Saturday, December 06, 2008

French TGV vs Amtrak's Acela Express

A comparison:

TGV from Paris to Lyon 245mi, 2 hours, ~345 seats, $64

Acela from Boston to NYC 229mi, 3 hours, ~303 seats, $100

Sure, this is not the whole picture. The TGV has coach seats, and the Acela only has business and first class; the Acela shares track with freight trains and the TGV got its own dedicated tracks. Still, is this the best that Amtrak can do? Of course, I don't know much about trains, but it seems like a high speed train should be able to make the trip from Boston to NYC in 90 minutes. It should also be cheaper than an airplane ticket. Right now, Acela costs about as much as airfare!

Here's a thought. A big benefit of trains is the scaling. Why not add a coach car to the Acela train that can fit 80 people (5 in each row). The costs to run the train should be almost the same. Charge $50 per seat, that's $4000 extra revenue per trip, just between Boston and NYC (add more for the DC leg). Assuming they make 40 one-way trips per week, that's more than $36 million in extra revenue per year. How much could an extra car possibly cost? I'm only talking about 20 extra cars in total, and no additional engines. An entire trainset including maintenance and engines only cost $40 million in 1996. Revenue in 2007 was about $400 million from Acela. An extra $36 million per year would be a big increase in revenue (I should actually double that number since the new cars could be used between NYC and DC as well). The new cars would quickly pay for themselves in ticket sales, leading the way for more rail investments like track upgrades to make Acela run at its full 150mph which make the Acela faster and even more attractive to travelers.

This kind of investment works whether or not the government does it or the private sector does it. What's worrisome is that our politicians keep talking about rebuilding the US "infrastructure", but they always say "roads and bridges". The reason to investment money is to get a return on investment. I have a hard time seeing what the return on investment is for rebuilding roads and bridges except that we won't have to repair them later on; it's still the same inefficient, car/truck based infrastructure that we have today, just less bumpy. Perhaps the problem with road maintenance is that there are too many roads; they're too big and too expensive to maintain. I would like to see smarter investment in efficient transportation like rail.

By investing in a passenger rail transportation network that is fast, practical, and affordable, the US can reduce the traffic on the roads, reduce the time people waste driving, reduce the costs to maintain roads, reduce road accidents, and more.

Monday, December 01, 2008

Sqirrels, Bikes, and Acorns

In the last couple years, I've read a number of Internet stories about cyclists who crash because a squirrel got caught in their bicycle's spokes. Often these stories come with a set of photos that show a squirrel's body, between two spokes of a low spoke count wheel.

Some long time cyclists remark that the increase in squirrel caused crashes is a recent phenomena and some blame it on the larger space between spokes in a low spoke count wheel. The logic is that the squirrel can fit through the larger space, get stuck, and lock up the wheel; in a 32-36 spoke wheel, the squirrel would just bounce off the spokes.

This seemed like a reasonable theory to me, except that there's also a story about an opera singer (I think) that crashed due to squirrel on his hybrid bike that probably had a high spoke count wheel. The existence of this crash suggested to me that squirrels can in fact get caught in high spoke count wheels, but it didn't rule out the possibility that squirrels are more likely to get caught in low spoke count wheels.

Today, I read a news story about a smaller than normal acorn crop this year, and how it is affecting squirrel populations. Then I read that this year's squirrel population has boomed.

According to biologists, acorn crops vary each year which causes fluctuations in squirrel populations, but the effect on squirrel population lags by one year, so when oak trees make many acorns one year after making few acorns in the previous year, there are not enough squirrels to eat all the acorns, and more of the acorns grow up to be trees:

"Squirrels eat acorns," he said. "These trees live hundreds of years and they make acorns to produce new trees, not to get fat squirrels. So if they produce regular seed crops, you get a buildup of squirrel population. But if they fluctuate the crop, there are some years where they're loaded with acorns and squirrels can't eat them all, and in other years there are no seeds, and the squirrel population goes down.

So what's my point? Perhaps it's just coincidence that there has been a large squirrel population the last couple years, and low spoke count wheels have become more popular at the same time. Maybe it's the high squirrel numbers that are causing more accidents, and not the wheels themselves. If this is true, perhaps next year after squirrel populations have fallen, we will see fewer squirrel caused crashes, despite even higher numbers of low spoke count wheels on the road.

Sunday, October 26, 2008

Fitting a K3 Focus Screen to a D70/D70s


Over the summer I modified an Nikon K3 focusing screen (meant for an FM3A body) to fit my D70s. This post describes how I modified the focus screen, and how I like the result. According to the Internets, the K3 screen is the same as the normal Katz Eyes screen so you can have the same screen for $30. This project was based on this discussion.

This procedure should work for many different DX format DSLRs such as the Nikon D40, D50, D60, D70, D80, D90, D100, D200, and D300. The only difference might be the dimensions of the screen. You can remove your existing screen and measure it to get the proper dimensions.

