IoTaWatt with SCT-013-030 - config info?

Hi.

Did anyone use the SCT-013-030 before?
Which values have you used to configure it?
I’m thinking about buying the IoTaWatt, and have quite a few of these CTs, so I was pondering on using’em…

Thanks in advance!

For anyone arriving here while trying to decide whether to get these voltage type CTs or the basic SCT013-000 current type, I would strongly recommend going with the current-type. You may be tempted by a low-price offer on eBay to get the 030 (or 020, 050) but if you look around a little, you should find similar offers for the SCT013-000. They are also available from OEM.

If you already have some 030’s and are a little handy, read on and I’ll fill you in on how to use them with your IoTaWatt.

Those are voltage type CTs. They are supported in the configuration tables, but you would need to remove the burden resistors inside IoTaWatt on the channels you intend to use them on and then set the burden value to zero in the device/burden config area. After that, you can just select the SCT013-030, or any other voltage CT like SCT013-050.

Alternatively, you can remove the burden resistor from the SCT013-030. The guts just slide out exposing a little circuit board with the TVS diode and the resistor(s) (sometimes they use two to achieve the needed burden). Remove them and reinsert the guts, carefull all the while not to chip the brittle iron cores.
Once you finish the “burdenectomy”, those CTs become current type that can connect to an unmodified IoTaWatt channel. Configure them as “generic”, with turns=1860, and phase = 3. When I do this to a CT, I usually mark the outside 1860/3.0 with a felt marker.

Aside from the physical difference, there are a couple of other diferences beween the two methods:

When you remove the internal burden in IotaWatt, the CT has the stated range of 30A and operates at full scale resolution, which theoretically means greater accuracy.

Removing theburden(s) from the CT will result in a CT with a range of about 77A and so resolution is a little less than half.

The resolution issue isn’t really an issue, as IoTaWatt will report even low power accurately either way.

Thank you very much for your extra detailed explanation!
Never tried to remove the burden resistor from the SCT-013-030, but it seems preferable than to mess with the IoTaWatt, as it would void it’s “generallity” - would be specific to these current sensors…

Regards!

Brand new and excited to get my IotaWatt going!

I pre-ordered a bunch of CTs before I understood well enough, it seems. I bought a load of SCT-013 15A and 20A CTs.

So I’m reading this thread, and I guess I will remove the burden resistors from my CTs (unless there is a better soln?). So I’ve opened up my 15A / 1V CT, I see 2 tiny components (they are the resistors?) and one larger component.

So I just remove the resistors and replace with a short?

Also, what are the turns and phase for the 15A and 20A CTs?

I’m excited to get going!

Thanks!

Edit:

I couldn’t wait - I opened up a 100A I have, and I see the resistors are removed and circuit left open, correct? Now I just need to know the turns/phase for the 15 and 20A - thanks in advance!

We believe the differences between the YHDC SCT-013-xxx c.t’s with a voltage output and the genuine current output types are:

  1. The number of secondary turns is slightly different, the current rating of the voltage type being determined principally by the value of the burden. It may well be that in order to use a standard value of burden resistor, the number of turns will be adjusted by a small amount for each different current rating,
  2. The 50 mA current output type has a voltage limiting Transient Voltage Suppressor across the output, to clamp the voltage at a safe value in the absence of a burden resistor,
  3. The 1 V output type does have the burden resistor, it might also have (but strictly does not need) the TVS.

If you do attempt a conversion, you’ll need to determine your own calibration data, because you have in effect created a new untested c.t. model.

See Better results up to 35 A - #3 by Simsala

It might be a good idea to reproduce some or all of this thread at The IoTaWatt forum, which is specifically about the operation and use of IoTaWatt.

The 30s and 50s are 1860 turns, phase 3.0. You can try that. Since you have a 100A SCT013 you can clamp it to the same circuit, preferably with a steady load of a few hundred watts, and compare the power. You can then adjust the turns to best match the SCT013.

