460 Seven Sensors tested: Measuring Current with Microcontrollers (Arduino, ESP32, ESP8266) 

Corrections and additions: . Ohms law was used wrongly (ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-Cg-4GjLeeAI.html ) - This video is about DC sensors (albeit some of them measure current in both directions). - The video about AC sensors is here: ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-Vb9-pbLdsfQ.html - The INA219 can measure currents in both directions

Thank you Andreas. You often save me from wading through the datasheets of inappropriate devices and board modules, and wasting my time looking for what I need. A 22 minute video brings me up to speed on the commonly available sensors, and their pros and cons. This is one such occasion.

The INA219 sketch you used does have the voltages back to front. Your rant at 13:35 - the 219 allows the difference between bus and load to be both positive and negative, so it can measure current in both directions - very useful. I think your comment at 15:26 is therefore not accurate. It’s one of my favourites.

Perfect timing and great overview and explanation! Exactly what I need for my smart home water level sensor 👍

The ACS712 is a wonderful chip and it helped me in a big project. I was measuring mains at 230 Vrms in up to 20 Amps (tested). The inner resistance of the ACS was so small (

Good catch on the variants of the INA3221 breakout board available from circuit designers with strange ideas. I agree with other comments though that some of the TI shunt sensors can handle reverse current sensing. From the datasheet: "VIN+ and VIN- can have a differential voltage of -26 V to +26 V; however, the voltage at these pins must not exceed the range of -0.3 V to +26 V". The datasheet shows an example of reading a -80mV shunt voltage from the CH1 shunt-voltage register.

French new embedded engineer here. I just discover your channel Andreas. I must have listien to this video 6 times by replaying back some portion of it. Great content. Subcribed!

I think that the INA219 and all similar modules can measure bi-directional current without any problem.

Very well done. One must also keep in mind temperature stability and drift and the concept of dithering in quantization. Resistors must all be 1% and for really important accurate measurement, consider a precision, temperature compensated voltage reference. Finally be sure to characterize the wiring/connector drops on any voltage measurement.

After being educated as an electrical engineer in the eighties I ended up in IT and now more low level stuff and Robotics. Obviously my knowledge is stale. Your back to basics with a modern twist is exactly what I need.

6:00 you dont need negative opamp for this, you can use a divider and bias input of the opamp with a positive voltage, then the negative voltage would add to that and the resulting voltag could always be poisitive

Ohms Law is I=U/R and not I=UxR as said in your video at 1:42

Watch out for a catch with the hall sensors - magnetised screws! I chased around for ages trying to find and null drift , only to find that every time I tightened the screws my magnetised screwdriver was changing the field around the screw and effecting the hall sensor

Your a wizard, Andreas. I've been playing with the idea of using a current sensor with LIFEPOv4 batteries to power and possibly then shutdown a Pi and been looking at creating a circuit to create an effective UPS with a couple of cells. So your video is very timely. I had already purchased a couple of the INA219 boards as this was what was used on a Pi UPS board that I managed to fry! So was particularly at using this chip. Thanks for your info and cross-sections of boards out there. I think may have chosen the right option! Thanks, as ever for your useful info.

Thanks for addressing all those points in the beginning, like implementation issues, high/low side monitoring, etc. Unfortunately, I think most of the viewers missed a lot of it, or just didn't understand. (Judging by comments) I thought you did a great job explaining it! Maybe they just need to watch again, or even a few more times? Good stuff. Thanks for the comparison of all these sensors. And for pointing out a couple that have "gotchas." I might not have noticed that before buying, because I would assume they made sense, like not having 3 sensors measuring the same load!

