Josef “Jeff” Sipek

OH-LCD

This post is part of a series named “Europe 2017” where I share photos from my adventures in Europe during the summer 2017.

When I attended the Kaivopuisto Air Show in early June last year, I learned about the existence of the Finnish Aviation Museum. It took me a month and a half, but eventually I found a free day to go check it out.

The museum itself is packed with all sorts of aircraft on static display. While they were interesting (and I certainly took plenty of photos of them), they aren’t what this post is about. This post is about Lokki—a retired Wikipedia article: DC-3 (registration OH-LCD) on display outside of the museum.

As luck would have it, the folks from the DC Association were there that day trying to see if they could start up Lokki’s engines—after 12 years of inactivity. After a lot of preparation, they managed to start them!

Without further ado, here are a few photos of Lokki (more photos can be found in the gallery).

Wikipedia article: Aero OY was the original name of Finnair:

One of the mechanics working on the left engine:

One of the people from the DC Association, seeing that I was obviously excited about the plane, asked me if I’d like to climb inside. I said yes, of course.

The inside was pretty bare-bones (which is to be expected of a static display that’s normally closed to public). I took a couple of photos inside, but most weren’t that interesting.

Throttle quadrant (note: most of the instrument panel was removed long ago):

It runs!

The livery is pretty simple—polished aluminum with dark blue lettering and a stripe:

I’m not really sure why they wanted to see if they could start the engines, but I’m happy that it worked out. Radial engines just have a unique roar to them.

Anyway, that’s it about Lokki. Hopefully I’ll get around to post processing the photos from the museum itself soon.

Kaivopuisto Air Show 2017

This post is part of a series named “Europe 2017” where I share photos from my adventures in Europe during the summer 2017.

In early June 2017, we attended an air show in Wikipedia article: Kaivopuisto. Unfortunately, we found out about it last minute, and so we missed the beginning which included a Finnair Airbus A350 flyby. Pity.

The show included a number of trainers and combat aircraft performing various maneuvers. Here are the highlights (for more photos visit the gallery).

Wikipedia article: Red Arrows:

A seagull joining in:

Wikipedia article: Finnish Coast Guard’s Wikipedia article: Turva nearby with Wikipedia article: Suomenlinna visible behind it:

Wikipedia article: Eurofighter Typhoon:

Wikipedia article: Saab 35 Draken:

Wikipedia article: Saab Gripen:

During one of the passes, I took a burst of images and then assembled them into a Southwest 737 “Airportrait”-style image.

Finnish Air Force Wikipedia article: F-18 Hornet:

A Finnish aerobatics team Wikipedia article: Midnight Hawks flying Wikipedia article: BAE Systems Hawk:

Even though this post has more photos than I typically share, there are many more in the gallery. So, if you are into airplanes, I suggest you peruse it.

2017-03-23

The million dollar engineering problem — Scaling infrastructure in the cloud is easy, so it’s easy to fall into the trap of scaling infrastructure instead of improving efficiency.

Some Notes on the “Who wrote Linux” Kerfuffle

The Ghosts of Internet Time

How a personal project became an exhibition of the most beautifully photographed and detailed bugs you ever saw — Amazing photos of various bugs.

Calculator for Field of View of a Camera and Lens

The Megaprocessor — A microprocessor built from discrete transistors.

Why Pascal is Not My Favorite Programming Language

EAA Video — An assortment of EAA produced videos related to just about anything aircraft related (from homebuilding to aerobatics to history).

The Unreasonable Effectiveness of Recurrent Neural Networks

Flying around Mount Monadnock

Last week I planned on doing a nice cross country flight from Wikipedia article: Fitchburg. Inspired by Garrett Fisher’s photos, I took my camera and the 70-200mm lens with me hoping to get a couple of nice photos of the landscapes in New Hampshire.

Sadly, after taking off from KFIT I found out that not only was there the stiff wind that was forecasted (that’s fine) but the air was sufficiently bumpy that it wouldn’t have been a fun flight. On top of that, the ADS-B unit was having problems acquiring a GPS signal. (Supposedly, the firmware sometimes gets into a funny state like this. The good news is that there is a firmware update available that should address this.) I contacted KASH tower to check if they could see my transponder—they did, so I didn’t have to worry about being totally invisible.

