0:22 Intro 1:42 LC LP filter (2nd order low pass filter) 3:40 RC LP filter (1st order low pass filter) 4:55 RCRC LP filter (2nd order low pass filter) 5:27 Disadvantages (voltage drop over R: V=R*I, e.g. 22V=22R*1A) 7:08 Solution (RC filter with low current draw) 7:32 Highlight 7:45 Emitter follower (Common Collector Amp) 10:18 Component size comparison 12:13 Issues (capacitor works not as energy storage, output voltage needs to be lower than input voltage) 13:22 Lower base voltage (with R) to drive the BJT at low input voltages 14:20 Conclusion 14:35 Historical use of the Capacitance Multiplier 15:15 Final conclusion (today's use) 15:50 Outro
This is a really great tutorial I like the fact that you did the circuit simulator and in real life. You broke it down in a way where anyone can understand your video.
Excellent and easy to understand video! Thank you. Sometimes it's difficult to find negative, or high voltage regulators, and they can be more expensive than just doing something like this. It's also nice to filter out switching noise.
Dude you are amazing. Also, the new layout with the circuits on the big screen and you in the small really makes the content experience a lot better. Keep up the great work.
It is a great circuit. I used it many years ago to completely get rid of any audible hum in an OP-amp power supply. That power supply was just a bridge rectifier and an electrolytic cap and then a capacitance multiplier.
I have never heard of this circuit before, actually, so thanks for the introduction. It looks best for providing filtering of substantial currents at low frequencies. Nice one Fesz!
Indeed, it does not make sense at high frequencies, since an LC filter would have fewer losses, but at low frequencies having a huge capacitor is not always practical.
Very nice tutorial.Would be nice to know how to size the components for a high voltage supply , 1kV perhaps with some simulation of the load current. Thanks !
Instead of adding a resistor in // with the capacitor, can't we use a PN diode in series with the resistor to drop the base voltage by 0.6V . The advantage is that you have a fixed offset independently of the average input voltage.
I guess that will also work quite well; but from a price point of view, the added resistor is cheaper - I guess that is why the diode implementation you mentioned is not that common. If you have the budget for a diode, its easier and better to just add a zenner in parallel with the capacitor, so you get proper output stabilization.
Wow, amazing again. History of how implementation depends on the cost of the devices. I mean, you are also an economist 😊. Keep going FesZ, you are far from another UA-camr. Brilliant explanation and work.
Thank you for the kind words! I wouldn't say I'm an economist, but I do have a day-job in electronics, and one of the topics every engineers is sick of is costs- "Why is the circuit so expensive? Make it cheaper!"
@@FesZElectronics You know how to do that because you know it is history. Mostly, nowadays "engineers" never care about that. I do want to bold these details. I'm sure that you make it cheaper without loose quality. Excellent man, excellent.
Hi FesZ, always a pleasure to see your videos. This one, I already saw it on 3~4 times. I still have some questions and I would appreciate any light you could bring: Q#1: As the Common collector (emitter follower) has an intrinsic “following” Base-Emitter response, could we assume it works as an OPEN LOOP circuit if seen at multi-component level (transistor, capacitor C2, resistor R1, etc)?
Cont. … Q#4: Instead of the voltage divider at the BJT’s base, if we have used a ZENER, wouldn’t it preserve the AC response from the RC filter, while using the lower impedance of the Zener (compared to R1) to further reduce the AC to the Emitter? In this case, would this “regulated” Capacitance Multiplier be a better option to the unregulated one? Possibly even better than LM78xx or LM317 units - in terms of RIPPLE only? P.S.: Because in terms of load & line regulations, surely the LM ones would outperform even the regulated Cap. Multi. Thank you for all Euthymios (aka: EJEuth)
Hi ! really a great tutorial. Would a darlington be the better device to be used as cap multiplier compared to a bjt ? if i have understood rightly the parameter to look at is the hfe ... some darlington have hfe= 5000 ! the kings of the cap multipliers ? Thanks a lot, gino
Cont. … Q#3: About Cap.Multi x Series Regulator = the latter being much more complex, with dozens of internal components and performing the feedback loop “externally” at component level (e.g. using diff. amplifier, etc) Wouldn’t a regulator be SLOWER and more prone to OSCILLATIONS than the “intrinsic feedback” provided by a simpler Cap. Multi.?
