Excellent placements. Currently we have some INAs (eg INA250) with built-in shunt and some interesting (good) features. When it arrives I will test it to see how they behave.
Thanks! The INA250s were new to me - and I like them: 4.5 mOhm in-line resistance and 0.3% accuracy. The only drawback are the relatively high offset currents in the mA / tenth of mA range. But then, they're made for measuring currents up to 10 A.
Software for drawing schematics is overrated 😅 No, seriously, in the time I need to start the program (DipTrace in my case), create the part I'm working with, create the circuit diagram and print it, I've drawn the whole thing several times by hand. Of course if I need a PCB layout down the road I do a schematic in DipTrace.
@@robertssmorgasbord It's like breaking your brain to write with a pen in the absence of gravity. With how easy it is to write with a pencil... 😅 Very good analysis of those amplifiers. Finally I decided to use the LT6105 to be able to measure consumption through the oscilloscope. Thank you very much for sharing!!!
Thanks very much Robert for covering the instance where the LT6105 power and the measured current were not connected. I wondered about that since it is not covered in the datasheet. We are planning on using the LT6105 in some test circuitry. It is easier to lay out our boards with the returns isolated and connect after the fact if needed rather than the other way around which would require a board respin.
Hi, good presentation. In reality , the currents sense amplifiers are designed for measuring currents with hi cmrr, not optimized for achieving low noise, thus it cannot replace all the designs
Thanks for the praise! And you're absolutely right: They are designed for extremely wide common mode voltage ranges and high CMRR, not low noise (or high precision). If you want low noise (and high precision) your first choice would be an instrumentation amplifier. But while those usually also offer a high CMRR, their common mode voltage ranges rarely extend beyond the power rails. Guess you can't have it all.
For the INA28x black magic, those switches represent switched-capacitors which has an equivalent resistance of 1/(C * f). In CMOS analog electronics capacitors are much more precise so the "resistors" are likely very well matched, hence the very large CMRR. Although there's an oscillator in there somewhere that they haven't told you about.
Thanks for the info! I wasn't thinking about that, but of course it should be possible to make very precise capacitors (and resistors?) in CMOS. I mean you can control the geometry down to 100nm or even less.
@@robertssmorgasbord maybe precise wasn't the best word to use. Process gradients cause resistors to mismatch while it's easier to match capacitors. I'm not certain on the technology reasons this is just what I remember from my analog design class :P
@@incinatus190 I guess that's the reason they laser trim resistors on chips when they need real precision. But wouldn't it be possible to laser trim capacitors too? Burn way a bit of the top layer?
nice video, thanks for sharing, I was about to ask you about having different grounds between mesuring circuit and the load, then I saw you tried it disconnected and worked just fine. Maybe with this the common mode issue is reduced dramatically since the common mode voltage is the voltage in the shunt resitance, only issue with different grounds is you can't touch the metering circuit if load voltage is high, I am planning to use this circuit to measure current in a VLF tester ad different grounds is just what I need. Regards from Chile
Thanks for the praise and you're welcome! But to make it absolutely clear: According to the LT6105 datasheet it shouldn't be used without a common ground (floating). See also 39:03 in the video. Though it was working in my setup, that's no guarantee it will in yours too. Anyway, good luck with your project!
Hello! Thank you very much for making this video. This was fascinating for me and also very relevant. For my dissertation (a software engineering degree) I am working on LoRa radio-related technologies. One of the things I have to do is to be able to accurately measure current consumption when the radio is sleeping, receiving, and transmitting. The amount of current drawn ranges considerably: From ~5uA when sleeping to ~140mA when transmitting at full radio power. Do you think the LT6105 would be a suitable device? I would feed the output of the LT6105 to an ADC on a microcontroller for sampling. Somehow I need to scale the output of the LT6105 from 0-5 or 0-3.3V for sampling. This is all new territory to a software engineer! Any advice from either yourself or your viewers of this channel would be greatly appreciated. I will watch the video a few more times as I need to let things sink in slowly! Again, many thanks for this wonderfully informative and very well presented video. Greetings from Scotland!
Hello, and greetings from Germany! To answer your question: No, the LT1605 is not suitable for this application (it does not really go down to 0 - I hopped through some hoops in my circuit design, so that won't be visible on the analog instrument). Your application got an extreme large dynamic range. If 140mA is your full scale, 5uA is 1/28,000 of that. So coming from the ADC side you would need at least a 15bit ADC (32,768 values) for the 5uA to show up as 1 LSB. On the analog side you'll need to achieve a precision of at least 0,0036% of dull scale, so your 5uA will show up at least in the range of 5uA. The next problem is the shunt resistor. To get even a 100uV signal at 5uA you would need a 20Ohm shunt. But that shunt would drop 2.8V at 140mA. In short, there's no single chip / single range solution for your problem. But fortunately there are (open source / open hardware) solutions for that (look for the "Current Ranger"): ua-cam.com/video/HmXfyLyN38c/v-deo.html
So my high side voltage can be anything from around 12V to 18V through an H-Bridge that creates a square wave of around 10kHz at 5A (maybe up to 10). I need to measure that current and convert it to a voltage from 0 to 3.3V into an analog input (a 10 bit ADC). It would be nice to have 10mA resolution and read all the way down to 10mA up to 5A, but I have 2 outputs with different needs, so could have one circuit measure 0 to 1A to measure down to 10mA (giving me more resolution and sensitivity in my 0 to 3.3V output), and then have the 5A circuit have less precision, especially on the low end since is is mostly to give an idea how much current is being used give or take a little bit and to provide short circuit protection if I go over 5A (or 10). But easier duplicating the same circuit if I can. I've even looked at using a Pulse Current Transformer. What would you recommend? It would have to be bi-directional. I also thought of just running something into the DC input to the board that creates the square wave and subtract out the current the board uses. There is an unknown, but high (2-4V drop) from The DC input to the square wave output depending on current. Many of the chips I have seen have a center point, zero current is 2V. The boards that give you 1V per Amp are really nice, but they are uni-directional for measuring the input instead of the high side output. Could you offer any ideas? Thank you.
Measuring the current in an H-bridge is not an easy feat. Have a look at these application notes: www.ti.com/lit/pdf/sboa174&usg=AOvVaw35atuliGHFP5Txoi8SBztE www.analog.com/media/en/reference-design-documentation/design-notes/dn407f.pdf www.st.com/resource/en/application_note/dm00653792-current-sensing-in-motion-control-applications-stmicroelectronics.pdf Regarding your two ranges: You've got a 10 bit ADC, you need a 10mA resolution between 0 and 1A and you need to measure up to 5A, yes? A 10 bit ADC (1024 counts) will give you a resolution of better that 5mA for a 5A range, so no need to fiddle around with two ranges. If you've come up with a solution and you're unsure about it, I recommend discussing it on the EEVblog. Hope that helps.
