My $70,000 thermometer

For those interested in studying the earth modes, teleseismic events, volcanic slugs and other “VLP” events, properly insulating your broadband against thermal fluctuations is easy. Let’s take the Nanometrics Trillium Compact as an example. This beer can-sized broadband unit has little mass and no thermal insulation. When installed in the nude it reads like a thermometer at the long periods. Solve this cheaply by pouring a small cylinder of concrete around the Compact to give it mass. Then cut some Styrofoam for insulation. The gains are HUGE. See below for a before and after screen shot from PQLX for one of Panama’s many Compacts. Notice how the long-period signal drops to follow the Nominal Low Noise Model (lower grey line) after a make-shift thermal insulator was applied!

– Branden

P.S. GO BOREHOLE for maximum thermal stability

Thermal Cap Eye Candy

Thermal Cap Eye Candy

The basic setup

The basic setup

7 comment(s) on “My $70,000 thermometer

  1. Hi,

    Won’t the thin layer of concrete cause ‘cracking noises’? This thin layer will dry out fast while settling… I think it is way too thin, or you should try to protect it from drying out too fast.

    I think you have better luck with a (much) thicker wall of cement.

    Another way (and even better one) could be to secure the styrofoam with a brick on top of the insulation, the lid is also a Styrofoam one. The inner walls of the styrofoam should be a couple of mm’s away from the instrument.

    • Thanks for your comment Marchal!

      Great points.

      We usually do not have problems with “crackling” because we either buy the concrete already poured (so it is already dry for many weeks) or pour it in a mold we create at OSOP and let it dry for many days before installation.

      I think that a thicker cement enclosure and even more insulation are good ideas. But notice the gains in the Power Density Functions at long periods just from this quick fix! As you provide more insulation and more concrete, the benefits are increasinly marginal compared to the initial benefit of just providing some basic mass and insulation.


      • Hi Branden,

        You are right about adding some basic mass and insulation. Out of experience I know it is better to stay away from concrete ‘walls’ like these, because they to vary alot with even the smallest temperature changes.

        My vaults are in deep clay and are very sensitive to temperature changes (temperatures varied from -18C to +35C in 2012 alone…. therefore PVC and styrofoam work alot better here.


  2. Make your ‘concrete’ out of equal proportions of cement and sand. DO NOT use gravel ! Gravel has a different expansion coefficient and CAUSES cracking noises as the temperature changes. Keep the casting wet for about a month till it is fully set.

  3. Hi Branden,

    Actually it’s not a thermometer, but a very sensitive temperature-rate meter. If you change the temperature and give it long enough to settle the output should come back very close to where it started.

    The vertical temperature-rate sensitivity is directly related to the temperature coefficient of the internal spring and the low corner frequency of a broadband instrument’s response. The temperature-induced output will be 6.9E-5 K T^2 um/s per deg C/hour. So for the Trillium compact having T=120 sec and K= -250? ppm/deg C you will see 248 um/s per deg C per hour. (big) Slowing down the rate of temperature change is everything. Your insulation + thermal mass makes a low-pass filter.

  4. Branden- Very Nice improvement…
    A few questions to bring me up to speed-
    Doesn’t the compact sell for ~$5K, 3-components?
    BH’s are always problematic….
    Concrete takes forever to cure- years…
    Buried in sand would be better…even if you bring your own sand.
    Now the Plot is rather interesting…
    Can you give a tutorial on it sometime??
    It appears 10-90% falls within center area between the H & LNM’s.

    • Hi Jim:

      Good day. Thanks for your message. The Nanometrics Trillium Compact sells for quite a bit more than that especially when you factor in the digitizer.

      70,000 is just a number I pulled out of the air. Vaults we build in Panama actually run for ~100,000 USD (broad band + digitizer + shipping + the vault + + power + communications). In many cases seismic observatories invest a lot more money in their broad band installations, especially when building the vault into bedrock or installing a really beautiful instrument like the Nanometrics Trillium 120.

      We are going to be blogging to our OTHER blog – the Company News blog on a 250 meter borehole installation currently underway in Panama. Next week Dr. Peter Malin and Marcos Alvarex of IESE in New Zealand will join us to install a “Sonde” downhole. I will be sure to blog about whether sand or cement or a combination are used. Plenty of photos too.

      A tutorial on PQLX? Sure thing. Click here for a presentation by Richard Boaz (the one who programmed PQLX- He will be moving to Panama to join us in one month). The presentation refers to “SQLX” or the commercially supported and much improved version. We will likely be blogging a lot about data quality issues.

      Just some quick comments on “10-90% falls within center area between the H & LNM’s”: The purple hump at the long periods are teleseismic events. We use PQLX a lot to access station data quality for clients. We can tell immediately if conditions have changed or if the instrument is broken. Below are too examples of how PQLX/ SQLX can be useful:

      1. In these examples, the dataless are clearly wrong. When reviewing the data for an entire network, these problems stand out easily.

      2. In this before and after example, the client accidently starting exporting data to colleagues of his from an Earthworm Ring containing filtered waveforms.

      I look forward to seeing you at the MASW/ REMI course in Panama City next January!



Leave a Reply

Your email address will not be published. Required fields are marked *


You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>