FesZ, thank you for that video. Question : the frequency domain for Cap Multiplier has a -20dB roll off ~ 100khz and then has +20dB gain after. Are there ways to solve this to make the frequency behavior the same as a single pole RC filter? What is contributing to this effect? it probably is the FETs/BJT parasitics coming into picture at higher frequencies but I'm not sure . Playing around with Cgs capacitance and adding Caps between Emitter/Source to GND helps but i was wondering if there is a better method
I guess you added pretty small values of resistor and capacitor in the base of the transistor if the roll of is at 100KHz, the nice thing about this circuit is that it can achieve very low filtration frequencies. Anyway, coming back to your question, I did see some schematics where a ferrite bead was added in the base of the transistor to improve high frequency behavior, but anyway, at frequencies >1MHz you can add an extra ferrite bead + capacitor filter to improve filtration proprieties. With any transistor, because of the E-C capacitance at some point, high frequency noise will pass trough.
What happens if I put one of these in the output of a buck converter and wire the feedback *after* the filter? Will it become a super clean buck converter or will it oscilate itself to hell?
I don't get it, at 13:52 you add a resistor to ground to lower voltage at base, thus lowering Vbe? How does that help?? I guess I also don't get why voltage output at C_multi drops immediately after input voltage becomes lower than at C_multi? Is this because Q2 somehow gets into cut-off region or something?
Hello AlanS. Vbe is a constant (roughly) around 0.6V, and that does not change. Now you can estimate the votlage on the emitter based on the voltage in the base by saying that Ve=Vb-Vbe; Ve in this case is the output of the circuit. By adding the extra resistor I push Vb lower, and that drags Ve also lower. Regarding the second point, the voltage at the output is coming from the input, so there is no way that the output will be larger than the input. Ve=Vc-Vce (Ve-output Vc-input Vce-voltage drop on transistor, a positive value usually >0.1V) Let me know if its a bit more clear or if you have any more questions!
When using a MOSFET in a capacitance multiplier, is the voltage output limited to less than max gate threshold voltage? For example I have a 70 volt transformer, rectified to 109VDC, and I want to filter the ripple, and use capacitance multiplier to amplify the current to 10 amps. 109VDC on the gate of a mosfet will destroy it, right? In this case, do I have to use a high voltage BJT?
Hello Josh! Well the Mosfet is in danger only if the Gate-Source voltage (Vgs) exceeds the rated value (usually its ~+/-15V). Just to be on the safe side you can add a zener diode between the G and S. The gate threshold voltage (Vgs_th)(usually 1 to 5V) is the voltage that is needed to get the transistor to conduct. Another parameter to consider is the Drain Source voltage (Vgss) - this tells you what is the maximum voltage difference you can apply between the transistor terminals. When the power supply is off, there is no voltage on the output. When you turn it on, in the first few ms there will voltage on the input and nothing on the output so the drain-source will see the full input voltage. I do recommend that you don't chose a transistor on the limit with this parameter - for the 109V you are interested in use a 150 or 200V transistor. Point is that it does not matter what voltage you are trying to stabilize. It will work with both a BJT and a FET, as long as the transistor is rated for that voltage. But since you are interested in 10A, FET might be the better option, or for the BJT version a darlington arrangement. Anyway you could try the circuit out in the simulator first just to be sure.
what does it mean for the filter to be second order filter rather than first order filter and why does it give a higher frequency ramp in the frequency response?
The filter order is related to the number of reactive components (an RC filter is order 1 because there is only 1 reactive component - the C); the attenuation of a filter is aprx filter order * 20dB/decade; a second order filter will be 40dB/decade
And you need to pay attention to discharging the capacitor voltage when the power supply is turned off, as the stored energy will discharge through the B-C junction causing a transistor failure if you are not careful!
I am driving an N-CH MOSFET with pwm signal (20-30 KHz). The voltage at the drain is the 60Hz signal from the wall outlet. Every time the gate turns on and off, there is a ripple at the Mosfet's source. The 7810 voltage regulator does not filter this ripple out. Can I use this technique to smooth out the ripple? I don't want to use an inductor at this stage of the circuit. Thank you.
What frequency is the ripple at? Another way that is used in power supplies is a "snubber". A series RC circuit between the noisy point and ground chosen so that is absorbs signal above a particular frequency (say 2R2 and 1nF). This should have no effect on the 50Hz but absorb high frequency stuff.