@@robertssmorgasbord Thank you. My situation is unique in that I am not directly driving a motor, therefore no inductive load. I am feeding it into a board that splits a bipolar 15V p/p signal into a rectifier to get voltage from it, and into an optoisolator to extract the signal from it. There are so many variables, it is difficult to find what I need. I want 5A, but may need a 10A sense. I will probably need 2 circuits, one to send the sense voltage to a 3.3V logic ADC and one for a 5V ADC. The LTC6103 seems interesting. Maybe even the INA240 or ACS724. I will have to spend quite a few days thinking through all of them. And then consider cost. The Arduino Motor Shield has an LMD298 on it, so it can only go to 2A. That has a .15 Ohm resistor connected to an Op Amp with a gain of 11. So at max current, it delivers 3.3, which works for 3.3V and 5V logic. I want to do the same thing for 5A and 10A. But I wanted to see if for the same or less money and parts count, I could do better. There are much better current sense chips available to day than when this board was designed. A shunt and op am will always be available though, while a fancy current sense chip could get discontinued tomorrow and then you have to deal with a board redesign.
@@DCCEX Wow, sound you've go a challenging project at your hands. A few more remarks from my side: I don't see any reason why you shouldn't go with low side current sensing in your application and stick with op amps. When you really need two ranges, use two shunts in series (e.g. 0.1 Ohms and 0.01 Ohms) and have one op-amp measuring across both and ground, and the other just across the smaller one and ground (the smaller one of course connected to ground).
No - and that's the great advantage of these devices. It's also stated in the text above the functional (!) block diagram: "[...] common-mode voltages that far exceed the supply voltage [...]" (26V vs 2.7V). The "trick" here is that it's just a FUNCTIONAL (!) block diagram. The INA181 doesn't really consist of 4 resistors and an ordinary OpAmp. There goes a lot of "special sauce" and "trickery" into the circuits of those devices. I would have to do some research first before I could really explain their inner workings ;-)
Hi Robert, than you so much for your lecture on the current sense amplifier principles. However I do have some questions if you may help me to solve my puzzles. How to detect a current leakage between the Shunt and Load and control the constant current flow into the load ?. Any advantage to monitor voltage drop between sense resistor and the Load (to detect in case of leakage along the path till the Load?.. Thank you for your help..
Hmm, interesting questions. The first thing you would have to answer is where is my current leaking to - obviously back to the ground of the power supply. If that's the case you would need a second second shunt as close to the load as possible with it's own current amplifier. But why do you expect some leakage current? What voltages are we talking here. Anyway, monitoring the voltage across the load through two Independent, non current carrying lines is a common feature in high end power supplies. It allows you to compensate the voltage drop across the current carrying lines between the power supply and the load. Basically the feedback for voltage regulation comes from these two lines, and not from the current carrying binding posts on your power supply. But it does absolutely nothing for detecting current leaks.
What if I want to use SMT current sense resistors to measure high bandwidth AC/DC current instead of using an expensive high frequency current probe? If we make the common mode voltage a little more challenging like mains input for a 25Watt (very popular) universal switching flyback auxiliary power supply. Can it be done?
If we're talking mains AC, there are specialized chips for that job, e.g. www.allegromicro.com/en/products/sense/current-sensor-ics/zero-to-fifty-amp-integrated-conductor-sensor-ics/acs730#PartTable . You don't even need a shunt.
Thank you for making these videos, they are fantastic and clear. Are there any low cost solutions for 120/240VAC? The common mode voltage always seems to be below mains voltages.
First, you're welcome! And no, to the best of my knowledge there are no current sense amplifiers suitable for mains AC. What immediately comes to mind are chips for mains AC energy metering (e.g. www.digikey.com/products/en/integrated-circuits-ics/pmic-energy-metering/765 ). They start as low as $1. However, these are completely different beasts, containing several ADCs (current & voltage), calculate peak and RMS values and usually sport some digital interface to read out the values.
@@robertssmorgasbord I'm surprised by that. I took a peek at a few of my multimeters (including Fluke) and they all seem to use a current sense resistor (Constantan) to measure mains current. I wasn't however able to figure out how they were measuring the voltage drop across it, I couldn't figure out some of the part markings to identify them.
@@robertssmorgasbord Perhaps the INA149 would work. It has a common mode voltage range of +-250V - the price is pretty steep though! I started with trying to build a low cost solution using a simple precise resistor however it's anything but at mains voltages. At that price, I might as well start looking at a hall effect sensor like ACS723 or something along those lines which is basically the same cost.
@@eyeTelevision [Reply to your second comment] Yes, the INA149 would be suitable for 120VAC (+/-170V), but won't survive 240VAC (+/-340V). But keep in mind that you still have to do peak detection and/or RMS calculation on the output signal of the INA149. The ACS723 would work for 120VAC and 240VAC. Its great advantage is that it offers isolation from the mains AC (mind your board design). Though It's not as precise as a current sense amplifier, and you still have to do the peak detection / RMS calculation on its output signal. The above mentioned energy metering chips do all that for you.
@@eyeTelevision [Reply to your first comment] I didn't regard current sense amplifiers suitable for your application because you want to measure AC (!) currents. So even if you find one that can withstand the common mode voltages of mains AC (which you have below ;-) ), you still have to do a peak detection / RMS calculation. Anyway, The above mentioned energy metering chips also require a current sense resistors. In a multimeter the main chip does all the peak detection and RMS calculation stuff.
Yup! Load and OpAmp do share a common ground. That said, there are circuits that don't require a common ground, but those also require the power supply for the OpAmp to be galvanically isolated from the power supply of your load. E.g. your multimeter in ammeter mode works this way, having it's own battery as a power supply.
You're welcome! To answer your question: Not really. The minimum output voltage of the LT6105 is unfortunately not very well specified. The datasheet just gives you maximum values between 35mV and 45mV, depending on the operating conditions. So you would need some kind of adjustable negative power rail. Also, the LT6105 is a current output device. Its internal circuitry assumes that Vout is connected via a resistance Rout to V-. If you connect V- to a negative rail and Vout via an Rout to ground, you might get a nice 0, but all non 0 output values will get distorted.
@@robertssmorgasbord Thanks Robert. I overlooked that it is a current output devices. I guess the "idle" current output is derived from the input offset voltage.
What is the output voltage at 400ma input current...I can see 200mv is that the output voltagte....if shunt resistor is 50mohm input voltagte is 20mv and if gain is 100 then it will be 2v right...but I see 200mv....
200mV is the voltage across the analog instrument which has a resistance of 5kOhm. There is a 45kOhm resistor in series with the analog instrument. 2V is the voltage across both resistances (45kOhm + 5kOhm = 50kOhm) and thus the output voltage of the amplifier. See also the circuit at 33:49
Great effort and equally great video. Just to know, can these ICs measure values less than 10mA? If then, how can it be measured using microcontrollers?
Thanks! Of course they can measure values of less than 10mA. But depending on your application you might wanna choose another type. If you want to read the measured value with an MCU you have three options: 1. read the output of the current sense amplifier with an analog input pin of your MCU (build in DAC), 2. use an external DAC with e.g. SPI or I2C interface to your MCU to read the output of the current sense amplifier - or - 3. use a current sense amplifier that has an build in DAC with SPI or I2C interface.
