Raspberry Jam appreciation post

[Ian Nesbitt]

Hello all,

I want to express how impressed I am with our Raspberry Jam so far. More than a decade ago, the University of Maine, Orono (USA) received a grant to purchase a RefTek RT-130 and Guralp CMG-40T as part of the New England Seismic Network (NESN, now headed by Alan Kafka). Several years ago the University network folks updated the campus DHCP configuration and our old RT-130 stopped being able to communicate with the outside world. When I started my Master's last year (in ground-penetrating radar...a completely different side of geophysics) no one had even thought about the seismic vault in years because it seemed such an insurmountable technical problem and would be so expensive to replace our hardware to function properly with the new network.

Despite that, it seemed to me a real waste to have a perfectly functional Guralp CMG-40T just sitting, literally gathering dust in the vault. When I started to troubleshoot however, I almost immediately found that there was very little hope left in using the RT-130. I learned about the Jam at an opportune time: as I was figuring out that the RT-130 had no DHCP capability and thus would not communicate with our updated network, and that the network technicians were surprisingly averse to both static IPs and devices other than their own acting as routers. That ruled out both the RT-130 acting of its own accord and having a Pi or other inexpensive device as a dedicated router. For both of these reasons, the Jam looked like the best option for us. Add to that the fact that the RT-130 only sent data to the repository every 10 minutes, and was functionally incapable of sampling at more than 40Hz in our configuration, and the Jam looked like quite an attractive alternative indeed.

Late this past spring, a faculty member (Dr. Alice Kelley) and I went to the department chair with a plan sketched out and asked for some funding for a solution. We were thrilled to have the request granted. It seems that having a seismometer on campus and a seismograph displayed in the lobby of the Earth and Climate Science building is a surprisingly effective selling point to get undergraduates excited about the department. The money wasn't much, but enough for a Jam, a Pi 3 B+, some waterproof connectors, and an IP67-rated waterproof/dustproof project box among a few other things.

We went with the customization option, mostly because I don't trust my electrical engineering skills enough to calculate proper resistor values. After struggling for a little bit to find the correct connector piece to interface with our existing RT-130 <-> CMG-40T cable, I was able to set up a neat, if slightly large, project box:

I knew just enough to get by, so wiring the 14-19 connector wasn't too difficult. After scouring both p. 11 of the RT-130 manual and Guralp's 40T manual for pinouts, I felt confident enough to solder the connections. Here is the Jam <-> 40T pinout diagram I sketched for our 14-19S connector:

And here is the final product. I ended up using a different 14-19 connector that I had ordered erroneously (oops) to pass the +-12VDC sensor power into the project box:

Side note here that will be obvious to some but not others: the ethernet cable is Cat. 5e and is wired as a patch (ie. it's straight-through). I was uncomfortable with the arrangement of the shorter patch cable in the first photo, which is now being used elsewhere.

And finally, it was time to put my work to the test. We were lucky that the seismic vault existed prior to this project, because it meant the digging had already been done! The vault is in an unfinished basement of a dorm addition. Temperature and humidity of the room are basically constant throughout the year, which is ideal. The 40T rests atop large glacial boulder pavement which is directly on bedrock. Orono is located in an area that was covered by 3-4 kilometers of ice at the peak of the last glacial maximum, which means that the boulder pavement is extremely tightly packed in lodgment till. Unlucky for us that the boulder pavements are so large as to be impossible to move, but at least the till makes for decent signal conveyance from bedrock to the device.

I didn't want to move the damping case from over the 40T for a photo, but I promise it's under there!

Shortly after installation, I added an APC 600Ah UPS backup both to protect from short outages, and provide simulated sine wave AC to the Pi/Jam and the AC/DC inverter that powers the sensor. The UPS was a must because I've been wary of the effects of under- and overvoltage here in Maine, which I'm pretty sure has claimed at least one of my other Shake SD cards over the past year. The UPS will regulate any power abnormalities going into the system and I think will dramatically extend the life of the device.

The Jam has been running for about two weeks now under the callsign AM.R6F62. I used another simple Earthworm installation on the Jam itself to remap to NE.ORNO.00.HH?, which is its official SNCL code under the New England Seismic Network (NESN). The station now provides data to Raspberry Shake and IRIS, and we're working on updating the NESN to accommodate it as well. We still need an updated dataless SEED file for IRIS which will contain response information appropriate to the 40T/Jam setup, but at long last the data from NE.ORNO is being archived again!

We've not had any detectable local earthquakes yet, but we've gotten a lot of really stunning long-period signals from distant quakes. These unfiltered heli plots are taken from my website, which has its own Earthworm installation and uses ObsPy to plot traces:

I encourage anyone with questions to contact me. This was quite an enjoyable side project to go alongside my Master's, and I'd be happy to share what I learned with folks looking to undertake something similar.

Greetings from Maine!

Ian

[Alan Kafka]

Thanks, Ian!!!!!

~ Alan Kafka

[Eddy]

Hi Ian,

Wow.. it's fascinating to read. great work! Good to see how a Jam helps to bring up the RT-130 back online and see how the vault is like! thanks.

Eddy

[Andres]

How nice it stayed.
if you can install a 4gb-8gb usb pen you will give more life to the card by not having to write the data files.

https://manual.raspberryshake.org/usbsds.html

Un saludo...