Current sensing in power electronics systems
ฝัง
- เผยแพร่เมื่อ 12 มิ.ย. 2024
- An intuitive explanation of various methods of current sensing including: sensing by a resistor, pulse transformer, Hall effect sensor and by applying the inductor's parasitic resistance.
![[FULL EP.13] เซียนหรั่ง พา ไททศมิตร ลงเรือฝึกสกิลหว่านแห | เฮ็ดอย่างเซียนหรั่ง | One Playground](http://i.ytimg.com/vi/EWFuf2Av0K0/mqdefault.jpg)
I haven't seen a person ever explaining complex things so easily and neatly. I really admire your knowledge.
Thanks Satyajit
This gentleman has certainly succeeded in illustrating and showing the beauty and the elegance in this area of engineering through how he relates current sensing to voltages in so many diverse manners. I appreciated all of them, but the manner in which he oriented "Vectors" to make a current passing through a path, reflect itself as a voltage," somewhere else" in the circuit, is the brilliant combination to get unity in the numerator divided by the denominator in the phase-shifting networks shown at 17:54.
Sir, I tip my hat to you, my congratulations. As an old engineer, I must say that seeing this video has ended the year 2020 with great admiration of this gentleman. Thank you.
Thanks Carmel for sharing the mental excitements and exuberance when uncovering the beauty of electronic engineering.
I'm an EE engineer with much experience and I can say your videos and explanations are awesome! Makes it fun to keep learning . Thank you.
👍😊
I totally agree, very good compilations of the must know and must do things to suceccfully have a dcdc running stable.
I've learned so much from you in just the last couple of days. I really appreciate your videos.
👍😊
Very interesting. Your intuitive teaching style is helping me understand dc converters much better. I'm still challenged by a lack of underlying math skills, but your helping me push my practical understanding. That's a tremendous talent for someone to communicate a subject like this without physical demonstration in a lab. Thanks.
Thanks for comment.
your videos are more than phenomenal Dr. Sam.
Thanks. Comments like yours keep me going.
Thank you very much, sir, for the info you have provided! Very awesome teaching, would certainly recommand to others.
Been many years since I was in school but this was a good refresher.
Thanks for a good presentation. And thanks to those below who commented on some of the details.
Thanks
Thank you very much Prof. Yaakov your active current sensing technic will help me to make a cheap measuring device on the oscilloscope !
Thanks for note. Happy to hear that.
Very informative! I like your videos because they show how to use the formulas, they have a mathematical approach.
Thanks for note
Thank you for your explanations. I'm learning so much from you.
Thanks for comment
Amazingly clear explanation! Thank you!
Thanks
Dr Sam, it is really cleared lot of my doubts that's so kind of you.
👍😊
Fantastic Explaination about real word things....
Thank you....
Thanks
Excellent. Especially in the final approach. Thank you.
😊
its nice and great videos Mr.Sam. Thank you so much for sharing info and its more useful
Thanks
Thanks a lot for the very intersting explanation, always very sharp and clear.
Thanks
You so good at what you do Prof that even the trolls can't touch the dislike button. I've hardly seen that happen before on youtube. Awesome. I subscribed.
Hi Emanuel, Thanks for comment.
Really very helpful, you are genius of power electronics. Not only this video but all of your videos are superb. Very good explanation 👍
Thanks
This is my first video of yours if seen and I loved it.. I'm going to subscribe straight away. 👍👍 brilliant
Thanks
I'm watching this video again because it is one of your best. It's a shame that there are only 22,404 views!
Well, there aren't that many smart people in the world. Thank.
Superb!
You explained the concepts well. Pls do more vids, it's really helpful.
Thanks. Will try.
Very informative, excellent sir...
Thanks
It's very useful, well-done for explanation
Thanks
Hello prof. Sam, it is intriguing how these circuits come to existence just by a certain desire to gain functionality. It's like we can throw components in any other way and still get any output that can be tweaked through math and then operate properly. I never thought CT could be pulsed. Thank you.
Thank you for kind note and for sharing your thoughts.
@@sambenyaakov Sir It is I who should thank you. I'm unworthy.
So nice you have made me informed may God reward you
Thanks
You have pretty good kind of english. i am understand all you said even without mega english knowledge. thanks
Thank you! 😃
RATED: 💡💡💡💡💡
9 out of 10 ELECTRONS liked this video.
Thank you Professor 👍😁 SUBBED!
Thanks, but can you decipher this for me?
Dear Prof. Sam, thank you for such a valuable education video in english! I made some conclusions after watching, but I'm not sure that they are correct.