@@robertssmorgasbord Thanks for your quick reply! So how much low value can the IC measure? Can you suggest a good IC for that? I have also seen that similar low-cost hall effect based current sensor ICs like ACS712, ACS758, and others which measures the current but not as low as 1mA or even low. Also they have total output error in their output voltage in the range of 1% to 2%. Does the ICs shown in your video have low error which can measure current in the range 1-10mA or even low? Also are there any alternative ICs out of the video? I use Arduino for interfacing BTW. I really appreciate the reply.
@@ajk0007 Hi, I need a bit more information to make a recommendation: Maximum current to be measured (I assume 10mA), value of your shunt resistor, voltage to ground at the shunt resistor, sampling rate and - of course - how much you're willing to spent for the parts. Maybe you could make a short video showing the circuit you've got in mind? Anyway, to answer you questions: IThe LT6105 can measure currents below 10mA, but that's really determined by the size of your shunt resistor. A LT6105 based circuit should have a maximum error of 1%. And yes, there are hundreds of different current sense amplifiers available ;-) BTW if you use an Arduino analog input and connect the its external voltage reference input to an external voltage reference (say 0.1%), you can measure voltages with 0.35% accuracy (the internal Arduino DAC is about 0.25%).
Thanks for your reply. I was trying to simulate a circuit to measure currents as low as 1mA to 5 mA or 10 mA. So only LT6100 is available in the software. In its typical circuit, a shunt of 3mOhm is used. Also I am using an Arduino to convert the output voltage of LT6100. But I am in doubt of the conversion. Is that a proportional change or is there any other formula for doing that? The supply for LT6100 is 3.3 V. The typical circuit is in its datasheet.
@@ajk0007 OK, staying with the LT6100 for now. I would set its gain to the max of 50V/V (Pin 6 (A2) and Pin 7 (A4) to VEE/GND). Forget about using the internal voltage reference of the Arduino and connect an external one its AREF pin. A cheap TL431 type with 2.5V (0.5%) should do. With the chosen amplification of 50V/V you can measure now a maximum voltage across your shunt resistor of 2.5V / 50V/V = 50mV. So at 10mA you'll need a shunt resistor with a maximum value of 50mV / 10mA = 5Ohm. Choose a 4.7Ohm 1% resistor (cheap and available) and you can measure currents up to 50mV / 4.7Ohm = 10.6mA with a total accuracy of about 2%.
I learnt a lot, thanks...Ive subscribed because I will be back here I think. The math was abit above me some times but I now have a much better understanding than I had at the beginning, so thats great!...I have a project in mind that I need to study in order to design it but basically I want to add to my tools a small but reasonably accurate way to measure a microcontrollers current draw when in "deep sleep" ….Arduino (ATMEGA328p-AU) at the moment. I was going to use an opamp but your vid has given me much to think about. My multimeter is a Fluke 79 series 2 and will measure down to 1uA (40mA scale) and I want to make a device on a small PCB to attach to the meter and use the mV range to measure currents from 1nA up to where my meter bottoms out (if that makes sense!) I would like to eventually make this "open source " so many could benefit from it. Arduino and other micro's are very popular, and of great importance in education but most mustimeters can't measure the Nanoamp range (ones I can afford!) and this is important in battery operated projects. Anyhow great vid and if you can offer any advice to me that would be great, I will now check out your other work and look forward to more......cheers. oh yeah nearly forgot I have a couple of INA105kp instrument amplifiers would they be any use in my project ?
Thank you and you're welcome! Interesting project you have going there. The INA105, being a unity gain amplifier (gain 1), might not be the right chip for your project. Measuring down to 1uA you want something with a high gain and very low input bias current/voltage. E.g. MAX9923F (fixed gain 250, input bias current 1pA, input offset voltage 0.1uV): With a 0.5Ohm shunt resistor that would yield a 0.125mV/uA output, a voltage drop of 0.5mV/1mA and at 40mA an output of 5V. Keep in mind that a designing a current measurement circuit covering 4 orders of magnitude (1uA to 40mA) is quite a challenge ;-)
@@robertssmorgasbordyeah think I need a lot of homework! but I'm gonna keep thinking, all I need is to be able to read about where common micros sleep (around 200nA ish) thanks for your kind reply will check out the MAX chip failing that save up for a meter!
Sure, but what are your other requirements? Especially voltage to ground at the shunt resistor, galvanic isolation, resolution/accuracy, sampling rate and - of course - costs? With a maximum signal of 75mV, the shunt at the low side (connected to ground) and no requirement for galvanic isolation you might go with a simple non-inverting OpAmp amplifying the voltage to e.g. 1.1V (Arduino ATmega328P internal voltage reference) and reading that with an analog input of your Arduino at 10 bit resolution and a sampling rate of max 9615Hz (for a 16MHz Arduino).
Yes, monitoring your currents has become more and more important. I remember the time when you just looked up the stall current for a motor in its data sheet and dimensioned the power supply / power electronics according to that. After adding a thermal cut-off switch you were done. I guess those times are gone, considering that your lithium battery can go bang when you draw too much current from it 🙂
Well, I need to know I wee bit more to make any recommendations: Are these 0 to 15V measured to the Arduino ground? What's the basic circuit? What Arduino are you using? What's the maximum value of the shunt resistor, respectively, the voltage drop across it you can tolerate? Anyway, 10% precision is absolutely no problem, even with the Arduino (ATmega328 for Nano, UNO) ADCs. Add an external voltage reference and you can precisions better than 1%. Simplest solution (the 0 to 15V are to the Arduino ground): Use an external 1.25V 0.5% reference (e.g. AZ432, $0.50) and a shunt resistor of 1.25Ohm. That solution has of course no protection and a high burden voltage.
@@robertssmorgasbord Thanks, I build this current-sink circuit (m.eet.com/media/1130553/13338-31705di.pdf) to log discharge curve of batteries and I just want to use an MCU to actively set the current sink current by monitoring it and setting it through vref. But meanwhile I think thats gets to complicated.. I'll simply adjust the current manually with my DMM.
@@paulg.3067 OK, you've got already a 0.33Ohm shunt resistor referenced to ground (which should/will be also your Arduino ground). 0.33Ohm yields 0.33V at 1A. We need to amplify that to 1.25V (AZ432 external voltage reference for Arduino), so an amplification of about 3.79. With your precision requirements and no common mode voltage issues (shunt resistor referenced to ground) there's no need for any special chips. Just use one half of a second CA3240 op-amp as non-inverting amplifier. The non-inverting input of that op-amp is connected exactly like the inverting input of IC(1A) to the shunt. Choose an amplification for the non-inverting amplifier a little below 3.79 (you don't want your ADC input voltage higher than your reference voltage). E.g. 10k and 27k (each 1%) would give you a 3.7 amplification.
Take a look at the Allegro ACS7xx series of Hall Effect Current Sensors. I have the ACS724. www.allegromicro.com//en/products/sense/current-sensor-ics/zero-to-fifty-amp-integrated-conductor-sensor-ics The ACS712 and ACS714 are popular, too. If you buy one of these, only buy it from an OEM vendor, as these ICs are highly counterfeited on Ebay/Amazon/Etc. Be careful of the full extension of the model number that you are interested in, because there are many variations within a single model number... Different amplitude and sensitivity ratings...
