereby 
reducing 
> standby current (minimum load to avoid stalling).
> 
> FWIW, 400 units, and assuming a $10 premium (wholesaler profit), 
that 
> comes out to $4000.  I think that if you approach a good power 
suppy 
> consultant, he would be more than happy to take $4000 to design you 
a 
> supply which could only cost you one half of what an off-the-shelf 
model 
> would be.
> 
> I know of one in my area (Pennsylvania) and deal with him from time-
time...
> 
> Regards,
> 
> TomW
> 
> -- 
> Tom Walsh - WN3L - Embedded Systems Consultant
> http://openhardware.net, http://cyberiansoftware.com
> "Windows? No thanks, I have work to do..."
> ----------------------------------------------------
>