Since I was already off the ground, I decided to do some nearby sightseeing, landing practice, and playing with the Garmin GNS 430 GPS.

First, I headed northwest toward Wikipedia article: Mount Monadnock. While I have seen it in the distance several times before, I never got to see it up close, so this seemed like a worthwhile destination.

As I approached it, I ended up taking out my camera and getting a couple of photos of the hills and mountains in New Hampshire. It was interesting how the the view to the north (deeper into New Hampshire) is hilly, but the view more east (and certainly south) is flatter. (Both taken near Mount Monadnock.)

While the visibility was more than good enough for flying, it didn’t work out that well for photography. In all of the photos, the landscape far away ended up being heavily blue-tinted. No amount of playing around with white balance adjustment in Lightroom was able to correct it. (Either the background was too blue, or the foreground was too yellow/brown.) That’s why all of these photos are black and white.

I made a full turn around Monadnock, taking a number of shots but this one is my favorite:

Once done with Monadnock, I headed south to the Wikipedia article: Quabbin Reservoir in Massachusetts. This is a view toward the south from near its north end:

At this point I started heading to KORH to do some landing practice. Since I was plenty busy, there are no photos.

I’ve never been to this airport before and landing at new airports is always fun. The first interesting thing about it is that it is situated on a hill. While most airports around here are at 200-400 feet MSL, this one is at 1000 feet. The westerly wind favored runway 29 which meant I got to see a second interesting aspect of this airport. The beginning of runway 29 is on the edge of the hill. That by itself doesn’t sound very interesting, but consider that the runway is at 1000 feet while the bottom of the hill (a mere 0.9 km away) is at 500 feet MSL. That is approximately a 17% grade. So, as you approach the runway, at first it looks like you are way too high but the ground comes up significantly faster than normal.

I am still hoping to do my originally planned cross country flight at some point. Rest assured that I will blahg about it.

Plane-spotting in Manchester, NH

Last weekend I got to drive to Wikipedia article: Manchester, so I used the opportunity to kill some time near the airport by watching planes and taking photos (gallery).

The winds were coming from the south, so runway 17 was in use. I think those are the best plane spotting conditions at KMHT.

It is relatively easy to watch aircraft depart and fly directly overhead:

Unlike all my previous plane spotting, this time I tried something new—inspired by Mike Kelly’s Airportraits, I decided to try to make some composite images. Here is a Southwest Boeing 737 sporting one of the Wikipedia article: special liveries:

It was certainly an interesting experience.

At first I thought that I would be able to use the 7 frames/second that the D750 can do for the whole departure, but it turns out that the planes move far too slowly, so the camera buffer filled up way too soon and the frame rate became somewhat erratic. What mostly ended up working was switching to 3 frames/second and taking bursts. Next time, aiming for about 2 frames/second should give me enough images to work with.

Even though I used a tripod, I expected that I would have to align the images to remove the minor misalignment between images due to the vibration from the rather strong wind and my hand depressing the shutter. It turns out that the misalignment (of approximately 10 pixels) was minor enough that it did not change the final image.

Here’s an American Airlines commuter taking off from runway 17. (I repositioned to get a less head-on photo as well.)

For those curious, I post processed each of the images in Lightroom, exported them as TIFFs, and then used GIMP to do the layering and masking. Finally, I exported the final image and imported it back into Lightroom for safekeeping.

As a final treat, as I was packing up a US Army Gulfstream took off:

As far as I can tell, they use this one to transport VIPs. I wonder who was on board…

2015-05-29

I’m going to try something new. Instead of sharing individual links per post as I come across them, I’m going to try to dump them whenever I have enough of them. It does mean that some of these links aren’t as “hot off the press”. Here’s the first batch.

How We’re Predicting AI — or Failing To

How Typography Shapes Our Perception Of Truth

Bitcoin mining on a 55 year old IBM 1401 mainframe: 80 seconds per hash

What is the difference between an “aggregate” and other kinds of “modified versions”?

SourceForge grabs GIMP for Windows’ account, wraps installer in bundle-pushing adware

Wikipedia article: Speed tape looks like duct tape but isn’t.

Concorde

I just came across someone’s blog post full of cool Concorde photos.

