Monday, August 9, 2010

August Already?? and Google Latitiude

I was figuring I'd spend a lot of time on the house and the automation system while I was job hunting, but it didn't turn out that way. It's been about 6 weeks and I really haven't done a whole lot. I did finish up most of the trim for the flooring, but I had to stop and find some flexible quarter-round to go along the bottom of our staircase (which has a full 180 degree curve on one side). I did pick some up from Trimster, but I haven't thought much about how to stain it, so it's sitting around.

As for automation stuff, I've been revamping our AJAX based floorplan GUI. Given all my free time, I've been reading up on CSS3 and jQuery and have completely re-written large portions of the code. My HTML page used to have markup, JavaScript and CSS all mixed together, but now I've gotten everything updated to typical coding standards, with the HTML page containing only the markup, and the CSS and .js files have grown significantly. It's definitely much neater to read now and should make maintaining and updating it easier as well. I figure at some point, I will migrate it to HTML5 and that will be another major overhaul, but for now, I'm very pleased with the new CSS and jQuery things I've learned.

When Google Latitude first came out, I put a placeholder at the bottom here, but had so many things going on, I forgot all about it. This thread reminded me about Latitude tracking and with the wife planning a 5-day roadtrip with a girlfriend, I finally decided to get off my butt and implement some tracking (with her permission of course! :)

First off, we need to get a public badge for Latitude - you can find out more here. We already have Google Maps on our cells so we just need to enable Latitude on the app. Next, we need an automatic trigger to start Latitude tracking. We use RFID tags in the cars and bluetooth on our phones for tracking, so when someone's phone and a car are gone, we can start the Latitude tracker.

The tracker is just a perl script that periodically fetches the Latitude badge's JSON feed and parses its information (latitude, longitude, timestamp, reverse geocode info). That information is stored in a MySQL table that is reset for every new trip. Finally, I've written an AJAX page that fetches the location data from MySQL via PHP and shows each data point on Google Maps with a line connecting the points. I plan to post more details as I debug this new feature.

Finally, I am about to close a deal to join an early stage startup. The company is in a very promising space and is at the right stage for me to be able to contribute a great deal to the product development. However, early stage equals long hours, but I'm ready for another startup. HA stuff will just have to be put on the backburner...

Wednesday, July 7, 2010

Charmed Quark Controller Going Open Source

Dean Roddey, the creator of CQC, has thrown in the towel, but is planning on releasing CQC as open source. It's very unfortunate as CQC is an incredibly powerful (and incredibly complex) software package. I am looking forward to checking it out for building touchscreen interfaces, but I really like the platform independence of my AJAX GUI. I will, however, try to add xPL support to CQC. In any case, it makes you wonder how the makers of HAL can possibly survive.

Update: The link above has been updated by Dean. He is no longer planning on open sourcing CQC...

Friday, June 18, 2010

I Am Still Here

No, I have not fallen off the face of the earth. I've been working on some upgrades, but nothing related to automation. I've been replacing our downstairs carpet with laminate flooring, and while that's been going on, the wife & kids have been helping me repainting too. I know, there are hardwood snobs out there who think laminate flooring is like linoleum. Granted it's not hardwood, but we already have engineered hardwood in the kitchen & nook areas, and it's been gouged to death by our previous and current dog. I'm not going to pay $20,000 to have the rest of the house done in hardwood, only to see it gouged up in days.

I've finished about 700 square feet and I'm down to the last 100-150 square feet, but have been delayed by the uneveness of the concrete subfloor. So, I've been using leveling compound on the floor, and it's been a 2 week hit to my project. I hope to finish it up this weekend.

One thing I discovered when looking into laminate flooring is that they all tend to have formaldehyde, which will emit gas (off-gas) after the floor is installed. One of the reasons I ripped out the carpet is that my son has asthma, so anything making the air quality worse isn't going to work. I discovered DuPont's RealTouch laminate flooring, which is GreenGuard certified as having a minimal impact on indoor air quality. It's held up great so far even with our Australian Shepherd racing around the house with the kids. Not a scratch or gouge.

We also went with GreenGuard certified paint, Dutch Boy Refresh. Again, this means it will emit a minimal amount of noxious gasses while it dries, while at the same time, minimizing noxious gasses in the bathroom ;)

Update 6/24/10
Finally, I was laid off yesterday as expected. As soon as I walked in, I saw facilities and HR people in the building, so I went straight to packing up the last of my stuff. It's good timing since the weather's great, I can spend time with my kids and besides, I've got a floor to finish and walls to paint.

