I have a CT that has a very long cable, depending on and where I tidy that extra cable the CT is picking up more or less noise. I didn't think the CTs were able to pick up noise from their cables, but apparently they can.
Anyone has any ideas on what the best way to deal with the excess cable of a CT, besides cutting it ofcourse!
Re: CT picking up too much noise
Are you certain it's getting in via the cable, and not capacitively coupled into the CT secondary winding?
If it's a shop CT, you'll find that the screen is not connected anywhere. While that doesn't normally cause a problem, if you're in a particularly hostile environment it might do, even though the impedance it's working into is only 18Ω. You could try stripping back the screen and earthing it (to a good earth, not to the plug sleeve). Or as you imply, if everything is installed in its final position, then cutting the cable to length (plus a bit!) and earthing the screen would be my approach.
Re: CT picking up too much noise
Cutting the ct cable to length seems to done the trick. I actually think the cable shield was touching the ground on the stereo plug in addition to the red wire, that could have been why it was picking up so much noise. Anyway cutting and resoldering fixed both issues. Now the readings are below the 4w mark.
Re: CT picking up too much noise
Something strange happened today. I had to restart the emontx after I moved things a little in the enclosure. Now my noise level goes negative as well as positive. Previously all noise readings were only positive. But since I restarted the readings go both sides of 0. Any clues as to why. No new sketch. Just a reboot and some minor physical movement.
Re: CT picking up too much noise
"some minor physical movement"
That's probably your answer.
Re: CT picking up too much noise
Robert, that's what I originally thought but here is where is getting stranger. My emonTX is connected to a Serial to Ethernet bridge. Both are powered by the same 5V input. Previously when I had just one wireless access point configured on my Bridge the noise level was only +ve. When I added all my other wireless access points, my noise level was both +ve and -ve. I then removed the other wireless aps and went back to a single ap configuration and the noise is back to being +ve only. Each time I have to power cycle. No physical movements at all because everything is done remotely.
So either the extra processing that the bridge has to do when its handling multiple aps is depriving the emontx from 5Vs or its just coincidental and actually its the power cycle that is resetting things.
Re: CT picking up too much noise
I think the noise is correlated with the 50 Hz mains and the bridge is drawing slightly different currents depending on what it's handling and that's putting it on the 5 V supply; or (highly unlikely but I never rule anything out where noise and interference is concerned) it's RF getting in somehow.
Can you filter the 5 V between the bridge and the emonTx?
Re: CT picking up too much noise
The bridge has got a wifi chip on it too, could that be causing the RF noise. What's the best way to filter the 5V before it hits the emontx? I thought that since the 5V was going through the emontx regulator that would be enough.
Re: CT picking up too much noise
"I thought that since the 5V was going through the emontx regulator that would be enough. "
I've been caught out by that before. You need to look at the data sheet for the MCP1702 regulator, Figures 2-25 & 2-26. At worst, at 4 kHz, the power supply rejection is only about 12 dB, so ¼ of the 4 kHz noise on the 5 V rail will get into the 3.3 V rail.
Re: CT picking up too much noise
Would using a regulator that has an ultra low noise output work better i wonder. e.g. ADM7150ARDZ, TPS7A4700 or LT3042
But I'm guessing these are an overkill for what we are trying to do here. So a certain amount of noise is to be expected and lived with!
Re: CT picking up too much noise
A certain amount of noise can actually help, because it adds jitter to the signal that, if it happens to be the correct amplitude, will improve the performance. But in this case, bigger is definitely not better.
You need to look for the best output regulation - particularly against input voltage but also against load changes (that's important as the radio turns on and off).
Re: CT picking up too much noise
I'm a big fan of dithering in DC applications. I think I've mentioned before that I've managed to turn a 10-bit ADC into an ammeter with 21mA of dynamic range and very close to 1uA of resolution. Without dithering, conventional ADC maths would tell you my resolution should be about 20uA. The secret is adding in white noise whose long term mean is 0 so that over enough averaged oversamples, the noise makes no difference to the result. Alternatively, if the noise has a non-zero but constant mean, then you can calibrate away the constant offset introduced by the noise.
Using dithering on an AC signal is a bit more complicated. Again you want the noise to have no correlation to the signal of interest; ideally the noise frequency would be well away from the bandwidth of interest in the signal. If your sole mission is to replicate a signal as in an digital audio application, then it all makes sense. But in a power meter application, invariably the first thing you do with a reading is multiply it by another reading, either by itself in the case of RMS calculations, or by its (I,V) buddy in the case of real power calculations. There in lies the rub. Even if you had managed to get the noise nice and random, as positive as it was negative over time, once you multiply it by itself it's going to be positive and will introduce an offset to the result.
The problems with any unstudied noise coming on the Vcc rail are that it's likely to:
Ideally you'd dither V and I each with their own unique noise source. The two noise sources need to be uncorrelated to each other, and uncorrelated to the mains. Even then, you'd need to calibrate out the resultant offset it produces in the RMS values.