1) As we need to reset core after we had pulse, we can use high enough Rr resistor. That givs us a big area under curve in a reset period, so we can reset core at wery short time. So that sensor could work close to 90% of Duty cycle. And Using HV diode i'll be safe.
2) As we should have a Rr - reset resistor the total current of the secondary will be IL*1/n=ILm+IR+IRr. So to ensure that the most part of the current is current through sensing resistor, we should have IRr and ILm as low as possible. In order to achive this, high value Rr will be nice too.
3) To ensure in low value of ILm and fulfillment of the condition: 2pi*fsw*Lm>R, i will use core material with highest "ue" possible.
Thanks again for the great stuff!
Vasiliy,. You are correct see minute 9 and on. Thanks.
Thanks. Excellent Sir!
Thanks
Thanks, very nice explanation.
Thanks
Very beautiful explaination sir. Thankyou.
😊
Nicely and simple explaination give easy understanding
Thanks
Awesome video. Thanks 👍😊
Thanks
i really the way you describe..
Thanks for this video
Thanks for comment
Thank you for a video!
Thnks
Really Interesting .Thanks a lot
Thanks
Great as usual
Thanks again!
It really helps. Thanks.
Thanks
Thank you!
Dear Prof. Sam, thank you very much for your video. I am wondering whether 'n' is missing in the equation of Bmax at 13:59. Thank you so much.
Dear Bin, Thank for comment and pointing out the omission. You are of course correct. I hope that viewers of this video will notice your remark.
Final expression for Bmax actualy do have 'n' term in numerator, corresponding to current Iav. Where else you suppose to put it? Seems correct to me now
Very interesting video! I have only thought about putting a single winding on the main power transformer, using the existing magnetic flux and the seperate 1:1 transformer using the existing input current to switching circuit. However, this brings a lot of new functionality to the table! Thank you. Also, what does j and omega stand for in 18:00 ?
As I think I have said in the video. I am doing the analysis on any one of the Fourier frequency components (omega) of the inductor current and using the complex representation (j) of the impedance and filter.
Oh that might be, I must have missed it, excuse me! Thank you!
All the best
I just learned about the last method and was exploring the time domain characteristics of it in LTspice, but I kept having to fudge the RC time constant value! After an hour of head scratching I realized LTspice was applying a default DCR of 1mOhm to my inductor, which i was not accounting for! Gees I thought I was doing my unit conversions wrong because I kept coming up being an order of magnitude off lol.
It's a really neat method, but looks like it can't be very accurate since the ESR of the inductor is a function of frequency of temperature, as is the inductance, causing the output of the RC filter to be under or overdamped with variation to L (like with large DC current bias or flux bias in the core)
Good points. Thanks for sharing. If not a member: welcome to join www.linkedin.com/groups/13606756
very good video, I nirmally use gall sensors with adequate bandwith, but yes once also used a current transformer.....just to had it done in oractice once.
Thanks for sharing
Thanks, I was wondering if you can simply use an extra winding on the output inductor in order to measure the current with a pulse transformer for a dc/dc converter?
A winding on an inductor does not sense current it senses voltage .
Thankyou very much
Thank you very much
Thanks
thanks
This is a wonderful video professor🙏. I'm watching it for the 3rd time and this time, I completely understood what exactly you're trying to explain. There are a few questions that arose in my mind while watching this video
1. In 1st slide: why don't we have a negative voltage spike when current fell to zero abruptly?
2. How exactly does ultra costly 50MHz (AC+DC) current probes work? I'm saying ultra costly coz atleast they are too expensive for individuals to own.
3. Can I make a low cost high frequency current probe by separating out the measurements for AC and DC signals? Using cheap low bandwidth hall effect sensor for DC component and using a pulse current transformer with a range of output resistors to avoid core saturation? Do we get accurate readings by adding these components? Also, is there a better way for doing this accurately at lower cost?
First slide? please indicate minute of video
In commercial current probe the HF portion is indeed based on the AC transformer action.
@@sambenyaakov Slide 9-4. Time: 01:01
Yes indeed there will be a negative voltage spike which is not shown. Thanks for pointing this out.
Question to 1:20 - should not this voltage spike appears at the end of the current pulse? not on the beginning?
It will appear at both the beginning and end (high di/dt). I neglected to show the negative spike. Thanks for pointing this out.
Very Good Video detailing the reset mechanism for CT. @7.10 What type of waveform we can expect across burden resistor whene Ct is not resetted propley or volt sec criteria not met. A sharp return zero instaed of sinusodal return? along with ringing?
Thanks. Good question. With no reset to core, the circulating current will keep rising, the inductance will drop and I suspect that there will be either a thermal run away, or the circuit will stabilize with a much higher resonant frequency.
