Measure current thru a coil test

newplumber

Senior Member
Hello world

I bought some coils from ebay and searched the forums for reading amps using a coil but again came up short
When I first tested this coil just using a multimeter and was using a heat gun for load (one wire thru center) my fingers holding the
wires to the prongs was getting a good shock and the voltage was like 41 ac
I noticed after I tested the voltage with a 1000 ohm resistor across the coil wires the voltage dropped to 2 volts
I tried this diagram to my picaxe 20m2 and seems to work okay ( i may have drawn the zener wrong but had it the right way when tested)
I just wanted to make sure this schematic i drew was okay to use and if there could be improvements I'd love to know
I could use more smoother caps because it bounces like my bank checks
thanks again

your neighbors friend
 

Jeremy Harris

Senior Member
Please add a load resistor directly across the transformer, as there is a significant safety issue in using an unloaded, or very lightly loaded, current transformer.

The load resistor will have a low value, a few tens of ohms usually, and is there to both make the current transformer safe and to convert the output to a fairly linear voltage that is proportional to current. Calibration of a current transformer is by varying the load resistor usually. Most current transformers are supplied with a pre-fitted load, because of the danger in using one without. I've changed the load resistor value to calibrate a 100 A current transformer, used to measure bi-directional AC power on our house main power feed (bi-directional because we have a big solar array, so we export power to the grid a lot of the time).
 

fernando_g

Senior Member
The schematic is wrong and dangerous.
Connect the load resistor, with the value indicated by the transformer's datasheet, right at the transformer's output leads.
The same datasheet will either have a graph or a formula indicating the developed voltage vs current.

The maximum voltage will be small...expect about 1 volt at maximum load. This voltage will be too low to properly rectify with a regular bridge rectifier. You will require a precision rectifier, which employs an opamp.
 

premelec

Senior Member
@newpb - are you trying to measure actual current or only wanting to see if a unit is ON or OFF?
 

newplumber

Senior Member
@ jeremy thanks I will add a load resistor

@ fernando thanks okay i will add a load resistor and your right about to low of voltage for 1n4007 bridge

@ premelec thanks yes I am trying to read amps but in a "safe" way with this transformer I have if possible

Okay I changed my schematic
I added a 1000 ohm 1/2 watt load resistor at transformer
the test results
Low heat = 5.6 amps AC transformer before bridge rec = 2.3 volts ac
high heat = 11.9 amps AC transformer before bridge rec = 5.0 volts ac

low heat after bridge rec/voltage divider = 400 Mill volts DC / 1.75 mill amps
high heat after bridge rec/voltage divider = 890 Mill volts DC / 4.05 mill amps

unless I need something else added... I like the results so far
I plan on using this CT for reading min amps = 4 max amps = 12 and it does not have to be exact
if it some how goes past 15 amps the circuit breaker will trip (in theory) could also put a 15 amp fuse on load
and sorry about the mistake of coil vrs CT I'm not a perfect word for object kind of guy
thanks again
 

Attachments

Last edited:

Reloadron

Senior Member
The resistor used with a CT (Current Transformer) is typically called a "Burden Resistor" and as was mentioned these burden resistors are typically a very low value, fractional parts of an ohm to a few ohms is normal. The subsequent voltage drop is very small, typically milli-volts making the voltage too low for using a typical bridge rectifier configuration. This is a good read on the subject. A current transformer can also have multiple primary turns. Current transformers are ratio metric in that you may see a 50:5 meaning 50 amps through the primary results in 5 amps at the secondary with the correct burden resistance. Looping the primary twice would result in a 25:5 transformer.

A Google of precision rectifier circuit for ct signal conditioning will give you an idea of how the low voltage developed across a CT burden resistor can be amplified and scaled. Unfortunately I have not been able to upload images or I would have posted a few CT images.

<EDIT> Now I see your circuit. If that works well enough for you then you can run with it. Sort of an unusual approach. :) </EDIT>

Ron
 

AllyCat

Senior Member
Hi,

Yes, the Load/Burden resistor still seems rather high. What if you were to put it after the bridge, but before the multimemeter connection (for safety)? Also, the filter capacitor might be better located after the 1k, i.e. across the zener and the 470 ohms (which may no longer be needed).

