Riding and Groceries

Tuesday in my world is date and grocery night.

Got to bike to work this morning; bike was an awesome ride into work. It was a little cooler than I wanted to be biking in; but it’s supposed to drop in temp most of this week and I wasn’t really wanting to be biking in the upcoming -10c temps.

The Fuji rides very well. The Sora groupset has a little more throw than I’d like when shifting up gears at speed, but nothing that I’m going to really complain about. I had to drop the seat a few mm; but I’m actually thinking I liked it up a little higher.

Work was fun as always; completed constructing some Yum repos for the upcoming mixed centos 6/7 deployment we’re working towards. Couple of new services on the go (One of the groups we work with is doing a HAproxy=>Logstash=>Telegraf=>InfluxDB=>Grafana for performance and stats collection; Looking forward to rolling that out!

Cycled home; almost died twice (people seem to think that bicycles stopping at stop signs means “I guess I can ignore that stop sign!”)

After work Brynn and I went and tried Chop. Pretty decent experience, great food, was a good time. (Menu is a little more limited than I’d like; but what we ordered was actually great!)

After that was the unexciting trip to the grocery store. Nothing overly exciting to write home about!

Nice quiet evening after that; was expecting to get the Moteino’s out and work on some transceiver code; but just didn’t. :P

Lots more on the go tomorrow, I do really enjoy work actually being exciting and fun!

Operating different

Exciting day; we’re starting to work with some CentOS 7 machines at work (we’re almost exclusively a CentOS shop at work)

We’re a bit of an odd shop, I’m one member of a three man operations team admining a good pile of servers mostly at a few datacenters in europe; and a few odds and ends in a number of other countries. One of these days, I’ll have to better document our existing stack!

We make pretty extensive use of Puppet and Ansible for configuration management and repeatable deployment. That means we define a server configuration using a declared collection of resources, and the software stack takes care of implementing those resources on the machines itself.

This means that when we replace a server, or move a site, we simply apply the same configuration to a new node; and take care of the little intricacies that don’t make sense to automate. (like the actual sites themselves; as there’s… lots of them. (some might argue that THAT’s what we should automate, that’s another discussion)

As a first expirement; we’re replacing an old legacy Gitlab server with a centos 7 box, and going to look at rolling out gitlab CI to speed up our deployments. That’s going to be ongoing for a while. (I’ll post updates as it happens!)

As a random aside; I’ve developed a STRANGE fascination with rulers. I’ve got a pair of FR-4 “electronics rulers” and an Architect’s Triangular, and randomly just ordered a new 34CM RF ruler. I have no idea where this comes from. Strangely; the things are actually useful and tend to feature interesting and useful tables of data or formulas.


Weather has been holding up as “nice” for a few days; almost all of the snow is gone, and I finally got a chance to take my new bike out for it’s first real ride. (rode it around at the office one day after work, but can only go so fast indoors!)

Did a 5.1KM loop from my building down the street, back down Wolseley ave, around twice, in about 20 minutes (only averaged a terrible 15KM/h; but there is a lot of stop-signs down westminster that I actually come to a complete stop at).

Excited to take the bike to work tomorrow for the first time; there’s presently ~3-4 people biking daily; and a few more like me wanting the streets to clear up a little more before we start biking.


Going to have to get a hold of the Winnipeg Amateur Radio Club to enroll in their summer classes; excited to try higher power stuff; (lol; “high”.. like.. under 5W..) (though enjoying playing around under 1 milliwatt at the moment, it does make for some SENSITIVE gear to detect down at the low -80+ range!)

I’ve got a pair of the http://lowpowerlab.com/moteino/#lora modules+units; that I’m simply setting down to their low power (5dBm) setting and throwing 10dB pads in front of. I’d really like to build some beam antennas and see how far I can reach on a pair down in the sub milliwatt range; LoRA is fairly well profiled down to the -120 to -130dB receive strength area.

Following; I’d love to try a multi-repeater, solar backed, ultra low power, low speed bandwidth link; though that’s going to be a multi-year project; as I’m hoping to buy a more permenant property to throw a tower up on and will be spanning ~12-15 KM across a minimum of three points (more likely 4-5 at the extremely low power levels I’d like to be in!)

I’m going to try my damndest to make it out to the April 17’th Flea Market if for nothing else than just saying hello and meeting some of the other Winnipeg HAM’s!

Photos and adventure

Pretty unexciting Sunday. What got done today:

  • Did a pile of research on FET’s and power dissipation
  • Started digging into determining heatsink suppliers for the potential DC Load project
  • Tested and reviewed a Fenix LD22 Flashlight
  • Finally got around to dumping photos off the camera
  • Went out for dinner with my partner (to her mum’s)

Power and Heat

Digging into the parts / theory behind building a DC electronic load; it’s quickly becoming apparent that dissipating 500+ watts is going to be a pain in the ass very quickly. Though do-able; it’s going to take driving 8-12 FET’s on a couple of forced-air heatsinks, temperature monitoring, additional gate drivers, and likely a relay switched path to select the sense resistor (I’d like to sense across a selectable 100,10,1,0.1,0.01Ω sense resistor so LOW mA or µA ranges are fairly accurate).