@@sambenyaakov thank you for the quick reply . My question was more like there is is burden resistor present and what type of wave dorm I can expect , my under standing from your explanation is that reset will still will be in sinusoidal nature but with higher frequency
Thanks Sir
👍🙏
Sir what do you think of measuring current through mosfets rdson? In very high current drive circuits, not adding one more lossy component and measuring through rdson is very attractive. any advice or contrary idea? thanks in advance.
This is a good approach. The down sides: temp and current dependence of Rds(on) and the extreme difficulty to implement it in a discrete circuit. Suitable for a monolithic design.
thabk you
👍🙏
Thank you for clearing sir.
Sir I am calculationg reset time of CT. Which parameters are important to calculate it.
What I understood is ( Secondary side inductance, shunt capacitance are enough to calculate reset time.) Shall I measure L & C at switching freq
Indeed.
Vicky, at 9:57 there is a very good schematic, comprising a Zener diode in series with a blocking diode shown. As professor Ben-Yaakov mentioned, this schematic is intended for the reset time control. The reset time can be defined as Tres=Lsecondary * Isecondary / (Vzener+Vdiode) The volt*second integral should remain intact. Therefore, for shorter Tres you should select a higher voltage Zener diode and don't forget about selection of a higher voltage blocking diode.
Say i have 2 in-phase AC signals, at 200 kHz. i would like to obtain a 200 kHz AC output with current equal to the current-difference between the two input currents. I don't desire a DC or proportional-voltage output. Can i use the CT or other methods? Thx!
Yes, you can definitely use a CT (one core) with two opposed windings. Remember that the load of the third output winding should be such that the voltage on it will not saturate the CT core. Good luck.
Sam Ben-Yaakov Thx! could I achieve the same using a transformer with 3 windings?
You need in any case three windings 2 for primaries 1 for output.
RC=L/Rs Very useful equation when the coil wire DC rezistance is steady. Maybe need to compensate the temperature coefficient of the copper wire for high current values?
Do you think a NTC in place of(or in parallel to) R would do ?
Thanks Mr. Yaakov for the nice videos.
You are correct there is a temp sensitivity. Inductor wires are normally made of copper which has an appreciable temp coefficient. R is NOT added! it is the resistance of the wire!
I meant R the resistor in serial with a capacitor across the inductor. 17:45
What correction method would you suggest to get accurate (free from temperature ) current feedback from the inductor? (other than larger wire diameter)
Thanks.
Sorry. I misunderstood. Good idea, perhaps somebody else already did it? I don't know. The problem of temp tracking is that in real systems the temp inside an inductor winding might be much higher than the ambient. At any rate, in most applications, there is no need for high accuracy in the measurement of inductor current. This signal is normally used for current feedback and there is no need for precise current measurements there.
Here’s an idea. The “differential amp” @19:44 could be implemented in the usual way with an op-amp and 4 resistors, where the gain is the ratio of resistors. Use copper (or iron/steel) wire-wound input resistors (with a tcr similar to the inductor) that are thermally connected to the inductor. This ought to decrease the amplifier gain as the inductor heats up.
Does the diode leakage current has any significant effect on the measurement results?
It is usually insignificant being much smaller that the measured currents.
@@sambenyaakov Okay, thanks for the reply.
Nice video but If it's open circuit CT how can you make a relationship between output current of the CT in terms of measured output voltage?
Open circuit voltage depends on magnetization inductance.
@@sambenyaakov Wow! you're a legend I didn't think you would reply coz the video was 3years ago. Thanks :)
why the dc current will flow though the diode but through the Lm, the resistor of Lm is zero to the dc.
Does power electronics designing domain have future.I am interested to work in this domain and suggest me which domain is best and have future.
Yes, definitely. EV area is growing fast.
could you please upload some videos of application examples in industry?
Will try
Sir
If the capacitor used for measuring voltage is connected to the output capacitor then the Common mode voltage of one of the terminal of the capacitor is constant (equal to output voltage). But what about the voltage at the other terminal of the capacitor. Is it pulsating .
Please indicate the minute in video you are referring to.
@@sambenyaakov 20:40
@@RAHULRAHUL-qy6cw the voltage at the other end is the reflection of the inductor current as is. There is a need for differential amp.
@@sambenyaakov Thanks sir
can you comment sometime on current measurement at very low values in nano and micro amps
Look up LPV821
Hi I need VTG. Sensing circuit per elect.
the last one with Vout = I Rs G sounds great... but Delta Rs is HUGE over delta Temperature. So now you have to compensate for that.
But quite predictable.
