Archived Forum

It may be helpful to check that the voltage and current sensors in your emonTx are giving sensible readings.  I recently posted a simple tool for doing this, it's called MinMaxAndRangeChecker and is the second entry on my Summary Page.

(any feedback re. this tool would be much appreciated)

Many thanks, I will give them a try (once I have had my dinner!)

How are you using your current transformer? If you have clipped it around a cable (with or without an earth wire) that has two cores carrying current in opposite directions, you will read nothing. The c.t. picks up the magnetic field surrounding the cable, and if the two fields cancel....

You must clip it around either the line or the neutral conductor only. It's all in the installation instructions!

Robin,

I'm not sure what to expect in terms of numbers here, but this is what I get when I run your sketch on the emonTx

503 504   1;   506 506   0;   508 509   1;   488 533  45
503 504   1;   505 506   1;   508 509   1;   488 533  45
503 504   1;   505 507   2;   504 512   8;   488 533  45
504 506   2;   484 532  48;   508 509   1;   510 511   1
505 508   3;   485 533  48;   508 509   1;   510 511   1

507 508   1;   488 533  45;   508 509   1;   510 511   1
508 508   0;   488 533  45;   508 509   1;   510 511   1
508 508   0;   488 534  46;   508 509   1;   510 511   1
508 508   0;   488 534  46;   508 509   1;   510 511   1
508 508   0;   488 534  46;   508 509   1;   511 511   0

You can see where I moved the sensor from one port to another when the readings change drastically. So I can see that on my single SCT013 sensor I'm getting 488 - 534, range of about 45.

Is this what you would expect to see? Does this give any clue as to why my readings are so low ? Do I need to adjust my calibration value?

Regards, Keith

Robert,

I have the CT clipped round the live supply cable to the meter, I have tried neutral as well and also tried after the meter on both live & neutral, but no real difference. I intend to spend some time in the cupboard tomorrow with the emonTx connected up to the laptop so I can read the serial output directly.

I did notice that if I connect the clip I do get a spike in the reading for a few seconds then it subsides and then settles to around 4 - 5 W.

If you have the 100 A c.t. and the standard sketch, the calibration value should be close - within 5%, so a tumble drier + kettle should give you close to 20 A. The standard emonTx gives a swing of 790 counts at 100 A, so you are seeing about 6 A (1.5 kW).

Have you got the emonTx_CT123_Voltage sketch by mistake? (emonTx_CT123 assumes 240 V, emonTx_CT123_Voltage measures the voltage using an ac-ac adapter plugged in to the emonTx).

Robert,

really appreciate your assistance with this. I definitely have the emonTx_CT123 sketch loaded. I'm planning to add an AC-AC adapter shortly, but not got it yet. When I powered up the emonTx this morning I was watching a realtime visualisation of the CT output and it spiked to around 850 then rapidly fell to 4 and now it jumps around between 3 and 7 on the display. Adding in additional loads (I just turned on both electric ovens on the range cooker) makes almost no visible difference to the realtime display. I will hook up the laptop later and capture some actual serial data from the emonTx.

I'm beginning to think that I may either have a faulty CT or my emonTx was put together badly, an entirely possible situation as I've not done a huge amount of soldering for ages, although the emonGLCD and the RFM12Pi seem to be working just fine.

Is there any way to determine if the CT is faulty?

deskyeti: Is there any way to determine if the CT is faulty?

Sure, the tool that I directed you towards yesterday should be able to check this.  While current is maintained through the CT, your ADC readings should follow the size of the signal that you're measuring.  If they don't, then something is clearly wrong with your measurement circuit.

I'm not familiar with the standard sketches for emonTx, but I gather that the one which you're using makes an assumption about the voltage level.  That's fine, providing that the CT continues to feed a good signal into your chosen analog input.  I don't know which analog input(s) you're using, but that should be obvious from the output from my checker tool.

