The biggest challenge to the project is of course getting
a location to mount the station as well as an Internet connection to upload the data to a server. Speaking
with Don, he volunteered his Internet connection to allow data to uploaded. To make this as straight-forward
and to require as little interaction as possible, a stand-along system which doesn't rely on Don's computer
is going to be ideal.
Next, was how to measure snow depth.
I found two commercially available Snow Depth Sensors. The first I found was the NovaLynx
unit. At the time I checked, it has a retail of $745.00. More research also found this sensor by
Judd Communications. It's listed on their web site at $652.00.
I'm not sure
if that includes a data-logger or not. But it's still pricey for a 'hobby-type' snow depth sensor.
In researching "home-brew", cost-effective options I first found this project put together by Steinar Midtskogen.
It uses a sensor called Mini-AE. His project did inspire me
to do additional research and really got me excited about the potential for putting my own sensor together.
Update: I have been in contact with Steinar. He was so nice to learn from
since he has already done this. I learned that he put his Mini-AE in
constant ping mode at 10Khz. That means that about 1 time a second you
hear a 10Khz click as the sensor ranges and reports back the voltage. Steinar's
install is in a remote area and he's not concerned about bothering anyone.
For my installation behind Don's house, I need to keep the sensor as quiet and
discrete as possible. There's a way to put the Mini-AE into external
trigger mode. More on that later.
The next well documented and intriguing sensor lead me to this project by Chris Nafis.
It uses a "Rabbit" processor-based board, That processor looked very interesting to me and was the direction I was planning on
heading. I would require the use of some C programming. No biggie. The author of the web site has the source code
listed too. I was excited until I found that the company that supplies the processor kit which includes the board
for the Ethernet connection, was only sold in lots of 50. That's about 48 or 49 too many for me.
I started some new research on new options which lead me to this other project which discussed
one-wire communications. This put me onto
a whole new level of research into 1-wire weather stations.
The author of the Weather Toys book uses a processor and board
called TINI, initially produced by Dallas (Maxium). I liked the idea of using this processor because this chipset
was designed with the 1-wire interface onboard. It's also designed to use embedded Java.
Next I learned that Tim, the author of Weather Toys, experimented
with a board called the TStik.
This is manufactured by a company called
Systronix. The author mentioned that the TStik
seems to have a more rugged design which has faired very well in harsh weather reporting conditions.
As is turns out, Systronix is based right here in Salt Lake City, Utah.
That's very convenient for me. I started speaking with Bruce over at Systronix and
he said that he would provide some support to me in this endeavor. They are also dealers for the
MaxBotix ultrasonic sensor which has a serial interface, and will we able to use the serial port on the TStik board.
This is now an all-in-one solution for a very scalable, pseudo professional one-wire weather station, that can also do the ultrasonic snow-depth sensing.
Some initial testing with the sensor hooked directly up to a PC's serial port, revealed an aspect to the sensor which I overlooked when I decided to use it.
While the MaxBotix sensors are very accurate and provide very stable results. They only have a resolution of 1" or 2.5cm. That's not great, but probably acceptable for this particular
I've been told that I REALLY need to mount the TINI processor in the house and
run the 1-wire network out to the weather station. To use the MaxBotix
serial sensor would require a VERY long serial cable. That's not going to
work. Also, the 1" resolution is a bit of a downside. I wound up
purchasing the Mini-AE that Steinar used in his 1-wire prototype station. The
next concern was controlling the sensor and using an external 1-wire trigger.
Steinar suggested using a very cool method by which you would use a 1-wire board
with 1-wire network activity LED. By tapping into the LED, you could use
the 5v signal from the LED to trigger the Mini-AE. Unfortunately the board
with the LED doesn't seem to exist any more.
I talked to Eric at Hobby-Boards.com he didn't have an immediate solution.
Posting my dilemma on a forum gave me a pretty cool result. Someone
suggest that I use Eric's HVAC monitor board. It uses the Dallas/Maxim
DS2450 chip and can be configured to use one of the 4 channels as an output.
One of the other channels will do the 5v A->D conversion from the Mini-AE.
Excellent. Eric volunteered to modify his HVAC monitor board to allow channel A
to provide the trigger signal.
I'm a bit disappointed with progress. I hooked up the Mini-AE to my
multi-meter and has some great success pinging and getting readings back.
It's great. Next I put the Mini-AE in trigger mode and decided I would
check to see how well the 5v trigger worked. No amount of coaxing would
get the Mini-AE to trigger using a 5v signal.
I emailed SensComp to find out what the problem was and this is what I
"In order for the external trigger
to operate, you need to take the external trigger enable (pin 4) low or to
ground. Then apply an external clock (TTL level square wave) from 1 HZ – 50HZ to
the external trigger (pin 3). "
So much for a simple 1-board solution. Now I think I
need to commission someone to make a board that will send a 1-50Hz TTL level
square wave signal that can be signaled by a modified 5v HVAC board.
Woo hoo! Success. I spoke with Eric at Hobby Boards and he said that the DS2450 output
should be 5v TTL compatible. With that information I worked on some code and managed to
get a snippet of Java code to control Eric's modified HVAC board.
By creating a loop
I was able to successfully cycle the ultrasonic sensor on and off. I chose channel C on the board
to receive the voltage returned from the Mini-AE and all works well.
I need to get my code integrated into Tim Bitson's TWS (TINI Weather Server) code. Once that's done,
it's just a matter of putting all the parts together and onto a weather station pole. Pictures soon to come!