It’s a hard choice, but my favorite photo is:

Concorde

(The black and white photography and the unusual camera position in these images remind me of the Wernher von Braun photo I posted years ago.)

Private Pilot, Honeymooning, etc.

Early September was a pretty busy time for me. First, I got my private pilot certificate. Then, three days later, Holly and I got married. We used this as an excuse to take four weeks off and have a nice long honeymoon in Europe (mostly in Prague).

Our flight to Prague (LKPR) had a layover at KJFK. While waiting at the gate at KDTW, I decided to talk to the pilots. They said I should stop by and say hi after we land at JFK. So I did. Holly tagged along.

A little jealous about the left seat

I am impressed with the types of displays they use. Even with direct sunlight you can easily read them.

After about a week in Prague, we rented a plane (a 1982 Cessna 172P) with an instructor and flew around Czech Republic looking at the castles.

OK-TUR

I did all the flying, but I let the instructor do all the radio work, and since he was way more familiar with the area he ended up acting sort of like a tour guide. Holly sat behind me and had a blast with the cameras. The flight took us over Wikipedia article: Bezděz, Wikipedia article: Ještěd, Wikipedia article: Bohemian Paradise, and Wikipedia article: Jičín where we stopped for tea. Then we took off again, and headed south over Wikipedia article: Konopiště, Wikipedia article: Karlštejn, and Wikipedia article: Křivoklát. Overall, I logged 3.1 hours in European airspace.

First Solo Cross-Country

A week ago (June 15), I went on my first solo cross country flight. The plan was to fly KARBKMBSKAMN → KARB. In case you don’t happen to have the Detroit sectional chart in front of you, this might help you visualize the scope of the flight.

leg distance time
KARB → KMBS 79 nm 47 min
KMBS → KAMN 29 nm 20 min
KAMN → KARB 79 nm 46 min
Total 187 nm 113 min

Here’s the ground track (as recorded by the G1000) along with red dots for each of my checkpoints and a pink line connecting them. (Sadly, there’s no convenient zoom level that covers the entire track without excessive waste.)

ground track

As you can see, I didn’t quite overfly all the checkpoints. In my defense, the forecast winds were about 40 degrees off from reality during the first half of the flight. :)

Let’s examine each leg separately.

KARB → KMBS

ground track

My checkpoint by I-69 (southwest of Flint) was supposed to be a I-69 and Pontiac VORTAC (PSI) radial 311 intersection. However when I called up the FSS briefer, I found out that it was out of service. Thankfully, Salem VORTAC (SVM) is very close so I just used its radial 339 instead. Next time I’m using a VOR for any part of my planning, I’m going to check for any NOTAMs before I make it part of my plan — redoing portions of the plan is tedious and not fun.

On the way to Saginaw, I was planning to go at 3500. (Yes, I know, it is a westerly direction and the rule (FAR 91.159) says even thousand + 500, but the clouds were not high enough to fly at 4500 and the rule only applies 3000 AGL and above — the ground around these parts is 700-1000 feet MSL.)

altitude

Right when I entered the downwind for runway 23, the tower cleared me to land. My clearance was quickly followed by the tower instructing a commuter jet to hold short of 23 because of landing traffic — me! Somehow, it is very satisfying to see a real plane (CRJ-200) have to wait for little ol’ me to land. (FlightAware tells me that it was FLG3903 flight to KDTW.)

While I was on taxiway C, they got cleared to take off. I couldn’t help it but to snap a photo.

CRJ-200

It was a pretty slow day for Saginaw. The whole time I was on the radio with Saginaw approach, I got to hear maybe 5 planes total. The tower was even less busy. There were no planes around except for me and the commuter jet.

KMBS → KAMN

This leg of the flight was the hardest. First of all, it was only 29 nm. This equated to about 25 minutes of flying. The first four-ish and the last five-ish were spent climbing and descending, so really there was about 15 minutes of cruising. Not much time to begin with. I flew this leg by following the MBS VOR radial 248. My one and only checkpoint on this leg was mid way — the beginning of a wind turbine farm. It took about 2 minutes longer to get there than planned, but the wind turbines were easy to see from distance so no problems there.

ground track

Following the VOR wasn’t difficult, but you can see in the ground track that I was meandering across it. As expected, it got easier the farther away from the station I got. Here’s the plot of the CDI deflection for this leg. The CSV file says that the units are “fsd” — I have no idea what that means.

