Reworked the PLEG code to work with the positive/negative power readings that were unlocked in the previous video. It's amazing what knowing the actual power and direction of flow makes in the code. Before, it was a guessing game and now it's a simple computation. This is my first time working with "conditional values" in PLEG along with logic expressions in actions. The Vera controllers use the Lua programming language and PLEG uses a Lua-like syntax as well. On thing that's "missing" is you can't have assignment expressions, like a=b+c. Instead, you need to make a condition called "a" and determine it's value as the result of (b+c). Then you can use the value in "a" as a variable in other expressions. I like to keep expressions simple and compute just one thing at a time. This helps as you can see the intermediate results. The key to this computation was to keep everything in floating point. My original code was integer-based and had a problem when it kept ignoring small ~1 watt changes in power. Over time those small changes would build up and until there was a big change of 7 or more watts, nothing would happen. With the floating point code, you keep accumulating the changes and once that rounds up or down to a different number, a change is made. The other key point is to have only ONE variable tracking the dimmer setting. Before I had tried to use a counter to handle incrementing/decrementing the dimmer setting, then writing that counter value to the dimmer. This was before I figured out the "conditional values" and was an easy way to add or subtract numbers in a variable. However, this created a problem in that the counter and the dimmer could get out of sync. Moving to just having a device property that follows the dimmer setting solved that issue. With only one copy, it can't get out of sync. It also allows manually adjusting the dimmer to see how the code reacts. Also tried to incorporate features into the code to reduce Z-Wave network traffic as well as spread that and the CPU load out in time. This should help reduce bottlenecks and synchronization issues. For example, the time delay between getting the power sample and then acting on any changes allows all the computations to take place before the logic that makes changes runs. More to come, seeing what other features are possible with this code... Subscribe for more content like this Comment, rate, share & click the bell icon And as always, thanks for watching Intro/Outro: Louisiana Fairytale by Austin Rogers https://web.archive.org/web/20170402222425/http://drfiddle.com/show_tune.php?id=94 ... https://www.youtube.com/watch?v=xXlhLOUYIow

Adding a self-powered USB hub to a VeraLite home automation controller. Time to integrate the DC-DC buck converter and USB hub with the VeraLite home automation controller. I used a 5.5x2.1mm coaxial power jack to plug the original Vera power supply into. Then I tapped off that running a jack over to the Vera power jack as well as to the buck converter. All those parts are attached with Velcro tape The main goal was to add no additional external connections to the Vera, so still just the power and Ethernet cables need connecting. This allows for easy disconnection of the Vera to take to a device for pairing, if needed. And with everything powered off a single 12VDC power jack, it'll be trivial to swap the device over to solar power when I am ready. Speaking of power, the VeraLite wall wart uses about 7-8 watts of AC power. That works out to around 62KWH/yr. If I were to add the ~2 watts of an AC powered USB hub, that's another 17KWH/yr. Add those together and that's around 80KWH/yr. of energy consumption or US$10-15 depending on your electricity rates. I didn't record it, but there was minimal affect on the Vera AC power consumption after adding the self-powered USB hub with the small USB memory stick and Arduino. That's because if you add a few mA of load to a power supply that's already putting out several hundred mA, its not much of a change. I saw at most 20mA additional load with the hub, memory stick and Arduino connected, out of nearly 300mA total. Be sure to rate, comment and share. And as always, thanks for watching! SOUNDTRACK: Lousiana Fairytale by Austin Rogers http://drfiddle.com/show_tune.php?id=94 ... https://www.youtube.com/watch?v=K_hlMJ0q1Ok

With the rebuilding cache, I was able to see the before and after performance of the cache. Since the cache had only been warming up for a day, I was able to reload some older e-mail messages that had a lot of image content to test the performance with and without caching. I use the Thunderbird e-mail client and to allow for access to the same e-mail content from multiple PCs, I locate the e-mail profile folder on the NAS file system. This is the folder where your inbox and mail folders are stored. I didn't realize how much data is stored in the profile folder, but for grins, here are the file properties reported on my e-mail profile folder: Size: 4.70 GB Folders: 451 Files: 99,988 Now one might not think that is a heavy workload, but I find it is, at least one where I notice the difference in performance. With the SSD cache, it feels like the profile folder is on local storage. Those image-rich e-mails likely come in as one large e-mail with many image attachments that result in many small image files. Then when that image is displayed, each little image tile is loaded from storage, one at a time. On a regular hard drive, chances are those little image files are all stored relatively close to each other so access times are quite low. On a RAID array, files get scattered across all the drives in the array, so you almost guarantee that at least one drive will need to perform a seek, increasing the latency when accessing those image files. With the SSD cache, as that e-mail message comes in from the e-mail server, it is written to the NAS and to the SSD cache. Then when it is displayed, it comes out of the SSD cache with minimal and consistent latency. ... https://www.youtube.com/watch?v=KpE13b_BuMw

Finally got all the mastic removed from the floor, time to pick a color for the sealer. It took 3 more stripping sessions from the last video, but at last all the mastic is removed! In all, it took about 6 weeks to remove all the mastic, trying various techniques, working around other projects and filming various aspects. With proper equipment and a crew of 3 or 4, it could all be done in one day. In all I used nearly 2 gal. of the Bean-E-Doo mastic remover, 1.25 gal. of the Emerge degreaser, 1 gal. of Zep Industrial Purple Cleaner/Degreaser and 2 brown 3M scrubbing pads on the floor machine to clean 160 sq.ft. of floor. Still need to make a final cleaning pass that should even out the look of the concrete before doing the concrete etch and sealer. Planning to use the Orange X-4 concrete sealer: https://stonetechnologiesinc.com/product/custom-colored-concrete-sealer-x-4/ Mexican paver tile floor: http://cbektilestone.com/wp-content/uploads/2013/01/photo242.jpg To be continued, we'll check on the concrete repair in the next video... Be sure to rate, comment and share And as always, thanks for watching SOUNDTRACK: Lousiana Fairytale by Austin Rogers http://drfiddle.com/show_tune.php?id=94 ... https://www.youtube.com/watch?v=GZDRduYeIhY