Will this work? temp and power monitor

[eric256]

I'm working on a project to monitor temps and KWH's. Peter Anderson makes a nice kit, that has optoisolators going to a pre-programmed PIC processor. http://www.phanderson.com/iom142/iom142.html.

I plan to use a bunch of DS18s20's for the temps, and then some small CT's for power.

The CT's are the main question. I plan to use these . http://www.controldepot.net/index.asp?Page...mp;Category=804, but I'm not sure which one is correct for this kit.

Anyone have any experience or opinion?

(yes - i've worked with electrical wiring a lot before - and know the dangers ................)

Aloha.

 

[BraveSirRobbin]

I would use the "True RMS" split core current transformers with the 4-20 ma output (better for long cable runs and is more noise immune).

You should also be measuring voltage as well as WATTS = VOLTAGE * CURRENT, but again, the voltage probably doesn't change all that much and you could get away with using a standard measured value.

You would have to provide a terminating resistor for the 4-20 milliamp output to convert it to voltage at the Analog to Digital input of the PH Anderson device. Note that you want to span the entire range of allowable voltage input in order to maximize the full scale bit resolution of the A-D; but, at the same time you have to make sure you don't over-range that allowable maximum limit. For instance if you will be reading 10 volts (which is probably near the maximum input for the Analog to Digital board) you will want 10 volts / 20 milliamps = 500 ohm resistor tied between the plus and minus inputs (use a 1 percent precision resistor). Note that you will have a "Y" intercept that will be something other than zero as your lower end will be the 4 milliamps x 500 ohms (my A-D guide mentioned below will take this into consideration).

I noticed this excerpt from the link you provided from PH Anderson:

The DS2438 A/D converter is a ten bit A/D and the device uses an internal band gap reference such that the readings are reported in VDC over the range of 0.0 to 10.23 VDC. The resolution is 10 mV.

The DS2450 Quad A/D is operated in a 12 bit mode using an internal badgap reference which quantizes the input voltage on each of the four channels over the range of 0.0 to 5.12 VDC. The resolution is 1.25 mV.

I have to admit I'm not certain what the voltage min/max is with that language . I'm not sure if he means the input range is 0 to 10.23 volts with 10 bit resolution for one and 0-5.12 volts with 12 bit resolution for the other model. Perhaps Email Professor Anderson (he is a very responsive/friendly person).

In any case once you determine the range of voltage output your current sensor will give you and after you made sure it is within the range of your A-D converter's input capability, you will then have to determine an equation to convert measured AC amps to the DC voltage input measured by this channel.

I have written a "How-To" guide on Analog to Digital Converters to help with just this purpose (as a matter of fact an example in that How-To shows how to use a current sensor).

You will then have to write some code to interpret the value of the serial interface output, then poll the device and store those values. You will then have to use that “voltage” value with your equation to convert it to AC amps, then multiply that by your AC voltage to get the Watts dissipated at that time. You will then have to look at total watts during a set amount of time to get Watt-Hours. This measurement's accuracy will greatly depend on the sampling rate/storage of the polled value. I'm not certain how much accuracy you will loose using a "set" AC voltage value (since you are not measuring it).

In cases like this I also like to get guidance from fellow Cocooners such as Guy Lavoie and Michael McSharry to make sure my methodology makes sense! . (Afterall I did a whopping ten minutes of analysis on this post, so I'm sure I missed/said something incorretly). Hopefully they will chime in and give some more guidance and suggestions.

 

[WayneW]

You should also be measuring voltage as well as WATTS = VOLTAGE * CURRENT, but again, the voltage probably doesn't change all that much and you could get away with using a standard measured value.

That is only true for ideal devices, like incandescent light bulbs. Anything with a motor or a power supply is likely to have a power factor less than one. So, you can measure the current with a sensor (current transformer) like those to get a relative reading for the load from day to day, but it won't match your electric bill. If you want to monitor your whole house and have it match the electric bill, you need to include the power factor, which means measuring the voltage and the current many times per second, then doing calculations.

http://www.alliantenergy.com/docs/groups/p...ub/p015093.hcsp

 

[BraveSirRobbin]

You should also be measuring voltage as well as WATTS = VOLTAGE * CURRENT, but again, the voltage probably doesn't change all that much and you could get away with using a standard measured value.