To the point previously mentioned about posting in the IoTaWatt forum, I’ve repeatedly tried to refute that CTs and burdens need to be full range matched to get accurate results with IoTaWatt. I don’t want to deminish the significance of that advice for the Emon devices as it no doubt is the result of lots of experience, but likewise that it is unnecessary with IoTaWatt is borne out with experience.

That said, the Echun 50 Amp CTs that we use are more appropriate for those circuits because they are smaller, more economical, and have less shift range than the SCT013.c&

@avp

Today I was working with something unrelated and wanted to change the burden resistor on one of the input channels. Before I heated up the soldering iron, I picked up an old SCT013-030 and pondered. That CT has an intenal burden of 62 ohms which when combined with the IoTaWatt 24 ohm burden yields an effective 17.30 burden. That was close enough for what I wanted to do, so I just went to the setup tab, selected “Burden Resistors”, and changed one of the inputs from 24 to 17.3. I configured the input as generic with 1860 turns and phase 3. Plugged in the unmodified SCT013-030 and it reads exactly the same as a standard SCT013-000.

The 30 A version is starting to push the envelope but it gets better as you go down. Here’s a quick chart that I made up:
image
So what I’m saying is that if you configure one of these voltage type CTs as generic 1860 turns phase 3 (wag), and set the burden for that input to the “Combined Burden” value in the table, you should get decent results.

This approach eliminates the need to do anything with burden resistors. I’ve not tried it with models other than 30, but it’s worth a look.

Newer IoTaWatt now in manufacture will have 20 ohm burden, so the table changes a little:
image

@overeasy Thanks for all the info.

So I finally got around to modifying my 15A unit. I removed the 2 burden resistors. I connected it and another 100A unit up to the same circuit (a circulation pump). The modified 15A read 25W, and the unmodified 100A read 26W. I was quite happy with that. I then jumped back on the forum here and saw your next post about modifying the burden values (therefore requiring no HW mods).

I tried another 15A, used 20.11 for the burden. It reads 23W. Tried another. It reads 24W.

All readings are very close. Is there anything you would suggest to make things even closer? Or am I in the margin of error of these sensors?

Any other advice/ideas?

Thanks!

All of the SCT013 series CTs are not particularly accurate in that very low power range, and to be honest the IoTaWatt isn’t built to do that with high precision either. Regardless of whether you are using an SCT013-000 (100A) or a voltage type SCT013-015 with no burden, or a hybrid as with the SCT013-015 with combined burden, the bottom line is that you have a ~2000 turns CT, and trying to measure 200ma with a 100A CT is trying to differentiate less than 0.1% of the range of a CT that has a 1% accuracy. If you need that precision, you will need to spend a lot more money on a more sensitive and accurate power meter.

If you want somewhat more accurate and consistent results at low wattage with IoTaWatt, you should invest in some quality 1000 turns CTs like the Echun ECS1050. The Echun ECOL09 is a solid core that is very stable at low current.

No doubt you’ll get a lot of other advice about putting multiple turns through the CT to amplify the current, and that’s all sound advice for those so inclined, but with IoTaWatt I try to stick to plug-and-play solutions.

I like this solution (changing the burden resistor values in the configuration) to avoid soldering in the sensors or the IotaWatt. I tried it on my system and somehow my readings seem to be off, generally they are around 2/3 of the expected values.

For example, I am monitoring our 3-phase solar setup and compared with readings from the inverter. The 3 phases measure more or less the same values (each about 1/3 of their sum), but the sum seems to be off. Power factor is very close to 1.

I work with 020’s and 030’s. I set the burden values to 19.08 and 17.3 respectively, windings to 1800 (from datasheet) and the phase to 3deg.

I hope someone can point me to something I may have done wrong. Thanks!

Couple of things you can try:

I know the data sheet says 1800 turns but every one I have measured was actually 1860. That won’t get you there, but will set off 3% or so.

If you have a recently purchased IoTaWatt from OEM, the burden resistors are 20 ohm. If so, look at the second table that I published above. The melded values for 020’s and 030’s would be 16.46 and 15.12.
You should have seen the factory burden values for all of the CTs when you set the modified values in the config app. Were the others 20 or 24? If they were 20, then use the melded values for 20 ohm.

If not, get back to me and we’ll take it from the top.