For the results you posted recently, it appears that you started measuring current on An 3 and then moved to An 1.  Consulting the emonTx Port Map on its Design Reference page shows that these inputs are connected as CT1 and CT3.

Recent versions of my Mk2 PV Router sketches include Tallymode, this being a comprehensive energy recording facility.  You don't need a triac, just a basic Atmaga-based system with a voltage and current sensor.  If you don't yet have a voltage sensor, a fixed value could no doubt be patched into place somewhere.  By porting the resulting data to a spreadsheet, you could get a nice hi-def graphical display of the power that your input sensor(s) is/are measuring.   Moreover, measurements taken with any Mk2 sketch are continuous rather than intermittent.

Even simpler and not depending at all upon the emonTx: if you have a multimeter that can measure alternating current down to 50 mA, just connect the multimeter to the c.t. jack plug - tip and sleeve. You should read 50 mA per 100 A of current.

Before you measure the current, you could check the resistance of the c.t. secondary - about 100 Ohms is what to expect. If you read anything stupidly different to that, like less than 10 or greater than 1 k, check the plug connections. There should be a red and a white wire connecting to tip and sleeve, and the screen possibly to the sleeve as well. It has been known for the solder tag to be insulated from the tip connection - easy to rectify with a blob of solder over the rivet. If that end looks OK, open up the c.t. Flip it open, and there are two lugs hooking over the bobbin cheek next to the catch. They can be pushed back and the bobbin slides out. Pictures of the inside here. Look for bad soldering and check continuity of both the secondary and the cable.

If you see 100 Ohms looking into the c.t, check that you can see the 18 Ohms of burden resistor across the socket (i.e. with the c.t. plugged in but with the cover off the plug, you read 18 // 100 = about 15 Ohms tip-sleeve).

Don't pay too much attention to the spikes, I think that is the software high pass filter settling.

Robert,

appreciate the help. I checked out the CT Plug, no obvious issues here, a bit messy inside but no shorts or breaks. Measured the resistance and I get 100 Ohms, when I plug the sensor into the emonTx I see 15 Ohms. So pretty much as described. Clip the sensor round the mains feed and I see measured across the plug around 1.55 mA background load (various things on in the house) when I load up by turning on one of the ovens I jump to around 5.9 mA, the kettle goes to about 7 mA or so.

So I think the sensor is working and picking up the signal OK, When I run Robin's MinMaxandRange checker on the emonTx I am recording values changing with the load as per my data here with my comments added to indicate additional loads

ready ...
The Min, Max and Range ADC values for analog inputs 0 to 3:
   0 507 507;     0 504 504;     0 508 508;     0 511 511
508 508   0;   498 525  27;   508 509   1;   510 511   1 Background usage
508 508   0;   498 525  27;   508 509   1;   510 511   1
508 508   0;   496 525  29;   508 509   1;   510 511   1
508 508   0;   498 525  27;   508 509   1;   510 511   1

508 508   0;   496 525  29;   508 509   1;   510 511   1
508 508   0;   497 527  30;   504 512   8;   510 511   1
508 508   0;   498 526  28;   508 509   1;   510 511   1
508 508   0;   497 526  29;   508 509   1;   510 511   1
508 508   0;   498 525  27;   508 509   1;   510 511   1

508 508   0;   464 558  94;   508 509   1;   510 511   1 Oven 1 turned on
508 508   0;   464 560  96;   508 509   1;   510 511   1
508 508   0;   463 559  96;   508 509   1;   510 511   1
508 508   0;   461 559  98;   508 509   1;   510 511   1
508 508   0;   463 559  96;   508 509   1;   510 511   1

508 508   0;   462 558  96;   508 509   1;   510 511   1
508 508   0;   461 558  97;   508 509   1;   510 511   1
508 508   0;   464 558  94;   508 509   1;   510 511   1
508 508   0;   465 557  92;   508 509   1;   510 511   1
508 508   0;   463 558  95;   508 509   1;   510 511   1

