Archived Forum

Welcome, Alan.

The c.t's where you have them will indeed only record the power (or more correctly current as you're not monitoring the voltage) consumed by their respective loads (unless of course there is a source of generated power downstream of the c.t. that you don't know about).

(Incidentally, you need to tick the "List" box when you attach a picture to your post).

Thank you Robert, so to be clear the ct's will only register the current flowing into the fuse boards and will not be affected by any power that is being exported. ( yes 'List' I will look out for that , I was having trouble posting the picture)

That's good, and so my next question is the best way to monitor the current

I trust it's ok to ask for advice on here despite the fact that I dont think it necessary, in this case,  to use an Arduino product

a) The CT is a SCT-013-000  which produces 50mV at 100A  It has transient voltage suppression built in

b) An external burden resistor can provide a wide range of voltages 

c) I am working with burden resistors in the 80 to 180 ohm range

d) The 180 ohm resistor produces about 2.5v pk to pk with 12A through CT

                                                                up to 16v pk to pk with 100A through CT

e) It is not necessary to monitor differences above 22A as this is the max output of the wind turbine and so it is irrelevant as to which fuse board the power travels at currents above 22A

f) I first tried simply rectifying the ac from the CT and feeding to an op amp comparator but the volt drop of the diodes prevented results at low outputs

g) I tried biasing the ct output above the diode voltage drop, should have been ok if both inputs were dc biased by the same amount into the comparator, it seemed to work to a degree but odd things happened at the comparator 

h) The monitoring doesn't require any particular accuracy, it's just a case of being able to detect which fuse board is drawing the most current (up to 22A) and switching a contactor to suit.

i) Capacitor rc times could be used as it would be better if the frequency of the contactor operation was limited , it's not necessary to switch back and fore for the sake of a couple of amps

j) So what do you think is the best approach,, a precision rectifier to comparator, amplify the ct ac and then to a comparator.

or any other ideas

I'm an electrician not an electronic so please keep it simple 

Regardless it is good to know that the exporting current will not effect the operation of the ct's , the AC current traveling in different directions is beyond me, I always thought it just oscillated back and fore at 50 cps

thanks

"a) The CT is a SCT-013-000  which produces 50mV at 100A  It has transient voltage suppression built in"

NO. It is a current transformer and outputs a current - 50 mA per 100 A. The "transient suppression" is there (I believe) to prevent dangerous voltages if you happen to open-circuit it. You can quite happily short-circuit a c.t.

"b) An external burden resistor can provide a wide range of voltages "

A c.t. is the dual of a voltage transformer, you overload it by asking for too much voltage. Therefore to preserve the fidelity of the output waveform, you need to keep the burden resistor as small as possible. Read this report on the c.t.

"c) I am working with burden resistors in the 80 to 180 ohm range"

You'll be close to maximum output and the transformer may exhibit signs of saturation.

d) The 180 ohm resistor produces about 2.5v pk to pk with 12A through CT

                                                                up to 16v pk to pk with 100A through CT

No it won't, and the output won't be anything like the input. The report contains pictures showing you what happens.  Therefore, depending on how you convert the ac to dc, there won't be anywhere like a linear relationship between output and input. You need to think carefully whether this will bother you. It might not. 

e) It is not necessary to monitor differences above 22A as this is the max output of the wind turbine and so it is irrelevant as to which fuse board the power travels at currents above 22A

Isn't it? Does the difference between the two import tariffs and feed-in tariff and what the meters read affect that decision? (i.e. don't you want to ensure the most expensive meter reads the smallest imported energy total, or don't I understand how the charging works?)

f) I first tried simply rectifying the ac from the CT and feeding to an op amp comparator but the volt drop of the diodes prevented results at low outputs

That is to be expected. You need a "precision rectifier" circuit, this includes the diode drop inside the feedback loop of an operational amplifier and so overcomes the diode drop.

g) I tried biasing the ct output above the diode voltage drop, should have been ok if both inputs were dc biased by the same amount into the comparator, it seemed to work to a degree but odd things happened at the comparator 

h) The monitoring doesn't require any particular accuracy, it's just a case of being able to detect which fuse board is drawing the most current (up to 22A) and switching a contactor to suit.

i) Capacitor rc times could be used as it would be better if the frequency of the contactor operation was limited , it's not necessary to switch back and fore for the sake of a couple of amps

j) So what do you think is the best approach,, a precision rectifier to comparator, amplify the ct ac and then to a comparator.

I would amplify and then rectify, then drive the comparator from the two outputs. Rather than just time (but you would probably need to set a minimum time for each state anyway) I think you need to build in some hysteresis into the comparator, so that it switches when the difference is a finite current (say 1 A) one way and only switch back when it is the same the other way.