Thanks for the info! These are some very nice chips for measuring high currents. Linear frequency response up to 10kHz (OK, they start to phase shift a little at 1kHz) and response times / slew rates in the microsecond range - I like it. The only drawback is the worst case accuracy of +/-3% (ACS724KMATR-20AB at -40°C to 25°C). But since the typical error is +/-1% you can live with that.
I need to measure a current .. the input would be 1mV to 200mV.... I would like to output 0-500mV... or 4-20mA.... would be better to use a instrument amplifier.... .. i need to avoid noise... so would pay more for better results...
Well, I would need some more information to make a suggestion ... 1) Is your shunt high or low side 2) How large is the shunt (you wrote "the input would be 1mV to 200mV", so I'm assuming that's the voltage across the shunt) Anyway, normally you wouldn't need an instrumentation amplifier, because its main advantage over a current sense amplifier is its high input impedance, and a shunt resistor is (normally) a very low impedance signal source (unless you want to measure pA currents and your shunt is in the MOhm range ;-) ).
@@robertssmorgasbord Hi Robert, thanks for your answer! I'm going to measure a leakage current from a secondary transformer, AC signal, from the high side, not ground side, but I'm going to use batteries for the measuring circuit, I'm going to isolate the ground basically transmitting the signal by wifi. The shunt I'm thinking to get 1 Ohm (10W) because my main range is 1mA to 200mA...(but should support short current 2.7A as it will be shorted frequently for testing) if I could measure 1uA would be great. So I'm looking for a good and precise measurement, low noise, that's why I thought to use Instrumentation amplifier for high CMRR, but I have looked many of these current sense amplifiers like INA219, INA226 (this one is easy to get here), INA282, INA181, INA190 (would be good if the INA190 had I2C because it has a low consumption..), INA240 (this last one looks really good but no I2C) , but not sure which would get better results than instrumentation amplifiers like INA114 etc.... I'm going to use an ESP32 so if it has I2C would be good to communicate directly, but the most important is better measurements. by the way ... your video was a great explanation... I wish my professor would teach like you.... I am inclined to use INA226 ...to try from 1mA to 200mA... and lower if possible....using more units with different shunt scales...
@@RF-yh3qh These are some very demanding requirements you have there ;-) Let's take it from the back end first: You want an I2C interface, your max current 2.7A and the desired resolution is 1uA -> You'll well need at least a 22-bit ADC wit I2C interface (we're talking here a 6 to 7 digit meter): A NAU7802 (about $2, 2 differential inputs, 24-bit, 80 samples/sec, Vcc 2.7V to 5.5V, available as SMD or DIP at Digi-Key) should do the job (if 80 samples/sec is fast enough). Just looking at the datasheet of that thing you won't need any pre-amplifier (your full scale voltage is 2.7V at 2.7A and 1Ohm). Maybe some overvoltage protection and RC filtering might be in order though. If the internal gain adjustment of that thing is not enough (1, 2, 4, 8, 16, 32, 64, 128) you can also vary your shunt a bit. And with two differential inputs you could implement two ranges (a smaller and larger shunt in series, on input measuring the smaller and the other both) - don't forget some overvoltage protection at the low range (over both shunts) input.
@@robertssmorgasbord Hi Robert, really thanks for your answer, yes you are right... too much requirements... hehehehe... I'm still studying and sometimes it's hard to choose. I took a look at the NAU7802, a really good one, I don't know if it would be hard to implement. I am thinking if 80 samples/s is enough........ not sure if I should get true RMS (if yes... then need more samples as the signal is 60Hz).... or if I could trust on p-p measurement to calculate it..... I'm afraid to get noise on the p-p measurement and screw up, no experience with that. I really appreciate your answer... I'm learning a lot ..... maybe I should stick first to 1mA to 200mA... and use another later to try to get lower measurements... I can use another one that is not I2C.... because I can use the analog input of the ESP32 to input the analog signal... I was just looking for I2C to try to avoid more errors or noise...
@@RF-yh3qh Hi RF, 80 samples/s if definitely too slow if you want to calculate true RMS. If your current waveform is a sine wave (!) at 60Hz you'll need at least 120 samples/s (Nyquist theorem). If it's another waveform you'll need even higher sampling speeds. Try and have a look at the ADS1219. Its features are similar to the NAU7802 and its 1k samples/s. it's about 6€ at Digi-Key, however it comes in a TSSOP SMD package (0.65mm pitched pins). And ... Texas Instrument datasheets are really good (with application examples, PCB layout examples etc.) :-) Anyway, it might be a good idea to start simple and use the ADC of the ESP32. Don't forget to implement a good input protection though ;-)
Well, at least the LT1605 is available (in stock) form all major distributors: www.digikey.de/en/products/base-product/analog-devices-inc/505/LT6105/61046 , www.mouser.de/c/semiconductors/amplifier-ics/current-sense-amplifiers/?m=Analog%20Devices%20Inc.&series=LT6105 , de.farnell.com/en-DE/c/semiconductors-ics/amplifiers-comparators?st=LT6105 .
Well, first of all thanks for the praise! Regarding spicing things up: Unfortunately I'm not very good at entertainment or humor - I'm German after all :-)
@@robertssmorgasbord Your delivery is fine. Easily understood by a mechanical engineer like me. Danke sehr. Ich habe drei jahre Deutsch gelernt, auf der schule.
@@outagas2008 Herzlichen Dank! Thank you very much! I'm just happy that you and other people (including HitAndMissLab) get something out of my videos. I also enjoy the technical discussions that ensue from these videos (see all the comments below). But believe me, I would love to make my videos more entertaining, it's just not something I'm very good at :-)
His delivery is just right. Great content with no distraction. The "slow" pace also allows me to think along as he speaks without having to pause the video here and there like so many other 3 minutes videos that make you feel like you understand something until you start thinking the subject deeper the next day. Perhaps it's just me but I like the way he teaches.
A very thorough treatment of this issue. Many thanks Robert 😊
Thanks for the praise! And you're welcome!
Very helpful series! Thank you for making them!
You're welcome!
I found this video series extremely useful, so well done and thanks!
You're welcome! And I thank you for the praise!
great video sir. very detailed and clear . It is a great help for me. Thank you.
Thank you very much! I'm always happy when my videos are helpful to someone.
Very thorough explanation. I learned a lot!
Thanks for the praise ;-)
Excellent placements. Currently we have some INAs (eg INA250) with built-in shunt and some interesting (good) features. When it arrives I will test it to see how they behave.
Thanks! The INA250s were new to me - and I like them: 4.5 mOhm in-line resistance and 0.3% accuracy. The only drawback are the relatively high offset currents in the mA / tenth of mA range. But then, they're made for measuring currents up to 10 A.
Love the pen and paper approach!
Software for drawing schematics is overrated 😅 No, seriously, in the time I need to start the program (DipTrace in my case), create the part I'm working with, create the circuit diagram and print it, I've drawn the whole thing several times by hand. Of course if I need a PCB layout down the road I do a schematic in DipTrace.
@@robertssmorgasbord It's like breaking your brain to write with a pen in the absence of gravity. With how easy it is to write with a pencil... 😅
Very good analysis of those amplifiers. Finally I decided to use the LT6105 to be able to measure consumption through the oscilloscope. Thank you very much for sharing!!!