CDI deflection

I can’t really draw any conclusions because…well, I don’t know what the graph is telling me. Sure, it seems to get closer and closer to zero (which I assume is a good thing), but I can’t honestly say that I understand what the graph is saying.

The most difficult part was trying to stay at 2500 feet. For whatever reason, it felt like I was flying in sizable thermals. Since there were no thunderstorms in the area, I flew on fighting the updrafts. That was the difficult part. I suspect the wind turbines were built there because the area is windy.

altitude

KAMN is a decent size airport. Two plenty long runways for a 172 even on a hot day (5004x75 feet and 3197x75 feet). I didn’t stop by the FBO, so I have no idea how they are. I did not notice anyone else around during the couple of minutes I spent on the ground taxiing and getting ready for the next leg. Maybe it was just the overcast that made people stay indoors. Oh well. It is a nice airport, and I wouldn’t mind stopping there in the future if the need arose.

KAMN → KARB

Flying back to Ann Arbor was the easy part of the trip. The air calmed down enough that once trimmed, the plane more or less stayed at 3500 feet.

altitude

It apparently was a slow day for Lansing approach as well, as I got to hear a controller chatting with a pilot of a skydiving plane about how fast the skydivers fell to the ground. Sadly, I didn’t get to hear the end of the conversation since the controller told me to contact Detroit approach.

As far as the ground track is concerned, you can see two places where I stopped flying current heading and instead flew toward the next checkpoint visually. The first instance is a few miles north of KOZW. I spotted the airport, and since I knew I was supposed to overfly it, I turned to it and flew right over it. The second instance is by Whitmore Lake — there I looked into the distance and saw Ann Arbor. Knowing that the airport is on the south side, I just headed right toward it ignoring the planned heading. As I mentioned before in both cases, the planned course was slightly off because the winds weren’t quite like the forecast said they would be.

ground track

You can’t tell from the rather low resolution of the map, but I got to fly right over the Wikipedia article: Michigan stadium. Sadly, I was a bit too busy flying the plane to take a photo of the field below me.

Next

With one solo cross country out of the way, I’m still trying to figure out where I want to go next. Currently, I am considering one of these flights (in no particular order):

path distance time
KARB KGRR KMOP KARB 239 nm 2h19m
KARB KBIV KJXN KARB 220 nm 2h08m
KARB KFWA KTOL KARB 210 nm 2h03m
KARB CRUXX KFWA KTOL KARB 210 nm 2h06m
KARB LFD KFWA KTOL KARB 225 nm 2h12m
KARB KAZO KOEB KTOL KARB 207 nm 2h01m
KARB KMBS KGRR KARB 243 nm 2h21m
KARB KGRR KEKM KARB 266 nm 2h40m

Plotting G1000 EGT

It would seem that my two recent posts are getting noticed. On one of them, someone asked for the EGT R code I used.

After I get the CSV file of the SD card, I first clean it up. Currently, I just do it manually using Vim, but in the future I will probably script it. It turns out that Garmin decided to put a header of sorts at the beginning of each CSV. The header includes version and part numbers. I delete it. The next line appears to have units for each of the columns. I delete it as well. The remainder of the file is an almost normal CSV. I say almost normal, because there’s an inordinate number of spaces around the values and commas. I use the power of Vim to remove all the spaces in the whole file by using :%s/ //g. Then I save and quit.

Now that I have a pretty standard looking CSV, I let R do its thing.

> data <- read.csv("munged.csv")
> names(data)
 [1] "LclDate"   "LclTime"   "UTCOfst"   "AtvWpt"    "Latitude"  "Longitude"
 [7] "AltB"      "BaroA"     "AltMSL"    "OAT"       "IAS"       "GndSpd"   
[13] "VSpd"      "Pitch"     "Roll"      "LatAc"     "NormAc"    "HDG"      
[19] "TRK"       "volt1"     "volt2"     "amp1"      "amp2"      "FQtyL"    
[25] "FQtyR"     "E1FFlow"   "E1OilT"    "E1OilP"    "E1RPM"     "E1CHT1"   
[31] "E1CHT2"    "E1CHT3"    "E1CHT4"    "E1EGT1"    "E1EGT2"    "E1EGT3"   
[37] "E1EGT4"    "AltGPS"    "TAS"       "HSIS"      "CRS"       "NAV1"     
[43] "NAV2"      "COM1"      "COM2"      "HCDI"      "VCDI"      "WndSpd"   
[49] "WndDr"     "WptDst"    "WptBrg"    "MagVar"    "AfcsOn"    "RollM"    
[55] "PitchM"    "RollC"     "PichC"     "VSpdG"     "GPSfix"    "HAL"      
[61] "VAL"       "HPLwas"    "HPLfd"     "VPLwas"   

