Thursday, February 23, 2012

Broken Brooks B17

Update March 15: I've written up some repair-related information in this post.

Last weekend, I did my first long (i.e. over 100km) ride of the year. The weather was alright at 3°C and some sun with clouds and I had a very nice ride out to the L'Île Perrot. Not so nice was the fact that about 85 km into the ride my Brooks B17 Imperial saddle broke: I heard a quick cracking sound and immediately knew that it had to the failure that I had been afraid of for a long time: one of the rails broke, right behind the clamp.

Brooks B17 Imperial: Still in good shape
This problem is widely known, as is the fact that Brooks claims that it's not their fault (and Grant Petersen agrees). Saddles with the tension too low, mounted too far back, or malicious clamping mechanisms supposedly cause the issue, but that seems to be only part of the story. For one, Brooks saddle rails are slightly thinner than standard rails, according to an article in the German bike magazine Fahrradzukunft, resulting in a significant reduction of strength. Due to their hammock principle --  you sit on a piece of leather that is only connected to the frame at the nose and the rear -- presumably a Brooks creates higher stresses on the frame. And it's easy to see how higher stresses plus weaker rails can in some cases lead to breakage. Of course, it is impossible to tell how frequently Brooks rails break. The company itself probably has a good estimate of it, but is not making any figures public. However, on the internet you come across a fairly large number of incident reports. (It's important to keep in mind that Brooks saddles are pretty popular and therefore the number of incidents has to be put into the context of the large number of Brookses produced.)

Just a little asymmetrical...
My saddle had been ridden for approximately 10 000 km and my weight is somewhere around 85 kg. I think it's unacceptable for a saddle to fail after only such a short period of use -- at my current annual mileage I'd have to buy a new saddle for well over 100 dollars every two years! Some people might say, "well, then why don't you just get a nice saddle from a different manufacturer?" Unfortunately, it's not that easy: first, I do need a leather saddle. Despite being vegan and having some weight-weenie tendencies, I can't really give up on leather in this case: I have a slight scoliosis, resulting in me sitting on the bike crookedly. My leather saddles can adapt to that while plastic saddles can't. Second, while Brooks is not the only manufacturer of leather saddles the number of alternatives is limited and they all have their drawbacks, too:
  • Selle An-Atomica: reportedly, the saddles are very comfortable, but the leather stretches very fast. In addition, they carry a hefty price tag of over 200 dollars shipped.
  • Gilles Berthoud offers a range of nice leather saddles and apparently they can be easily repaired at home. Presumably due to their hefty price tag, they are not very common and it's therefore difficult to find reviews. 
  • Velo Orange sells a range of Brooks knock-offs. They're certainly cheap, but who knows long they will last.
So I'm not really sure what to do. I feel very reluctant to spend money on yet another Brooks, especially since breaking rails are not the only problem one can have with a Brooks (my father-in-law, for example, has a B17 that absolutely does not break in).  For now, I have fixed the saddle by using the advice from the Fahrradzukunft article mentioned above: move the saddle forward, so that the broken part is covered by the clamp, and then tighten the clamp bolt to 25 Nm. We'll see how long that is going to last and how long I'll be willing to put up with the less than ideal saddle position.

Repairing the saddle is another option. Since prior to the defect I was very happy with the B17 Imperial and the leather is probably good for at least another 10 000 km, it would be nice to repair the saddle. While Brooks offers all the necessary spare parts, unsprung saddles like the B17 require you to re-rivet the saddle onto the frame. This is pretty tricky and Brooks recommends to have it done by a professional. Which, of course, is pretty expensive. I haven't talked to any of the local Brooks "dealers of excellence" in Montreal yet, but what I've found online is not that promising: at this place, the cost would be over 100 dollars, and I imagine other places will charge a similar amount.
After the temporary repair

So I'm not sure what to do yet, but I'll keep you updated. If you have experience with Brooks saddles breaking or being repaired, please comment.

Monday, February 13, 2012

Planning for spring: Quebec to Montreal

Last year was a bad bike camping year for us. We only managed to do one overnight trip, and even that one had to be cut short due to health issues. Well, there's always the next year, and the cold of winter is a great time for daydreaming and actual planning of tours to come. Very high on the list of tours to do is a one-week ride from Quebec City to Montreal via Sherbrooke on the Route Verte 1. Using OpenCycleMap and the official guide I have mapped out a 445km route. The idea is to start in Quebec City and then ride back home to Montreal. This has the advantage of allowing us to ride as fast or slow as possible, without the pressure of having to catch a train. It will probably take about a week at about 65km per day.

