YHDC current sensor - the theory please

Hi All,

I'm resuming a project I started more than one year ago; it is about to use a YHDC current sensor.

Attachment 10007


For now, I have a 16F18446 carrying a 12 bits ADC that I hope sufficient just to put me on the track and avoid the usage of an external higher resolution ADC chip adding "difficulties".

The YHDC sensor I have on my bench is a 100A:50mA model.

I found a lot of these drawings like this one below but I can't make them work or better said, I don't know how to "read" their output.

Attachment 10010



Do I have to sample let's say a 30 to 50 times or so to grab an alternance, find out both zero-crossing values (or as close as possible) and then extract the higher value in between?

What is the correct way to do this?

Thanks for any explanation on this one :wink:

If you are trying to read the magnitude of current flowing in the primary of the current transformer(CT), you would like to first amplify the AC output of CT using an op-amp. Then do a DC peak conversion on the output of the op-amp. Read this value. It will be proportional to the current flowing in the primary. The ADC on the PIC can now read this DC value in a poll.

a simple digital peak detector will give a reasonable result,

eg
2v p-p input offset to approx 1.5v for a vcc of 3.3v for adc

Attachment 10011


[arduino ide serial plotter]
at a 200us sample rate
Attachment 10012

at 1000us sample rate , its still not bad
Attachment 10013

Code:

---

'****************************************************************'*  Name    : peak_detector_ADC_16F18446.pbp
'*  Author  : richard
'*  Notice  :
'*          : All Rights Reserved
'*  Date    :
'*  Version : 1.0
'*  Notes  :








 '====== PIC 16F18446 FUSES ========================================================================
  #CONFIG
    __config _CONFIG1, _FEXTOSC_OFF & _RSTOSC_HFINT32 & _CLKOUTEN_OFF & _CSWEN_ON & _FCMEN_ON
    __config _CONFIG2, _MCLRE_ON & _PWRTS_PWRT_64 & _LPBOREN_OFF & _BOREN_SBOREN & _BORV_LO & _ZCDDIS_OFF & _PPS1WAY_OFF & _STVREN_ON
    __config _CONFIG3, _WDTCPS_WDTCPS_31 & _WDTE_SWDTEN & _WDTCWS_WDTCWS_7 & _WDTCCS_LFINTOSC
    __config _CONFIG4, _BBSIZE_BB512 & _BBEN_OFF & _SAFEN_OFF & _WRTAPP_OFF & _WRTB_OFF & _WRTC_OFF & _WRTD_OFF & _WRTSAF_OFF & _LVP_OFF
    __config _CONFIG5, _CP_OFF
  #ENDCONFIG


' ====== REGISTERS =================================================================================


    ANSELA  = % 00100000
    ANSELC  = % 00000000
    TRISC  = % 11101111
    ADCON0  = % 10000100 'ADC is ON, RIGHT justified result


' ====== DEFINES ===================================================================================
    DEFINE OSC 32
    DEFINE ADC_BITS 12    'Number of bits in ADCIN result
    DEFINE HSER_TXSTA 24h ' Enable transmit, BRGH = 1
    DEFINE HSER_CLROERR 1 ' Clear overflow automatically
    'DEFINE HSER_SPBRG 138 ' 57600 Baud @ 32MHz, -0.08%
    DEFINE HSER_SPBRG 68  ' 115200 Baud @ 32MHz, 0.64%
    SP1BRGH = 0
    BAUDCON.3 = 1        ' Enable 16 bit baudrate generator
    pause 500
    RC3PPS = 15          ;tx            PORTC.3
    RA0PPS = 15          ;tx            PORTA.0      ;PK3
    RC4PPS = 0          ;rx            unused
    LED    VAR LATC.4
    clear
    new_read    var word
    State        Var BYTE 'serial Bd (baud per second)
    peak        var word
    last        var word


