I installed the MB7534 Maxbotix Rangefinder for snow depth measurement today. The specs are pretty good, it has a resolution of 1mm.
I connected the serial output to the Linux box. minicom is logging the data to a file. The rangefinder puts out an "R" followed by the distance in millimeters. The height is 1731mm. I use an awk expression to subtract the reading from 1731 to get the depth and convert to Imperial units.

Here is a photo of the setup. The APRSWorld shield is for a thermal sensor which the rangefinder uses for temperature compensation. (no data is available, the rangefinder reports either an "I" at boot up or an "E" for "Internal" or "External" sensors. I get the "E" so it is working)

I cleared the ground under the sensor, so it is showing "0" until we get some snow. I added it to the snow web page, but it is kind of boring. I might need to find some glitzy display widget for it or something....

Looks cool 

This is an area of observation that has few solutions and not many experimenting as you are.

Is there another page that you have explaining the reason for the somewhat elaborate appearing support system and all?

And looking over the data sheets and specs from the manufacturer, the one piece order seems to be in the range of $200 for just the gizmo, plus the support stuff you have to incorporate.

For those of us interested, is that the going rate, or are there places one can get quantity one for less than that?

Dale

The support system is based on a "swing arm" that I bought from swingarmusa.com. It is the "Original" and I picked that one because I wanted a level mounting and all of the other ones were angled, probably to support planters etc. It turned out that there was a slight angle.

Although the swing-arm is OK, it is basically just ¾" conduit with end caps, eye-hooks, and standard mounting clamps. If I owned a ¾" conduit bender, and had known that it was just conduit I would have made it up myself. For the cost of the conduit bender, it was just as good to order the swing arm. One word of caution is that despite the photo on their web page there is a slight angle on the outer section. I decided to take a chance of a kink and bent it a bit to make it more level. Then I shimmed the upper mount with a couple of ¼" washers when I lag-bolted it to the deck upright.

I wanted to mount it on the deck so I could run the power/RS232 wiring directly into my basement near the Linux box. I bought a Hoffman (Bud box) and a Lambda/TDK 5V power supply (10W) to run it. I probably have a good $300 total into the project.

The sensor itself is that funnel-shaped gizmo on the end of the PVC conduit. PVC conduit is kind of like a "wet noodle" and the swing-arm is rigid, hence the piggy-back mounting. I decided to use the junction box as a mount for the APRSWorld shield. I like it because it is small and relatively cheap compared to a Davis shield.

As far as sensor cost, I bought the "self cleaning" version which is $50 more than one without the feature. There is a discount schedule for qty 5 and 10. Although I bought it directly from MaxBotix, I see that Amazon has a smattering of them, mostly the bare unit. They do not seem to have the snow-measurement optimized version that I bought.

I hope that this helps, if there are any other questions I'd be happy to answer. I was originally going to get the Ambient Weather wireless one, but it keeps showing as out of stock and the queries about availability received vague answers like "2015". June? September? Now that I have built this, I think that it is a more "industrial grade" robust solution.

Greg H.

Another update:

I was having a lot of problems due to pulseaudio consuming a lot of CPU cycles. I have since disabled it. The script that runs cat just runs every five minutes now under cron. It runs cat for five second, then kills the process. Between these two "fixes" the computer seems to be more stable.

Today I was getting bizarre numbers, several inches negative snow. The raw sensor data was in agreement and the sensor was reporting much further distances than the baseline was set at. I found this paper, that says the beam must be perpendicular to the mounting surface. After "eyeball" checking it and noticing that the bottom of the sensor was slightly off angle, I confirmed this with a pocket level. I leveled the sensor in the axis of the mounting, and it improved the situation giving believable numbers.



Update on progress on this project:

I was having all kinds of problems with minicom and or kswapd consuming a lot of CPU cycles. This would happen after several hours of running. This would apparently freeze the computer, but I could get into another console and kill the process. Apparently there are some old bugs with kswapd, but I really suspect minicom.

Last night, I used a much simpler solution than minicom.

cat /dev/ttyUSB0 > snowlogfileI then parse the snowlog file to separate the data from the preceding "R" and get the last few bytes of the file.

Interestingly, when I was using the minicom-generated snowlog file, I used the last 4 bytes, with this simple "cat" technique, the last 5 bytes are needed.

Greg H.

It is good to see interest in this type of project, thanks for posting these links.

Since I last posted, I made a number of changes. I found that there was a postfix problem causing accumulating errors. It was a permissions issue. Now the system is stable.

I also switched to grabserial to acquire the data instead of using cat. The final issue was that I needed to let grabserial continuously run instead of start/stop. The awk script that captured the data would often have zero-byte length and cause errors. Continuous running solved this.

The only remaining problem that I have is dead tree leaves swirling under the sensor causing false readings. I have raked multiple times, but they keep re-appearing with our current windy conditions.

I am looking at placing a snow board under the sensor. This has been a learning experience but I think that I am making progress.

Greg H.

- optional isolate ground, can use insulating plates and artificial grass (good soft interface surface and camouflage measuring zone)

I have been thinking of elevating the snow board above the grassy surroundings, hoping that the leaves would stay beneath this surface. The insulation might help with warmer ground melting the snow.

Thanks for the links and ideas.

Greg H.

My expanded PVC snow board arrived today. It is 24" x 24" and ¾" thick. (60cm x 60cm x 2cm)
I put the old painted compressed wood snow board (same size) under the sensor and am using the new board for CoCoRaHS.

I already saw a leaf blow off of the board, where the other leaves get caught on the grass.
The variability looks really good now. We are expecting snow tomorrow, so it will be interesting to see how it performs.

The snow board has made a big improvement already. I had to re-normalize the zero snow distance so that the calculated depth is correct.

Greg H.

from experience, I think you'll find your board responding favorably, and consistently.  It doesn't "chill" as fast as a blade of grass, but
some experimenting I did last year with 2 similar, but thinner (½") boards, one elevated 3" and the other on gravel, all three in a 15' area,
showed no significant difference, with the limited snowfall last year.  If anything, the elevated board showed slightly less accumulation. 

Very nice.

Randall