@@Pepe-ry8pm Thanks for the praise! And you're very welcome!
Thanks very much Robert for covering the instance where the LT6105 power and the measured current were not connected. I wondered about that since it is not covered in the datasheet. We are planning on using the LT6105 in some test circuitry. It is easier to lay out our boards with the returns isolated and connect after the fact if needed rather than the other way around which would require a board respin.
You're welcome! And yes, it's really never mentioned in the datasheet, not even implicitly.
Really enjoyed your video. Really informative. Learned alot.
You're welcome!
Great video. thanks!
You're very welcome! And thanks for the praise 🙂
Hi, good presentation. In reality , the currents sense amplifiers are designed for measuring currents with hi cmrr, not optimized for achieving low noise, thus it cannot replace all the designs
Thanks for the praise! And you're absolutely right: They are designed for extremely wide common mode voltage ranges and high CMRR, not low noise (or high precision). If you want low noise (and high precision) your first choice would be an instrumentation amplifier. But while those usually also offer a high CMRR, their common mode voltage ranges rarely extend beyond the power rails. Guess you can't have it all.
For the INA28x black magic, those switches represent switched-capacitors which has an equivalent resistance of 1/(C * f). In CMOS analog electronics capacitors are much more precise so the "resistors" are likely very well matched, hence the very large CMRR. Although there's an oscillator in there somewhere that they haven't told you about.
Thanks for the info! I wasn't thinking about that, but of course it should be possible to make very precise capacitors (and resistors?) in CMOS. I mean you can control the geometry down to 100nm or even less.
@@robertssmorgasbord maybe precise wasn't the best word to use. Process gradients cause resistors to mismatch while it's easier to match capacitors. I'm not certain on the technology reasons this is just what I remember from my analog design class :P
@@incinatus190 I guess that's the reason they laser trim resistors on chips when they need real precision. But wouldn't it be possible to laser trim capacitors too? Burn way a bit of the top layer?
it is very helpful, thank you so much
You're welcome!
Thankyou Sir. Great explaination.
You're welcome!
Fascinating...cheers.
Ah, you're watching my videos faster than I can reply to your comments :-D
And as always, you're welcome!
nice video, thanks for sharing, I was about to ask you about having different grounds between mesuring circuit and the load, then I saw you tried it disconnected and worked just fine. Maybe with this the common mode issue is reduced dramatically since the common mode voltage is the voltage in the shunt resitance, only issue with different grounds is you can't touch the metering circuit if load voltage is high, I am planning to use this circuit to measure current in a VLF tester ad different grounds is just what I need. Regards from Chile
Thanks for the praise and you're welcome! But to make it absolutely clear: According to the LT6105 datasheet it shouldn't be used without a common ground (floating). See also 39:03 in the video. Though it was working in my setup, that's no guarantee it will in yours too. Anyway, good luck with your project!
Hello! Thank you very much for making this video. This was fascinating for me and also very relevant. For my dissertation (a software engineering degree) I am working on LoRa radio-related technologies. One of the things I have to do is to be able to accurately measure current consumption when the radio is sleeping, receiving, and transmitting. The amount of current drawn ranges considerably: From ~5uA when sleeping to ~140mA when transmitting at full radio power. Do you think the LT6105 would be a suitable device? I would feed the output of the LT6105 to an ADC on a microcontroller for sampling. Somehow I need to scale the output of the LT6105 from 0-5 or 0-3.3V for sampling. This is all new territory to a software engineer! Any advice from either yourself or your viewers of this channel would be greatly appreciated. I will watch the video a few more times as I need to let things sink in slowly! Again, many thanks for this wonderfully informative and very well presented video. Greetings from Scotland!
Hello, and greetings from Germany! To answer your question: No, the LT1605 is not suitable for this application (it does not really go down to 0 - I hopped through some hoops in my circuit design, so that won't be visible on the analog instrument). Your application got an extreme large dynamic range. If 140mA is your full scale, 5uA is 1/28,000 of that. So coming from the ADC side you would need at least a 15bit ADC (32,768 values) for the 5uA to show up as 1 LSB. On the analog side you'll need to achieve a precision of at least 0,0036% of dull scale, so your 5uA will show up at least in the range of 5uA. The next problem is the shunt resistor. To get even a 100uV signal at 5uA you would need a 20Ohm shunt. But that shunt would drop 2.8V at 140mA. In short, there's no single chip / single range solution for your problem. But fortunately there are (open source / open hardware) solutions for that (look for the "Current Ranger"): ua-cam.com/video/HmXfyLyN38c/v-deo.html
So my high side voltage can be anything from around 12V to 18V through an H-Bridge that creates a square wave of around 10kHz at 5A (maybe up to 10). I need to measure that current and convert it to a voltage from 0 to 3.3V into an analog input (a 10 bit ADC). It would be nice to have 10mA resolution and read all the way down to 10mA up to 5A, but I have 2 outputs with different needs, so could have one circuit measure 0 to 1A to measure down to 10mA (giving me more resolution and sensitivity in my 0 to 3.3V output), and then have the 5A circuit have less precision, especially on the low end since is is mostly to give an idea how much current is being used give or take a little bit and to provide short circuit protection if I go over 5A (or 10). But easier duplicating the same circuit if I can. I've even looked at using a Pulse Current Transformer. What would you recommend? It would have to be bi-directional. I also thought of just running something into the DC input to the board that creates the square wave and subtract out the current the board uses. There is an unknown, but high (2-4V drop) from The DC input to the square wave output depending on current. Many of the chips I have seen have a center point, zero current is 2V. The boards that give you 1V per Amp are really nice, but they are uni-directional for measuring the input instead of the high side output. Could you offer any ideas? Thank you.
Measuring the current in an H-bridge is not an easy feat. Have a look at these application notes:
www.ti.com/lit/pdf/sboa174&usg=AOvVaw35atuliGHFP5Txoi8SBztE
www.analog.com/media/en/reference-design-documentation/design-notes/dn407f.pdf
www.st.com/resource/en/application_note/dm00653792-current-sensing-in-motion-control-applications-stmicroelectronics.pdf
Regarding your two ranges: You've got a 10 bit ADC, you need a 10mA resolution between 0 and 1A and you need to measure up to 5A, yes? A 10 bit ADC (1024 counts) will give you a resolution of better that 5mA for a 5A range, so no need to fiddle around with two ranges.
If you've come up with a solution and you're unsure about it, I recommend discussing it on the EEVblog.
Hope that helps.
@@robertssmorgasbord Thank you. My situation is unique in that I am not directly driving a motor, therefore no inductive load. I am feeding it into a board that splits a bipolar 15V p/p signal into a rectifier to get voltage from it, and into an optoisolator to extract the signal from it.
There are so many variables, it is difficult to find what I need. I want 5A, but may need a 10A sense. I will probably need 2 circuits, one to send the sense voltage to a 3.3V logic ADC and one for a 5V ADC. The LTC6103 seems interesting. Maybe even the INA240 or ACS724. I will have to spend quite a few days thinking through all of them. And then consider cost.