We'll be taking the train to Quebec City, which is slightly complicated: the only regular train with checked luggage service (and therefore with bicycle transportation) is the Ocean, going from Montreal to Halifax. Unfortunately, it doesn't stop in Quebec proper, but on the other side of the St. Lawrence, in Charny. In addition to that, it gets into town rather late. The alternative to that is to take the "Bike Train," special VIA Rail trains that don't require you to box up your bike. The downside is that these trains only run on the weekend (and maybe only in the summer? Not sure.). So we'll have to see what works best. We'll probably do the fully-loaded camping kind of trip, and judging from a POI file of campgrounds in Quebec that shouldn't be a problem. Since I'm vegan, we'll probably do most of our cooking ourselves and I hope there will be sufficient shopping infrastructure for me not having to eat pasta and tomato sauce each night...

Have you been on this ride or portions of it before? It would be great to get feedback on the route, camping choices, and what the infrastructure is like.

Monday, February 6, 2012

Testing a Penny Stove in the Cold

I am a big fan of DIY alcohol stoves, especially the penny stove variety. I like DIYing in general, and in this case it's also about the fact that a homemade penny stove is actually better than many if not most of the commercial offerings in terms of weight and heat output.
Heineken Penny Stove and stand


In a recent thread on the German Radreise-Forum about equipment choices for an overland trip from Germany to China, the question of the ideal stove came up. Alcohol stoves were discounted by some, based on the availability of fuel in many countries along the way and due to their questionable performance at high altitudes and low temperatures. I was somewhat surprised by the latter objections, as on the penny stove website there are a number of testimonial about the cold weather performance of the stove, including a quotation by Reinhold Messner who supposedly made tea with a jet-based alcohol stove on one of his expeditions to the Himalayas. However, the information there was not very specific -- how much longer does it take to boil water at cold temperatures and how much more fuel does it take -- and my curiosity was sparked. I decided to test it myself.

Air temperature of -9°C
Stove, pot, windscreen
This winter has been pretty mild in Montreal and so it took some time until it was cold enough for a meaningful test. Last Sunday morning had been predicted to be about -10°C and so I put my stove, fuel, and thermometer out on the porch right after waking up and ran the test after about an hour. So this is the experimental setup:

Altitude: 10m
Amount of water: 1l
Air temperature: -9°C
Water temperature at start: 5°C
Pot: 2.2l REI coated aluminium pot with slightly to big stainless steel lid (I couldn't find the original plastic lid)
Stove: Heineken-based penny stove with wind screen
Wind: almost no wind
Fuel temperature: same as air temperature


Burner at full power
I started the timer right before lighting the stove. The lighting itself was unproblematic: I just dropped a burning match into the stove. The total time until reaching a full boil was about 14 minutes and 30 seconds. I didn't measure the fuel consumption, as my kitchen scale recently broke, but the stove still had quite a bit of fuel in it.



5°C water right from the tap

Now what's the bottom line? Is the penny stove an appropriate choice for winter camping or riding through the Himalayas? I don't think you can definitely answer that question based on my test. Unfortunately, I don't own any other types of stoves and therefore couldn't do replications of the test on other equipment. But there are some hints: first of all, the stove does work in the cold. Even if it may not be the fastest, it will still boil your water in a not totally unreasonable amount of time. Second, we can compare the stove's cold weather performance with some of the published data at higher temperatures: in this test, a penny stove boiled 32 oz (0.95 l) of water in 7:50 at 20° degree air temp with 18°C degree water temp and 670m elevation. It would be interesting to calculate the estimated burn time adjusted for the different starting temperatures of the water, but I haven't done that yet.

Full boil
Assuming that it will get colder again here at some point, I'll try to repeat the test at something closer to -20°C and I might also do a test indoors to get a good baseline to compare my results to. If any readers can contribute data on the cold weather performance of alcohol or other stoves, please feel free to comment.

Thursday, January 26, 2012

Experiments with measuring bike geometry

The current weather in Montreal has kept me indoors more than I wanted. Inspired by an article in the current issue of Bicycle Quarterly, I decided to measure the geometry of my Gunnar Roadie.The article in BQ describes how to measure bike geometry from a photograph, and an older article gave instructions how to measure directly on the bike. As a quick summary: measuring on the bike works better for almost all dimensions but trail and fork rake. Trail and fork rake can be measured on a picture, but it's hard to get them accurately.

