Weeknotes 11

  • Car Battery Upgrade
  • Well, we have one trip to LA and back in the car.
    In short, it went great.
    We charged to 100% Friday night and headed up and back over the weekend.
    Got to LA (the half way mark) with more than half a charge left, so had a good feeling we would be Ok.
    The traffic on the way up was rather free flowing, good for making time, but 70mph (112 kph) is not the best battery economy. Point is, the trip up was under the worst conditions so I was very happy to see the half and some charge left.
    Left Sunday morning and wouldn’t you know it, unlike pretty much every other time we have driven home in stop and go traffic, it was wide open, so yeah, 70 mph all the way home as well (and just to be clear, the freeway speed limit is 70, the fast lane is doing 80-85 (~135 kph), I was in number 2 lane).
    Here is how the dash looked when I pulled in the drive when we got home.

    1. The guessed max range you could get (downhill with a tail wind).
    2. The worst range (70+mph into a headwind).
    3. Super hard to see, but there is a few pixels of orange under the middle range, Freddy and I call it the worm. It goes down in orange if your current driving trend is getting worse and it goes up in green if your trending to better economy. Its the most useful part of the gauge as you can quickly see if your current driving conditions are more likely to give you the better range or the lower range and by what amount. At 70 mph, it’s pretty much orange all the way down to the bottom range.
    4. How much power you are currently using. Sitting in the drive with the AC on, its 1kw. It goes up in orange when taking off from the lights or down in green if you are slowing down using regen and thus adding some charge back in to the battery.
    5. Trip meter. I reset this when I pulled out of the drive so I know how many actual miles I have done.

  • New Tires
  • Got some new tires put on the car late in the week.
    Went with Pirelli Cinturato P7 All Season Plus II. Same size as whats on the car 215/50R17.
    I don’t have anything against the stock Michelin Energy Savers other than they were total garbage. Almost zero traction in the dry, crazy dangerous in the wet. Pathetic tire life, I got 12,800 miles (20,600 clicks).
    So far, all of a few drives to work and back, I love the new shoes. They are super quiet, seem to actually grip and the car is just feeling a lot more planted.
    I wish I could have gone with 225/50’s but since I have put the lowering springs on, I don’t think BB has the wheel well clearance.
    Now that we have 100% charge back, Freddy and I are talking about doing a few drives now and then like we used to.

  • Chrome.exe
  • I use a screenshot node in Node-RED to, uh, take screenshots of websites.
    Every now and then for some as yet unknown reason, the screenshot node wont close and so leaves a chrome.exe process running that uses both CPU and RAM.
    After a day or so my Windows PC is just about on its knees crying. Windows is too dumb to kill that cruft off and so it just keeps building up and maxing out everything.
    The mouse wont really move and not even task manager will launch.
    I happen to catch it at this point one day this week and found I could use tastklist in a CMD window. I had around 70,000 chrome.exe processes.
    Quickly found that the command taskkill /IM "chrome.exe" /F did the job in a few moments and tada. My computer fan stopped its jet impersonation, the mouse started to move freely and things were back to what ever running Windows on a PC is like.
    So, now my task is to find out why this has just started to happen after 5+ years of using the node with zero issues.
    I did a Node-RED update and a few Windows updates, so its going to be hard to pin down, but for now I have a graph that plots CPU and memory and when it gets a bit high/low, I issue the kill code. Not great, but a work around none the less.

  • Python depends
  • To install most Python libraries or apps you use an installer called pip. pip is a recursive acronym that stands for ‘pip installs packages’ or ‘pip installs python’.
    Google clearly states that we need to use pip3 for python 3 packages and pip for python 2 packages.
    EPIC comes with python2 out of the box and we have instructions for installing python3.
    The problem arises in that the version of python3 in the Opto22 repo is 3.2 and most machine learning packages need python3.7 or better.
    To build python3.7 from source needs some other libraries installed, but thankfully all those are in the Opto repo they can be installed and (after a few failures and many hours work) you can get python3.7 installed on EPIC.
    The problem now is that python3.7 uses a new version of pip3 and to install that, you need to install some new libraries, those new libraries are not included in the current Opto22 repo.
    So, I started building them from source.
    The first package depended on another package that is not included, so I started to build that second package from source, but it depended on a third package that was not in the Opto repo, so I started to build it from source and it depended on a forth and that is when I stopped.
    I have been using Linux (Thanks Dan) for around 30 years now and have a personal rule that when I get three dependent packages deep, it’s time to stop and step back and rethink things.
    So yeah, had a ‘fun’ few hours in the Python dependency rabbit hole this week.