This is going ro require fairly beefy sense resistors, though they would only be dissipating ~0-10W after the FET droppers, they will still need to carry up to ~45A.

This whole proceedure is going to produce a lot of waste heat. I’m mostly fine with that when testing / discharging into the low tens of watt hours; but after that point all I’d be doing is creating heat needlessly.

Considering reapproacing the task by instead using the programmable load to charge an intermediate battery (though this would need to be done in such a way as to not prevent the battery under test from discharging naturally; part of the entire idea behind building a load is characterizing batteries!)

I also though about looking into Backfeeding, though that’s going to limit the long-term usefulness of the project IMHO.

All and all; I might just continue using the IRF630‘s I’ve got lying around and build a poor-mans-copper heatsink for them (might also use a copper busbar as the common-drain connection at the same time; as it shares the mounting tab). I’ll need to get a hold of some half-decent instrumentation amplifiers to make better use of the MCP4921 DAC’s I’ve got sitting around. (the LM324 is a ‘meh’ quad op-amp; but the variable input offset voltage is killing me down in the lowest bit range. I REALLY should just offset the lm324; but that’s going to require calibration rather than just getting it right to begin with.


I’ve been in a bit of a rut for photography lately. I generally love taking photos, though I might have to swap back down to the lighter camera bag for a few months. (Carrying a Lowepro Flipside 20L does get a little tiring as a daily-carry bag!) The issue I have at the moment is the workflow.

Many years ago; I signed up as a Flickr Pro user. I’ve enjoyed the unlimited storage (legacy pro accounts got unlimited storage) and the simplistic API they provide; but I’ve just never been a huge fan of Flickr. (for no real discernible reason).

I’ve got a Piwigo app running that hosts most of the “private” events I shoot (mostly just events that I want to casually shoot without getting waivers from the participants). That’s worked fairly well; but I think I might look at replacing it with a node.js app (which might be fun to build)

I’m a Lightroom user; I’ve been using Lightroom since version 3 back in ~2007. I do enjoy lightroom; but since upgrading to 4K desktop displays; I’m finding lightroom is just getting clunky to use. I only REALLY use it for crop/exposure corrections these days (and the odd bit of spot removal when circumstances request it)

I MIGHT just look at a straight “shoot and edit in camera” and just auto-dump EVERYTHING to the Piwigo site. That might allow me to simply “shoot and share” without having to deal with going though / removing photos (which is the part I rarely enjoy)

REALLY hoping that I get some willpower to get out and take my gear biking with me. 10-20KM on a bike is a fairly short trip at 30-40KM/H; but really expands what all I shoot.


I’ve got the 650 size unit JUST about ready to fly; it’s waiting on a mount for the FPV camera before I go take it out. I’m planning on 3D printing a mount; but am waiting on parts coming in before I can get the 3D printer going again. (it’s having throat problems!)

Weather has been clearing up early; wasn’t really expecting to be ABLE to get outside and fly at all until late April / early May at the earliest.

I’ve got to charge/discharge one of the 3S 4AH lipo’s a few times to determine; but it’s LOOKING like one cell might just be dead. I’d hate to have to convert the thing to a 2S (though that might be nice to have a spare 2S 4AH pack laying around..) We’ll see what happens with it!

Now it’s time to go shave the ol’ head + shower; then back to work tomorrow!

Random Electronics

Today was a bit of a strange day. Spent some time:

  • Constructing some receiver antennas
  • Sorted out some 50Ω Dummy loads
  • Tested charged and discharging a 4AH; 3S battery
  • Realized I need to build an electronic Dummy load

Radio Stuff

=Antenna testing photos= (excuse the mess!)

Built a pair of 70CM Dipole antennas both tuned fairly well; nice and straight forward to build + tune (make them too long + trim!) Both are modestly low SWR (~1.35:1), and close-enough to 50Ω for my receive only purposes.

Experimented with some multi-element array setups; but without having the space to really clear the immediate area I’m seeing all sorts of intermittent and unexpected behaviour that I’m having to chock up further work on multi-element. Would love to get working on some low-gain Yagi’s, but they’ll have to wait!

My Test setup is FAR from ideal. I’ve got countless things within range of the antennas I’m testing causing reflections and nulls that don’t exist. Without space to really TEST the things; I’m not going to have much luck getting much better than that though.

At the moment; I’m mostly expirementing in the 70cm, 33cm, 23cm and up in the 5cm world. MOSTLY I’m space constrained keeping me out of HF/VHF, But unitl I have a HAM licence to get into TRANSMITTING I’m somewhat less interested in skip and LONG range comms.

I had picked up a RF Explorer and 6GHz Signal Generator a while back (early 2015) as a toss-around Spectrum Analyzer. The Explorer itself is tolerable; from 15-2700MHz the unit features:

  • -105dB ADNL
  • 10 ppm frequency accuracy + stability
  • +/- 1 dBm amplitude stability
  • +/- 6 dBm amplitude accuracy (this has burned me a few times)

The unit I have is the 6G Combo unit; which also features a 4.85-6.1GHz front end as well. I intend to extend some of the software included with the unit and building a simple X-Y RF-transparent gimbal for directional antenna testing at some point. (though going to need a LOT of space for that!)