As you can see, there are lots of columns. Before doing any plotting, I like to convert the LclDate, LclTime, and UTCOfst columns into a single Time column. I also get rid of the three individual columns.

> data$Time <- as.POSIXct(paste(data$LclDate, data$LclTime, data$UTCOfst))
> data$LclDate <- NULL
> data$LclTime <- NULL
> data$UTCOfst <- NULL

Now, let’s focus on the EGT values — E1EGT1 through E1EGT4. E1 refers to the first engine (the 172 has only one), I suspect that a G1000 on a twin would have E1 and E2 values. I use the ggplot2 R package to do my graphing. I could pick colors for each of the four EGT lines, but I’m way too lazy and the color selection would not look anywhere near as nice as it should. (Note, if you have only two values to plot, R will use a red-ish and a blue-ish/green-ish color for the lines. Not exactly the smartest selection if your audience may include someone color-blind.) So, instead I let R do the hard work for me. First, I make a new data.frame that contains the time and the EGT values.

> tmp <- data.frame(Time=data$Time, C1=data$E1EGT1, C2=data$E1EGT2,
                    C3=data$E1EGT3, C4=data$E1EGT4)
> head(tmp)
                 Time      C1      C2      C3      C4
1 2013-06-01 14:24:54 1029.81 1016.49 1019.08 1098.67
2 2013-06-01 14:24:54 1029.81 1016.49 1019.08 1098.67
3 2013-06-01 14:24:55 1030.94 1017.57 1019.88 1095.38
4 2013-06-01 14:24:56 1031.92 1019.05 1022.81 1095.84
5 2013-06-01 14:24:57 1033.16 1020.23 1022.82 1092.38
6 2013-06-01 14:24:58 1034.54 1022.33 1023.72 1085.82

Then I use the reshape2 package to reorganize the data.

> library(reshape2)
> tmp <- melt(tmp, "Time", variable.name="Cylinder")
> head(tmp)
                 Time Cylinder   value
1 2013-06-01 14:24:54       C1 1029.81
2 2013-06-01 14:24:54       C1 1029.81
3 2013-06-01 14:24:55       C1 1030.94
4 2013-06-01 14:24:56       C1 1031.92
5 2013-06-01 14:24:57       C1 1033.16
6 2013-06-01 14:24:58       C1 1034.54

The melt function takes a data.frame along with a name of a column (I specified “Time”), and reshapes the data.frame. For each row, in the original data.frame, it takes all the columns not specified (e.g., not Time), and produces a row for each with a variable name being the column name and the value being that column’s value in the original row. Here’s a small example:

> df <- data.frame(x=c(1,2,3),y=c(4,5,6),z=c(7,8,9))
> df
  x y z
1 1 4 7
2 2 5 8
3 3 6 9
> melt(df, "x")
  x variable value
1 1        y     4
2 2        y     5
3 3        y     6
4 1        z     7
5 2        z     8
6 3        z     9

As you can see, the x values got duplicated since there were two other columns. Anyway, the one difference in my call to melt is the variable.name argument. I don’t want my variable name column to be called “variable” — I want it to be called “Cylinder.”

At this point, the data is ready to be plotted.

> library(ggplot2)
> p <- ggplot(tmp)
> p <- p + ggtitle("Exhaust Gas Temperature")
> p <- p + ylab(expression(Temperature~(degree*F)))
> p <- p + geom_line(aes(x=Time, y=value, color=Cylinder))
> print(p)

That’s all there is to it! There may be a better way to do it, but this works for me. I use the same approach to plot the different altitude numbers, the speeds (TAS, IAS, GS), CHT, and fuel quantity.

You can download an R script with the above code here.

Powered by blahgd