I am running into some issues with the Bar100 sensor. I swapped out the stock Bar30 sensor for the deeper rated Bar100, and and getting a 'Depth Sensor Error" message in QGroundControl periodically. This message pops up every few minutes; between the error messages, the depth readings that I see in QGroundControl seem to be fine. This error mostly manifests at surface, but does come up periodically at depth as well. I can calibrate the sensor without issue in QGroundControl.
I have read in some other topics that the “Depth Sensor Error” message is a hardware fault; however, I’ve tried two different Bar100 units and the error persits. Recalibrating the pressure sensor in QGroundControl also does not seem to solve the issue.
I am running the latest BlueOS release, on a RPi 4B. QGroundControl V.4.2.3, ArduSub is the latest available stable release.
Are there any setup steps that I may have missed going from a Bar30 to a Bar100? Any advice would be welcome.
Note that BARO_PROBE_EXT should generally include both the 256 and 512 bits, to enable the KellerLD-based Bar100 and the MS5837-based Bar30 respectively
I’m currently running a BlueROV2 with a Bar100 sensor. I’m using the Navigator autopilot with firmware 4.1.1. Basically, I’m getting QGC to throw me a “Depth Sensor Error” message seemingly randomly every few minutes or so. I’ve seen the error both on the surface and underwater, and between error messages, I seem to get reasonable depth estimates. I also seem to be able to calibrate the sensor. I’m able to operate in depth hold mode, but when the error comes up, I’m kicked into Manual mode. I’ve seen other people in the forum run into this issue, and I’ve verified my baro settings (BARO_PROBE_EXT and BARO_EXT_BUS). I’ve also seen other people say that ferrite beads fixed their problem, but I don’t suspect that’s my issue if I periodically get this error when motors aren’t running (haven’t tried the ferrite bead yet). My settings for the baro should be in the image below.
Hi @nickz21 -
The motors don’t need to be running for the I2C bus to have issues. I’d recommend re-routing your pressure sensor cable, and installing ferrite-cores if possible. Sorry for the hassle!
I appreciate the recommendation, but it doesn’t appear to have helped. I bought a second Bar100 sensor that is also consistently giving me the same depth sensor error. I’ve wrapped the wires around ferrite-cores for both sensors and have re-routed the wires. The new sensor isn’t even installed in the end cap yet; I have it loose, so I can move the cable around. I’m starting to suspect this is a software issue. I’ve tried changing Baro_ext_bus to -1, but this doesn’t appear to have helped. I see the same behavior of the sensor showing up but then periodically having a depth sensor error.
Bump! Exact same issue here, even with the ferrite bead. All stock in the autopilot bottle, regularly bumped out of closed loop control due to depth sensor error, very frustrating.
Hi all, looks like we still have the issue @nickz21 has, even when totally stock with the upgraded Bar100. We’ve replaced the baro, installed the ferrite bead, and we still periodically get the depth sensor error. Is there a way to prevent this error from kicking us out of closed loop control (depth/position hold etc.), especially if we just need to live with this periodic nuisance? It prevents us from doing any significant closed loop operations. Any other insight? @EliotBR ?
I’ve tested with a Bar30, and I don’t appear to be getting any periodic “depth sensor errors” thrown at me. We’d like to use the Bar100, but getting these errors thrown at is is hindering our ability to test reliably. We can use the Bar30 as a short term fix, but we’d prefer to be able to use the Bar100 moving forward. Any insight you have as to why the Bar30 works seamlessly and the Bar100 throws errors every once in a while would be greatly appreciated.
Interesting find! So potentially this is a defect with the Bar100?! We’ve replaced the 100 and continue to get the fault. Any thoughts from BR? @tony-white@EliotBR . We’ve been hitting our heads against a wall on this since March.
Sorry this ended up slipping through. I had a half-written response to the last comment that never got posted!
Coincidentally @tony-white was digging into this problem recently, and managed to find a possible cause in the ArduPilot driver code, which he’s raised a fix for.
It’s currently under review, but does seem to have fixed the issue for him. If you’re willing to help test I can build a firmware for you to try running. Are you using a Navigator on an RPi4, or some other setup?
You should be able to install it using the “Upload firmware file” functionality in the BlueOS Autopilot Firmware page, after which hopefully the spurious readings (and corresponding depth sensor errors) will stop occurring.
Note that this firmware is built from the latest development state of ArduPilot, so isn’t considered stable. If this gets confirmed as working then we can merge it in as part of the ongoing 4.7 beta program, and include it in the upcoming 4.7 stable release
Thanks all for bringing this issue back to life. It was such a frustration of an issue that we uninstalled the Keller sensor and used the more basic baro because we couldn’t work on autonomous control of the vehicle without getting bumped out into manual mode regularly. Excited to see how this progresses.