Although the PICaxe is relatively "slow", you might still be able to make 10 to 20 "instantaneous" ADC measurments of the "raw" ac in a 16.7ms "window" (one cycle of 60 Hz), or an interger multiple thereof. Then calculate either the peak, or the "r.m.s." value. You might use either the ac rectified signal (as above), or directly-coupled to the pin (but with a series resitor, and beware that there can be issues with the negative-going part of the cycle), or biassed up by an additional resistor to the supply rail.

For r.m.s. (root mean square), a simple average of the sampled points will indcate the "bias" level (if there is one) and then the sum of the squares (always positive) of each offset instantaneous value (i.e. ADCvalue minus bias or bias minus ADCvalue) can indicate the square of the current. Then the square root could be estimated if necessary, to give a linear current value. That's the principle that nearly all electronic wattmeters and "smart" meters use (after measuring the "true" voltage/phase as well).

Cheers, Alan.
 

newplumber

Senior Member
Thanks Reloadron first of all thanks for explaining in very simple detail I'm learning like a burden resistor and I think to understand clearly a CT scale 50:1 2000:1 etc
"Sort of an unusual approach." <- lol yes true I always seem to do things backwards and then you forum friends straighten it all out for me (I'm still writing all my IOU's in a safe BIG box)
I also know this approach is less accurate but its just to see if a ac motor remains stable.

wow didn't see your added info Ally but I will look into that again thanks for the idea and its awesome I don't have to write it down (you did) I will try it out next week and put the results
with a new schematic thanks again

well have a Happy Easter.
 

AllyCat

Senior Member
Hi,

Beware, I didn't write it all down. There are potentially some complications with timing a single cycle, negative voltages (and numbers), calculating squares (numerical overflow) and square roots (if required), etc.. And it can get quite complicated with reactive (inductive or capacitive) loads. But we're here to help. ;)

A Happy Easter to you (all) as well.

Cheers, Alan.
 

Reloadron

Senior Member
You also want to keep in mind the forward voltage drop on a 1N4007 or 1N4000 series diode is between about 0.7 and 1.0 volt so with a full wave 4 diode bridge you will be dropping between 1.4 and 2.0 volts on alternate half cycles of your 60 Hz. I also doubt with the circuit you have measuring current will be linear.

Oh yes, in keeping with the season you all have a nice Easter Weekend. :)

Ron
 

darb1972

Senior Member

Attachments

newplumber

Senior Member
@ allycat thanks again ...when you say "There are potentially some complications with timing a single cycle, negative voltages" I understand how that can be a problem
and how the cycle is bouncing ( my word for dc cycle) and if I am on the same page I understand when I am reading the adc when the cycle is at 0 it would show no data so my reading amps schematic is useless for accuracy
and like you saying i believe is to read it many times and get the average which depending how many times reading it could in be some what accuracy or i'm in outer space so I will draw up a different schematic later

@reloader thanks i'm looking at using this brdge rec instead of the 1n4007 but haven't looked into it much still reading it to understand
https://www.fairchildsemi.com/datasheets/DF/DF005M.pdf

@darb1972 thanks for the idea and info I glanced over the 3700 and need to read lot more so I will test it when I get some in I ordered a few of them
BTW if I order from digikey since there main hub is in MN usa I get it in 2 days normal shipping which i love (everywhere else is > 5 days)
but sometimes their price and its importance isn't friends. So in a few days I am excited to try out all these new ideas if I can download them to my brain

but thanks again

@darb1972 LOL wow that is some information about the hcpl3700 in the pdf thanks for adding it
I've read it twice and still am in the fog even tho it seems to be very neatly explained
I'm also trying to see if you can read ac amps or is it just used for a AC/dc switch but I guess 12 more times reading it i should
get closer to figuring it out
 
Last edited:

fernando_g

Senior Member
WoW!!!
The HCPL-series of optocouplers used to be Hewlett Packard-Opto, which then became Avago, which then became Fairchild, which now has become OnSemi?
 

hippy

Technical Support
Staff member
I understand when I am reading the adc when the cycle is at 0 it would show no data so my reading amps schematic is useless for accuracy
If you mean zero as in when the mains cycle crosses through zero volts then that is not a problem; you are rectifying the incoming AC and have a capacitor-based peak detector. That is not quite so problematic as trying to read a signal without a peak detector.

