From: Radochonski, Wayne
Sent: Wednesday, August 08, 2007 7:25 AM
To: Massion, Gene
Subject: "Low Power" Microcontrollers for acoustic data logging

Gene

Unfortunately, the microprocessor that I hoped would meet you requirements does not support a USB host interface, only a USB device port.  I'm checking on a few others.  

On the seismic data logger, a local SRAM buffer was used to keep from spinning the hard drives unless absolutely necessary.  The majority of power was spent turning the drive platters.  You may find that doing this kind of power management via USB just isn't possible -- it isn't on the 1 TB Maxtor external USB drive that I use for backups.

I'm curious... (You don't have to answer these questions if they're a bother.)

How long is a deployment?

Why don't you use the MARS cable to send the data back via Ethernet and store it on the beach?  (My guess is that the cable don't go near the area of interest -- bummer).

Four 24 bit ADCs channels running at 48K samples/sec generate almost 50GB of data per day.  Who is going to listen to that much noise?  (Why collect data you can't or won't analyze?)

Can you use software to detect an interesting "event" and only collect samples, when necessary?

I will have some power #s for "leading edge" micros in the next day or two but I'm certain the power for the hard drives will make the electronics look miserly.  BTW, 100 GB 1.8 volt hard drives are available from Toshiba, they use about 250 milliwatts when spinning.

Wayne