Measuring angles

For measuring angles, using a digital angle finder or an angle finder app for your smartphone is the best option. According to the interwebs, the angle finder apps are pretty accurate. The angles you need are that of the seat post, the head tube and the top tube. The top tube angle is only important for calculating the effective top tube length, as described below. So if you plan on measuring the top tube length directly, you can skip that step. For most bikes, seat tube and top tube angles will be somewhere between 70 and 75 degrees. Top tube angles used to be 0 on most bikes, but after the sweeping success of compact geometry with sloping top tubes, there now is a large range of angles. In order to get good measurements, measure multiple times and make sure that your angle finder is flush with the frame. Many frames have butted tubing that will be thicker near the intersections of the tubes and you should thus measure in the middle of the tubes.

Measuring lengths

The seat tube is probably the easiest thing to measure. The cranks are somewhat in the way of putting a ruler on there, but it's still easy to get a good measurement. Just make sure to note which measurement you're taking: either center (of bottom bracket) to center (of top tube intersecting the seat tube) or center to top (of seat tube). In addition, you will probably already know this number, as this still is the most common way in which manufacturers size their frames.

The top tube is a little bit more tricky, at least if it is a sloping top tube. In order to measure the effective top tube length you have to put a level on your ruler and then measure the distance from the center of the head tube to the intersection with the seat tube (or the seat post as its extension). I find this pretty difficult to do properly, and fortunately there is another way by employing the powers of trigonometry: You'll need the length of the actual seat tube and the angles of the seat and top tube. Using the laws of sines, you can calculate the effective top tube length x as follows:

x = sin γ ⋅ a sin β = sin ( 180 - α - β ) ⋅ a sin β

Another relevant dimension is the bottom bracket drop, that is the distance from a line between the two wheel dropouts and the center of the bottom bracket. Jan Heine suggests tying a string around the front and rear quick release skewers and then using a ruler to measure the distance from the string to the bottom bracket. You have to correct that measurement with the radius of the skewers.

Next is the front-center distance, i.e. the distance from the front wheel axle to the bottom bracket. The problem is that the fork has to be completely straight in order to get an accurate measurement--something not easily to achieved. Jan Heine suggests measuring on both sides and then making sure to get identical readings, but I found that rather difficult.

Now we get to the tricky part: measuring fork rake and trail. I won't go into the discussions if high trail or low trail are TEHBESTTHING(tm) and instead only show how to best measure trail. Measuring trail directly is almost impossible. Trail is defined as the distance between where the front tire touches the ground and the extension of the steerer axis, as shown in the following figure.
Fork rake r and trail t
Projecting the head tube and the vertical axis through the hub onto the ground accurately just doesn't work properly. [1] Fortunately, you can use the fork rake and wheel diameter to calculate the trail. Measuring the fork rake is easier than measuring trail, but it's still tricky. Bicycle Quarterly use a custom made tool or they measure rake from a digital image. If you only have access to the image of a bike, it's even possible to get all of the above measurements from the image; however, I had a hard time producing a distortion-free image with the lenses at my disposal.

First, take a picture of the front part of the bike. Using your image editor's (I'm using Photoshop) measuring tool, rotate the bike so that the head tube angle matches the value you got from the angle finder. Measure some other tubes, too, to see if and where there is distortion. In addition, draw a circle and see if it is well aligned with the rim. If that shows too much distortion, take another picture and repeat the process. Now you will determine the scale of your image. Draw a vertical line through the hub and measure the distance from the center of the hub to the outer rim. It doesn't really matter which unit of measurement your software uses. On your bike, this distance is a known variable, and dividing the real value through the measured value will give you the scale of your image. With my bike, I measured 1.729 in the image and by dividing 316 (622mm bead seat diameter plus 10mm, then divided by 2) by 1.729 I got a scale of 1:182.8. It is a good idea to repeat the measurement a couple of times on different sections of the rim to once again control for distortion. I measured 1.729, 1.758, 1.651, 1.524, showing that there is still a good deal of distortion. One might be tempted to just take the average of these measurements, but since distortion is not random, this is not advisable. It is preferable to keep taking more pictures, with different lenses and at different angles to achieve a minimum of distortion. For now, we'll just assume that the 1:182.8 scale is correct.



Now we can actually measure the rake: Draw a thin line through the steerer axis, extending it all the way to the end of the fork blades. The draw a line parallel to the steerer axis through the hub. Parallel lines are easiest to get with copying and pasting the first line or drawing a rectangle. Measure the distance between the two parallels. In my case, this was 0.240, yielding a rake of 0.240 * 182.8mm = 43.9 mm. Common values for fork rake are between 40mm and 75mm, with most forks being in the lower range of this spectrum.

Using fork rake, head tube angle, and wheel/tire size you can now use the excellent online trail calculator to get your bike's trail. For Gunnar that would be 54mm trail with a 25-622 tire. Since small measuring errors in any of the measurements can throw the calculation off easily it's hard to say how accurate this value is. It would put Gunnar in the medium trail range (Heine gives 25-80mm as the common range).