' ====== setup PROGRAM =============================================================================
    led = 1
    ADCIN 5, last 
    pause 1000
    HSEROUT ["ready ",13,10]
    led = 0   
    ;707 -3670
' ====== MAIN PROGRAM =============================================================================
MAIN:
    ADCIN 5, new_read ' ADC
    gosub getpeak
    HSerout ["ADC:",dEC new_read,",peak:",dEC peak,13,10]                    '    PLOTTER
    pauseus 200  ;  sample rate                       
GOTO MAIN


getpeak:
    if state==0 then    ;rising
      if  new_read < last  then
        state = 1;falling
        peak =  last
      endif
    else        ;falling
      if  new_read > last  then
        state = 0;rising
        peak =  last
      endif
    ENDif
    last = new_read
return

---

Peak detector! That's it!

Thanks a lot for explaining and the code I can't wait to try when I'm back from office :biggrin:

Though, why do I always read about 3.3V measuring circuits and not 5V?

Also, a "peak detector" can be as simple as a resistor and a capacitor in parallel "smoothing" the output of the sensor, or like Jerson wrote, of the op-amp, right?

If "yes", how do I dimension them? Are trials and readouts on the scope the best way to go?

Quote:

---

why do I always read about 3.3V measuring circuits and not 5V?

---

my dev board has a 3.3v reg, otherwise can't see that it makes a difference

Quote:

---

a "peak detector" can be as simple as a resistor and a capacitor in parallel "smoothing" the output of the sensor, or like Jerson wrote, of the op-amp, right?

---

not really , it needs a rectifier too. if the wave is not symmetrical the neg and pos peaks may be needed. plus it still needs adc code to read it in anyway

Quote:

---

Are trials and readouts on the scope the best way to go?

---

almost essential, especially depending on the nature of the current being monitored,
if the current is not sinusoidal and/or varies rapidly then its going to need filtering and other compensation to get any sort of reasonable calibration

Quote:

---

a "peak detector" can be as simple as a resistor and a capacitor in parallel "smoothing" the output of the sensor, or like Jerson wrote, of the op-amp, right?
not really , it needs a rectifier too. if the wave is not symmetrical the neg and pos peaks may be needed. plus it still needs adc code to read it in anyway

---

May I kindly ask for a principle circuit to start with the correct foot please?

I don't see how to "rectify" a very low AC signal - I'm an electrician and used to easely handle voltages over 230VAC but when it comes down to millivolts.... 😬

Why do want a nasty old analogue peak detector, When you can create a digital one as per my code ?
Using the very cct you posted.
If you realy need to try analogue google ""peak detector circuit"



An accurate analogue peak detector uses an opamp to create an "ideal" diode [overcome the diodes vf ]

...sorry Richard, didn't try your code by now - still in office.... 😬

Richard,

I've connected my YHDC sensor one pin to ground and the other to RA5. Got readings (see hereunder).

Because I'm not sure what I'm doing right now with the current sensor, I tried to add a 1Mohms resistor and a 100nF polarized cap between GND and RA5. Same readings.

Just for info, my multimeter shows 0.157mVAC - I'm measuring a 230VAC 42W light bulb for now.

I get readings on my serial monitor but the values don't make too much sense to me, for now - there's going to be some kinf of scaling to do, no?


This is what I get for now with bulb "ON":
Attachment 10014


And bulb "OFF":
Attachment 10015

It can't work wired the way you described. The ac signal must be biased by vcc/2 and be less than vcc peak to peak.
If the serial baud rate is too low it will ruin the sample rate.
If the burden resistance is incorrect extremely high voltage may be possible

Post code and schematic

Have you connected your CT as shown in your first post? That is, the bottom to 2.5V (or VRef/2) and the top to input pin AND - VERY IMPORTANT - a load resistor shown as 33ohms in diagram? You must have a load resistor or the output voltage across the secondary of the CT can get very high, possibly damaging the ADC input.