Friday, May 21, 2010

Goodbye AT&T You SUCK

At our 2nd house, we've had AT&T DSL and phone for 4 years. It has steadily gone up in price every year and the service still sucked just as bad as the year before. Three and a half years ago, they raised the price for caller ID to $8/month from an already inflated $6/month. That was practically as much as the phone service itself cost. A little over 2 years ago, they jacked up the cost of DSL by 25%. More increases later, our latest bill for 1.5 Mbps service hit $28/month and with the land line, the total ended up over $43/month.

I gave them a call to see if they had any promotions for existing customers. Nope. Then I asked to cancel the landline and run naked DSL. The customer service rep tried her best to sell a load of bullshit. She said "Oh don't do that. This is what all of us here have - the landline and the DSL." "Well, to do that we'd have to disconnect your DSL and landline and we're not sure if we'd be able to reconnect your DSL to the same circuit." What the f*ck does that mean? Do they have some 5 year old kid pulling and connecting DSL lines for them? She tried to discourage that option. I said "So I'd be better off just canceling all my AT&T services.", to which she replied "No, you're better off leaving everything alone."

I gave AT&T the boot yesterday after we got Comcast 12Mbps service. To replace the landline, I picked up a second Magic Jack. Of course, I'm just trading one evil (AT&T) for another (Comcast), but then paying $35/month to run 1.5Mbps naked DSL would be just stupid - if AT&T would have actually been able to figure out how to connect it in the first place.

Friday, April 30, 2010

First Full Month of Solar

As April ends, I can now assess the results of our first full month of having solar power. Specifically, I was very interested to see if my projections for solar production were even close. If you remember, I wrote this blog entry back in October 2009 as I was trying to determine how big a system to get, what sort of output I could expect and what my savings would be. I created this Google Docs spreadsheet to help with all these calculations.

For the month of April, the spreadsheet predicted for the 3.68kW system we got, production of 513 kWH (this actually comes from the PVWatt tool mentioned in the spreadsheet). It didn't look like we'd get there as the 27th and 28th were overcast and rainy days, but the last 2 days of April turned in the highest production so far. Actual production came in at 538 kWH. Based on this, it looks like the analysis I did in the spreadsheet turned out very accurate and as a result, I'm very pleased with our purchase so far.

Monday, April 19, 2010

Dog Sitters: Don't Shortchange Customers With Automated Homes

We recently went on vacation for a few days and had a dog sitter/walker take care of our dog while we were gone. We had the sitter visit our dog during the day to walk, play with, feed and clean up after him. We chose the duration of the visits and were billed accordingly. To get to our dog, the sitter had to open the side gate, which of course, is monitored and logged by our system. I used that information to gauge when a visit started and when it ended. Our security cameras, which are recorded, could also provide an extra confirmation if needed.

Unfortunately, the results are not good. There was not a single visit where the sitter stayed for the duration we paid.

Visit #1: 18% short
Visit #2: 50% short!
Visit #3: 33% short
Visit #4: 26% short
Visit #5: 23% short
Visit #6: 28% short

Overall, we were billed for 28% more than the actual number of minutes. My wife and I were stunned by how the sitter consistently shortchanged our pooch and not just by a couple minutes. For example, a 30 minute visit was really only 15 minutes and a 1 hour visit was cut short by 20 minutes. All those 5 star reviews on Yelp must have been written by friends or completely clueless people (AKA normal people without automation systems).

Our dog is a very important part of our family, and it bothers us more that he was shortchanged on attention than us being ripped off. Needless to say, rather than confront the sitter, we are going to cut our losses and find a new service (and put up a review on Yelp). This brings up the question, is it really possible to find a sitter who actually stays for the agreed upon durations? You are really at the whim of the sitter's honesty and it seems really easy to cheat. How do you guarantee you get what you pay for?

Thursday, April 1, 2010

Overcast Day Has Greater Solar Intensity???

It seems counter-intuitive, yet on a cloudless, sunny day, our panels peaked at a lower electricity output than on a day with overcast. In fact, on the cloudy day, our panels maxed out the inverter, causing it to clip. See the graph below where I combined the charts of 2 different days.



The red line shows the nice production curve on a cloudless day, but the green line shows the output on a cloudy day. The green line maxed out earlier in the day and even later in the day than the red line.