I am also not sure how using a HCPL-3700 logic output opto-isolator is going to help when it appears you want to measure the level of signal rather than just its presence.

What you have shown in your schematic may not be perfect or ideal but I do not see that there is anything fundamentally wrong with it. It is almost exactly what I had intended to use in my own current transformer experiments. The only difference that, if it showed promise, I was going to tweak the CT load resistor to keep the ADC input below 5V rather than use a resistive divider.

That may not be great, it may not be responsive, accurate or linear, may have other limitations, but it seems it is premature to start changing things or trying to improve the circuit before seeing how well it works for the intended purposes and identifying what the actual problems are.
 

westaust55

Moderator
As mentioned earlier, the 1N4000 series diodes have a forward volt drop of the order of 0.75 V with around 50 mA of current.

An alternative is to consider Schottky diodes though some are better than others:
1. A 1N5819 will have a forward volt drop of around 0.5 Volts at 50 mA
2. a SS14 SMD diode will have a forward volt drop of around 0.3 Volts at 50 mA

I have been using the SS14 diodes as a means to minimise the volt drop for a low power circuit and have been achieving ~0.25 volt with 15 mA current.

Less voltage drop across the diodes should give better accuracy at lower current levels
 

darb1972

Senior Member
Maybe ignore my suggestion. The HCPL-3700 is possibly not suitable. I thought you wanted to measure presence rather than level. The HCPL-3700 can detect threshold voltage levels but I gather this isn't what you need.

I'd delete the post if possible but I don't think I can.
 

newplumber

Senior Member
@ hippy Yes true thanks I am trying to keep it simple and read just a average of current of the ac motor and I am still in the process of using allycats idea (testing but waiting for a better/worse bridge rec I ordered)

@ westaust55 thanks I will plan on ordering some of them (ss14 smd)... which again I didn't know existed.... I've been here for (wow time flies) 4 years and
you /hippy/many more make this the best forum ever and I appreciate that ...specifically keeping it clean

@ darb1972 LOL hey your idea is awesome because I understand .001 % of it but thanks for the info trying to help
I will try to learn/test more of the hcpl-3700 if/when I become smarter or more knowledgeable

@ everyone .. lol every time I come up with a schematic (since I'm no schematic maker) you all are probably like "OH GREAT not again!" (pulling hair out)
but I'm having fun and learning from the best :)
 

newplumber

Senior Member
Hi
I designed another schematic trying to follow allycats suggestion (i'm still learning the RMS)
I tested this plan and the results are perfect for what I am going to use it for

Thanks again for all your info
 

Attachments

Last edited:

AllyCat

Senior Member
Hi,

Move the 47uF capacitor so that it is directly across A-B , and a single PICaxe ADC reading will probably be all that you need. In practice, the 1k then could be larger (up to 10k) and (then) the capacitor smaller (down to a few uF), but it shouldn't make much difference.

Don't worry too much about rms, etc.. Most multimeters only use a simple mathematical conversion from a rectified "average" to "rms" (IIRC the "magic number" is 1.11).

Cheers, Alan.
 

Reloadron

Senior Member
As long as what you have works for you it's fine. Much of current sensing gets a little more complex but if you are only concerned with looking at a sine wave and can get a number that is about what you want then it's good. As AlleyCat mentioned when we just get a rectified value it is like many of the lower cost AC voltage or current meters. They are average responding and RMS indicating, the better more expensive devices are RMS responding and RMS indicating. The relationship is as follows:
For a pure sinusoidal waveform ONLY, both the average voltage and the RMS voltage (or currents) can be easily calculated as:

Average value = 0.637 × maximum or peak value, Vpk or Ipk
RMS value = 0.707 × maximum or peak value, Vpk or Ipk

As long as you are working with a sine wave anything Average Responding RMS Indicating is fine. Generally the output of a burden resistor is amplified using what is called a "precision rectifier" circuit because normally the burden resistor is sub one ohm, typically with a voltage drop in the milli-volt regions. If you want to accurately measure current over a wide range with linearity then it gets a little more complex.