That is only true for ideal devices, like incandescent light bulbs. Anything with a motor or a power supply is likely to have a power factor less than one. So, you can measure the current with a sensor (current transformer) like those to get a relative reading for the load from day to day, but it won't match your electric bill. If you want to monitor your whole house and have it match the electric bill, you need to include the power factor, which means measuring the voltage and the current many times per second, then doing calculations.

http://www.alliantenergy.com/docs/groups/p...ub/p015093.hcsp

Yes, I realize this which is also why I stated the following in my reply above:

This measurement's accuracy will greatly depend on the sampling rate/storage of the polled value. I'm not certain how much accuracy you will loose using a "set" AC voltage value (since you are not measuring it).

These are EXCELLENT points (and link) though and I should have been more descriptive in my initial reply.

 

[BraveSirRobbin]

I'd like to expand a little on Wayne's point above. You can actually monitor/see devices that have a power factor less than one by using a Kill-A-Watt meter (which a lot of Cocooners already own). Just plug it in and then plug in whatever you want to measure (will not do 220 volt appliances though).

Turn the appliance on and then start pressing buttons. Note that you can measure voltage directly. Turn the appliance on and off and see if this voltage "sags". You can then press the "PF" button to monitor power factor. You will note that if the device is reading less than one you will probably have a voltage change when the unit is on and off.

Also note that you can measure power in watts and also the voltage * current (VA). Again, if the power factor is less than one you will notice a difference between these readings.

 

[Michael McSharry]

I have to admit I'm not certain what the voltage min/max is with that language

This is a description of how the PIC firmware has been configured based upon the assumed input range of the DS2438 or DS2450. I think you will find the DS2450 a better choice for your application as it has both higher precision and four channels per device. The HVAC board from hobby-boards.com may be your cheapest source for a DS2450 (surface mount device) already mounted on a PCB. I think the HVAC board is setup for 24 Volts rather than the 10 volts expected by the PIC.

I question the selection of the IOM142A since your application does not need the relay outputs or the opto inputs and you still need to add additional hardware for the A/D. It looks like you are selecting 1-wire as your interface, so why not just use a basic 1-wire interface such as the DS9097U/DS9490D. This way you will have off-the-shelf software available to use the DS2450 and DS1820 and will not need to develop new software to interface to the IOM142A.

The use of 1-wire for power monitoring may or may not be approriate for what your objectives are. If you are looking for ballpark power utilization then it will be fine. If you are looking for actual power consumption then you will need faster voltage and current measurements than the 1-wire will provide. I think (but not positive) that a typical A/D conversion for 1-wire is about 1 second and if you have multiple devices for A/D (which includes the temperature A/D conversions) then you will have several seconds between each sample. There can be considerable power fluctuation over a period of 10 seconds or so.

Along the same line of reasoning, there is no need to measure voltage if you are not measuring current in millisecond ranges. The average voltage will provide just as good a measurement. The use of split core makes for easier install, but typically the accuracy is decreased by a factor of 2. True RMS is fine for most any applicaiton and becomes more important as you venture from monitoring power company input which is a nicely controlled sine wave to monitoring downstream switching circuits were the waveform is no longer clean.

There are several companies that provide power monitoring interface hardware with a variety of interfaces used for data collection. They have already dealt with the power interface issues. Some provide sufficient quality for general awareness which may in the 5% range and others provide accuracy similiar to that used to measure utilization by the power company.

 

[BraveSirRobbin]

Michael, thanks for your reply.

I also wanted to point out something and that is the overall "objective" of the project. If it is a matter of power conservation and a rough cognizance of how your consuming power over various times of the day, then perhaps a general current only reading may suffice. (For instance you see that you are drawing a large amount of current between the hours of one to four in the morning, you might want to investigate what device is causing this consumption). If the objective is to truly create a measurement of overall power consumption, then all the considerations mentioned above will come into play.

 

[eric256]

Thanks for you many replies. I haven't had time to reply to all of them, but yes - the objective is to monitor the power at a gross level. My voltage is pretty steady. I have a meter that will measure power factor, and it rarely drops at all. No big motors in my house.