Here’s an idea. The “differential amp” @19:44 could be implemented in the usual way with an op-amp and 4 resistors, where the gain is the ratio of resistors. Use copper (or iron/steel) wire-wound input resistors (with a tcr similar to the inductor) that are thermally connected to the inductor. This ought to decrease the amplifier gain as the inductor heats up.
sir any simple way to identify the phase sequence of three phase supply
Maybe connect a motor and see which way it turns?
Beautifull...at 19:40...
👍😊
12:55 "the impedance of (the magnetization inductance) must be much lower than (the load resistance). I think you mean higher, not lower.
Hi Graham, thanks for the the correction of the slip of the tongue. Indeed, the impedance of transformer need to be larger than the resistor as written on slide, and the current lower, which is also indicated.
Ehh.... this is quite interesting but i'm not sure about some things....
My main question is what happens when you use a step down transformer instead of a step up transformer?
Most of the compensation/error mechanisms mentioned herein are caused by measuring a voltage that has been stepped up and the resulting reverse voltage conditions that naturally occur because of that.
I suspect that the answer is "impedance", but changing the voltage seems like a guaranteed way to mis-match said impedance, so perhaps not.
Hi. You seem to have missed the point. No voltage is stepped up. The purpose of the pulse transformer is to measure current. If it will be a step down then: 1. The output current will be LARGER than the primary current, 2. The impedance reflected to the primary will be large, disturbing the current path and dissipating power.
Sam, the secondary voltage v is a derivative of the primary current. If the primary current has a DC component (it is hard to imagine a conventional converter that does not), it should be counted too, since it magnetizes the magnetic core, moving the setpoint on the B-H curve closer to the saturation. The derivative ignores the DC current component. Integrating the volt-second product disguises the real process in the core, since it deals with the voltage value obtained AFTER taking the derivative and thus missing the effect of the core magnetization by the DC current. Therefore the statement "Pulse transformer operation is limited by voltage NOT CURRENT" 14:28 is incorrect since it actually is limited by the total primary current - AC + DC, flowing through the primary winding.
Dear Gregory, The pulse current sensor is intended to measure a pulsed current. For example in the Drain of a switching transistor, or in the line of a diode. This is the application. In this case, the core is magnetized during pulse time and demagnetized during the off time. What matters then, in terms of saturation is the volltsec imposed on core core, conveniently measured at secondary. Being a transformer, the primary and secondary currents during pulse are canceling each other so what is left is the magnetization current, not the actual pulse current.
Regards
Sam
One more point (added to first reply which may appear as second) . The output voltage is NOT the derivative of the primary current! During the pulse time I2=I1/n (save the magnetization current). So V2=I1*R/n
@@sambenyaakov Dear Sam, the secondary voltage v is a derivative of the primary current. If the primary current has a DC component (it is hard to imagine a conventional converter that does not), it should be counted too, since it magnetizes the magnetic core, moving the setpoint on the B-H curve closer to the saturation. The derivative ignores the DC current component. Integrating the volt-second product disguises the real process in the core since it deals with the voltage value obtained AFTER taking the derivative and thus missing the effect of the core magnetization by the DC current. Therefore the statement "Pulse transformer operation is limited by voltage NOT CURRENT" 14:28 is incorrect since it is actually limited by the total primary current - AC + DC, flowing through the primary winding and defining present flux density B."
I have a MathCAD analysis of a current transformer from the point of view of the pulse top transfer fidelity. If you have MathCAD I can share this file with you through a personal email.
@@sambenyaakov Dear Sam, yes, it is intended but the DC current is still present even if you put the current transformer in the drain circuitry directly.
Dear Gregory
It is a common mistake to think that a transformer can not transfer an average DC current. This notion is incorrect when it comes to switched circuits. See for example a forward coverter. There is an average DC current component in the primary! What is not allowed is an average DC VOLTAGE on any of the windings.
As for the pulsed current transformer explained in the video, it is widely used in the industry. Challenging it is like a person seeing a camel and exclaiming: there is no animal like this😊 In a switched NONLINEAR circuit you need to consider each segment by itself. If you will do that, then during the pulse time the primary current is balanced by the secondary current and then you have a magnetization current which is discharged during the off time. So nothing is violated here.
You should add a practical demo for this ... showing the advantages at different frequencies and selecting actual parts , hence giving a reasonable design pattern for people to implement this in their designs... good presentation though thanks !
David, Your the first! Out of the 54K or so who have watched this video, you are first and only to complaint: "You should add". I guess all others appreciate the time I take to prepare these videos sharing them for free to all.
Hard to understand. Please explain in detail and speak a little slower. Thanks.
Play it at 0.75 speed. I am sure that watching it again will help you out.