Danke viel mal Andreas This video came at precisely the right moment as I am trying to implement overcurrent sensing on a 2 cell lithium ion powered project.

informative and accurate as always. i would have liked to see a little more explanation of “high side” & “low side” measurements.. and when /why you might use each…. also a slightly better explanation of “common mode” voltage specs and where/why this is V important if using it on high side. ive heard a lot of users on forums burning up chips because they used a “low side” spec chip for high side measurements…. not realising that its NOt the shunt voltage thats critical when spec’ing it for high side.. i use a lot of INA226 (in “24V systems”) … a lot of users “forget” that a 24V system will commonly see 29V when batteries are charging … (similarly for 12V systems … 14.5V ++)

Beautiful. I had in the past looked at all these types of sensors, and it’s always very difficult to determine which ones are appropriate. Thanks for sharing

I work for a long time on a project (so far: ESP32 I2S MCLK and current sensing), figure out some things, and only afterward do I come across your excellent videos on the exact same topic. One of these days I will look for a video from you first! FWIW, I appreciate the reviews of the ICs specifically. I make PCBs so don't need breakout boards, though they can be useful to test and dissect. I ended up with a hall sensor for isolation (TMCS1101-A4, 5V version for max sensitivity, my load is ~100-200mA, 24VAC). Also had to figure out an RC filter (R1K, 300nF) and a better ADC than the ESP32. Hall sensors have a lot of HF noise, the RC filter is required to make any use of it. I don't think being inline means a hall sensor has a burden voltage. It's just how the IC routes the trace under test very close to the hall sensor. High amp ICs require special consideration for thermal dissipation. I found many users on the TI forums talking about their burnt ICs. It should be possible to do it right though. Anyway, nice video! Thanks!

As always, interesting subject, explained precisely, maximum informations popularized with tact and generosity while remaining short and precise. Thank you Mr.

Interesting - I was just thinking that if you were measuring small currents, and you wanted to use the inductive clamp method, maybe you could wrap the wire 10x around the lead and then just divide the measured current by 10? (Or however many loops) Also, this is awesome! I’ve tried measuring voltage and current with an Arduino in the past, and have had mixed results. This should help.

The INA228 and INA229 are very nice sensors measuring current in both directions also having voltage , current , power , energy and temperature registers working up to 85v.

Another good choice is the Texas Instruments AMC1302. It's a galvanically isolated op amp designed to sense current through a resistor with ±50 mV burden voltage. It needs two power supplies, one isolated, but that can be done with a small transformer, a square wave (Arduino TONE using a spare pin) diode and capacitor. Working isolation voltage 1500 VRMS. They also have an AMC1311, with a 2 volt high impedance input designed to measure voltage with the same galvanic isolation. Both have ~100kHz frequency response. I've used them to make efficiency measurements on high power audio amplifiers by monitoring the voltage and currents of the mains supply and amplifier output.

Around year 2000 I had to measure current in a project and I used a similar approach shown between 5:00 and 6:00. Back then I didn't have the option to buy a INA219 module. After the rectification and filtering of the power supply, only the op-amp was powered. Then I put a diode on the negative rail creating a voltage step (0.6 - 0.65 V) to the remaining circuit. This way I created a negative supply for the op-amp. It was not a symmetrical supply, since the positive rail is the same for the entire circuit, and the negative rail was only a diode drop from the main circuit. But it worked since the voltage drop on the shunt resistor "fitted" inside the voltage drop of the diode.

Very fortunate for me, I was just researching this.

Nice Pointer!

I'd add SCT013 clamp sensor, it's easy to use and you can choose from many versions, depending on the current and precision.

I've used a current transformer to measure AC . That clips a ferrite core over a power cable of an external device and provides an AC voltage signal with amplitude proportional to the current... Usually DC biased so that no negative voltage is seen as the microcontroller. It's instantaneous current so the AC has to be measured rapidly to determine the average current. With some software trickery, and assuming that the load is sinusoidal, you can play software tricks to measure the voltage at a certain angle after zero crossing. That avoids lots of number crunching. One can adapt to AC frequency at startup and any drift in phase offset over time by monitoring the zero crossing voltage at the expected time... Sort of phase-lock. If you need true RMS then it's calculation intensive. Most of the time you don't need it for control purposes where you wish to avoid damage through over-current. The phase-point measurement could be sufficient in many application where it can react at 4 times the AC frequency, if you choose.