@@FesZElectronics Thank you for replying. I finally fixed it by placing a 0.1uF between the drain and the source of the mosfet. And another one between source and ground. The frequency of the ripple was so high that I could only see the waves when I zoomed down to 20ns on the oscilloscope. But the voltage was over one volt.
The inductance of an inductor is not a constant; when there is a magnetic core, the inductance is proportional to the magnetic permeability of the core; The core permeability is magnetic flux dependent - stronger flux reduces the permeability; at high inductor currents, the core permeability drops and so does the inductors inductance
Thank you for the suggestion! Before I get to do that though, if your main interest is in using the pi filter for switching power supplies, I highly recommend - www.analog.com/media/en/technical-documentation/application-notes/AN-1368.pdf If nothing else, this might clarify some topics regarding the ferrite bead.
@@FesZElectronics I will take a look. I am more interesting pi filters at transmission lines, I have seen sometimes chokes and others beads. I will take a look in the AN. Thx
It's great. Would you mind to record something for real begginers? Something like analysis of common emiter amp? How dc and ac circuit flow, why do behave this way, etc? Greetings from Poland
Nice video again! But why not using a RCRC filter to this type? That way you even filter a lot more! I actually find it strange that not a lot of people seem to use a rcrc filter with these kind of circuits. Another useful addition is to mention that for quite some circuits (mostly audio), noise and ripple is far more important than perfect voltage regulation. So in this case a simple zener will do the job. In the end you'll have a circuit that has much better PSRR than a standard 78xx regulator (and probably also better than a lm317). I like your videos, they have a good pace and are well put together 👍🏻
Well I covered that(RCRC) in the beginning, I just refereed to it as a "second order RC filter", and afterwards I did not include it in the schematic to keep things simple. I fully agree, its more efficient than the simple RC filter and depending on the frequency you are trying to filter and where the corner frequency is, it can be even better to have 2x(R/2+C/2) rather that R+C so the size of the 2 options is comparable.
0:22 Intro
1:42 LC LP filter (2nd order low pass filter)
3:40 RC LP filter (1st order low pass filter)
4:55 RCRC LP filter (2nd order low pass filter)
5:27 Disadvantages (voltage drop over R: V=R*I, e.g. 22V=22R*1A)
7:08 Solution (RC filter with low current draw)
7:32 Highlight
7:45 Emitter follower (Common Collector Amp)
10:18 Component size comparison
12:13 Issues (capacitor works not as energy storage, output voltage needs to be lower than input voltage)
13:22 Lower base voltage (with R) to drive the BJT at low input voltages
14:20 Conclusion
14:35 Historical use of the Capacitance Multiplier
15:15 Final conclusion (today's use)
15:50 Outro
This is a really great tutorial I like the fact that you did the circuit simulator and in real life. You broke it down in a way where anyone can understand your video.
Good lesson I learned from this video: the bias down using the 33k. Nice work, thanks!
Excellent and easy to understand video! Thank you. Sometimes it's difficult to find negative, or high voltage regulators, and they can be more expensive than just doing something like this. It's also nice to filter out switching noise.
Dude you are amazing. Also, the new layout with the circuits on the big screen and you in the small really makes the content experience a lot better. Keep up the great work.
Glad you enjoy it! this was actually a suggestion from a viewer! I also think its much better this way.
It is a great circuit. I used it many years ago to completely get rid of any audible hum in an OP-amp power supply. That power supply was just a bridge rectifier and an electrolytic cap and then a capacitance multiplier.
Thanks for the video. This was great, I like how you compared other traditional methods to the CM circuit
Thanks to UA-cam algorithms for pushing to me your video this morning, most interesting!
I have never heard of this circuit before, actually, so thanks for the introduction. It looks best for providing filtering of substantial currents at low frequencies. Nice one Fesz!
Indeed, it does not make sense at high frequencies, since an LC filter would have fewer losses, but at low frequencies having a huge capacitor is not always practical.
Great tutorial
Very Informative and detailed lesson.
Thank you for the video, I didn't know this circuit in so much detail
Most voltage regulators aren't good at rejecting high frequency ripple, so this circuit still has some uses.
very true for IC regulators in particular
Excellent video again!
Another great tutorial
Thank you!
