Fin MacDonald

Friday was my final day with NSCC Applied Research working full time. When I left Friday the monitoring system was not live. There are still some challenges to overcome with this very large project. When we replaced the wire with CAT5 network wire the system worked for a little while but then we started getting shorted bus errors. Because of the number of connections the resistance on the wire was too high. We will need to solder the connections instead of using quick connect clips in order to improve the signal. The WEL sends 5V over the 1-wire bus and because of the amount of quick connects we used for sensors runs we were getting too much resistance and noise on the line. We reduced the system to include only one of the evacuated tube collectors and the signal returned to normal. Once the soldering is done we can include the others again.

Our network connection is browning out as well. We are able to get a wireless signal on the roof with the equipment we have but we can’t keep it connected. This meant we couldn’t go live with the system. We will need to replace the antenna with something stronger. We have an access point on the roof for the solar photovoltaic monitoring and we may need to try connecting through that as well. That would require significant changes to the way the photovoltaic monitoring system communicates though so we would consider that a last resort.

I leave the project in the very capable hands of Dr. Alain Joseph and they rest of the crew. They will continue to move the project forward and I expect to see the system live this summer. I prepared a manual for them documenting the process of solar hot water monitoring using the WEL. I’m sure that will provide them with the knowledge transfer required for someone else to step in and continue the project. I have also agreed to offer them support on a part time basis as needed.

The experience of working on a large project like this one has been very valuable to me. We’ve worked through a number of different problems, and learned to anticipate things to go wrong. I know now that I shouldn’t expect anything to work right the first time. I also learned the importance of documenting your work properly, so that others can pick up where you left off.

Getting a pyranometer to work with the Web Energy Logger (WEL) was a long process for us in the lab. A pyranometer is a device that measures the solar intensity in watts per meter squared. We used a 0-5V pyranometer from Apogee (http://www.apogeeinstruments.com/pyranometer/). This was a challenge because the WEL does not have a 0-5V input. We needed to convert the signal to 4-20 mA so that it could be read by the WEL. There is a benefit to 4-20mA though since it is a current signal and not voltage it doesn’t drop off when you use a long wire run. Originally we had an electronics student design us a signal converting circuit board because we had the tools to design and mill circuit boards here at the college. We found a company that manufactures boards like this (http://controlsignalconverter.com/) and it is much more practical to purchase them then make them in most cases.

When testing the pyranometer we noticed that the 4-20 mA signal would drop off around the 12 or 13 mA mark and saturate. The WEL would never receive the full 20 mA. We ran numerous tests with our equipment and determined that the problem was within the WEL itself. We spoke to the manufacturer of the WEL and he explained the problem and how to fix it. It requires some cutting to fix it. There are two zener diodes inside the WEL that need to be snipped off. To do this you need to remove the rabbit board network chip by wiggling it and lifting it at the same time. It is fairly snug so you will need to work at it to get it off. Once it is off you will have access to the zener diodes and you can snip them out with some wire cutters. You will see them next to the green 4-20 mA plug on the WEL sandwiched between two resistors. I took a picture of our board with them removed below and have identified where they used to be with a red circle. Once we removed the zener diodes it worked perfectly.

You need to scale the numbers in order to get the proper engineering units. We used the formula f(x)=78.125x-312.5 to convert the 4-20 mA signal into watts per meter squared. This was specific to the equipment we used so if you use a different pyranometer you will need to do the math yourself.

Pyranometers are important for research because they allow you to see how solar panel output changes with solar intensity. We are particularly interested in how solar performs in the winter time in our climate. It is cold but often the sun is shining. With a pyranometer in our system we will be able to draw more meaningful conclusions from our data.

The solar thermal monitoring project is moving along despite a couple of setbacks. We received the wireless equipment in the mail and were able to set up the wireless bridge to the schools wireless network. We used a directional antenna to improve the reception because the school’s wireless signal on the roof is weak.

We put a 12V deep cycle battery on the roof and we did a test run. It was able to power both the router and the Web Energy Logger (WEL). The battery is rated at 90 amp hours and the load for both the WEL and router is 0.4 amps combined. This means we are able to get over 3 months of power off the battery before we will need to swap it out for charging. This is good news because the battery is very heavy!

Testing with the pyranometer hit a bit of a wall initially. As I mentioned in my previous post the pyranometer measures the solar intensity in watts per meter squared. Since we are converting the 0-5V signal on the pyranometer to a 4-20 mA (milliamp) signal for the WEL we needed to scale the results. We did the math calculation and did a test run. We weren’t getting to the proper numbers in the upper ranges. Once we were confident the math was correct we tried numerous 4-20 mA devices with the WEL and had the same result. It really pays to have an electrical engineering student around when you are trying to troubleshoot electrical problems. Travis Keeping is our electrical expert at the lab and he a bunch of tests for us. In the end a call to Phil Malone from OurCoolHouse.com who designed the WEL was all it took to discover our problem. Turns out that the WEL has a defect and there are two zener diodes that we will need to snip out of the circuit and it should work fine after that. Phil also told us that new versions of the WEL will have a voltage port so we won’t need to convert the signal in the future. We have another WEL on order and it will have the voltage connection on it.

The wire we chose for the sensor wire run was cheap wire and it started giving us grief. We used a 75m run and it had too much resistance because it wasn’t twisted pair wire. Twisted pair wire helps to reduce the electrical noise on the signal. We started getting “shorted bus” errors on the WEL. I took the wire run down and we will be replacing it with CAT 5 wire, which is what is used for computer network cables. Once the new wire is installed we should be able to go live.