INA226 can measure negative shunt voltages ie negative current.

I seem to remember one old application note describing use of a floating differential sampling and storing front end. In principle, it had a 2-pole, 2-throw relay and a capacitor that stored the sample, then delivered the stored voltage to the amplifier input, when the relay changed state. Of course the relays are not a favored thing in the era of all solid state systems. Yet, an interesting addition to the collection of different topologies.

Something you did not address here is frequency response. Normally all you need to measure is low frequency like 50/60 Hz, but sometimes you need to measure much higher frequencies and the resistor types are the only ones with any real higher frequency response. And the stuff I was measuring was only in the kHz range. Not really RF frequencies.

Voll cool, danke Andreas, Ich bin seit zeiten auf der suche nach einem guten Modul, ein guter Vergleich hat mit noch gefehlt.

Hello, nice video and very well explained as always. Just to confirm, you mention that the sensor doesn’t support reverse current, but at least for for INA219 I know that it does measure both directions. I have a setup where I use a solar charger to a Li-ion, where the sensor is in series to the battery, and I can see positive current values when it’s charging, and negative when it’s discharging the battery.

After going over a lot of options I ended up going with the INA237 due to availabilty, cost and simplicity. Coupled with an Attiny212 and i2c LCD made a nice USB-C volt/current meter with PC data link. It's a nice chip and only a small programming change from the INA219 but with greater common mode and resolution. Worth having a look at.

Hi, my teacher always recommend your vídeos

So that's where I messed up. I tested the shunt version but did not like the accuracy. I am not 100% sure but I probably did not connect to both GNDs. Learned something. Thank you.

Working with a Raspberry Pi Pico W I had issues with the fluctuation from a voltage divider as well, but I stabilised it in the software by using the average of the last 10 measurements as the value for decision making. It worked well and produced a value that was very stable in comparison with a bench multimeter. The example is in python: # Calculate the average of the last 10 Batv values batv_average = sum(batv_values) / len(batv_values)

The Hall effect sensors like the ACS721 also have the problem of hysteresis, in other words, the "zero current" output voltage changes depending on what was the last current you measured. Measure +1A, then note the "zero" voltage (nominal Vcc/2). Now measure -1A. Now check the "zero" voltage, and it will have changed by several % of full scale.

Use piecewise linear interpolation to calibrate your values. //used on esp32 // Coefficients for the piecewise linear interpolation //Raw ADC value's float coefficients_ADCread[] = {2 ,149,270 ,363 ,452 ,550,637, /* Add other ADCread values here */}; // corosponing mA measured externally float coefficients_mA[] = {3.2,8.3,12.3,15.3,18.3,21.4,24, /* Add other mA values here */}; float convertToMilliAmps() { // Read ADC and convert to mA adccounter=0; ADCread=0; //take 1000 mesurements for more avarige value while (adccounter coefficients_ADCread[index]) { index++; } if (index == 0) { // Ensure the index is within bounds index = 1; } else if (index == sizeof(coefficients_ADCread)) { index = sizeof(coefficients_ADCread) - 1; } // Perform linear interpolation within the segment float mA = coefficients_mA[index - 1] + (ADCread - coefficients_ADCread[index - 1]) * (coefficients_mA[index] - coefficients_mA[index - 1]) / (coefficients_ADCread[index] - coefficients_ADCread[index - 1]); return mA; } void setup(){ analogSetClockDiv(255); analogSetPinAttenuation(35,ADC_11db); adcAttachPin(35); //Analog read pin (pin 7 IO35) } void loop(){ Serial.print(convertToMilliAmps()); }

for clarification, the INA219 can be used to measure current in both directions. I have used it before to measure positive amps (discharging battery) and negative amps ( charging battery) in a project before. Measuring the current is a matter of identifying which side the voltage is lower... thats the direction of flow.

