Bicycle Seismograph Measures The Streets

Riding the streets of the Netherlands on a bicycle is a silky-smooth experience compared to doing the same on those of Germany. So says [Kati Hyyppä], who made the move with her trusty Dutch bike. The experience led her to record the uneven cobblestones and broken asphalt of the German roads on a home-made seismograph, a paper chart recorder driven by the bike’s motion and recorded upon by a pen free to vibrate as it passed over any bumps.

The resulting instrument is a wooden frame with a ballpoint pen mounted in a sliding holder weighted with some washers and kept under some tension with elastic bands. The paper roll is driven from the motion of the bike by the drive from a mechanical speedometer feeding a set of FischerTechnik gears, and the whole unit is suspended from the crossbar.

You can see it in action in the video below the break, and if you would like to build one yourself she has put the project up on Instructables as well as posting the description linked above.

There aren’t too many previous articles in this vein on Hackaday, but we did feature a vertical seismometer in a post-quake round-up back in 2011, and a very clever design using a modified loudspeaker in 2014.

We’ve gone with [Kati]’s description of the device as a seismograph rather than an accelerometer, but forestalling the inevitable native-speaker Anglophone language discussions in the comments we recognise it’s probably a more appropriate word when applied to vibrations in the earth.

25 thoughts on “Bicycle Seismograph Measures The Streets

  1. This is cool! I love that this is a mechanical solution to the issue.

    I now have a practical reason to keep messing with connecting an IMU to an ESP8266. It’s not as mechanically elegant, but I’m better with code than woodworking :)

  2. The paper spooling rate changes with the diameter of the roll, and the graph isn’t normalized for speed so the result doesn’t really tell you anything.

    You can’t tell where along the route a specific bump was because the paper is advancing faster and faster as it gets picked up on the spool, and you can’t actually tell how bad the bump was because you don’t know how fast the bike went over it.

    The first problem could be solved by an second pen that draws a dot on the paper every so many rotations. The second problem is a bit more difficult to solve. You could try to pedal along at a constant speed.

    1. Well, you obviously don’t need the spool position indicator, because you have the spool and can simply measure it (each point of the seismogram obviously is defined by his own xy position). You could maybe do a speed diagram along with the seismogram and to then somehow normalize the graph to it (both for time and bumpiness™). You could actually calculate the roughness from speed and acceleration.

      1. Measuring the paper is attempting to backtrack exactly how thick the roll was when the original measurement was made, to determine how many inches on the paper correspond to how many yards on the ground at each different point along the roll.

        The practical accuracy will be poor and reading the paper will be extremely tedious – it’s simply more convenient if the mechanism puts a little dot on the paper every time the wheel rotates N-number of turns, because it automatically corrects for the spooling rate. You just count the dots to see how far along the road the bump was.

        The spool and paper will also stick and slip, causing the paper to be wound alternating tight and loose, causing the speed of the paper to vary by another mechanism which you can’t measure or replicate afterwards. Only the little dots would reveal that the paper speed changed.

        1. “…it’s simply more convenient if the mechanism puts a little dot on the paper every time the wheel rotates N-number of turns…”

          True. But then again, these devices were built in a workshop setting, where you want to keep the mechanical design as simple as possible, so that every workshop participant can successfully build a working seismograph. The good thing is, the design of the gadget was so simple – and hackable – that one participant added exactly the function that you described. Just go on the project website and have a look at the last photo in the image gallery.

  3. This reminds me of an episode of Monk where he is locked in the trunk of a car. They try to find the secret hideout of the bad guys by putting him in the trunk again as he shouts out directions!
    It was hilarious. :)

  4. I hope they compensated for the low natural frequency of the recorder’s spring-mass system – one of the perpetual bugaboos in shock/vibration recorders that were used in shipping analysis.

    1. Put nails on the seat. Even cruder and louder.
      At the end of the race, if is a long one, the medium amplitude may get lower. Like getting the bottles smashed.

  5. Why not use a [ Geophone or an Accelerometer + ESP8266 ] Data Log to phone using it’s GPS / speed estimation.

    http://www.ebay.com.au/itm/SM-4-Geophone-Element-SM-Sensor-HUB-10-375-/371705344678
    http://www.geo.uib.no/eworkshop/co2-2010/uploads/Main/121013.pdf

    if you’re capable to do so you could possibly place 2 sensors front and back to see the readings and get a better image of the terrain chuck in a BMP180 pressure sensor and for a better image as well.

  6. One could of course just make an app for a smartphone utilizing the available accelerometer/gyro devices combining it with GPS location data.
    But a good ol’ mechanical hack like this is just cool anyway :-)

  7. I love the analog approach! The old school speedometer cable to advance the paper is a particularly clever way to create a tighter coupling between bump size & location. It’s not very practical from an engineering point of view, but certainly would make for some good art installations to help people see the value of automated geospatial data tagging.

    Sometime ago (years? not sure how many) I read an article about potentially putting GPS logging pothole detectors (accelerometer) on municipal garbage trucks. The theory being that they travel down nearly every street in a city every week all year long. Over time that mass of data would start to show where potholes form, how quickly (and how effectively) they are repaired. The project didn’t seem to go anywhere, but I thought it was particularly clever.

      1. The relationship doesn’t work for bicycles because the wheels are completely different.

        A truck will ride quite happily on sand. Try cycling the same way, and you’ll see the point. Same thing as with regular shoes and stilettos.

    1. I’m sorry, are the bicyclists property, land, income, and sales tax not enough to fix the damage to the roads? If you believe bikes cause a proportional amount of damage to the roads, bring it up at a city/town/county council meeting.

      Because that’s actually what happened in my state. Enough 49cc and smaller mopeds were being used, and the state considered them bicycles for bike lane and traffic law purposes, that they changed the law to require a license plate style tag and yearly fee. It started in just one city, then the state capitol, now the whole state. Personally I avoided the bike lanes (scooter and motorcycle style without pedals have that choice here because they can do 40mph) because I knew I would cause more wear on it and because I could keep up with traffic. But knowing that I’m paying to keep the roads nice, and paying gas tax for the same, doesn’t bother me a bit.

      Only time I would use the bike lane was on stretches of road with 45 speed limits because, while those are the fastest roads a sub-50cc is allowed on, I had opened mine up full throttle and felt the governor kick in at just over 42; back-pressure and the loss of spark for ignition hitting suddenly at that speed was like a sudden gust of wind trying to slow only the bike while I’m left leaning back to avoid doing an endo.

  8. Nice mechanical hack!
    Another set of data needed to be gathered is how much swerving you do to avoid bumps, potholes, sunken manholes, … This swerving smooths out the ride when in fact the problem still exists. I guess monitoring the amount of steering done while moving a distance between some minimum and maximum between steers when moving in a generally straight line should do it. On a cobblestone road you’d probably be doing small adjustments in direction constantly so short distance between steers shouldn’t count.

Leave a Reply to ehud42Cancel reply

Please be kind and respectful to help make the comments section excellent. (Comment Policy)

This site uses Akismet to reduce spam. Learn how your comment data is processed.