Hello Everyone.
This is a nice site!
This is my first post.
===
Please note that I have been an EE for 20 years, but only in the digital (computing) world. I have done some analog work, but this was RF and digital audio. Incredibly, I have never worked with power lines!! And my schooling details for power lines are really relegated to the my mind's backside!! Yes, I know, I can study up, but I thought I would throw this out, and see what you might think...
===
I would like to ask you all to see if I am on the right track with this project, which is likely very easy to you, but not for me!
===
I need to measure the power line full wave with an ADC to find out how much AC is on the line relative to rectified, but not filtered DC. The algorithms involved are not a problem, nor is the output.
Here's the gotchas:
1. The input impedance must be >=1meg (2+meg preferable).
2. This unit is battery powered and the battery must last for 7 years minimum!! Note that I want to run only a single supply.
3. The sample time is once per hour.
===
What I am doing so far:
1. I am running the raw 60hz AC inputs through an AC voltage divider (2 meg each leg in, and 22k between), to bring the input (Potentially 200vac RMS because of the AC on top of the rectified DC) down to internal +/-1.5v levels.
2. Remove flat DC with 100nf series caps.
3. Connect one leg to internal ground through a low value resistor.
4. Connect the other leg to a unity gain Jfet buffer with an input voltage divider between 3v and gnd. 10k resistor between output and gnd to set the impedance. This obviously shifts the AC to between 0 and 3v.
Note that the ADC inputs of either PIC's or AVR's is 10k nominally.
5. Connect this output to the ADC of either a PIC or AVR.
The input buffer power will have a power switch that the processor will turn off when it is not sampling.
My questions are: Am I doing things reasonably well with the input as I described it, considering I'm dealing with a raw AC line?
I forgot to mention that there will be TVS's in appropriate places.
Thank you,
Kent
This is a nice site!
This is my first post.
===
Please note that I have been an EE for 20 years, but only in the digital (computing) world. I have done some analog work, but this was RF and digital audio. Incredibly, I have never worked with power lines!! And my schooling details for power lines are really relegated to the my mind's backside!! Yes, I know, I can study up, but I thought I would throw this out, and see what you might think...
===
I would like to ask you all to see if I am on the right track with this project, which is likely very easy to you, but not for me!
===
I need to measure the power line full wave with an ADC to find out how much AC is on the line relative to rectified, but not filtered DC. The algorithms involved are not a problem, nor is the output.
Here's the gotchas:
1. The input impedance must be >=1meg (2+meg preferable).
2. This unit is battery powered and the battery must last for 7 years minimum!! Note that I want to run only a single supply.
3. The sample time is once per hour.
===
What I am doing so far:
1. I am running the raw 60hz AC inputs through an AC voltage divider (2 meg each leg in, and 22k between), to bring the input (Potentially 200vac RMS because of the AC on top of the rectified DC) down to internal +/-1.5v levels.
2. Remove flat DC with 100nf series caps.
3. Connect one leg to internal ground through a low value resistor.
4. Connect the other leg to a unity gain Jfet buffer with an input voltage divider between 3v and gnd. 10k resistor between output and gnd to set the impedance. This obviously shifts the AC to between 0 and 3v.
Note that the ADC inputs of either PIC's or AVR's is 10k nominally.
5. Connect this output to the ADC of either a PIC or AVR.
The input buffer power will have a power switch that the processor will turn off when it is not sampling.
My questions are: Am I doing things reasonably well with the input as I described it, considering I'm dealing with a raw AC line?
I forgot to mention that there will be TVS's in appropriate places.
Thank you,
Kent