@overeasy

Thanks for the help. Good to know that the CT’s have 1860 turns. The burden resistors were originally set to 24 Ohm, so I would guess I have the “old” version (is there any other way to check this?). Nevertheless, I treated the system as a 20 Ohm one and changed the burden values accordingly. After also updating the CT turns value my measured values probably went up by a factor (1860/1800)*(24/20) = 1.24.

The weather today was too cloudy to test with the solar measurements. Instead I took the water kettle (2kW). This is a group measured by a 020 CT. I compared the measurement of an optical pulse sensor on the meter with the reading of the IotaWatt. I still see a rather large deviation: 360Watt difference between the measurements. That is 360/(230*20)=7.8% of the sensor scale. Still a rather large deviation, isn’t it?

OK, that was a first pass going on very little information. So lets start at the top.

I don’t follow your description of the water kettle experiment. Unless I’m missing something, the optical sensor on the meter counts Watt-hours or energy. When you say the difference between the IoTaWatt and that sensor is 360 Watts, I get lost. It’s apples and oranges. The IoTaWatt should be within 1% of your inverter when you get this right.

You say you have three-phase, so I am assuming you are using the IoTaWatt “derived three-phase” reference method.
Can you confirm that and show:

A screenshot of your input configuration display?
A screenshot of your status display when the sun is shining?
A picture of your IoTaWatt with VT.
A picture of your installed CTs on the inverter.

Assuming that guess was correct, let me explain the problem: There are 48 possible combinations for a set of three CTs on three different phases when you allow the possibility any CT can be reversed. Only one of those combinations is correct. That’s why it’s necessary to have a methodology to get a correct result. There a quite a few incorrect combinations that will yield the wrong result but be equally wrong on all three phases, and we haven’t yet even proven that it’s a phase problem.

If I were there I could probably get it right in about ten minutes, but it becomes a very tedious iterative process when done this way. It could take dozens of iterations, so the total time will depend on how quickly the iterations occur. So if you could answer the leading question and post the information I’ve asked for, we can get started.

@overeasy

I will start with clarifying the water kettle story. The system with the optical sensor counts Wh’s indeed, but it also takes the time derivative and shows power. So I am really comparing Watts numbers. With each measurement system I looked at the differences when switching the kettle on and off (other appliances were active as well, seemingly with constant power). These differences gave me the power readings. The optical sensor system gave a reading of about 2000W and the IotaWatt reading was 360W lower.

Yes

For a status screenshot I need to be home during the day when there is good weather. It may take a while before I have this. I will add a screenshot of the 3-phase power data of such a day. I guess you are interested in the power factor? This is always >0.90 for all phases.

Sorry for the spaghetti (“solar phases” are coming from the inverter):

The VT:

I think I should explain how I went about configuring the phases. I figured the VT has to be phase A (although it is plugged into an outlet of the 3rd phase in my electrical cabinet). Then the phase B and C only needed to be placed at the right phase, I thought. I chose them such that the power factors of the measurements (when some serious load/consumer is active) looked right. To be honest, I did not pay attention to the orientation of the clamps. I just inverted the values if the signs were wrong. Perhaps I messed something up in this process?

Let’s hope for the best :slight_smile:
In any case I really appreciate your help!

P.S. In the meantime the discussion has drifted away from the original topic. I am not sure how to take it to a new one.

I used to have super powers here, but no longer. I would actually rather do this on the IoTaWatt forum. So if you would like to register there and post a quick message, I’ll jump on with the next step.

These three phase installations can be tricky, and you may be interested in some of the others that we’ve worked through there.

UPDATE: Hello @jascha, I believe I have found the basic problem. The good news is that although I can see at least one backward CT, I think the problem is a bug introduced with recent changes to the burden resistor settings. It is easily fixed. Please register on the above forum and I’ll get you going straightaway.

Just a note to thank you for outstanding product support :slight_smile:

I was getting strange readings with my new IotaWatt, then I realized I had purchased SCT013-050 CTs by mistake. Using the combined burden method I am now getting perfect readings on each of my three phases :slight_smile:

Franck