Edit: according to comments from Dennis below, the D40, D50, D60, and D70 all share the same screen dimensions. For other cameras, you should measure your original screen before you cut the new one. If you do measure them, please add a comment below with the dimensions; I and other camera hackers would appreciate it!

1 x Nikon K3 focus screen from
1 x Roll of Scotch tape

- Utility Knife
- Calipers
- Pliers
- Steel Ruler
- 6-Inch Bastard File


Here are the target dimensions of the D70s focus screen. These figures are not to scale, and the prisms should be pointing at you, out of your computer monitor:

Here's the dimensions of the K3 focusing screen:

Here is the amount to cut off the K3 screen to get the D70s dimensions (close enough).

Cutting the Screen

Stick a layer of scotch tape on both sides of the focus screen. This should help you avoid scratching the focus screen.

Repeat for each of the three sides with big cuts:

Set your calipers to the amount of screen you want to cut off the current side and lock using the caliper's set screw. Slide the calipers along the side of the screen, scoring the tape. Place a ruler against the screen, and score with the utility knife (make two passes). Flip the screen over, and score the opposite side with the knife and ruler. Press the screen against your table, and grab the portion you're trying to cut off with the pliers; snap it off.

For the forth side, score and snap off the tab that says K3. Use a file to remove the .022".

Clean up the edges using the file, and sand paper.

Remove the tape. If there's tape residue, take the sticky side of a fresh piece of tape, stick it lightly to the residue, and pull it off. This should pull off the residue.

Remove the Existing Screen

Take a junky mini-screwdriver and file a notch in the tip. Turn your camera upside down and find the wire bail near the focus screen, by the lens mount.

Use the screw driver with the notch to press the bail towards the film plane and unhook the wire bail. Gently push the bail away from the focus screen.

Let the screen fall out of the camera. You might find a shim like the brass shim pictured. Keep it because it sets the focus screen in the same focal distance as the film plane. You will reuse this shim if your camera has one.

Rotate the screen so that the split prisms stick out towards the top of the camera (towards the penta-prism). Rotate the screen so that the split prisms are closer to the film plane. Drop the new screen into place. Close the wire bail.

Finally, test the focusing screen. Here's my setup: I focused on one hash mark on the ruler using the split prism and exposed a frame. Then I examined the frame, and the focus screen is dead on.


I've probably taken a few hundred exposures with the new focus screen. It is much easier to focus with the new screen, but it's not as good as my old FM. The viewfinder is smaller, and darker, but the K3 screen helps.

The K3 also shows the correct depth of field which is great. When I press the DOF preview button, I can actually trust what is in the viewfinder.

One thing is weird about this K3 screen compared to my FM and EM bodies is that the images in the split prisms are out of focus when the subject is out of focus. In my old cameras, the images in the split prisms is always in focus. I prefer the way the old screens work because it's hard to see what's going on in the split prism when it goes out of focus.

In the end, it was worth $30 and the hour or two I spent messing cutting the screen and installing it!

Wednesday, May 14, 2008

SLR Focusing Screens and Split Prisms

A couple of years ago, I upgraded from a 35mm film SLR (Nikon FM) to a DSLR with an APS-C sized sensor. It has been a good experience; I've been very happy with with output from the camera, and recently, I sold my film camera equipment.

One thing that does bug me about my new DSLR (a Nikon D70s) is the viewfinder. You can read all about it on the Internet. Everyone complains that the D70 viewfinder is small, dim, and "the worst viewfinder Nikon ever made". I'm ok with that part of it. I have decent vision, and I can see into the viewfinder with my glasses on... The problem I do have is that the focusing screen shows everything in focus all the time, even if it's actually out of focus!

This is most noticeable with a fast lens like an f1.4, but did notice something funny even with an f2.8. I thought I was going crazy, but when I looked through the D70s viewfinder, I had none of the defocus I saw in my FM's viewfinder. Then I learned that some focus screens make a compromise between brightness and the ability to manually focus (which means you can see when it's in focus and when it's out of focus). It looks like the D70s focus screen chooses brightness rather than viewing accuracy. You can confirm this by opening up lens aperture, looking in the viewfinder, taking a shot, and comparing the viewfinder image with the photo. You'll see a big difference in the defocus. I don't believe this is specific to the D70s, I think many new cameras make this trade-off, but I think it's wrong.

I think that half the point of using an SLR is to actually SEE what's going to hit the film. If the focus screen is just going to show me some distorted, version of it, then why wouldn't I just use a rangefinder? Smaller, lighter, quieter, less shake... Plus, I think it does a disservice to beginners who don't know how to use defocus. They won't know what they're missing, because they'll never see it in the viewfinder.

Anyway, now that I know what's going on, I'm going to change the focus screen to one that comes from a manual focusing film body. It has the added bonus of manual focusing aids which I also like.

In my investigation of this problem, I stumbled upon this article
Principle of the Split Image Focusing Aid and the Phase Comparison Autofocus Detector in Single Lens Reflex Cameras, written by Douglas Kerr about split prism focusing aids and AF techniques. It's pretty interesting and explains how these focusing techniques work.