In message <[EMAIL PROTECTED]> you wrote:
> > I just re-read the entire Capstone report, and I can't find any
> > justification for the use of a separate power supply.
> Aaaargh, sorry. I clearly remember discussing the power supply voltage 
> issue in detail during the capstone project. Switching the entire rocket 
> avionics bus to 5V would of course make us pretty standard USB accept 
> for the connector.
> We decided against this, however. Unfortunately, the only record I could 
> find of this decision is that the capstone *requirements* that say they 
> must use the 14V avionics bus. See:
> Let me quickly summarize what I remember to be our arguments:
> 1. Many of our systems require higher voltages, in particular 12V. These 
> 12V loads tend to be our big power hungry loads, like RF power 
> amplifiers. Efficiency is critical here, and boosting 5V to 12V means 
> twice the current flow and thus four times the resistive power loss 
> which is a minus. Switching supplies are more efficient when they're 
> switching small voltages, and the 14V Li Ion back is ~ 12V for most of 
> its life, so that's a plus. Of course, it's less efficient for the new 
> 3.3V nodes, but they draw very little current compared to the high power 
> 12 V loads, so it's still a win to have the power bus ~ 12V.
> 2. Making power distribution 5V means putting a switching supply between 
> the battery pack and the rest of the rocket. This is a big time minus 
> because:
> a. A failure in this main supply kills *everything* on the rocket that 
> doesn't have backup power.
> b. There is now a tremendous amount of switching noise on the main power 
> bus that was otherwise pretty quiet. In particular, RF loads, other 
> switching supplies, etc., will all cause fluctuations on the power 
> supply bus. This is particularly bad for our sensor systems, where we 
> want very, very quiet power supplies.
> ... I think there's a few other things I missed, but I think that 
> summarizes the main points. Anyone remember anything I forgot?

None of which (excepting 2a, which is perhaps a big deal,
but I'm having a hard time imagining this being that hard to
get solid) suggests to me that we shouldn't supply both raw
battery voltage *and* usb-ready +5V to the nodes, since
we're using a non-standard connector anyway.  I don't see
the extra pair of wires as a huge drawback, and this would
allow devices to choose what they needed.  In particular,
this would let the microcontroller run off USB power, which
means that groups like PARTS could reuse our layout and PCB
for the microcontroller half-board as-is.

Am I missing something?


psas-avionics mailing list

Reply via email to