To conclude, measuring a bicycle's geometry, either on the bike or from a picture, can be tricky. For many bikes you will be able to get the correct values from the manufacturer (please note, though, that most likely there will be minor aberrations from the values on your frame as artefacts of the manufacturing process). If you want to measure your own frames it might be a good idea to start with a bike with known dimensions to give you an idea of how good your own measurements are.


[1] I wonder if one could use a laser pointer attached to a concave piece of metal that matches the head tube's curvature. I've ordered a laser pointer and tinker around with it a bit.

Review: Innova 700x35C Studded Tires

Back in November, I published a "first look" article on my newly mounted Innova snow tires. After having ridden them for only a few kilometer they seemed like reasonably nice tires: cheap, not too heavy, not too much rolling resistance. Since then I have ridden the tires for approximately 1500 km and my verdict is very different now. The Innovas turned out to be a bad investment for me.
Before...
...and after

The main reason for this are the steel studs. In the initial review I pointed out that steel studs have a reputation for not lasting long, and now I can say for sure: yes, they really don't. I admittedly rode more than usual this winter and a lot of it was not on snow or ice but just bare pavement. But the degree of wear just isn't acceptable. On the pictures you can see that a lot of studs have completely sunken into the tire. In addition, I have lost two studs in the front tire, despite following the manufacturer's advice to break in the tires by careful asphalt riding for the first 50 km1.

The studs on the rear tire have completely sunken into the tread


The problem is exacerbated by the overall low number of studs. 110 studs make for light weight and low rolling resistance, but on a 622mm tire they limit the performance on ice. I got a taste of that when -- somewhat foolishly -- I endeavored to ride over a bike path covered in snow that had been turned into ice by freezing rain. I quickly realized that this probably was not a very good idea but once you're on the ice there is not much you can do. So I kept going at a very slow pace, but after maybe ten meters it was over. Wham! The front wheel slipped and down I went. I landed on my behind and aside from a painful bruise no physical damage was done. The mental damage, a loss of trust in my tire, however, was quite severe. In my post-accident analysis I've come to the conclusion that 110 studs is not enough for riding on uneven ice on a 622mm tire. A 35-622 tire has a circumference of 2168mm, meaning that the studs are about 40mm apart on each side. On uneven ice that's definitely enough to take you down.

Do I have anything good to say about the tires? Well, they have a nice tread which works well in snow. And as long as the studs were still reasonably new they worked well enough for Montreal conditions. Because of the wear issues I can definitely not recommend this tire; however, if Innova would produce the same tire with carbide studs, like the very similar Suomi Tyres W106, it would make a pretty good commuting tire for urban conditions with infrequent ice. For next season, I will probably switch to something like the Schwalbe Marathon Winter with 294 studs. 

Product details:

Innova TR5263 Snow tire
110 steel studs
700x35C (37-622), also available in 700x45C
Price 40-75 USD

1 This is fairly common, however, even with more expensive tires from Schwalbe or Continental.

Wednesday, December 21, 2011

Measuring the SP PD-8 hub

UPDATE January 2015: You can now find a technical drawing with measurements here.
CORRECTION: Please note that an earlier version of the post claimed that the center-to-flange distance is 32mm. I've corrected that to 22mm. Sorry for the confusion.

I'm getting read to build up new wheels with my recently acquired dynamo hubs. In my previous post, I had posted the manufacturer's stated dimensions, but I realized that they only provided the pitch circle diameter and the flange diameter--the latter being irrelevant for determining spoke length--but not the center-to-flange distance and also not the spoke hole diameter.

The best instructions for measuring a hub come from Roger Musson, author of a popular book on wheelbuilding. Fortunately, he has made this part of the book freely available. All you need to take the relevant measurements are calipers and a ruler. Musson suggests to drill a piece of wood, so that you when you insert the hub's axle the wood will be flush with the locknuts. If you're building a lot of wheels this probably makes sense. As an alternative, you can just measure from the end of the axle and then subtract the easily measurable distance between axle end and locknut. Then use the ruler to measure the distance between the wood/axle end and the inner side of the lower flange (what Musson denotes as x). Then turn around the hub and repeat the measurement on the other side, giving you the y value. With the SP disc hub that's a bit tricky, as the diameters of the two flanges are different. Musson's suggestion is to use a small piece of cardboard, mark it, and then measure on the cardboard. I was in a bit of hurry and thus just eyeballed it with the ruler after seeing that x and y would likely be the same for this hub. X and y are not the actual values needed for calculating spoke length, but they allow you to calculate the desired value by subtraction from half of the over-the-locknut dimension of the hub, z in Musson's system. On modern front hubs for regular bikes you don't really have to measure this, as it is a standardized dimension at 100mm. For the PD-8, x and y equal 28mm (32mm-4mm for the distance from axle end to locknut), yielding a center-to-flange dimension of 22mm on each side.