Measuring current on my solar system has been a recurring interest for me for several years. This year I have implemented a circuit based on Stuart Pittaway's excellent RU-vid channel. I'm using 3 INA229 chips (charge controller load, and battery current) on a custom PCB. The chip uses an SPI interface. Current measurement are by directional, with very low zero offset. Therefore coulomb counting for battery charge level is extremely accurate and repeatable from day to day. I show the board in a short overview RU-vid Video if interested.

in my opinion one more chip( ADS1115) can use as currert meter in positine or negative power cable. General this I2C chip include 1)4 positive analog inputs with PGA in 16 bit ADC In PGA have 6 range A)PGA: +/-6144 mV Resoultion: +/-187.5uV B)PGA: +/-4096 mV Resoultion: +/-125uV C)PGA: +/-2048 mV Resoultion: +/-62.5uV D) PGA: +/-1024 mV Resoultion:+/- 31.25uV E)PGA: +/- 512 mV Resoultion: +/-15.625uV F) PGA: +/-256 mV Resoultion: +/-7.8125uV 2) alert output pin for high voltage in analog input 3) differential mode convert 4 Analog input in 2 differential channels channel 1 (ADC 0-1) channel 2 (ADC 2-3) ================================================================ Then use differential mode ADS1115 analog input MUST be between 0 and the ADS1115 supply voltage ================================================================ if connect in posstive power supply cable the max range is the power supply of ADS115 if connect in negatige power supply( ground point) the range is larger because the voltage in resiston is very smail. - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - My design i use 2 differential Channel with R0.05/10W connect to "negative"( ground point) to get resoultion in six range's 1)PGA: +/-6144 mV Resoultion: +/-7.5mA => 245 A 2)PGA: +/-4096 mV Resoultion: +/-5mA => 163 A 3)PGA: +/-2048 mV Resoultion: +/-2.5mA => 81.62 Α 4) PGA: +/-1024 mV Resoultion: +/- 1.25mA => 40.96 Α 5)PGA: +/- 512 mV Resoultion: +/-625uA => 20.48 Α 6) PGA: +/-256 mV Resoultion: +/-312.5uA => 10.24 Α keep mind with 10W resistor max current is about 14.14 A so for safety limit is 10A. with 50mV MAX loss of Voltage in resistor - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -

Another great video Impressive and interesting test as always 😊

Thank you Andrea. There is also the INA226 which can handle voltages up to 36V against 26V for the INA 219 (usefull for my Oscar 100 final PA...) 73 de HB9IIU

Schöne Übersicht über Stromsensoren ..... Weiter so ... Auch mit dem Akzent 😉😉😉 Echt tolle Videos

I=UxR isn't right Andreas! U/R is what you meant! A neat video is so much more human with the odd slip-up. We know we're not being bamboozled by AI!

Good info... BTW...There in no 'W' in "circuit". The 'U' is silent....😉

This episode is awesome as always - many thanks. Would you mind covering the *new* INA3221 that you just mentioned in this video (since you did not have it back then)? It seems to be a great fix over the old INA3221 design and capable of measuring different voltages. I searched all day but just found „mods“ to fix the old design yet no docs about the new board design and the mysterious two GND pins. The new INA3221 seems to be perfect for monitoring solar charger systems and apparently is used for this purpose already. Covering it would be of great interest to many, especially since this topic seems not be covered yet anywhere, and INA3221 are both readily available and cheap. I simply can’t wrap my head around how to wire up the new version (lacking your experience). Any hint greatly appreciated. Many thx!

Great stuff ... Andreas do you care to present some longer distance radars (100-200m) like used for collision avoidance in cars ?

I would expect that when measuring mains, being AC, you would been some code and a bit of bandwidth on the mcu to calculate the RMS equivalent current value? It looked like all your tests were done with a DC source.

The ina226 can be used both on high and low side.