I want to zero in on the big hitters in my power bill, AC, hot water, dryer, etc. I plan to use 4-6 CT's, maybe more as the project move forward. I also need to submeter a apartment running off the main panel, but I don't know if this will be accurate to measure the total KWH's over a months time. There are some nice retail products that can do that for me.

Thanks again - and I will let you know how it is going. Peter Anderson seems to have a ton of interesting PIC's to work with. Wish I had been in his class.............

Aloha

Michael, thanks for your reply.

I also wanted to point out something and that is the overall "objective" of the project. If it is a matter of power conservation and a rough cognizance of how your consuming power over various times of the day, then perhaps a general current only reading may suffice. (For instance you see that you are drawing a large amount of current between the hours of one to four in the morning, you might want to investigate what device is causing this consumption). If the objective is to truly create a measurement of overall power consumption, then all the considerations mentioned above will come into play.

 

[eric256]

I have a Fluke 43B just to test that point. I measured a number of devices in my house, and as I expected, with a "residential" load, the PF never dropped below 0.97.

With that being the case, for this adventure I will ignore the PF. Being 3% or closer to my "meter" reading is good enough for me. I have also measured my voltage over a 24hr period, and found less than a 5% change. I have a transformer on the pole out side my house, and the substation is close as well.

I just wish my power wasn't $0.20 a KWH. We are in Hawaii and everything is expensive here .........................

I've used a KillAWatt and it is a great device for monitoring single items, like the fridge, or a window AC unit.

Aloha

I'd like to expand a little on Wayne's point above. You can actually monitor/see devices that have a power factor less than one by using a Kill-A-Watt meter (which a lot of Cocooners already own). Just plug it in and then plug in whatever you want to measure (will not do 220 volt appliances though).

Turn the appliance on and then start pressing buttons. Note that you can measure voltage directly. Turn the appliance on and off and see if this voltage "sags". You can then press the "PF" button to monitor power factor. You will note that if the device is reading less than one you will probably have a voltage change when the unit is on and off.

Also note that you can measure power in watts and also the voltage * current (VA). Again, if the power factor is less than one you will notice a difference between these readings.

 

[cactsbob]

I have been using a system built from off the shelf parts to monitor kwh usage for about 6 months now. I got the power aquisition devices from Wattnode at:

http://www.ccontrolsys.com/products/pulse_output.html

I also got the CTs from the above website.

Currently, the pulse output from the Wattnode is read by a counter that I got at:

http://www.embeddeddatasystems.com/Merchan...Product_Count=2

The counter is read by a PC process through an ethernet bridge at:

http://www.embeddeddatasystems.com/Merchan...Product_Count=0

This system parts cost me about $300 dollars and after 6 months of using it I would pay the $300 again. It is great.

I wrote a .NET windows service that collects the kwh usage every minute & stores it in a SQL Server Express db. I created a website that gets data from the database & presents it in tabular or bar graph charts for selected date ranges.

I am actively working on replacing the ethernet with a zigbee interface where I can read kwh through the pc as well as a wall mounted LCD screen.

Also, my billing has been within $10 of the wattnode data for each month.

 

[jim3342]

That looks aweseome and what I have been looking for the type of solution I would like to try. What are the model numbers of the CT's you bought from ControlSys?

 

[cactsbob]

I got the CTS-0750-150. It takes two CTs if you monitor both legs of your service. I installed these on the load side of the main in the breaker panel.

http://www.ccontrolsys.com/products/current_trans.html

Let me know if you install this. Like I said earlier, I'm working on a Zigbee interface to replace the ethernet interface. But, the ethernet has worked flawlessly since March 07.

Here's another screen shot. This has daily usage for a selected date range.

That looks aweseome and what I have been looking for the type of solution I would like to try. What are the model numbers of the CT's you bought from ControlSys?

 

[jim3342]

I'm going to order the hardware this week. Do you have a picture of what your install looks like in the breaker box?

 

[cactsbob]

I'm going to order the hardware this week. Do you have a picture of what your install looks like in the breaker box?

I'll take a picture tonight & post it.

 

[toymaster458]

As another option to the 1-Wire interface is using a DataNab Ai8_R13 which has pulse inputs. This would replace the need for the GP1 Counter and the Ethernet 1-Wire Host Adapter. For Temp sensing depending on the number of probes you need the Ai8_R13 would still have 7 inputs left for Analog 10k Temp Sensors