Ron
 

newplumber

Senior Member
@ Allycat thanks okay I switched the cap in my drawing and to learn rms can go infinity with calculus (way over my head) so i'm just trying to keep it simple.
I put the wrong A-B test voltage in the first one so I changed them to the right ones. 8th schematic is the charm for not real accurate poeple like me.

@ Reloadron thanks for that ... just to make sure when you say peak value that its 0 to peak times .707 so in a example
Vpk = 169.73 volt * .707 = 120 ac measuring half cycle or to get peak from rms you can divide rms volt into .707
Of course you can throw in many different (way over my head complications) but for me I think I winning here for a simple way.
BTW when you mention lower cost voltage meters .... I think your pointing to me cause I use harbor freight free ones and they can be off +/- 99%
but when I test my picaxe related items I use a good one well probably junk for the pros and maybe someday I will get a fluke and be cool.
 

techElder

Well-known member
OT Reply: RMS

From a Fluke web page:

... a true-rms meter is widely preferred because only it can accurately measure both sinusoidal and nonsinusoidal ac waveforms. ...

  • Sinusoidal (sine) waves: Pure, without distortion, with symmetrical transitions between peaks and valleys.
  • Nonsinusoidal waves: Waves with distorted, irregular patterns—spikes, pulse trains, squares, triangles, sawtooths and any other ragged or angular waves.

As mentioned previously, rms = root mean square. Though its formula can be challenging to grasp, rms essentially calculates the equivalent direct current (dc) value of an ac waveform. More technically, it determines the “effective,” or dc heating value, of any ac wave shape.

An average-responding meter uses averaging mathematical formulas to accurately measure pure sinusoidal waves. It can measure nonsinusoidal waves, but with uncertain accuracy.

A more sophisticated true-rms meter can accurately measure both pure waves and the more complex nonsinusoidal waves. Waveforms can be distorted by nonlinear loads such as variable speed drives or computers. An averaging meter attempting to measure distorted waves can be up to 40% low or 10% high in its calculations.
The last I remember on this technical RMS description is that the method of conversion was to create a heating element inside a special IC to obtain the "dc heating value".
 

Reloadron

Senior Member
@ Reloadron thanks for that ... just to make sure when you say peak value that its 0 to peak times .707 so in a example
Vpk = 169.73 volt * .707 = 120 ac measuring half cycle or to get peak from rms you can divide rms volt into .707
Of course you can throw in many different (way over my head complications) but for me I think I winning here for a simple way.
BTW when you mention lower cost voltage meters .... I think your pointing to me cause I use harbor freight free ones and they can be off +/- 99%
but when I test my picaxe related items I use a good one well probably junk for the pros and maybe someday I will get a fluke and be cool.
That covers it. The RMS value of a sine wave is 0.707 * Epk to get to Epk knowing the RMS value it is RMS * 1.414. So exactly as you mentioned 120 VAC RMS * 1.414 = 169.68 Epk. Anyway, yes, I would just work along as you have been and less the advanced AC theory or math. What you have works for what you want so there is no sense in getting into building an active circuit to measure a RMS Current Value. The majority of hobby enthusiast do not need a true RMS meter, the Harbor Freight meters work just fine for their applications. :)

Ron
 

techElder

Well-known member
The only point to keep in mind is that measuring the "simple" way will not give you correct/the same values when evaluating a phase controlled load, because by definition the current waveform will not be sinusoidal.
 

newplumber

Senior Member
@ reloadron thanks .. sweet i finally am winning here somewhat , I just think its sad (i'll cry later) this site don't have 1 million hits a day with the info it puts out
I tell all my friends "Picaxe is easy (+ forum) to learn ...if monkeys were some what smarter they would be using picaxe controlled banana peelers"

@ Texasclodhopper Thanks for the info/pointer someday I plan on learning a lot more about "phase controlled loads/sinusoidal" but I need to do a lot more homework after i get done
with my current picaxe projects I'm working on.
 
Top