Overdue system diagram for power monitor

6 Responses to Overdue system diagram for power monitor

1. Felix Rusu says:

   April 23, 2012 at 2:51 pm

   Looking at your overall system design, a suggestion comes to mind – why not use a 3-phase IC like ADE7758? You could drop the PICs for MUXing, and could then just measure two phases if that’s all you care, and perhaps chain a few to obtain the desired 30 circuits of measurement capacity. You could then use an serially controlled analog matrix switch which you can use to route the sensor signals into any of the ADEs. I think that way it’s possible to mix/match whichever channels you want, and then be able to play around with circuit sums and such. This make sense or am I smoking crack?
   • Skye Sweeney says:

     April 23, 2012 at 3:18 pm

     No clue what you are smoking but you are making sense.

     A poly phase chip might make more sense. But I started out thinking one ADE7763 single phase chip per branch circuit and never revisited the selected chip after I went to the two ADE7763 for the whole house approach. Now with all the code written, its hard to go back and look at a new chip. But perhaps I should.
     • keith t says:

       May 10, 2013 at 10:39 pm

       i know i am very late t the game, but did you ever look at the Maxim 78M6618 (i think) it is a nine channel energy monitoring chip. 1 voltage in and 8x current in.
       love to see how this is going.
       • Skye Sweeney says:

         May 16, 2013 at 8:38 am

         I did not know this chip existed. I did an extensive search at the time and did not run across this part. I wonder if it is new. In any case, I will look into it. Thanks for the warning.
2. Felix Rusu says:

   April 24, 2012 at 7:47 pm

   I saw your code and it would be a bit of an effort to migrate.
   I also checked digikey prices and 1 ade7758 = $15, whereas 1 ade7763 = $4.3, so … individual 7763s might make more sense financially if you want to monitor individual circuits or a whole phase, with the tradeoff being that three 7763s take more space and increase component count. An idea is to have multiples of 4/8 ADE7763s per phase (say 8 on phase1, and 8 on phase2 each set hardwired to the respective phase’s voltage), and then hookup the current sensors to the two sides, the limitation here is that you can’t mix the two branches (ie measure current on a phase2 circuit while the ADE’s voltage sensing is on phase1) or the readings will not be legit.
   Then you would use daisy chained serial to parallel buffers to expand one CS line to all ADEs while addressing each individual one.
   I’m starting to think the ade7758 is for meters that measure whole buildings, or settings with 3 phases, and less suitable for chaining to measure individual circuits. For just 2 phases or many circuits, ade7763 makes more sense and is cheaper.
   • Skye Sweeney says:

     April 25, 2012 at 8:29 am

     The whole ADE line was designed for power meters. They all meet the specifications for accuracy set up by various agencies. I just happen to be using an ADE because it was the simplest way to get an accurate RMS reading. Other people have simply sampled the current as fast as possible to make the computation, but my experiments indicated that although OK for resistive loads, the error would be way too large for reactive loads. I looked for just an RMS converter chip, but did not find anything I liked. The ADE with its SPI interface and the ability to integrate energy was a slam dunk for ~$5.00

Comments are closed.