The Arduino Motor Shield has an LMD298 on it, so it can only go to 2A. That has a .15 Ohm resistor connected to an Op Amp with a gain of 11. So at max current, it delivers 3.3, which works for 3.3V and 5V logic. I want to do the same thing for 5A and 10A. But I wanted to see if for the same or less money and parts count, I could do better. There are much better current sense chips available to day than when this board was designed. A shunt and op am will always be available though, while a fancy current sense chip could get discontinued tomorrow and then you have to deal with a board redesign.
@@DCCEX Wow, sound you've go a challenging project at your hands. A few more remarks from my side: I don't see any reason why you shouldn't go with low side current sensing in your application and stick with op amps. When you really need two ranges, use two shunts in series (e.g. 0.1 Ohms and 0.01 Ohms) and have one op-amp measuring across both and ground, and the other just across the smaller one and ground (the smaller one of course connected to ground).
Thanks for the info
4:45 shouldn't Vs minus Ground be larger or equal to Vcm max ?
No - and that's the great advantage of these devices. It's also stated in the text above the functional (!) block diagram: "[...] common-mode voltages that far exceed the supply voltage [...]" (26V vs 2.7V). The "trick" here is that it's just a FUNCTIONAL (!) block diagram. The INA181 doesn't really consist of 4 resistors and an ordinary OpAmp. There goes a lot of "special sauce" and "trickery" into the circuits of those devices. I would have to do some research first before I could really explain their inner workings ;-)
Hi Robert, than you so much for your lecture on the current sense amplifier principles. However I do have some questions if you may help me to solve my puzzles. How to detect a current leakage between the Shunt and Load and control the constant current flow into the load ?. Any advantage to monitor voltage drop between sense resistor and the Load (to detect in case of leakage along the path till the Load?.. Thank you for your help..
Hmm, interesting questions. The first thing you would have to answer is where is my current leaking to - obviously back to the ground of the power supply. If that's the case you would need a second second shunt as close to the load as possible with it's own current amplifier. But why do you expect some leakage current? What voltages are we talking here. Anyway, monitoring the voltage across the load through two Independent, non current carrying lines is a common feature in high end power supplies. It allows you to compensate the voltage drop across the current carrying lines between the power supply and the load. Basically the feedback for voltage regulation comes from these two lines, and not from the current carrying binding posts on your power supply. But it does absolutely nothing for detecting current leaks.
What if I want to use SMT current sense resistors to measure high bandwidth AC/DC current instead of using an expensive high frequency current probe? If we make the common mode voltage a little more challenging like mains input for a 25Watt (very popular) universal switching flyback auxiliary power supply. Can it be done?
If we're talking mains AC, there are specialized chips for that job, e.g. www.allegromicro.com/en/products/sense/current-sensor-ics/zero-to-fifty-amp-integrated-conductor-sensor-ics/acs730#PartTable . You don't even need a shunt.
Thank you for making these videos, they are fantastic and clear. Are there any low cost solutions for 120/240VAC? The common mode voltage always seems to be below mains voltages.
First, you're welcome! And no, to the best of my knowledge there are no current sense amplifiers suitable for mains AC. What immediately comes to mind are chips for mains AC energy metering (e.g. www.digikey.com/products/en/integrated-circuits-ics/pmic-energy-metering/765 ). They start as low as $1. However, these are completely different beasts, containing several ADCs (current & voltage), calculate peak and RMS values and usually sport some digital interface to read out the values.
@@robertssmorgasbord I'm surprised by that. I took a peek at a few of my multimeters (including Fluke) and they all seem to use a current sense resistor (Constantan) to measure mains current. I wasn't however able to figure out how they were measuring the voltage drop across it, I couldn't figure out some of the part markings to identify them.
@@robertssmorgasbord Perhaps the INA149 would work. It has a common mode voltage range of +-250V - the price is pretty steep though! I started with trying to build a low cost solution using a simple precise resistor however it's anything but at mains voltages. At that price, I might as well start looking at a hall effect sensor like ACS723 or something along those lines which is basically the same cost.
@@eyeTelevision [Reply to your second comment] Yes, the INA149 would be suitable for 120VAC (+/-170V), but won't survive 240VAC (+/-340V). But keep in mind that you still have to do peak detection and/or RMS calculation on the output signal of the INA149. The ACS723 would work for 120VAC and 240VAC. Its great advantage is that it offers isolation from the mains AC (mind your board design). Though It's not as precise as a current sense amplifier, and you still have to do the peak detection / RMS calculation on its output signal. The above mentioned energy metering chips do all that for you.
@@eyeTelevision [Reply to your first comment] I didn't regard current sense amplifiers suitable for your application because you want to measure AC (!) currents. So even if you find one that can withstand the common mode voltages of mains AC (which you have below ;-) ), you still have to do a peak detection / RMS calculation. Anyway, The above mentioned energy metering chips also require a current sense resistors. In a multimeter the main chip does all the peak detection and RMS calculation stuff.
Thanks.
You're welcome!
nice job
but do you connect ground of load to ground of opamp?
Yup! Load and OpAmp do share a common ground. That said, there are circuits that don't require a common ground, but those also require the power supply for the OpAmp to be galvanically isolated from the power supply of your load. E.g. your multimeter in ammeter mode works this way, having it's own battery as a power supply.
Nice video ,thanks ! Is it possible to get the output down to 0V if we give a negative voltage on the V- instead of connecting to ground ?
You're welcome! To answer your question: Not really. The minimum output voltage of the LT6105 is unfortunately not very well specified. The datasheet just gives you maximum values between 35mV and 45mV, depending on the operating conditions. So you would need some kind of adjustable negative power rail. Also, the LT6105 is a current output device. Its internal circuitry assumes that Vout is connected via a resistance Rout to V-. If you connect V- to a negative rail and Vout via an Rout to ground, you might get a nice 0, but all non 0 output values will get distorted.
@@robertssmorgasbord Thanks Robert. I overlooked that it is a current output devices. I guess the "idle" current output is derived from the input offset voltage.
@@mumbaiverve2307 You're welcome! And you might be right about the input offset voltage causing the minimum output voltage / "idle" current.
Can you provide the link to the shunt resistor you used in your circuit?
It's a Riedon MSR series. I got it from Digikey: www.digikey.de/products/en/resistors/through-hole-resistors/53?FV=-5|18778
What is the output voltage at 400ma input current...I can see 200mv is that the output voltagte....if shunt resistor is 50mohm input voltagte is 20mv and if gain is 100 then it will be 2v right...but I see 200mv....
200mV is the voltage across the analog instrument which has a resistance of 5kOhm. There is a 45kOhm resistor in series with the analog instrument. 2V is the voltage across both resistances (45kOhm + 5kOhm = 50kOhm) and thus the output voltage of the amplifier. See also the circuit at 33:49
Great effort and equally great video. Just to know, can these ICs measure values less than 10mA? If then, how can it be measured using microcontrollers?
Thanks! Of course they can measure values of less than 10mA. But depending on your application you might wanna choose another type. If you want to read the measured value with an MCU you have three options: 1. read the output of the current sense amplifier with an analog input pin of your MCU (build in DAC), 2. use an external DAC with e.g. SPI or I2C interface to your MCU to read the output of the current sense amplifier - or - 3. use a current sense amplifier that has an build in DAC with SPI or I2C interface.