Very nice tutorial.Would be nice to know how to size the components for a high voltage supply , 1kV perhaps with some simulation of the load current. Thanks !
Love your channel. Could you also incluse how you connect the oscilloscope in your demostrations. Thank you.
Thank you. This was a big help to me
Hi Fesz,
Great work. Learned a lot from this video.
Can you also make a video on Hold Up Circuits please.
Cheers !!
I'll have to look into this, seems like an interesting topic! Thanks for the suggestion!
That is an awesome video. Thanks for making it!
Instead of adding a resistor in // with the capacitor, can't we use a PN diode in series with the resistor to drop the base voltage by 0.6V . The advantage is that you have a fixed offset independently of the average input voltage.
I guess that will also work quite well; but from a price point of view, the added resistor is cheaper - I guess that is why the diode implementation you mentioned is not that common.
If you have the budget for a diode, its easier and better to just add a zenner in parallel with the capacitor, so you get proper output stabilization.
Wow, amazing again. History of how implementation depends on the cost of the devices. I mean, you are also an economist 😊. Keep going FesZ, you are far from another UA-camr. Brilliant explanation and work.
Thank you for the kind words! I wouldn't say I'm an economist, but I do have a day-job in electronics, and one of the topics every engineers is sick of is costs- "Why is the circuit so expensive? Make it cheaper!"
@@FesZElectronics You know how to do that because you know it is history. Mostly, nowadays "engineers" never care about that. I do want to bold these details. I'm sure that you make it cheaper without loose quality. Excellent man, excellent.
Hi FesZ, always a pleasure to see your videos. This one, I already saw it on 3~4 times. I still have some questions and I would appreciate any light you could bring:
Q#1: As the Common collector (emitter follower) has an intrinsic “following” Base-Emitter response, could we assume it works as an OPEN LOOP circuit if seen at multi-component level (transistor, capacitor C2, resistor R1, etc)?
Cont. … Q#4: Instead of the voltage divider at the BJT’s base, if we have used a ZENER, wouldn’t it preserve the AC response from the RC filter, while using the lower impedance of the Zener (compared to R1) to further reduce the AC to the Emitter?
In this case, would this “regulated” Capacitance Multiplier be a better option to the unregulated one? Possibly even better than LM78xx or LM317 units - in terms of RIPPLE only?
P.S.: Because in terms of load & line regulations, surely the LM ones would outperform even the regulated Cap. Multi.
Thank you for all
Euthymios (aka: EJEuth)
Hi ! really a great tutorial. Would a darlington be the better device to be used as cap multiplier compared to a bjt ? if i have understood rightly the parameter to look at is the hfe ... some darlington have hfe= 5000 ! the kings of the cap multipliers ? Thanks a lot, gino
Excellent video. That is topic well explained.
Cont. … Q#3: About Cap.Multi x Series Regulator = the latter being much more complex, with dozens of internal components and performing the feedback loop “externally” at component level (e.g. using diff. amplifier, etc) Wouldn’t a regulator be SLOWER and more prone to OSCILLATIONS than the “intrinsic feedback” provided by a simpler Cap. Multi.?
❤❤❤❤
Cont. …. Q#2: Or the Cap. multi. should be seen as a CLOSED LOOP circuit, probably because there is a feedback on the hybrid model (base & emitter)?
FesZ, thank you for that video. Question : the frequency domain for Cap Multiplier has a -20dB roll off ~ 100khz and then has +20dB gain after. Are there ways to solve this to make the frequency behavior the same as a single pole RC filter? What is contributing to this effect? it probably is the FETs/BJT parasitics coming into picture at higher frequencies but I'm not sure . Playing around with Cgs capacitance and adding Caps between Emitter/Source to GND helps but i was wondering if there is a better method
I guess you added pretty small values of resistor and capacitor in the base of the transistor if the roll of is at 100KHz, the nice thing about this circuit is that it can achieve very low filtration frequencies. Anyway, coming back to your question, I did see some schematics where a ferrite bead was added in the base of the transistor to improve high frequency behavior, but anyway, at frequencies >1MHz you can add an extra ferrite bead + capacitor filter to improve filtration proprieties. With any transistor, because of the E-C capacitance at some point, high frequency noise will pass trough.
What happens if I put one of these in the output of a buck converter and wire the feedback *after* the filter? Will it become a super clean buck converter or will it oscilate itself to hell?
