UPAMD - goals, considerations, energy

[Bruce Nordman <bnordman@xxxxxxx>]

UPAMD folks—

Greetings.  I come to the UPAMD process with energy consumption
and efficiency as my primary interest, though I am also quite mindful
of how the technology can be most functional for people.

I recognize that the prime goal of the process is to support notebooks
and devices derivative of them, and that extensions to support other types
of devices must not compromise this.

I do think that consideration of energy efficiency should be part of the
process, though not necessarily part of the normative result.  It would be
useful to have an informative annex that discusses efficiency issues.
One example of this is that efficiency regulations on power supplies
typically are based on measurements with no load, as well as 25%, 50%,
75%, and 100% of rated output.  Consideration of how this applies to
UPAMD and whether any different approach would be better for assessing
efficiency should be done.  Also, text for standard test conditions for UPAMD
devices should be created, e.g. in specifying characteristics of the load device
(voltage, etc.).  Again, this doesn’t have to be in the normative part of the
standard.  I expect that California, U.S., European Union, and other regulations
on external power supplies will apply to UPAMD adapters from the very
beginning, so being proactive on this will help everyone.

10 – 130 W is more than an order of magnitude in range, and with today’s
technology, it seems likely inefficient to power a 10W device with a 130W
supply (this could change in future), and also, there will be size/weight
penalties for more capacity.  Thus, it seems inevitable that there will be
some variation in adapter capacity, with some supporting the full 130 W,
and others less (those powering multiple devices—see below—may well
go above 130W, at least in total (possibly not per port)).  Whether there should
be a series of fixed intermediate capacities, or whether it should be continuously
variable, I don’t know.  Apple adapters provide a relevant example: my Macbook
adapter can either power or charge a Macbook Pro, but not do both simultaneously,
as it has lower capacity than the adapters for the Macbook Pro (but both share
a common connector).

While today notebooks are powered by EPSes that move with the device,
a useful feature of UPAMD is to leave the power supply stationary, as for
in cars, conference rooms, and a UPAMD “outlet” in the wall of any building
(or perhaps in devices like PCs and Set-top boxes). This argues for having the
cable separable from the adapter.

While the initial case is certainly a single device being powered by an adapter,
the idea of multiple devices already noted seems valuable. This could be an
optional extension (possibly not even in V1.0 of the standard) for adapters,
but should have the ‘hooks’ in the first version for supporting this.  I think the
USB model is appropriate, enabling multiple ports off of an adapter, and multiple
devices to hang off an individual port (subject to the max per-port power draw
which I am not suggesting to change).  I have called this a “NanoGrid” – see:
  http://eetd.lbl.gov/ea/nordman/
and propose that there be a common “meta-architecture” for nanogrids to enable
interoperability between them.  A key feature of a nanogrid is that it is a distinct
administrative domain of electricity supply that may have its own local electricity
price – taking the grid price if there is one, or generating a local price if it the
nanogrid is off-grid (vehicle or during power outage) or if is short on capacity
(there could even be higher prices on ports that are over-subscribed as a way
to allocate the limited capacity).  The paper describes other features of nanogrids
I think UPAMD should implement.  We will want to have gateways between
nanogrids of various sorts: UPAMD, PoE, USB, eMerge, 12V car technology,
new car voltages, 380V DC, and of course, traditional AC power.  As above, I
think this should be discussed for a version 2.0 in parallel to those discussions
required to enable the basic single adapter-single device functionality.

I am not an engineer but my instinct is to have UPAMD operate at a single voltage.

Finally, mention was made of LED indicators of adapter status.  This should
probably be optional, but if implemented, then the labeling, color selection,
and dynamics (e.g. blinking) should be based on a standard, either part of UPAMD
or referenced from elsewhere.  A related standard that could serve as a model
and reference for this is IEEE 1621.  See:
  http://eetd.lbl.gov/controls/1621/1621index.html
This could be discussed in an informative annex for V1.0 so as to not be a
distraction to the core effort.

That is plenty for now.  Thank you for your time and consideration.

--Bruce

-- 
Bruce Nordman
Lawrence Berkeley National Laboratory
BNordman@xxxxxxx -- 510-486-7089
eetd.lbl.gov/ea/nordman/