The sunlight hitting earth must be reflecting back up to the sky, but the overcast is causing it to reflect back down again. This results in more solar energy hitting the panels. I would assume that the glass over the panels is like this as well - allowing the sunlight to pass, and some light that reflects off the panels would get reflected back to the panels by the glass.

Tuesday, March 30, 2010

Waiting for the Axe to Fall

It's been a quiet several weeks. There have been rampant rumors at work that the company is planning on shuttering my group. It's something that we've been expecting at work for the last 3-6 months, but as quarter end nears, it's possible that it may be right around the corner. So I've been a little preoccupied to do any HA stuff other than mindlessly monitoring the production of our solar panels. According to the latest rumors, the axe was to fall last Friday, but upper management had to delay it since they're waiting for our group to finish up some work before they'd be comfortable canning us all. As I sit here contemplating impending joblessness, I wonder what my next career move will be. I'm an ASIC design and verification engineer, but I'm burnt out on it and a change of career may be in the cards. I've got an unused MBA, some .NET programming skills and experience with AJAX, PHP and MySQL, so I have some options to explore. In the meantime, feel free to donate! ;).

Wednesday, March 17, 2010

xPLBattWatch - A Battery Charge Monitor

This is the first release of xPLBattWatch. This app is used to monitor the battery level and charge status of a laptop. It issues sensor.basic messages based on changes in the charge level and status. It may be a little rough, but I thought I'd throw this out there and let you play with it. More details and download here.

Friday, March 12, 2010

Solar Array Powered Up and Some Graphs

The installer powered on the solar array on Tuesday after the inspection. After he left, I installed current sensors for the solar breakers and ran those along Cat-5 back to the ECM-1240. I was debugging the charting and got it pretty much sorted out by Wednesday night. Here's the first few RRDTool graphs from Thursday.


You can see it generated nearly 21 kWH yesterday. I'm really happy with that as we typically use 22-25 kWH/day. I didn't expect to get that much production in March. Does this mean the assumptions I made in my Solar Spreadsheet may have been a little pessimistic? There's still a lot of rain & overcast in March (of which we're seeing today), so it will probably average out. We'll have to wait for April to get a good full month to compare.


With a southeast facing roof, you can see we catch the morning sun very early. The array woke up around 6:22AM and peaked around 11AM. Once the sun passed over the top of the roof around noon, production started to drop and around 2PM it started to decline rapidly. By then, it produced about 80% of the total for the day. I wish I had room on the west facing roof to catch the afternoon sun, but it would only fit maybe 4 panels - good for another 1kW. Maybe when we get an EV ;) Production stopped almost exactly at 6PM. You can see some of the spikes in production as the sun was behind some clouds.


This is just a net usage graph. Solar production for the day caught up to our daily usage by noon. Between 3-4PM, production fell below our usage and about 9PM, all our credits were used up. For yesterday, we basically paid the utility for electricity for the 3 hours from 9PM-12AM. Pretty cool.

Sunday, March 7, 2010

xPLWebControl Released

I released an xPL interface for CAI's WebControl. The app communicates with the WebControl via its web interface since it has no other means of control. You can read more about it and download it here.

Friday, March 5, 2010

New Toy: CAI WebControl and Random Updates

So I finally broke down a bought a WebControl device. It's a network enabled IO device that's been discussed heavily at CT and HS message boards. I got it from this seller on eBay. I'm surprised it actually powers on. It was shipped in a USPS Flat Rate Priority envelope. The device itself was in a small bubble pack pouch - just a single layer of bubble pack to protect the sharp corners of the device's components. The corners of the ethernet jack poked through the bubble pack, puncturing the envelope in several places. For $8.71 in shipping, is it too much to ask for 2 or 3 layers of bubble pack????

Anyway, I powered it on, assigned it a static IP address and new login info, but didn't get much further than that. I hope to test out the digital IO and 1-Wire interface this weekend before I'll believe it's undamaged. I have no idea what I'm going to use it for. I am planning on building an xPL app to interface to it, but I'm not in any big hurry. Still feeling a bit burned out on HA.

It's been 10 days since the solar installation was completed and I haven't heard from the city about inspections nor has the utility swapped out my meter. It was frustrating waiting nearly 4 months from signing the contract to the installation (due to weather & backlog), but this is even more frustrating because I just want to flip the switch!

This week I did update xPLTweet and Jab2Twit to use OAUTH for Twitter authentication. That way there's no stored username & password and has the added benefit of my apps showing up as sources for tweets :)