@@robertssmorgasbord Thanks for your quick reply! So how much low value can the IC measure? Can you suggest a good IC for that? I have also seen that similar low-cost hall effect based current sensor ICs like ACS712, ACS758, and others which measures the current but not as low as 1mA or even low. Also they have total output error in their output voltage in the range of 1% to 2%. Does the ICs shown in your video have low error which can measure current in the range 1-10mA or even low? Also are there any alternative ICs out of the video? I use Arduino for interfacing BTW. I really appreciate the reply.
@@ajk0007 Hi, I need a bit more information to make a recommendation: Maximum current to be measured (I assume 10mA), value of your shunt resistor, voltage to ground at the shunt resistor, sampling rate and - of course - how much you're willing to spent for the parts. Maybe you could make a short video showing the circuit you've got in mind? Anyway, to answer you questions: IThe LT6105 can measure currents below 10mA, but that's really determined by the size of your shunt resistor. A LT6105 based circuit should have a maximum error of 1%. And yes, there are hundreds of different current sense amplifiers available ;-) BTW if you use an Arduino analog input and connect the its external voltage reference input to an external voltage reference (say 0.1%), you can measure voltages with 0.35% accuracy (the internal Arduino DAC is about 0.25%).
Thanks for your reply. I was trying to simulate a circuit to measure currents as low as 1mA to 5 mA or 10 mA. So only LT6100 is available in the software. In its typical circuit, a shunt of 3mOhm is used. Also I am using an Arduino to convert the output voltage of LT6100. But I am in doubt of the conversion. Is that a proportional change or is there any other formula for doing that? The supply for LT6100 is 3.3 V. The typical circuit is in its datasheet.
@@ajk0007 OK, staying with the LT6100 for now. I would set its gain to the max of 50V/V (Pin 6 (A2) and Pin 7 (A4) to VEE/GND). Forget about using the internal voltage reference of the Arduino and connect an external one its AREF pin. A cheap TL431 type with 2.5V (0.5%) should do. With the chosen amplification of 50V/V you can measure now a maximum voltage across your shunt resistor of 2.5V / 50V/V = 50mV. So at 10mA you'll need a shunt resistor with a maximum value of 50mV / 10mA = 5Ohm. Choose a 4.7Ohm 1% resistor (cheap and available) and you can measure currents up to 50mV / 4.7Ohm = 10.6mA with a total accuracy of about 2%.
I learnt a lot, thanks...Ive subscribed because I will be back here I think. The math was abit above me some times but I now have a much better understanding than I had at the beginning, so thats great!...I have a project in mind that I need to study in order to design it but basically I want to add to my tools a small but reasonably accurate way to measure a microcontrollers current draw when in "deep sleep" ….Arduino (ATMEGA328p-AU) at the moment. I was going to use an opamp but your vid has given me much to think about. My multimeter is a Fluke 79 series 2 and will measure down to 1uA (40mA scale) and I want to make a device on a small PCB to attach to the meter and use the mV range to measure currents from 1nA up to where my meter bottoms out (if that makes sense!) I would like to eventually make this "open source " so many could benefit from it. Arduino and other micro's are very popular, and of great importance in education but most mustimeters can't measure the Nanoamp range (ones I can afford!) and this is important in battery operated projects. Anyhow great vid and if you can offer any advice to me that would be great, I will now check out your other work and look forward to more......cheers. oh yeah nearly forgot I have a couple of INA105kp instrument amplifiers would they be any use in my project ?
Thank you and you're welcome! Interesting project you have going there. The INA105, being a unity gain amplifier (gain 1), might not be the right chip for your project. Measuring down to 1uA you want something with a high gain and very low input bias current/voltage. E.g. MAX9923F (fixed gain 250, input bias current 1pA, input offset voltage 0.1uV): With a 0.5Ohm shunt resistor that would yield a 0.125mV/uA output, a voltage drop of 0.5mV/1mA and at 40mA an output of 5V. Keep in mind that a designing a current measurement circuit covering 4 orders of magnitude (1uA to 40mA) is quite a challenge ;-)
@@robertssmorgasbordyeah think I need a lot of homework! but I'm gonna keep thinking, all I need is to be able to read about where common micros sleep (around 200nA ish) thanks for your kind reply will check out the MAX chip failing that save up for a meter!
@@andymouse Again, you're welcome :-)
Hi sir, I'm using shunt resistor 0.025milliohms (300A 75mv), I want to measure current using Arduino can you help me out?
Sure, but what are your other requirements? Especially voltage to ground at the shunt resistor, galvanic isolation, resolution/accuracy, sampling rate and - of course - costs?
With a maximum signal of 75mV, the shunt at the low side (connected to ground) and no requirement for galvanic isolation you might go with a simple non-inverting OpAmp amplifying the voltage to e.g. 1.1V (Arduino ATmega328P internal voltage reference) and reading that with an analog input of your Arduino at 10 bit resolution and a sampling rate of max 9615Hz (for a 16MHz Arduino).
Timely subject in era of battery powered DC motors that when heavily loaded can suck your battery dry in seconds.
Yes, monitoring your currents has become more and more important. I remember the time when you just looked up the stall current for a motor in its data sheet and dimensioned the power supply / power electronics according to that. After adding a thermal cut-off switch you were done. I guess those times are gone, considering that your lithium battery can go bang when you draw too much current from it 🙂
Hi Sir, maybe I'm not ok today, but one question!! How Imax=40mA, Vmax=200mV goes to R=5kOhm??? 22:35
It's Imax=40μA, not Imax=40mA (microampere not milliampere)! Please excuse my poor handwriting.
Any recommendation how to measure current from 0.01 to 1A from 0-15V with an Arduino ADC with 10% precision ?
Well, I need to know I wee bit more to make any recommendations: Are these 0 to 15V measured to the Arduino ground? What's the basic circuit? What Arduino are you using? What's the maximum value of the shunt resistor, respectively, the voltage drop across it you can tolerate? Anyway, 10% precision is absolutely no problem, even with the Arduino (ATmega328 for Nano, UNO) ADCs. Add an external voltage reference and you can precisions better than 1%. Simplest solution (the 0 to 15V are to the Arduino ground): Use an external 1.25V 0.5% reference (e.g. AZ432, $0.50) and a shunt resistor of 1.25Ohm. That solution has of course no protection and a high burden voltage.
@@robertssmorgasbord Thanks, I build this current-sink circuit (m.eet.com/media/1130553/13338-31705di.pdf) to log discharge curve of batteries and I just want to use an MCU to actively set the current sink current by monitoring it and setting it through vref. But meanwhile I think thats gets to complicated.. I'll simply adjust the current manually with my DMM.
@@paulg.3067 OK, you've got already a 0.33Ohm shunt resistor referenced to ground (which should/will be also your Arduino ground). 0.33Ohm yields 0.33V at 1A. We need to amplify that to 1.25V (AZ432 external voltage reference for Arduino), so an amplification of about 3.79. With your precision requirements and no common mode voltage issues (shunt resistor referenced to ground) there's no need for any special chips. Just use one half of a second CA3240 op-amp as non-inverting amplifier. The non-inverting input of that op-amp is connected exactly like the inverting input of IC(1A) to the shunt. Choose an amplification for the non-inverting amplifier a little below 3.79 (you don't want your ADC input voltage higher than your reference voltage). E.g. 10k and 27k (each 1%) would give you a 3.7 amplification.
