Thursday 18 November 2010

Testing and recording bicycle bells

In most countries of the world it's a legal requirement to have a method of warning people that you are approaching on a bike. In some countries, such as the UK, the human voice is considered to be adequate. In other countries, such as the Netherlands, a bell is a legal requirement on a bicycle.

I've got a bunch of bicycle bells here at the moment as we've wanted for some time to have these available in the webshop, and as with other things in the shop, I've specifically bought in those that we like. However, rather than just listing the bells, I decided we should do this well. Let's test them to find out how loud they are, and also have sound samples so that people can hear what the bells sound like. And here are the results, ranked in order of how loud they are from one metre away, and with recordings of each bell. Click on their pictures if you'd like to buy them:
Widek 80 mm Ding Dong bell (€9.88)
106 dBA
"Ping" bell (€2.50)
100 dBA
Gazelle handlebar grip bell (€11.90)
98 dBA
Spanninga "Turning" bell (€3.09)
96 dBA
Classic Brass bell (€5.83)
96 dBA

If you have space for it on your handlebars, the Widek Ding Dong bell gives the most sound. It's rather subjective, but to me it also gives by far the best sound of the bells. Just listen to that sustain. There are now cheaper copies available, many of which look very similar, but they do not have the same resonant sound. The Ping bell, which is smallest and weighs just 20 g, is best for those with limited handlebar space or who are concerned about weight. It's surprisingly loud, but note that because it has little sustain it's not really more effective than a traditional bell.

Free Bells !
We have a freely downloadable Android application which includes all these bell sounds.

How I tested and recorded the bells

For the geeks amongst us, this is how it was done.

I used my old Radioshack sound-level meter to find out how loud the bells were. Each bell was held one metre from the SPL meter. These tests were all done in the same way, so they are directly comparable with one another. However, because this was done indoors and with reflective surfaces (the wall and table) nearby, you would get different figures if you measured out in the open. Each bell was rung a few times in succession, as you might if you thought you were about to collide with someone. I think it's reasonably representative of what you might do if you were really trying to get someone's attention. With a single ring, the slow rise-time of the analogue meter missed the peaks. As a result, this helped the ping bell result more than the others. A single "ping" doesn't really sound louder than some of the quieter bells. The loudest bell subjectively with a single ring was definitely the ding dong bell.

In any case, the peaks were quite a lot louder than the figures here. By the end of the test, it was not only the bells, but also my ears that were ringing.

A Sony WMD6C "pro" walkman was used to make the recordings. I'm not a big fan of Sony, but this was always a splendid product. I found the quietest room in the house, and used a long microphone lead so that I could be at the opposite end of the room from the recorder. I did in in this way for two reasons. First because I wanted to get away from the fan noise of the computer - the walkman is much quieter and easily portable - and secondly because I liked the idea of playing with the walkman again. It's a great piece of equipment which I now rarely use. Yes, an analogue recording on a cassette, but not any old cassette deck. The recordings give a pretty good idea of the sound of each bell, though reproducing the sound at a level anything like so loud as in real life is actually quite difficult to do.

I'm not going to pretend I don't care about technical issues, around cycling or anything else. I don't think it's cool not to know stuff. I'm proud to be a geek...

Monday 25 October 2010

Improving velomobile braking efficiency

My Mango velomobile has been a really wonderful thing to ride, but for one thing: the brakes. The Sturmey Archer drum brakes never seemed to quite have the stopping power required. An attempt to stop suddenly was met with a slightly too gradual deceleration and locking the wheels to skid was impossible.

Almost all velomobiles use these same brakes. The Alleweder, Go-One, Milan, Quest, Strada and WAW all also have the same brakes. So do many open recumbent trikes. It's been obvious to me for a while that there was a lot of variation in the performance of the brakes amongst different bikes fitted with these drum brakes. The problem was clearly not inherent in the Sturmey Archer drums themselves as at their best they work very well. I first assumed that the problem was variation in the brakes themselves, and tried changing them, but it made little if any difference. Other people reported improvements by changing the brake cable outers. As a result of this I took a close look at how the cables behaved on my Mango when I pulled the brake lever.