Then, your CT produces 50mA with a 100A primary current. 1A primary give 0.5mA and so yout 42W load gives about 100uA secondary current. Keep that in mind when selecting the load resistor.

EDIT: OK, Richard beat me to it while I was typing.

You might want to specify the model number of the sensor since there are a number of different types, and the circuit shown may not even apply

adjust resister to for the mid current used incase of some nonlinearity... shotkey diode has .3 volt drop
Attachment 10016

Sorry guys for not reponding earlier - I'm still waiting for 20A:50mA current sensor to arrive; my 100A:50mA are far too low and "out of range" for what I intend to measure.

BUT I could grab a 5A:1V one from a friend today so I'm back on my workbench.

To have "better" values -measures with FLUKE 177- , I now use my 230VAC hairdryer:
Pos 1: fan only = YHDC:54mVAC - FLUKE: 1.1A
Pos 2: fan + low heating = YHDC: 170mVAC - FLUKE: 3.5A
Pos 3: fan + high heating = YHDC: 320mVAC - FLUKE: 6.6A


Since I now use a voltage output type sensor, I leave the burden resistor out, right?

Attachment 10017


And here are my readings:

OFF:
Attachment 10018


Pos 1:
Attachment 10019


Pos 2:
Attachment 10020


Pos 3:
Attachment 10021


I can't still figure out what I'm reading.

you cannot use a current transformer that way without a burden resistor.
by voltage sensor do you mean ACS712 or similar ?
if so that circuit is incorrect, typical use
Attachment 10022

Quote:

---

I can't still figure out what I'm reading.

---

the output is intended to be plotted in the freely available easy to use arduino plotter, the average human will have nil chance of interpreting the data with such a small number of samples. if your load current is sinusoidal is should look similar to this

Attachment 10023

sample data

Code:

