DC to DC Converter

Wow.  Thank you guys.

Okay, I was just going at this project laissez-faire, but with all this help and to be fair to you then I really should spec out the requirements and at least the high-level use cases.  That is not something I have done.  I've got a big project now at work and it may be several days before I can even find some free time at home to do this.  But let's keep the dialog going, as knowing what is reasonably possible allows me to write requirements that don't have to be revised for workarounds.  Anyway, let me know if I'm on track here...

• Solar Panel (will most likely need to increase the capacity).
• Minty Boost to charge LiPo batteries.
• LiPo batteries
• Battery Eliminator Circuit to prevent over discharging the LiPo batteries.
• L2596 module to control the voltage to the Circuits 1 & 2.
• Circuit 1 - PiPower to power the RPi (later wireless something).
• Circuit 2 - Other peripherals (at least the web cam  and probably also some IR LEDs for night).

Perhaps we need to consider a method that will allow the entire system to restart in the case where the batteries have discharged to the point of shutdown but then after time the solar has recharged the batteries enough to support operation again.  (I'll put this in the requirements.)

Roger, yes, I'll take you up on the beta PiPower.  But I'm thinking I should wait until we have the full strategy defined and parts identified before I start ordering.  What do you think?

Thanks again guys.  Aside from me needing to properly define the project and write the requirements, what are the next steps?

Regards,

Andrew

High level goals of the project -> strategy (what components). -> box diagram.

What are you planning for "power strategy"? Are you going to be able to conserve power by shutting down the 'pi for some lenghts of time?

Farnell's Q1 financial results are out (covering Feb-Apr)

http://www.premierfarnell.com/content/results-first-quarter-financial-year-ending-03-february-2013

RPi is mentioned under "Strategic Highlights" as bringing in new customers, but no mention of

bringing in profits.

• "Continued progress as a digital enterprise, with 2.7m visits in the quarter to our element14 community as it played an important role in the global launch of Raspberry Pi which will add in excess of 60,000 new customers for the second quarter."

Roger Wolff wrote: What are you planning for "power strategy"? Are you going to be able to conserve power by shutting down the 'pi for some lenghts of time?

Power is certainly something that will need to be considered.  The objective is to create yet a another motion monitor and recorder.  Rather than using a PIR, I was hoping that with the low power consumption of the RPi that I could use a web cam.  (The web cam and appropriate software allows for a much finer level of motion filtering than a PIR.)  Ultimately I want the unit networked with other similar nearby units and I want the 'primary' unit to be able to either receive calls or make cellular calls to transmit the images (or maybe with a cheaper cell phone at least indicate that images have been recorded).  However, I'm finding that I may indeed need to go with a PIR strategy that would let me turn off the RPi for extended periods.

Thought:  I could use 1 to 4 PIRs to wirelessly trigger the RPi to turn on.  That would achieve the level of control I need over what gets recorded and is more likely to only capture warm bodies (which is good).  Having somewhere around a 100 yard trigger range would be very good.  But this brings another element into the project.  I don't even know if such a wireless trigger is available at any reasonable price.  My goal is to have the cost of any individual unit around $100 (sans cell phone).

Thanks again,

Andrew

Another thought (option to above):  Instead of using a radio signal to remotely trigger the RPi to turn on and record, perhaps invisible light could be used.  It would limit the range and force line of sight, but it might keep the costs way down (and require less power).  But it might also eliminate problems with cross-signals between the RPi units.  (With radio, there would need to be a unique signal so only the designated RPi is turned on.  This is another complexity.  With invisible light, this issue is eliminated.)

Regards,

Andrew

This is only tangentially relevant to the question, but it's on topic for "DC to DC Converter" --- http://nathan.chantrell.net/20120430/the-magic-of-the-boost-converter/#more-2675 .

The Pi's full current requirements are somewhat on the heavy side for the capabilities of the Maxim MAX757, but other than that, it's well suited to being driven by a 3.7V Lithium-Polymer cell such as is charged by my Olimexino-STM32 --- http://www.olimex.com/dev/olimexino-stm32.html .

I could in principle charge the Li-Po from the Olimexino's mini-USB input or from the barrel connector (which accepts 9-30V and hence would be great in-car), and the Li-Po cell could drive the boost converter to power the Pi for mobile use.

As long as one isn't too demanding on HDMI and USB power requirements, this should work, and in effect gives the Pi a 5-30V input power range and mobile capability, while at the same time massively expanding its I/O potential through the Cortex-M3 board (which costs only 17 quid).

Morgaine.

I have the impression that boost is a bit more tricky than step-down. On the other hand charging a dual-cell is slightly more tricky than a single cell.

A model-A should be somewhat more appropriate for this. You're not going to have wired-ethernet in the field. That means the lan chip won't be powered. And the lan chip is a power-hog as it uses mostly 3.3V which is inefficiently generated from the 5V.

Which leads to the following suggestion: When battery-powering your 'pi, you could generate 3.35V using a buck converter from whatever your powersource is and deliver that to the 3.3V pin on the GPIO.

That's a good point about the Model A.

Model A would also combine well for mobile use with the Olimexino-STM32 charging its Li-Po cell, plus a Slice of Pi board and XRF radio module for occasional low-power messaging, as we were discussing in Simon Monk's Pi+Sensors thread.

Can the HDMI be turned off completely, as well as the graphics parts of the VideoCore?

Yes and no respecively: HDMI can be turned off by not connecting an HDMI compatible screen.

No the graphics part will be generating low-bandwidth composite video.....