Thanks for this great series! I was working as a BMS Trainee Engineer a few months ago and was working on the battery voltage isolated sensing when I came across this problem. I even searched for 1-to-1 ratio flyback converters but they also didn't seem like the solution. THIS SERIES INSPIRED SO MANY IDEAS.
27:50 according to companies datasheets optocouplers are slow, and even higher end ones do not stand a chance against MEMS inductive coplers. At least for digital couplers like high-speed general purpose ones and purpose built ones for e.g. specific MCU bus
I have a system with 70 - 120V output at around 40 to 100W - but the frequency (I can vary) is around 10-40 KHz. Load is an incandescent bulb. I need to measure the output power at that frequency (especially around 10 kHz. It looks like your system will work... but what do you think? thanks
Unless you need a very fast response a low-pass filter before of after this would work nicely. If you're wanting to measure both voltage and current you can use a current sense amplifier to get a decent voltage out of a shunt, or could just go for one of the nice LEM sensors the current ones aren't especially expensive. if its pure AC you could consider a much cheaper current transformer too. If the load has a constant resistance I'd go with just measuring current then calculating power as that is cheaper and simpler, although an incandescent bulb would probably not be. In which case for voltage measurement you'd probably be best going for a commercial isolation amplifier, which hides all this tricky stuff in a small and affordable package.
you do have cores with a gap they are called a pice of bent iron it will work even at highrates of changing flux because of the air gap saturation should not happen .
I used a gapped ferrite core. Saturation is not the problem it’s just the inductance attenuating the current/flux at higher frequencies. Only a higher input amplitude would help with that, and in this case the frequency would still be limited by parasitics causing unwanted resonance.
I think at 100 kHz your system starts ringing. Since conductors resist changes in current and you're trying to change the current fairly quickly at 100 kHz. The coil introduces a delay, which means that your feedback circuit corrects for the current too late, either increasing or decreasing the current, depending on the phase shift. Basically, as you already suspected, you probably need a coil with less inductance. Which of course also decreases the sensitivity, so as usual it's a trade-off... Anyway, so what's up with the German flag in the background?!
Ah yes that would explain why it starts so suddenly at around 100kHz. And the flat is because I appreciate German engineering, and it was stolen from a real German engineer at a solar powered car race! (They won of course)
@@electrarc240 Ah ok... I am in fact a German (electrical) engineer. But Germans tend to be a little suspicious about people (particularly Germans of course) having a German flag in their apartment, cause you know, history. :D Anyway, I like that your videos show and explain practical examples of reasonably complex circuits. Not just the basics that RUclips is kinda full of already.
@@electrarc240 Well, that's up to you of course. I guess a flag could potentially be misunderstood by some. But then again, it's also kinda your thing how you choose to decorate your apartment. I was just curious what that was about. Anyway, keep up the good work. I'd definitely be interested in your attempt at optical voltage measurement using the Vishay optocoupler. :-)
The IL300 matched opto is more accurate (about 4x more) and way faster (with fast op-amps, hundreds of kHz could be done) than this hall thing. It must have something about it that the IL300 does worse, otherwise why?
Because sometimes it is fun to make things even though you know they won’t be optimal. This is part of a series comparing techniques for isolated voltage measurement, wouldn’t be much of a series if I just said “this is the best” and left it there.
Thanks for this great series! I was working as a BMS Trainee Engineer a few months ago and was working on the battery voltage isolated sensing when I came across this problem. I even searched for 1-to-1 ratio flyback converters but they also didn't seem like the solution. THIS SERIES INSPIRED SO MANY IDEAS.
Great to hear I’ve inspired you!
Great video! I thought I had a good understanding of how hall effect current measurement works.
The life of an electrical engineer “I have to be quite speedy to take measurements; otherwise, everything will start burning and melting.”
I am so pleased with your candor. “The capacitor clickbaity thing for more subs” is hilarious.
I just discovered your channel and your videos are very good! Keep it up! I'm looking forward to more videos
Wow thanks!
Magnetic measurements also are sensitive to the surrounding magnetic field, so that if you move your device you will see a change in the output offset
Yes that’s a good point. The commercial units do a good job to minimise these effects but mine would probably be very susceptible to that!
27:50 according to companies datasheets optocouplers are slow, and even higher end ones do not stand a chance against MEMS inductive coplers. At least for digital couplers like high-speed general purpose ones and purpose built ones for e.g. specific MCU bus
Yes they do seem very "slow", a shame because I like their simplicity
👍👍
I have a system with 70 - 120V output at around 40 to 100W - but the frequency (I can vary) is around 10-40 KHz. Load is an incandescent bulb. I need to measure the output power at that frequency (especially around 10 kHz. It looks like your system will work... but what do you think? thanks
Unless you need a very fast response a low-pass filter before of after this would work nicely. If you're wanting to measure both voltage and current you can use a current sense amplifier to get a decent voltage out of a shunt, or could just go for one of the nice LEM sensors the current ones aren't especially expensive. if its pure AC you could consider a much cheaper current transformer too. If the load has a constant resistance I'd go with just measuring current then calculating power as that is cheaper and simpler, although an incandescent bulb would probably not be. In which case for voltage measurement you'd probably be best going for a commercial isolation amplifier, which hides all this tricky stuff in a small and affordable package.
@@electrarc240 Thank you for those ideas!
you do have cores with a gap they are called a pice of bent iron it will work even at highrates of changing flux because of the air gap saturation should not happen .
I used a gapped ferrite core. Saturation is not the problem it’s just the inductance attenuating the current/flux at higher frequencies. Only a higher input amplitude would help with that, and in this case the frequency would still be limited by parasitics causing unwanted resonance.
I think at 100 kHz your system starts ringing. Since conductors resist changes in current and you're trying to change the current fairly quickly at 100 kHz. The coil introduces a delay, which means that your feedback circuit corrects for the current too late, either increasing or decreasing the current, depending on the phase shift. Basically, as you already suspected, you probably need a coil with less inductance. Which of course also decreases the sensitivity, so as usual it's a trade-off... Anyway, so what's up with the German flag in the background?!
Ah yes that would explain why it starts so suddenly at around 100kHz. And the flat is because I appreciate German engineering, and it was stolen from a real German engineer at a solar powered car race! (They won of course)
@@electrarc240 Ah ok... I am in fact a German (electrical) engineer. But Germans tend to be a little suspicious about people (particularly Germans of course) having a German flag in their apartment, cause you know, history. :D
Anyway, I like that your videos show and explain practical examples of reasonably complex circuits. Not just the basics that RUclips is kinda full of already.
@@phinok.m.628 Ah, maybe I should take it down so I don't give the wrong impression. Don't want that!
@@electrarc240 Well, that's up to you of course. I guess a flag could potentially be misunderstood by some. But then again, it's also kinda your thing how you choose to decorate your apartment. I was just curious what that was about.
Anyway, keep up the good work. I'd definitely be interested in your attempt at optical voltage measurement using the Vishay optocoupler. :-)
@@phinok.m.628 I'm designing a PCB for it right now! Might be about a month before it's here though unfortunately.
The IL300 matched opto is more accurate (about 4x more) and way faster (with fast op-amps, hundreds of kHz could be done) than this hall thing. It must have something about it that the IL300 does worse, otherwise why?
Because sometimes it is fun to make things even though you know they won’t be optimal. This is part of a series comparing techniques for isolated voltage measurement, wouldn’t be much of a series if I just said “this is the best” and left it there.
@@electrarc240 :):):)