@AndreasSpiess Have you considered making a video about how to program an ESP32 C6 with zigbee to make your own zigbee-devices, such as plant sensors, remotes and other sensors? This would be a game changer for DIY smarthome.

🌟 I need a sensor that detects a mAmps current, just detect it, I don't need to measure it, the current is about 20mA, can you advice? Nice video as usual, thank you 👍

Right on time! I'm on a project where I want to measure voltage and current of a 90V solar panel. The INA219 only goes up to 26V, so I'm curious to see what alternatives you're reviewing. Going to watch it now!

Hi Andreas..very interesting. What IC would you use to check if a smartphone is charging or not (an IC in the middle between charger and smartphone)?

Could we get a video on powering microcontrolers with redundant power supplies? What type of powersupply is the most reliable? What makes power supplies fail, which component fails first from heat, cold, moisture etc. How to protect board from the elements and lightning etc. Kinda obscure toppic, but I think it would be interesting for people making IOT stuff that is a royal pain to get to when needs replacing.

The big hall sensor board is bad for mains as it has copper fills up along the side of the chip. This reduces insulation.

Grape Video ! 🍇

Thank you Andreas, But i have one question, could you please give information regarding AC current sensing that which sensor is best for 85khz Ac current application and it should be connected to MCU, how can I achieve it ?

Yaay i was wanting an update on this :)

Hello. Awesome video. Please what sensor would you recommend for a large scale use, say for instance measuring all appliances in a house

Which of these sensors are appropriate to measure reactance, the phase difference between voltage and current, to differentiate apparent power versus real power in inductive loads such as motors, solenoids, coils, antenna and to perform power factor correction of digital devices power supplies ? Soaking of power supplies, what happened to the "super power" power supply ?

Thanks!

Ina226 can be used high side or low side, but it does still add a burden voltage.

Nice video for current sensor, I am from ME major. I don’t understand why I need to measure a current linked to arduino? Why can’t they separate in two loops? Say I want to measure if my water pump is turned on from line voltage I could use hulls sensor in it’s loop and don’t bother my low DC supply to Arduino. Even for the shunt sensor at 5 A range, can’t design them in a different loops? Like you have example with resistors? Also if I use a transformer to supply say 7 and 5 volts DC then I have more choices to handle my application in a different loop and don’t bother normal working voltage for Arduino? Thanks

You put your restisor divider in the beginning of the video on the wrong side of the sense resistor. The divider would only measure the supply voltage which should remain constant. I got what you were trying to show though but it should have been on the other side of the sense resistor so you would be measuring the drop rather than the supply.

The ina3221 realisation is not good. The channels has crossover error.

Thanks Andreas! A great overview on current monitoring modules and very timely too! I'm looking at a project to remotely monitor and control some 12V(nominal) fused circuits for an amateur radio setup. Are you aware of any modules that can reliably/safely switch 12VDC up to 50A for the main, and as much as 25A for a branch that may have a 100W transceiver on it? I might just have to go with automotive relays and a driver circuit, or figure out a suitable high side MOSFET switch circuit, but I was hoping for a nice easy to use module! 😂

Hi Andreas, very good overview, also you use Hall sensors, but I'm missing XMR sensors like TMR sensors. They are much more accurate. Any Plans for TMR? Best regards Thomas

Analog measuring devices measure currents and then the voltage via the internal resistance. Digital measuring devices measure the voltage and then the current via the shunt. Others measure the magnetic field and thus the current. The analogue input of the ESP8266 is not very good and is disturbed by the WLAN.

Thank you, Andreas. Have you ever used a Rogowski coil to measure alternating current?

Very interesting ! Will you make a video where you explain how to program the esp with grafana and co to give a nice current reading over time in a webbrowser ?? ;) Or is there already one ?😅

15:27 wrong. some of them can measure reverse currents, for example ina219/ina226.

what would you use for 2x 100W solar panels with 20V (running parallel)? I get max 10A. Can I use a i2c board???

Thx

Isn't there a Hall sensor in the ESP32 also?