@@robertssmorgasbord Thanks a lot , I will try that for sure :)
@@paulg.3067 You're welcome :-)
❤
😆
Take a look at the Allegro ACS7xx series of Hall Effect Current Sensors. I have the ACS724. www.allegromicro.com//en/products/sense/current-sensor-ics/zero-to-fifty-amp-integrated-conductor-sensor-ics The ACS712 and ACS714 are popular, too. If you buy one of these, only buy it from an OEM vendor, as these ICs are highly counterfeited on Ebay/Amazon/Etc. Be careful of the full extension of the model number that you are interested in, because there are many variations within a single model number... Different amplitude and sensitivity ratings...
Thanks for the info! These are some very nice chips for measuring high currents. Linear frequency response up to 10kHz (OK, they start to phase shift a little at 1kHz) and response times / slew rates in the microsecond range - I like it. The only drawback is the worst case accuracy of +/-3% (ACS724KMATR-20AB at -40°C to 25°C). But since the typical error is +/-1% you can live with that.
I need to measure a current .. the input would be 1mV to 200mV.... I would like to output 0-500mV... or 4-20mA.... would be better to use a instrument amplifier.... .. i need to avoid noise... so would pay more for better results...
Well, I would need some more information to make a suggestion ...
1) Is your shunt high or low side
2) How large is the shunt (you wrote "the input would be 1mV to 200mV", so I'm assuming that's the voltage across the shunt)
Anyway, normally you wouldn't need an instrumentation amplifier, because its main advantage over a current sense amplifier is its high input impedance, and a shunt resistor is (normally) a very low impedance signal source (unless you want to measure pA currents and your shunt is in the MOhm range ;-) ).
@@robertssmorgasbord Hi Robert, thanks for your answer! I'm going to measure a leakage current from a secondary transformer, AC signal, from the high side, not ground side, but I'm going to use batteries for the measuring circuit, I'm going to isolate the ground basically transmitting the signal by wifi. The shunt I'm thinking to get 1 Ohm (10W) because my main range is 1mA to 200mA...(but should support short current 2.7A as it will be shorted frequently for testing) if I could measure 1uA would be great. So I'm looking for a good and precise measurement, low noise, that's why I thought to use Instrumentation amplifier for high CMRR, but I have looked many of these current sense amplifiers like INA219, INA226 (this one is easy to get here), INA282, INA181, INA190 (would be good if the INA190 had I2C because it has a low consumption..), INA240 (this last one looks really good but no I2C) , but not sure which would get better results than instrumentation amplifiers like INA114 etc.... I'm going to use an ESP32 so if it has I2C would be good to communicate directly, but the most important is better measurements. by the way ... your video was a great explanation... I wish my professor would teach like you.... I am inclined to use INA226 ...to try from 1mA to 200mA... and lower if possible....using more units with different shunt scales...
@@RF-yh3qh These are some very demanding requirements you have there ;-)
Let's take it from the back end first: You want an I2C interface, your max current 2.7A and the desired resolution is 1uA -> You'll well need at least a 22-bit ADC wit I2C interface (we're talking here a 6 to 7 digit meter): A NAU7802 (about $2, 2 differential inputs, 24-bit, 80 samples/sec, Vcc 2.7V to 5.5V, available as SMD or DIP at Digi-Key) should do the job (if 80 samples/sec is fast enough). Just looking at the datasheet of that thing you won't need any pre-amplifier (your full scale voltage is 2.7V at 2.7A and 1Ohm). Maybe some overvoltage protection and RC filtering might be in order though. If the internal gain adjustment of that thing is not enough (1, 2, 4, 8, 16, 32, 64, 128) you can also vary your shunt a bit. And with two differential inputs you could implement two ranges (a smaller and larger shunt in series, on input measuring the smaller and the other both) - don't forget some overvoltage protection at the low range (over both shunts) input.
@@robertssmorgasbord Hi Robert, really thanks for your answer, yes you are right... too much requirements... hehehehe... I'm still studying and sometimes it's hard to choose. I took a look at the NAU7802, a really good one, I don't know if it would be hard to implement. I am thinking if 80 samples/s is enough........ not sure if I should get true RMS (if yes... then need more samples as the signal is 60Hz).... or if I could trust on p-p measurement to calculate it..... I'm afraid to get noise on the p-p measurement and screw up, no experience with that. I really appreciate your answer... I'm learning a lot ..... maybe I should stick first to 1mA to 200mA... and use another later to try to get lower measurements... I can use another one that is not I2C.... because I can use the analog input of the ESP32 to input the analog signal... I was just looking for I2C to try to avoid more errors or noise...
@@RF-yh3qh Hi RF, 80 samples/s if definitely too slow if you want to calculate true RMS. If your current waveform is a sine wave (!) at 60Hz you'll need at least 120 samples/s (Nyquist theorem). If it's another waveform you'll need even higher sampling speeds. Try and have a look at the ADS1219. Its features are similar to the NAU7802 and its 1k samples/s. it's about 6€ at Digi-Key, however it comes in a TSSOP SMD package (0.65mm pitched pins). And ... Texas Instrument datasheets are really good (with application examples, PCB layout examples etc.) :-)
Anyway, it might be a good idea to start simple and use the ADC of the ESP32. Don't forget to implement a good input protection though ;-)
Sadly these ICs are not available for most cases
Well, at least the LT1605 is available (in stock) form all major distributors: www.digikey.de/en/products/base-product/analog-devices-inc/505/LT6105/61046 , www.mouser.de/c/semiconductors/amplifier-ics/current-sense-amplifiers/?m=Analog%20Devices%20Inc.&series=LT6105 , de.farnell.com/en-DE/c/semiconductors-ics/amplifiers-comparators?st=LT6105 .
are you okay? why do you sound like you are constipated?
Yes, I'm OK, thanks. Maybe I had cold when I shot that video. I really don't remember.
great content, but such a boring delivery. Maybe you can think of a way to spice it up a bit if you want your channel to grow.
Well, first of all thanks for the praise! Regarding spicing things up: Unfortunately I'm not very good at entertainment or humor - I'm German after all :-)
@@robertssmorgasbord Your delivery is fine. Easily understood by a mechanical engineer like me. Danke sehr. Ich habe drei jahre Deutsch gelernt, auf der schule.
@@outagas2008 Herzlichen Dank! Thank you very much! I'm just happy that you and other people (including HitAndMissLab) get something out of my videos. I also enjoy the technical discussions that ensue from these videos (see all the comments below). But believe me, I would love to make my videos more entertaining, it's just not something I'm very good at :-)
His delivery is just right. Great content with no distraction. The "slow" pace also allows me to think along as he speaks without having to pause the video here and there like so many other 3 minutes videos that make you feel like you understand something until you start thinking the subject deeper the next day. Perhaps it's just me but I like the way he teaches.
@@robertssmorgasbord That explains it 😂If you were at least a Dutch HitAndMissLab would have been so entertained 🤣