Skip to main content
S
Smithson
Associate III
May 12, 2026
Question

Feedback from "Flint" - an accelerometer system in the Smithsonian Museum

  • May 12, 2026
  • 1 reply
  • 238 views

Flint is an accelerometer system in the Smithsonian Museum.  

It is a MKI109 and MKI 210 pair.  It has been in place for about three years after a recommendation from your David Reynoldson, who is a really great guy and helps a lot. 

We have about six to ten such systems that were used and have now been pulled, but old Flint just keeps on working. 

It has a partially bespoke C# program.  The PI designed and manufactured some small boxes. But to hold the 210 onto the 109 I used a rubber band before it was boxed.  It worked well for three years and then broke. Flint had some minor problems:

1. You have to use a USB Managed HUB to turn the device on and off if you reboot the computer.  The HUB is as expensive as the accelerometer pair.  Or you unplug the computer, the device does not reset to home if the computer is switched off - major design flaw in a complex system.

2. You have to put a watcher program to monitor the accelerometer program that can just stop running with no warning.   The new Sensor TileBOX Pro has the same problem.  I have not yet put in the watcher program nor got another USB Hub, so the PRO is a manual pull out the cables.  

3. The frequency of the output is not constant, so it has to be monitored every 8 seconds and recorded with the data

and the FFT step adjusted.  

4. They have no serial numbers but thankfully the PRO does, the number is so big it cannot be held as int in C#. The number is larger than the number of grains of sand in the world, statistically speaking. 

5. It took me a month to work out Flint was not working because of the rubber band and not the electronics.  It now works again. 

It will pick up a daily change less than 15 microHz on a degrading structure, which is all structures.  PS Some degrade much faster, even if they are newer.  It will pick up a subway train at half a mile. 

Upgrading the programs over a cell modem that runs as a 3 is a pain. 

But it is a complex system with only a few annoyances.  

The only negative response from ST was when I asked David if I could buy ten of the sets and he complained that it would clean out North America, we left at least two I think.  

I know it easier to talk to the Pope's personal secretary than one of your mem's people, lol, personal experience.  

 

 

Screenshot 2026-05-12 164854.png

The PRO's revised program 

Screenshot 2026-05-12 172807.png

The PRO is great on a USB cable, but it is not as good as the other pair, it is much harder to program because of the C program that is mostly locked.   Of course they are sales toys for ST but for us they are mission critical devices.  Put up the C code for the dll's and allow us to use MSVS and not g++.  I gave up make type files in 1988.  

PS Captain Flint is named for the uncle in the Arthur Ransome Book, Swallows and Amazons, which is an excellent read for children and has a lot of scientific method and unstructured play in it.  

 

1 reply

Peter BENSCH
Peter BENSCH
Technical Moderator
May 13, 2026

Thank you very much for your detailed feedback and the positive comments on our products. We are very pleased that your application has been running successfully for such a long time.

At the same time, we would like to kindly point out that the evaluation board used is, in accordance with the Terms of Use, intended exclusively for evaluation and testing purposes in a research and development environment and must not be used as part of a production application.

Your local contact person or local distributor can review the requirements with you and will be very happy to support you in finding ST products suitable for continuous operation. If you are not able to carry out such development yourself, I am also sure that they can help you find a suitable company for this.

Regards
/Peter

In order to give better visibility on the answered topics, please click on Accept as Solution on the reply which solved your issue or answered your question.
    S
    SmithsonAuthor
    Associate III
    May 13, 2026

    David Reynoldson is my contact person.  

    He and I have discussed this topic and the use of these devices.  

    Over the years, he has sent me devices for testing purposes, which is all that they can be used for until they reach TRL 7.  (Refer Nasa Guidelines.) To meet TRL 5+ you need a remote location for the test, with a year of continuous operation, my office is not a remote location to assure the client we are at TRL 7, which we achieved using the CX1 accelerometer in 2018, remotely in Iowa. 

    David recently sent me three Tile Box Devices that are marked NRND and so I purchased three Sensor Tile Box Pros, so I was not in NRND phase.    

    I explained in the email what the shortcomings of these devices are, and so they are about TRL3, no where near production use.  I research on monitoring methods.

    If you would care to answer the other questions, which have been asked several times, I would be quite happy.  But locked C DLL's that crash after 40 seconds by the worst case is not going to get anything past TRL3.  We use CX1's for real work.  

    I trust David will keep me informed if there are other products to test.  

    Smithson_0-1778698761135.png

    This is the full reboot sequency for the Sensor Tile Box Pro that is sitting on a brick under my desk.  When the period passes one year, allowing for electrical blackouts, then it can be used usefully, otherwise it is just a nice research tool.  Nothing more.  Each dot is a reboot, the period is ten days. 

    One does not do a Gaussian analysis with one sample, one uses not the engineering three, but at least ten, hence my purchase.  It is a simple analysis to understand this number, 50 is the real number, but life it too short. 

    I am happy to look at any suggestions David offers. 

    Warm regards

    John

    S
    SmithsonAuthor
    Associate III
    May 17, 2026

    The CLI_EXAMPLE locked DLL's are throwing random errors, I cannot fix these errors in your code because I cannot get to the C code.  

    It is hard to evaluate when they crash at three iterations.  

    Please help me help everybody who wants to evaluate these devices.  

    Terms & ConditionsAccessibility statement