Regardless it is good to know that the exporting current will not effect the operation of the ct's , the AC current traveling in different directions is beyond me, I always thought it just oscillated back and fore at 50 cps

Yes, but it's the phase of the current with respect to the voltage that changes. See 'Direction of Power Flow' here.

I know you've said you don't want an Arduino product, however the emonTx with a bit of add-on gear will do all you need. Get an emonTx kit less the RF (you don't need it), 2 c.t's, a programmer, a triac driver and a triac to drive your contactor on an output, a 5 V power supply and a suitable box and it's all done (given you have a computer to program it). You should see change out of £100 for the hardware. Read up here on Calypso-rae's Mk2 and MartinR's development of it - the hardware you need is broadly speaking very similar (small external pcb's for the triac & driver are available). The emonTx input is good down to below 1 A without any extra amplification, so you can forget amplifiers and comparators. The software would take care of switching times and hysteresis.

Thanks Robert, you have given me plenty to think about.

Yes I understand , mA not mV , i typing error 

Yes you are correct to question the tariff details, as one tariff is half the price of the other I intended to compensate by making one burden resistor half the size of the other, the outputs of the ct's would then be skewed to take into account the different tariffs. 

As to the output voltages I had better go and double check my results, I have that many results on scraps of paper from different trials, i'll get back with that.

thanks

Spoot:Yes you are correct to question the tariff details, as one tariff is half the price of the other I intended to compensate by making one burden resistor half the size of the other, the outputs of the ct's would then be skewed to take into account the different tariffs. 

My preference would be to have both current sensors measuring their respective circuits as accurately as they can, and then apply any comparison logic in the software, rather than the hardware.

That's how I would do it with a µP solution, but Spoot's original preference was for a purely analogue hardware (op.amps & comparator) design, which makes a variable gain somewhere a necessity. That's the technology I was brought up on, so it's not entirely unfamiliar. He says he's an electrician, not an expert on electronics, so I think he'd be a lot safer, and find an easier, more versatile and probably cheaper (certainly cheaper if he factors in his time) solution in an emonTx / Arduino, than trying to design and prove an analogue hardware solution from scratch.

If he does go for the emonTx, then with the addition of the RF module and a GLCD (well under £100 extra even when you add the 3rd c.t., voltage input, another box, cables & power supply) he could sell his customer a real-time display of consumed & generated power.

Attached is a graph of my results with the ct's  I used a combination of light bulbs from 60W to 500w to vary the current through the ct  The ct was wound with 50 turns of wire, this scaled up the current to simulate 12.5 A to 105A

You can see there is a plot of the pk-pk output for both the 82 ohm and 180 ohm burden resistors.

The waveform is reasonably undistorted up to 7v Pk-Pk  (82 ohm burden)  and 10.5v Pk-Pk (180 ohm burden)

The two plots are relatively proportional to the input current ( good enough for my purpose,I think)

At the 105 A point there is distortion of the waveform, nevertheless even the distorted waveform can be rectified to give a reasonable indication of the current flowing through the ct

DC measurements after the bridge rectifier produced approximately halve the Pk-Pk readings

You can see in the sketch exactly how I made the readings, a voltage devider gave me 6 volts to overcome the diode drops.

Unfortunately I couldn't manage to connect this successfully with a voltage comparator , when the comparator switched it's inputs + -  also seemed to change, the hysteresis feedback resistor perhaps the cause, but I don't know enough about op amps to figure it out.

I am happier to continue with the analogue circuit 

I know nothing about the digital circuits you recommend

An analogue circuit would cost less than £10

The enclosure containing the contactor , comparator and power supply needs to be as small as possible as the unit will be mounted in the fuseboard cupboard of a domestic dwelling which are often short of space

thanks for your help

"An analogue circuit would cost less than £10" with a box and power supply and 2 c.t's?  Are you sure?

A comparator is quite simply an amplifier that has far too much gain and is allowed to saturate! That's all there is to it. To add hysteresis, you are actually adding a controlled amount of positive feedback in order to make it remain latched in the previous state even though it wants to flip.

This is just a rough outline - it's not designed, just thrown together! The voltage sources are your rectifier outputs.

Nice pic, Robert.  Why doesn't the output briefly go low at around 67 seconds?  Maybe this amplifier runs out of steam when close to the supply rail ...

It's due to the hysteresis caused by the feedback resistor R3

OK, it's just that at T=25 it looks to switch sooner than at T=67

deleted 

I wouldn't believe that simulation too much. I didn't check the op-amp properties. In a real case, you'd have an undefined condition if the inputs went out of their permissible range, and I'd suspect that could be what happened.

If the LTSPICE performance alters when the inputs are close to the voltage rails, then I'm impressed.  In my limited experience to date with "Livewire", I've never found their modelled components to be a particularly good match to real ones.

It all depends on the accuracy of the models, and how the simulation is run. You can have highly detailed models that are highly accurate but take a lot of processing power, or less accurate models that run faster and use less resources.