Flutterby™! : 2nd Project

Next unread comment / Catchup all unread comments User Account Info | Logout | XML/Pilot/etc versions | Long version (with comments) | Weblog archives | Site Map | | Browse Topics

2nd Project

2007-11-13 15:52:05.044802+00 by meuon 9 comments

I'm offically working two large projects for a while. My hope is that the second one gets beyond 'Phase I' and ends up being something that requires hiring a handful of talented uber-geeks and up and coming wannabe's to do once it has some paying customers. I violated a rule of mine: unpaid development/phase I work. I've been bumping into the client for about two years, he's established in his niche and trying to evolve past it. He's also "good people" and pretty bright and has paid for a development system. The payment processing and collections side is rather boring, but something I'm good at. the interesting part is a reason to play with ZigBee, and the fun of turning on and off a houses electrical power/meter as well as some other kewl toys. So, for the record, other than Nancy being #1 top priority, my next two are: NextKnowledge (great system, website sucks, not my job), and UtiliJuice (website sucks, name may change soon, and site IS my job).

Now if only I was good at juggling.

[ related topics: Interactive Drama History Work, productivity and environment Heinlein California Culture Clowns Real Estate ]

comments in ascending chronological order (reverse):

#Comment Re: made: 2007-11-14 00:25:31.805707+00 by: Dan Lyke

Hmmmm... interesting.

The last few conversations I've had with my dad have been about palying with current measurement, and I'm looking for the spare few hours to build a little current meter. I need to find a little ferrous ring (to run one leg of an extension cord through and wrap some lead wire around to build a little current transformer) and a zener diode for a voltage limiter so I don't blow a hole through the ADC in my microcontroller when I get spikes/overvoltage. The issue I see is that my favorite microcontroller has a 10 bit ADC, which, if I managed to get everything else calibrated exactly perfectly, would only get me 150mA resolution.

Down at Apple today, said "Hi" to Steve (he parks in the visitor's lot, apparently), gonna be here tomorrow and Thursday, so it ain't gonna happen this week.

#Comment Re: made: 2007-11-14 01:51:59.060365+00 by: Larry Burton

Thanks, Meuon, and I thought I'd gotten over my X10 addiction and you point me to this. :)

Dan, what current range are you looking at reading and outputting?

#Comment Re: made: 2007-11-14 16:22:04.99046+00 by: Dan Lyke

Larry, I was standing in our living room, and we had some number of our computers running (between the two of us we've got six in active use) and our various battery and cell phone chargers, and one or two lights, and I thought it'd be great if current measurement devices could be cheap enough that I could actually see the result of making changes in that room. So that'd be a standard house circuit, 110v, 15A max.

Right now I've got a 1 month latency (end of my billing period to when I write the check) and sample period for any changes I make, it's hard to tell the impact of a computer sitting idle versus one sleeping, or a light turned off after dinner versus when we go to bed, in that period. For about $70 I could get a dedicated current measurer (or I could build a rig to plug a multimeter into), but that's a process of actually measuring the devices, rather than having some more passive "why is that number higher than it normally is?" situation, and it's not cheap enough to just start throwing them at wall sockets (let alone have an aggregate for the room...).

It'd be cool if these things were cheap enough to put into light switches or power strips, and I think I could get get quantity prices to a few bucks. The markup on power strips is exhorbitant, but if you could get one with a display for five dollars more.

However, I think my server eats between 9 and 25 watts, and a few years ago there was that "discredit the environmental movement" thing about wall wart and idle device power usage going around, so to make this thing useful it seems that it needs to be sensitive to less than 100mA (and that's still pretty high, I'll bet modern cell phone chargers and such waste less than a few thousandths of a Watt at idle), but able to measure up to 15A.

11v*15A/2^10 means that 10 bits on my A/D converter gets me to just over 1.6 Watt resolution, if I get all the other ranges right (need to see what accuracies on my various components are). Might be able to use an amplifier to subdivide that range a little bit, but those ratios get back to resistor accuracies and stability and a possible need for a calibration procedure, and increased parts counts.

And then I get into issues of aliasing on my measurements with noisy loads...

(Further musings on power supplies: For a production device, I'd have to do the numbers on display power consumption, but what about two coils, one for measurement, and one that's parasitic that I can power the microcontroller and a display off of only when there's some basic draw on the circuit.)

#Comment Re: made: 2007-11-14 17:45:48.940203+00 by: Larry Burton

Okay, I understand your problem now. I thought you were looking at current loops for control. The standard, in this country, is a 4-20mA range and the ADCs we use have been 12 bit but we are starting to use some with even higher resolution. Speaking of all this reminds me that I have even less time this afternoon to get to the airport than I thought. I've got to stop and pick up an amp meter for this job. See ya.

#Comment Re: made: 2007-11-14 20:22:15.720776+00 by: meuon [edit history]

Zigbee is interesting for it's mesh networking capabilities and some other features (low power, etc..), and it's the defacto standard in that industry.

What the smart metering world is alos seeing is that people with an electric meter display inside of the house use about 20-33+% less juice, just because they are aware that they are burning money. It's essential for Prepaid Electric Metering/Payment to work, but it's also a benefit for intelligent consumers that just want to control either their bill, or their carbon footprint, or both. I don't know what the resolution of the Echelon's smart meters is, but it's good enough for billing... so I think that's good enough for what Dan needs as well.

Utility companies are interested as it effectively lessens electrical load which reduces capital expenditure for building new plants.

#Comment Re: made: 2007-11-14 20:46:06.924379+00 by: Dan Lyke

Yeah, Zigbee may have some potential issues, but cheap wireless chipsets and lots of reference designs are good! Even if the software layer of the stack might not be the best idea.

What do those meters cost? Maybe the thing to do is to put the panel someplace obvious in the house and have that networked, rather than doing lots of sub-displays.

#Comment Re: made: 2007-11-14 22:12:54.232844+00 by: meuon

I think around $250 each.. but I'm not sure what "retail" is. Maybe we need to send you in exchange for some code? A 100amp single phase unit could be very useful.

#Comment Re: made: 2007-11-16 10:55:51.431334+00 by: meuon

Hey Dan: Would one of these power cost monitors work? Turns out UtiliSol is a US Distributor.

#Comment Re: made: 2007-11-16 14:05:50.140977+00 by: Dan Lyke

Hey, that's really cool! I love that installation is a simple mechanical clamp! It doesn't isolate to a room, but I think it solves the meta issue of "I turned off this light, what changed?".

If the execution is as cool as the concept, that's awesome!