The brake cables on the Mango take the most direct route they could under the circumstances, but relative to the average upright bike it is still quite a circuitous route with a lot of relatively tight bends. The curvature of the cables changes quite noticeably when you pull on the brake lever, and that represents lost braking effort.

Also, these tight bends lead to more friction than normal between the inner and outer cables.

Together, these two things add up to an effect where the inner cable moves a greater distance at the brake lever than at the brake. Much of your effort in pulling on the lever goes into compressing the cable and overcoming friction rather than applying the brakes. They also mean that the springs in the brakes themselves are only just about strong enough to pull the cable back when you let go of the brake lever.

Potentially this could happen on any bike with any type of mechanical brakes, but because cable runs are typically less direct on recumbents than upright bikes, and less direct again on velomobiles vs. open recumbents, I suspect it is more frequently a problem on velomobiles, recumbent trikes and recumbent bikes than on uprights.

A couple of weeks ago I made a change to the brake cables on my Mango which significantly improved this situation. I can stop very quickly with my front wheels skidding. This is how it was achieved:

Good cables are the key to the transformation. I used low compression outer cables to solve the first problem, combined with Teflon coated inner cables for low friction.

Note that while gear shifter outer cable offers low compression it should never be used for brakes. It is constructed differently from brake cable and is not strong in compression. For this reason, it can fail under hard braking, resulting in little or no braking just when you need it.

A tip given to me by a colleague at the Ligfietsgarage a year ago was to remove the rough swarf on new brake cables so that they can rotate more easily within the brake levers. This reduces the chance of the brake cables breaking in the levers, as they often do.

Before and after use of the file.

And the other side...

The wheels are removed with a 5 mm Allen key, giving access to the brake mechanisms. It is a good idea to lubricate the pivots in the mechanisms, but make sure that no oil or grease gets onto the braking surfaces as this will drastically reduce the power of your brakes.

Low compressibility outer cable, with a low friction liner. To make a good job of cutting this you need the proper tool. For the Mango or Quest, the new outer cable needs to be 1.03 m in length.

The easiest way of making sure that the new outer cable takes the same route as the old, including through the hard to reach parts under the bridge, is to use the old inner cable as a guide.

I used a little white grease on the adjusters for the brakes, and also an additional nut to fit against the brake lever for additional rigidity at the lever end. These adjustments should be screwed all the way in. The wheels should then be re-fitted and the brakes adjusted so that they are just free when the brake lever is not pulled. This means that later adjustments can be made, even while riding, if the brake performance drops a little in the first few km after fitting the new cables.

Finally, refit tyres (in this case I have fitted Schwalbe Marathons ready for winter) and pump the tyres up to pressure (the Schwalbe pressure gauge gives an accurate reading of the pressure in your tyres).

Afterwards, it is much easier to pull on the brake lever, with much less obvious friction, the brake lever returns much quicker to being fully "off" when I release it, and the braking is very much improved. I can now lock the front wheels and skid to a halt, stopping in a much shorter distance than I could stop previously. This removes a slight doubt I had with the Mango over whether I would stop in an emergency situation. Here's the evidence:

A slow motion view of the skid. Note how in every frame the label on the tyre is in the same place:

The total list of parts that you need for doing this job is as follows:
All the parts can be bought individually in our webshop, or we can send you a kit of the parts you need with the brake outer cable already cut to the correct length, at a slightly lower price than buying the parts individually.

Total cost for a Mango or Quest (1.03 m outer cables) is €16.50 + postage (+ 19% tax if you live within the EU). If you've a different bike/trike/velomobile with weak brakes, please check what lengths of cable you require and let me know. I'll add them to the menu.