The flange diameters can directly be measured with calipers. Just make sure to measure from the centers of the spoke holes. The PD-8 comes out to 52mm and 58mm. The spoke hole diameter is a bit tricky to measure: The inside jaws of regular calipers tend to be too big for such small holes. I measured 2.7mm, but as most hubs use 2.5mm holes I'm assuming this is a measurement problem.

Now you can plug your values into the spoke length calculator of your choice and voilá! I recommend Karl Stoerzinger's Freespoke or the DT Swiss calculator.






Wednesday, December 14, 2011

First impressions: SP Hub Generators PD-8 and PV-8 (updated)

Wouter Scholten's website is a great resource for all hub generator lighting issues. He extensively tests almost all models of hubs and lights on the market in real world settings, and he is also unusually well-informed with regard to the introduction of new products. A while ago, he wrote about a new player on the hub generator market, Shutter Precision Ltd. from Taiwan. As the name implies, they started as a supplier for digital cameras and then later moved into bike component manufacturing. While the company itself has little name recognition, they are (presumably) the suppliers for two better known companies, Velo Orange[1] and Supernova.

After having introduced their switchable dynamo (in my opinion a solution in search of a problem), at the 2011 Taipei Cycle Show SP announced what would be the lightest generator hub, the SD-8 (disc brake)/SV-8 (rim brake). These two hubs, however, only are rated to a power output of 2.4W in a 622mm wheel, making them not street-legal according to the German StVZO requirements. However, they also announced a slightly more heavy model rated at 3W, the PD-8 and the PV-8. The technical specs and the prototype tests by Wouter sounded promising, and when Wouter's shop, Gambiet Fietsen, offered them for pre-order in September I went in for two generators, one disc, one rim. The disc one will replace my Shimano DH3-N80 on the Cross-Check which I unfortunately messed up while trying to re-grease the bearings, and the PV-8 will be part of a second front wheel for Gunnar, finally allowing me to turn him into a real brevet bike.

The hubs arrived yesterday, at a total price of 180 Euro shipped. The actual wheelbuilding will have to wait for a little while longer, but I wanted to share some pictures and impressions.
PV-8 in its original packaging. I'm not sure if all the color and spoke count variants are or will be actually available.

The hub looks quite similar to the current SON 28...

...but it does have the same terminal as Shimano hubs

Here you can see the German K-certification wave, making the generator street legal in Germany



The weight pretty much exactly matches the claimed weight.

Both the disc and non-disc version of the hub match their claimed weight of 390 g and 410 g, compared to 440/460 for the new SON 28, 490 g for the Shimano DH3N80, and 399 g for the Supernova Infinity S (center lock disc mounts). Only the SONdelux at 390 g (rim) and 395 g (disc) can match this, but at a significantly higher price point.

SP claims that the new generators do not only excel in weight, but are also "the most efficient" generators (this refers to the 2.4W ones) and their product sheets has a couple of graphs that show the generators to be more efficient than the SONdelux. I'm somewhat skeptical about those claims and am very much looking forward to some independent testing.

Measuring the dimensions of the hubs is a bit tricky with the tools I have and I've only been able to find the manufacturer's numbers for the 2.4W version. Probably they're going to be the same for the 3W, but I'll confirm this with Wouter or measure myself. The bearings are sealed cartridges and I'm not sure if they are user-replaceable.

I think that's all I can say for now. Once I have the wheels built up and ridden them a bit, I will post an update.

Update 2011-12-16:
In the meantime I've found the website of SP's distributor, and their catalog lists the relevant dimensions as follows:
PV-8 PD-8 (non-disc/disc side)
Spoke P.C.D. 52mm  52mm/58mm
Flange diameter 60mm 60mm/66mm

They also have "replacement instructions" on their site that might contain instructions on how to replace the bearings. Unfortunately, the document is corrupted, though, and can't be opened. I've sent them an e-mail and we'll see if I hear back.

[1] If I remember correctly, the Velo Orange switchable dynohub got some bad reviews. At any rate, they no longer seem to sell it.
[2] There is also a 2.4W version, called SD-8/SV-8, suitable for those on small-wheeled bikes or those who don't care about the extra power/street legality in Germany.