Seeing Dropout (don't think it's network related)

[adamslab]

Hello,

I’ve been working for sometime on getting my shake setup in it’s vault, I just finished a revamp of the power system and got it back online. Everything seems ok, except I’m seeing frequent dropouts of data on Station/Data/and interestingly, the Helicorder.

I’ve attached the log files, since I’m not familiar with the innerworkings of the Shake software, I’m not sure what’s normal or not. I did run the log files through ChatGPT and it noted the following:

RSH.R8B94.2025-03-25T19_31_32.logs.tar (1.9 MB)

Any thoughts on what might be causing the dropouts is appreciated.

[Stormchaser]

Hello adamslab,

Thank you for the logs and the screenshots. Indeed, if the helicorder is also displaying these gaps, then the issue is a local one, and not a data connection/transmission one.

In the logs, while the Shake starts correctly, a lot of “gibberish” mixed with other errors can be found, indicating potential power supply or corruption issues:

2025 084 19:28:59>>	No Data has been received from the MCU in 12 read attempts.It appears the MCU is not transmitting data.  This is a fatal condition and should be investigated if this condition persists!
2025 084 19:28:59>>	Data has been successfully received, fatal condition resolved.
2025 084 19:28:59>>	internal error: buffer overflow!  cannot process read data...
2025 084 19:28:59>>	buf: {gF
`F.DF$cb` sA;F`§rFwctEæDjæ'`cd3'kkB`wdbd32F'FD²dfvVdfvV`§vFsg% sfjnE`@jfA`'jnF`
d'|FDd72FU#D`cdd;f
2025 084 19:28:59>>		16063	44

In such cases, we recommend our standard troubleshooting path, which I have attached below.

My first thought (and the probable cause of the gaps) would be to check and see if the current power supply you are using is continuing to deliver a stable voltage between 5.0 and 5.2V and a current of at least 2.5A at all times (3.0A if the Raspberry Pi board that is being used is a RPi4), as a decrease in power could lead to data services interruption. If possible, you can check this by temporarily removing the Shake from the vault and connect its standard power supply to a single wall socket. And, if you have another Pi power supply you know is in working condition, please try to exchange the current one with that and see if the Shake now correctly works for extended periods.

A second check that you can do is to see if all the connections between the sensor, the blue Shake board, and the Pi board are still solid and free from dirt or any other element that could compromise transmission. And, as usual, if you decide to disassemble the Shake when doing this, please ensure you are using proper ESD (ElectroStatic Discharge) protection (such as gloves, etc.).

As a last resort, and if all these checks come out as positive and the Shake is still not working, then I would recommend re-burning the microSD card again (or using a different microSD) after formatting and erasing all its data/partitions first (you can use DISKPART for this as it is very efficient), and see how the Shake behaves with the newly installed system, removing potential issues derived from corrupted files. I will leave the burning instructions link here for your convenience: microSD card topics

Thank you for your collaboration.

[adamslab]

Thanks for the reply.

After some more checking (mostly the logs on the shake, and with a multimeter) I found it was an under voltage issue. The under voltage notices in the logs in the Shake corresponded exactly with the dropouts I was seeing in the helicorder.

Fortunately easy enough to solve, I took voltage drop into account on all but one section and used too small a gauge of wire and it was enough to cause the issue.

In case anyone else is curious, the regular Raspberry Pi command to check for voltage issues isn’t available in the default Shake image, I didn’t want to risk mucking things up by installing anything. Easy enough to look at the system logs though.

dmesg -wT | grep -i voltage

This will follow the system log, showing any events that have “voltage” in them.

[Stormchaser]

Great troubleshooting adamslab!

No trouble at all; I’m glad to read that everything works as it should now.

Do you have pictures/specifications of the power supply to your vault, and the vault itself? I’m sure many would be interested in those (I sure am!).

[adamslab]

Oh, once I’m fully done, I will! I’ll put together a full report .

Actually, Mike reached out to me after I posted about the enclosure I built and asked if I’d be interested in doing a blog post about my setup. I’m a bit (okay, a lot) behind after having to revamp the power delivery system a bit, plus juggling other projects, work, etc., but I’m almost there.

The Shake/vault setup is solar-powered. Initially, I was powering most components (including the Shake) via PoE, but for efficiency and better control, I switched to powering things more directly from the solar system. I’m now sending 5V from the solar setup (located above ground) to the Shake in the vault.

This is where the low power issue came up. To minimize voltage drop, I’m using a DC-to-DC converter to step down the nominal 24V from the solar system to 5V for the Shake. The converter outputs 5.1 volts, and I ran 12-gauge wire from the solar system to the vault to reduce voltage drop as much as possible.

For convenience, I switched to a 22-gauge pigtail inside the vault. I figured the short length wouldn’t matter much, but I was wrong. Under load, it caused enough of a voltage drop to make the Shake unstable. I replaced the 22-gauge pigtail with a 12-gauge one, and now everything’s running smoothly.

[Stormchaser]

Amazing news, adamslab! Really looking forward to finished post!