508 508   0;   465 557  92;   508 509   1;   510 511   1
508 508   0;   465 559  94;   508 509   1;   510 511   1
508 508   0;   465 559  94;   508 509   1;   510 511   1
508 508   0;   431 592 161;   508 509   1;   510 511   1 Oven 2 turned on
508 508   0;   431 592 161;   508 509   1;   510 511   1

508 508   0;   429 593 164;   508 509   1;   510 511   1
508 508   0;   430 592 162;   508 509   1;   510 511   1
508 508   0;   428 591 163;   504 512   8;   510 511   1
508 508   0;   430 592 162;   508 509   1;   510 511   1
508 508   0;   428 595 167;   508 509   1;   510 511   1

508 508   0;   430 593 163;   508 509   1;   510 511   1
508 508   0;   430 593 163;   508 509   1;   510 511   1
508 508   0;   431 592 161;   508 509   1;   510 511   1
508 508   0;   465 559  94;   508 509   1;   510 511   1 Oven 2 turned off
508 508   0;   463 558  95;   504 512   8;   510 511   1

508 508   0;   464 559  95;   508 509   1;   510 511   1
508 508   0;   461 559  98;   508 509   1;   510 511   1
508 508   0;   463 557  94;   508 509   1;   510 511   1
508 508   0;   496 525  29;   508 509   1;   510 511   1 Oven 1 turned off
508 508   0;   498 525  27;   504 512   8;   510 511   1

508 508   0;   451 570 119;   508 509   1;   510 511   1 Kettle switched on
508 508   0;   452 572 120;   507 509   2;   510 511   1
508 508   0;   449 572 123;   508 509   1;   510 511   1
508 508   0;   451 571 120;   508 509   1;   510 511   1
508 508   0;   451 571 120;   508 509   1;   510 511   1

508 508   0;   453 572 119;   508 509   1;   510 511   1
508 508   0;   452 571 119;   508 509   1;   510 511   1
508 508   0;   452 571 119;   508 509   1;   510 511   1
508 508   0;   453 571 118;   508 509   1;   510 511   1
508 508   0;   452 573 121;   508 509   1;   510 511   1

508 508   0;   450 569 119;   508 509   1;   510 511   1
508 508   0;   451 570 119;   508 509   1;   510 511   1
508 508   0;   451 571 120;   508 509   1;   510 511   1
508 508   0;   453 573 120;   508 509   1;   510 511   1
508 508   0;   450 571 121;   508 509   1;   510 511   1

508 508   0;   497 524  27;   508 509   1;   510 511   1 Kettle switched off
508 508   0;   496 525  29;   508 509   1;   510 511   1 Background usage
508 508   0;   497 525  28;   508 509   1;   510 511   1
508 508   0;   497 524  27;   508 509   1;   510 511   1
508 508   0;   497 527  30;   508 509   1;   510 511   1

508 508   0;   499 524  25;   508 509   1;   510 511   1
508 508   0;   498 525  27;   508 509   1;   510 511   1
508 508   0;   497 526  29;   508 509   1;   510 511   1
508 508   0;   497 525  28;   508 509   1;   510 511   1
508 508   0;   498 525  27;   508 509   1;   510 511   1

508 508   0;   497 526  29;   508 509   1;   510 511   1
508 508   0;   497 525  28;   508 509   1;   510 511   1

and this is a sample of data from the emonTx with no sensor plugged in running the emonTx_CT123 sketch

emonTX CT123 example
OpenEnergyMonitor.org
Node: 10 Freq: 433Mhz Network: 210
12704 12633 12669 3293
45 33 33 3293
32 0 0 3293
27 0 0 3293
29 0 0 3283
27 0 0 3293
2 0 5 3283
0 10 29 3293
0 6 36 3283
0 6 25 3293
0 5 22 3283
 

So it was this point when I started to realise that I might have been a little bit of an idiot! When I was plugging in the CT at first I was basing the location on the labeling shown on the picture of the enclosure on the emonTX page instead of looking at the board and the schematic. I was using the "top" socket when I should be using the "bottom" socket. How can I have been so dumb? 