The 6G SigGen unit is by far the more useful unit of the two; featuring 24 to 6000 MHz output range; 0.5 ppm frequency stability from 0-45°C, internal programmable attenuator, and features frequency sweep and CW output. Shame it doesn’t offer any modulation methods; but for the price of the thing, I can’t complain.

The Antenna analyzer I primarily use is an older JDSU/GenComm GC7105A, which is limited from 25MHz - 3GHz. (there’s a 4GHz software upgrade option, but I can’t find a code online, and the company doesn’t sell them anymore.) The built-in Tracking Generator and directional couplers make it a wonderful blend of a network analyzer + spectrum analyzer + power meter, though the maximum 250 points, the slow sweep speed, and the fairly high noise floor (~-100dB depending on span + range) limit the long-term interest I have in keeping the thing around. I Managed to pick this one up used from a Korean test equipment depo back in ~2014 for an insanely low price (as there’s scratches on the screen, and some physical damage on the tracking-generator N-connector, none of which affects operation).

While working on that; and having just put the “test fixture” together; I figured I should test some of the various inexpensive 50Ω loads I’ve picked up; and build a low-quality “50Ω axial resistor dummy load” to test at the same time. I’ve got:

All of the “rated” loads tested just fine in the 25-3000MHz range that I’m testing them in.

For anyone who isn’t farmiliar with why carbon-film resistors with axial legs aren’t used at high frequency; lead inductance is a pain in the ass!

For my own sanity; I built a 0.5W 50Ω terminator using a pair of parallel 100Ω 0.1% carbon film resistors to verify the inductance vs frequency does in fact lead to a climb in reflection.

Built the test circuit in a hammond metal test box; and threw it in the fixture and confirmed that my understanding was true. (only swept from 140-550MHz; as I was too lazy to recalibrate the analyzse)

RC Batteries

The HobbyKing (rebrand) X120 Battery charger is just terrible. To the point I might just need to crack the thing open and see what I can do about modifying it.

When disipating a load; the display is reading about 55% of the actual total consumption value (assuming the HobbyKing power meter is correct) I’ll have to pass this through an external meter later to see which one is lying. (The power meter I feel like I’ve checked in the past and found it to be within 1%) I’ve always taken the displayed “charging mAh” to be correct; as it generally matches my expected values (if I draw 80% out of a 1.3AH lipo, I expect to see ~1300mAh * 0.80 => ROUGHLY 1AH go back in; and it’s usually pretty close) I suspect the load sense resistor is likely the wrong value; or damaged. (or simply too low that the sense circuit is burying the value in noise.)

Dummy Load

Further to the above topic on Batteries; the “load” portion of the HobbyKing X120 is just useless for large batteries. It takes over 6.6 hours to discharge a 4AH 3S battery; the load being limited to a measly 10 watts. (two orders of magnitude lower than just putting it into an RC copter and FLYING the thing)

I’m really hesitant to simply BUY an electronic load; they’re SUPER expensive and I always hate test equipment software. (though as long as the thing provides data over USB/LAN; that can always be fixed!) I’m tempted to try building one of the Arduino Dummy Load projects; but may go for a somewhat larger (~300 - 500W) range for battery discharge testing.

I’m also somewhat tempted to look at acquiring or building a four quadrant source-measure-unit; Though that get’s even more expensive fast!

All that; and now on with the evening! Might play some computer games; might continue working on the electronic load; might take a look at Brynn’s bike to see what it needs to be temporarily rideable. (She’s convinced she doesn’t want a new one “yet”)

Hexo Startup

Well: this is me getting started with a new blog. Figured I’d give something half-decent looking a try!

This is Hexo; I’ve gotten into Node.js over the last few weeks; and am looking forward to trying my hand at some larger applications!

Going to TRY my damndest to keep a daily journal / blog going here; we’ll see how well I can stick to it!


New Bike

Got a new bike!

My old faithful bike was stolen in the fall. Was saddened to go down to the storage room (of the building I live in) expecting to GO biking with my partner; only to find that it wasn’t were I left it locked to itself. (Shared storage room; need to be able to move stuff around, so can’t lock things to the wall/fixtures)

I had just bought new tires and wheels; converted the flat to a narrow drop bar, converted down to a single speed freewheel, replaced drive chain, replaced the stem, new seat and seatpost, and torn down and regreased all the bearings. (including replacing a SINGLE ball bearing in the head joint)

That bike had seen a good pile of kilometers; LIKELY in the neighborhood of 25,000 - 35,000 Kilometers over the 12 + years I rode it.

I ended up buying a Fuji Tread from the great guys over at Olympia.

Photos of the new bike

I’ve picked up a new set of pedals; going to try going back to riding clipless again this summer (it’s likely been 4-5 years since I’ve ridden clipless) Still on the hunt for a set of biking shoes; having gigantic feet does make finding shoes a pain in the ass.

Hoping to get out and start riding; weather has been pretty shitty for the last two weeks.