Many thanks - please consider an episode on how an MPPT controller dynamically changes the impedance that is ‘seen’ by a solar panel depending on the load, where load seems to be a combo of battery charging needs , and the inverter input . I have looked on the internet for months and still can’t find a clear technical explanation . The question arose when my inverter showed 400 watts solar panel output on a sunny day with a 4.6 Kw array - the battery was fully charged and most appliances turned off.

Wouldn't it be possible to just use a hall effect sensor and measure the magnetic field induced by the current? Would propably work better with a small coil

Nothing appeared in the top right corner for me

In future videos like this, could you include the approximate prices for the alternatives, please?

There is a typo at 1:40

I want to measure mains current and direction. I've been told I need to put two inductors in quadrature to determine...what does that mean? Possibly a future video?

Thanks again, I am looking for affordable DC current clamps for the PV strings, to integrate it in my home automation system.

I've been battling with this current monitoring problem for a while now and still don't really have a solution after watching your excellent video. The INA219 can measure current in both directions, and operates at VCC 3.3v (even lower actually), however it is limited to max 3A current measurement. The ACS712 can measure higher currents, but only operates at VCC 5V, which is no use for ESP8266 projects. What I need is a current sensor that operates at 3.3v and is capable of measuring 5A or more bi-directional - so I can see when current is being drained from the battery or fed into the battery from the solar panel. Is there anything out there that can fulfill this?

Why this old video again?

Why the Reupload?

No, there is mal so garnix appearing in the top right corner....

I have several INA219 boards but they are only rated @3.2A making them unusable for solar which can run up to 10Amps (in my basic setup case). I am a bit confused over that issue.

Does anyone know of an open source equivalent to the Emporia energy monitor? It has up to 16 clamp-on current transformers to monitor home energy usage. It must use an analogue multiplexor to service so many inputs rather than multiple ADCs - but I have no experience with analogue muxes…

I missed the popular PZEM 004T in your lineup. For anyone using this sensors i am interested in your experiences. Cheers.

no video in top rtight cortner.

The ina226 have higher sensitivity tha the ina219

1:42 Your Ohms law equation is incorrect. should read I = U/R or U = I x R

I think I=UxR is wrong, it should be U=RI

🌟🌷🌟

Ahem... 01:43 shouldn't U and I be swapped? R=U/I, so U=I*R...

I know you are on well deserved holidays brother, but repeating the same video 3 years later gives almost nothing to anybody UNLESS it is i.e. updated or so. Believe me, nowadays there are much more sophisticated sensors available. Otherwise it is only to... increase the click count ;-(

You put a picture of the holocauster Churchill who murdered 4.3 MILLION of my people. Think about that. I don't think you realize how offensive that is.

Cool, i just put together a battery capacity tester using arudino and was wondering/ looking into this problem. Ended up using a shunt resistor following this guide here ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-zgrMHbX9UC8.html I was going to upgrade to a more elaborate circuit/sensors but not sure if i realy need it becaues the discharge test seems to match the rated capacity, some are above some are below, what you'd expect, so how much more accuracy are we getting to warrant a more elaborate circuit? hmm I was convinced it wouldnt be any good after learning the theory(needing low value shunt etc) but the tests seems good and consistent, at leaste relative capacity is there which is most important in my applicataion anyway as im trying to match cells for a battery pack not determine absolute capacity. Im using a 1 ohm resistor in series with a 2.6 ohm load and simply the voltage across the 1 ohm shunt is the current since the math works out easy I was able to verify my calcuated results with a multimeter which matched almost perfectly. So how much more accuracey am i really getting with a more elaborate circuit? hmmm I feel the circuit is not as good as it could/should be (it powers on the arduino when the battery is hooked up and the arduino is not) but im no expert. I heard knowledgable youtubers criticise even more elaborate capacity tester circuits as "useless" so i dont know what to think.

Evening Andreas, summer holidays I know. It is freezing here.