---

ADC:1632,peak:608ADC:2447,peak:608
ADC:3224,peak:608
ADC:3704,peak:608
ADC:3722,peak:608
ADC:3312,peak:3722
ADC:2528,peak:3722
ADC:1632,peak:3722
ADC:880,peak:3722
ADC:576,peak:3722
ADC:704,peak:576
ADC:1248,peak:576
ADC:2032,peak:576
ADC:2880,peak:576
ADC:3520,peak:576
ADC:3776,peak:576
ADC:3568,peak:3776
ADC:2928,peak:3776
ADC:2032,peak:3776
ADC:1180,peak:3776
ADC:644,peak:3776
ADC:608,peak:3776
ADC:992,peak:608
ADC:1656,peak:608
ADC:2528,peak:608
ADC:3296,peak:608
ADC:3714,peak:608
ADC:3718,peak:608
ADC:3296,peak:3718
ADC:2451,peak:3718
ADC:1551,peak:3718
ADC:864,peak:3718
ADC:576,peak:3718
ADC:736,peak:576
ADC:1272,peak:576
ADC:2079,peak:576
ADC:2928,peak:576
ADC:3568,peak:576
ADC:3776,peak:576
ADC:3552,peak:3776
ADC:2880,peak:3776
ADC:1984,peak:3776
ADC:1138,peak:3776
ADC:632,peak:3776
ADC:608,peak:3776
ADC:1008,peak:608
ADC:1760,peak:608
ADC:2608,peak:608
ADC:3338,peak:608
ADC:3736,peak:608
ADC:3696,peak:3736
ADC:3168,peak:3736
ADC:2316,peak:3736
ADC:1423,peak:3736
ADC:766,peak:3736
ADC:576,peak:3736
ADC:792,peak:576
ADC:1392,peak:576
ADC:2240,peak:576
ADC:3056,peak:576
ADC:3616,peak:576
ADC:3776,peak:576
ADC:3462,peak:3776
ADC:2752,peak:3776
ADC:1840,peak:3776
ADC:1031,peak:3776
ADC:608,peak:3776
ADC:632,peak:608
ADC:1088,peak:608
ADC:1856,peak:608
ADC:2690,peak:608
ADC:3424,peak:608
ADC:3760,peak:608
ADC:3680,peak:3760
ADC:3082,peak:3760
ADC:2240,peak:3760
ADC:1344,peak:3760
ADC:736,peak:3760
ADC:576,peak:3760
ADC:864,peak:576
ADC:1472,peak:576
ADC:2310,peak:576
ADC:3136,peak:576
ADC:3680,peak:576
ADC:3760,peak:576
ADC:3424,peak:3760
ADC:2656,peak:3760
ADC:1760,peak:3760
ADC:992,peak:3760
ADC:608,peak:3760
ADC:646,peak:608
ADC:1120,peak:608
ADC:1904,peak:608
ADC:2752,peak:608
ADC:3440,peak:608
ADC:3776,peak:608
ADC:3648,peak:3776
ADC:3056,peak:3776
ADC:2168,peak:3776
ADC:1295,peak:3776
ADC:704,peak:3776
ADC:576,peak:3776
ADC:880,peak:576
ADC:1520,peak:576
ADC:2368,peak:576
ADC:3168,peak:576
ADC:3696,peak:576
ADC:3744,peak:576
ADC:3376,peak:3744
ADC:2594,peak:3744
ADC:1692,peak:3744
ADC:934,peak:3744
ADC:608,peak:3744
ADC:655,peak:608
ADC:1150,peak:608
ADC:1934,peak:608
ADC:2800,peak:608
ADC:3462,peak:608
ADC:3760,peak:608
ADC:3610,peak:3760
ADC:3008,peak:3760
ADC:2144,peak:3760
ADC:1264,peak:3760
ADC:678,peak:3760
ADC:576,peak:3760
ADC:906,peak:576
ADC:1566,peak:576
ADC:2416,peak:576
ADC:3202,peak:576
ADC:3696,peak:576
ADC:3738,peak:576
ADC:3334,peak:3738
ADC:2544,peak:3738
ADC:1648,peak:3738
ADC:899,peak:3738
ADC:576,peak:3738
ADC:704,peak:576

---

sample wave, note peaks and mid point values
Attachment 10024

Oh, I get it now, the "plotter".

I just need to monitor a current level that will trigger a relay.

I do have ACS712 but the problem is, the current has to pass inside (though) the component meaning I would need to modify the cabling inside an electric board where it is intended to be installed. But I want to avoid this - it has to be like SHELLY energy counters or so.

I just received a module, an analog AC Current Sensor SEN0211 unsing a YHDC current transformer: DFROBOT. I'll give it a try.

Quote:

---

you cannot use a current transformer that way without a burden resistor

---

Does this mean there is no resistance inside the sensor?

Attachment 10026

Quote:

---

Does this mean there is no resistance inside the sensor?

---

can't tell, the picture is too blurry to read properly but i see no reference to a resistor mentioned it the bits i can read
post a link to data sheet


edit
ps, found this so no resistance inside the sensor

Attachment 10027

if you don't need instantaneous current sensing but can react within 10 or 20 milliseconds..... you can 'filter' the output to DC with the circuit I posted and constantly take a few samples and scale the a/d reading to the amps

Quote:

---

no resistance inside the sensor

---

Again as always: R T Full M :D

Quote:

---

Originally Posted by amgen

if you don't need instantaneous current sensing but can react within 10 or 20 milliseconds..... you can 'filter' the output to DC with the circuit I posted and constantly take a few samples and scale the a/d reading to the amps

---

I'll go this way, thank you :wink:

...and from the same supplier:

Attachment 10028



Saint Thomas taps me on my shoulder!

Attachment 10030


In the 20A/1V model, there are two resistors and a TVS diode. The first resistor is 1k2 and the second is 1k.

I also checked a 100A:50mA model and there are no resistors in there, only a TVS diode.