Date: Mon, 15 Oct 2012 18:42:34 -0400
Reply-To: David Beierl <dbeierl@ATTGLOBAL.NET>
Sender: Vanagon Mailing List <vanagon@gerry.vanagon.com>
From: David Beierl <dbeierl@ATTGLOBAL.NET>
Subject: Re: doc wattson or watts up?
In-Reply-To: <EC6A626D-C70E-4C01-904E-E68EBC3F555F@shaw.ca>
Content-Type: text/plain; charset="us-ascii"; format=flowed
At 11:34 PM 10/14/2012, Alistair Bell wrote:
>I didn't see the voltage drop you mention, but did note that the
>unit's current draw is 7mA (which is less than my stereo head unit
>even with face plate off).
One millivolt per amp in the current shunt. That's a commendably low
number, much lower than a typical digital multimeter.
Per s7.2, about 5 millivolts per amp including supplied leads, total
dissipation about 28 watts at 75 amps. Since the lead lengths aren't
specified anywhere I can find, you can't tell whether that's using
the easy four-wire or the preferred three-wire connection,* but it
represents a total length of 14-ga wire of about twenty inches based
on their quoted example in s4.4 of 2.5 milliohms per foot.
*Load current has to pass through the black lead because that's where
the current shunt is; but putting it through the red lead as well is
merely a convenient way of making up the connections that doubles the
voltage and power insertion loss and the heat load in/near the
meter. For our type of use a completely external shunt would be far
superior but you can't simply move it off the board without spoiling
the factory calibration in the EEPROM.
Notes from reading and poking around:
1) If the meter is to be powered by the system, the black lead MUST
form the sole connection between the battery and chassis ground. If
powered by its own battery the meter black lead could be placed in
the system positive lead with the source connection toward the
battery (and the red lead carefully isolated from everything). Just
as with a panel ammeter, wiring choices could then dictate whether
the device was measuring strictly consumption or integrating
consumption and charging. This needs careful consideration, because
on the one hand the meter was designed with consumption-only use in
mind and does not seem to have any provision for compensating for
charging efficiency (or for the Peukert equation relating rate of
consumption to usable capacity); and on the other hand,
system-powered operation dictates that it will be integrating charge
and discharge, under which conditions the consumption figures will
quickly get out of whack with effective reality in a cycling
application. This is where you get to pay your money in skull-sweat
and knowledge instead of the several hundred bucks for a system
monitor that takes these things into account.
2) The 50 amp continuous current rating is a little
complicated. It's based on several factors, and I'm going to pick
them apart because this isn't a panel meter in a metal case, it's a
bunch of electronics on a circuit board in a little plastic box with
a heater inside; and if you exceed its ratings it just might turn
into a sparkler. Says so right in the book.
a) Only one power-carrying lead passing through the meter
("three-wire" connection).
b) Free air circulation around the meter and its leads. Out in the
open, in other words; air on all sides and plenty of ventilation.
c) Maximum air temp 35C/95F.
d) Maximum temperature of the connections to the meter leads
35C/95F. In other words, they're expecting the connections to do
part of the heat-sinking job.
e) Maximum temp of the meter leads near the case 70C/160F (that's
right, the wires are silicone-insulated and can run comfortably at
that temperature, AND EXPECT TO).
N.B. Heat dissipation is roughly proportionate to lead length, but
varies with the square of the load current. 3.125 watts at 25 amps,
12.5 at 50, ~28 at 75, 50 watts at 100 amps. That I-squared term in
the power law can sneak up on you. Copper also has a positive
temperature coefficient of resistance, so the warmer it gets the
higher its resistance and hence power dissipation. Not large in this
context, takes over 175C change to double the resistance; but in the
wrong direction.