So thanks to Robert and Robin for your assistance and patience. I have however learned a little bit more about how this all works, and also to have a little more faith in my soldering abilities, but reinforced the old maxim of RTFM !!

That's great, another system up and running!

A single CT will never record zero power simply because any small amount of signal will be squared so the sign disappears. Once you get a voltage sensor in there, your power/energy contributions will gain a sign bit and an overall zero value then becomes possible.  When no real current is flowing, I would not expect your system to be seeing more than a Watt of leakage current.  If it does, try running any of my Mk2 PV Router sketches which measure energy flow continuously.

If there's a mistake on the page, we'll get it corrected! Which page did you misinterpret?

deskyeti's display from MinMaxAndRangeChecker nicely shows how very small signals can be significantly affected by precisely where the neutral point sits on the ADC's scale.  The first circuit (which I presume is live!) never moves from the value of 508.  Its energy content will therefore be reported as zero.

The fourth circuit appears to be sitting close to an ADC boundary in that the range during each 3-second measurement period is always 1.  If this movement (between steps 510 and 511) were to be in synch with the mains voltage, which is very likely, then this signal would be seen as having a power of roughly one thousandth of full scale.  As full scale is around 20 kW, this one-step background signal would be seen as having a power of 20 Watts.  There's more about this topic in a Building Blocks page.

Both of these channels are equally good (in fact, with so little noise in evidence, they are both amazingly good!) but they are likely to record quite different power values at very low currents.

If there's no need for a system to record up to 20 kW, performance can undoubtedly be improved by increasing the burden resistor so that the CT's sensitivity is increased.  For a PV Router application, there seems little point in extending the system's measurement range beyond the PV's ability to generate power, or the dump load's ability to consume it, whichever is the greater.

Robert,

it was me having the picture on the emonTx main page showing the schematic diagram of the whole system that threw me.

http://openenergymonitor.org/emon/emontx

I realise now that it is not an accurate representation of the actual device, but the top-down ordering of the CT sensors CT1, CT2, CT3 had stuck in my head and when I plugged in my own I assumed top-down. As I did more debugging using Robin's tools and then the serial output from the sketch it suddenly clicked that I had the plug in the wrong socket! When I initially tried the other sockets I did not have the correct sketch loaded (no TX so no data sent) and after that when the TX sketch was loaded I did not try the bottom socket.

Thanks again for your input.

Regards, Keith

Robin,

the data I posted from your MinMaxRange tool only had a single CT sensor connected at any one time. The annotated data was with the CT connected in to the "top" socket which is actually CT3 (a0 in your terms I believe) not CT1 (a3). The second set of serial data from the emonTx_CT123 sketch was with no sensors connected, this is displaying CT1, CT2, CT3 and battery voltage.

Thanks again, Keith

Hi Keith, Yes, I think I'd pretty much worked out what was connected at the time of your readings.  Your annotated display was excellent, it's good to know that some of the tools that I've produced are being used.  I think that this latest one, which displays the ADC data from each of the first four 'analog' ports, should be particularly useful.

The association between the analog pins on the processor and the analog ports on the emonTx is not of my making.  It's fixed by the layout of the board, but is not very intuitive and may well lead to further confusion. 

When I first started to look at the emonTx, I was baffled by the connectivity, hence my cry for help thread.  The Port Map on the emonTx Design Reference page is well worth bookmarking.

Keith:
I'll mention this to Trystan. I suspect he won't want to change it, because as you say it is just an overview (and in his defence in the photo below I can actually read the port designations!). A new picture is likely to appear shortly with all the I/O ports labelled, so hopefully that will avoid someone else falling into the same trap in the future (this one).

Anyway, I'm glad to know you are up and running.