Given the above, since running the second power lead through the case
will close to double the heat load inside the case, max continuous
rating with a four-wire connection and all the above limits is
probably down around 30+ amps. So likely you're not going to use a
four-wire connection. But look at the ambient limits and compare
them to summer automotive interiors. You're going to have to de-rate
anyway, and unfortunately the manual does not give the necessary
information to calculate by how much. You'll have to find out by
experiment. And of course you'll have to de-rate further if the
device and its wires are enclosed or mounted to a surface. If the
surface is cool metal it will probably help instead of hurt. If you
run fat wires right up to the case that will help.
This puppy is meant to run hot so as to save weight and stiffness in
its wiring, but it's not going to degrade gracefully and the manual
is refreshingly candid about the dangers you're working with. It's
clearly written by someone who's been there (and his lawyer, who is
careful to remind you that nothing in the manual is to be relied on
and the entire responsibility for the thing not going up in smoke is
on You-Know-Who). Like every manual/spec sheet I've ever seen, it
doesn't give you everything you need to know, which is why the
manufacturer has applications engineers who will consult with *your*
applications engineers while you develop the beast that includes this
component. You say you don't have applications engineers? Look in
the mirror, pal. Depending on the manufacturer, you may have to look
in the same place for theirs. Given the info that *is* in the
manual, someone on the engineering side knows how to communicate; but
they likely don't have the staff to help much with onesie
applications, and there may or may not be a language problem as
well. In fact all those spec numbers may have been generated
entirely by a simulation and never tested in the real world, because
that kind of testing is expensive. But I digress. It's a very good
manual and well worth studying on its own merits.
Getting back to the point, and reinforcing what Dennis said earlier -
this isn't an appliance, it's a specialized device that's been
carefully optimized for specific application in the recent
development of high-energy lightweight electrically powered
radio-controlled models. Low safety margins, no protection. The
development of Li-Poly batteries has made possible large
battery-powered aircraft models, for example, that would have been
inconceivable a few years ago; and they charge and discharge at very
high rates with cycle times of a few minutes - and someone's eye is
clamped firmly on them every second they're operating/crashing. I
was looking at an 11.1 volt, 6.8 amp-hour battery that weighs under
600 grams and is rated for charge at 54 amps and continuous discharge
at over 400 amps, which is to say full discharge in one
minute. Burst rate is 50% higher. And in case you don't have an
empty paint can handy, you can buy a fireproof bag to charge it
in. It can power a 25 pound helicopter for a few minutes. Like
this one: http://www.youtube.com/watch?v=s9d3hRExTiw&feature=related
(the actual flying starts about five minutes in). Hundred dollar
battery, $800 motor controller and $400 motor in that bird,
incidentally. $2000 for the shell and $700 for the rotors plus a
month or two construction time - before all the bells and whistles
got added to this one. Apparently having the pilot salute on command
is either the zenith of ecstasy or the nadir of wretched excess
depending on how you view the hobby....
In this context the Watts Up is a pretty conservative device, but
it's a different world than we live in; so think carefully and use it
conservatively in unattended operations. Pity they haven't one with
an external shunt, because that would get rid of most of this
concern. It still wouldn't be ideal for RV-type use but that would
be a matter of trading features against price, instead of worrying
about setting light to the thing.
3) The basic electrical abilities of both units are the same. I'd be
inclined to consider the Watts Up for better precision at low drains,
and I'd inquire about their statement that the Watts Up has higher
reading capacity than listed at lower resolution, which would be
excellent if true.
4) Fascinating dip (for me) into the modern R/C hobby. Maybe a
little climb will clear my head, into another place where heights and
electricity mix... http://www.youtube.com/watch?v=tgO4Gd4RhvM . Be
sure to watch full screen for the proper effect.
Yours,
David
ps - found the lead length spec on another page. The leads are 3.6"
long, so the four milliohm lead resistance includes both sets. I've
worn myself out with this turkey so I'm not going to try and rework
the numbers...exercise for the reader. Other editing lapses please
forgive, parts of it are starting to feel like a fever dream.
| 