I don't get it, at 13:52 you add a resistor to ground to lower voltage at base, thus lowering Vbe? How does that help?? I guess I also don't get why voltage output at C_multi drops immediately after input voltage becomes lower than at C_multi? Is this because Q2 somehow gets into cut-off region or something?
Hello AlanS. Vbe is a constant (roughly) around 0.6V, and that does not change. Now you can estimate the votlage on the emitter based on the voltage in the base by saying that Ve=Vb-Vbe; Ve in this case is the output of the circuit. By adding the extra resistor I push Vb lower, and that drags Ve also lower.
Regarding the second point, the voltage at the output is coming from the input, so there is no way that the output will be larger than the input. Ve=Vc-Vce (Ve-output Vc-input Vce-voltage drop on transistor, a positive value usually >0.1V)
Let me know if its a bit more clear or if you have any more questions!
When using a MOSFET in a capacitance multiplier, is the voltage output limited to less than max gate threshold voltage? For example I have a 70 volt transformer, rectified to 109VDC, and I want to filter the ripple, and use capacitance multiplier to amplify the current to 10 amps. 109VDC on the gate of a mosfet will destroy it, right? In this case, do I have to use a high voltage BJT?
Hello Josh! Well the Mosfet is in danger only if the Gate-Source voltage (Vgs) exceeds the rated value (usually its ~+/-15V). Just to be on the safe side you can add a zener diode between the G and S.
The gate threshold voltage (Vgs_th)(usually 1 to 5V) is the voltage that is needed to get the transistor to conduct.
Another parameter to consider is the Drain Source voltage (Vgss) - this tells you what is the maximum voltage difference you can apply between the transistor terminals. When the power supply is off, there is no voltage on the output. When you turn it on, in the first few ms there will voltage on the input and nothing on the output so the drain-source will see the full input voltage. I do recommend that you don't chose a transistor on the limit with this parameter - for the 109V you are interested in use a 150 or 200V transistor.
Point is that it does not matter what voltage you are trying to stabilize. It will work with both a BJT and a FET, as long as the transistor is rated for that voltage. But since you are interested in 10A, FET might be the better option, or for the BJT version a darlington arrangement.
Anyway you could try the circuit out in the simulator first just to be sure.
what does it mean for the filter to be second order filter rather than first order filter and why does it give a higher frequency ramp in the frequency response?
The filter order is related to the number of reactive components (an RC filter is order 1 because there is only 1 reactive component - the C); the attenuation of a filter is aprx filter order * 20dB/decade; a second order filter will be 40dB/decade
And you need to pay attention to discharging the capacitor voltage when the power supply is turned off, as the stored energy will discharge through the B-C junction causing a transistor failure if you are not careful!
I am driving an N-CH MOSFET with pwm signal (20-30 KHz). The voltage at the drain is the 60Hz signal from the wall outlet. Every time the gate turns on and off, there is a ripple at the Mosfet's source. The 7810 voltage regulator does not filter this ripple out. Can I use this technique to smooth out the ripple? I don't want to use an inductor at this stage of the circuit. Thank you.
What frequency is the ripple at? Another way that is used in power supplies is a "snubber". A series RC circuit between the noisy point and ground chosen so that is absorbs signal above a particular frequency (say 2R2 and 1nF). This should have no effect on the 50Hz but absorb high frequency stuff.
@@FesZElectronics Thank you for replying. I finally fixed it by placing a 0.1uF between the drain and the source of the mosfet. And another one between source and ground. The frequency of the ripple was so high that I could only see the waves when I zoomed down to 20ns on the oscilloscope. But the voltage was over one volt.
Big thumbs up (Y)
Please tell the name of the 1966 electronic book having Radio related circuit's
I added it in the video description now, hope that helps
@@FesZElectronics Sir thanks for the reply , sir the book is in different language , there is no English version of it
@@FesZElectronics Sir please suggest some good books , if u can explaining basic amplifiers , to advance amplifiers
@@FesZElectronicsAre you actually Hungarian?
what does it mean for the inductor to saturate and why is that important?
The inductance of an inductor is not a constant; when there is a magnetic core, the inductance is proportional to the magnetic permeability of the core; The core permeability is magnetic flux dependent - stronger flux reduces the permeability; at high inductor currents, the core permeability drops and so does the inductors inductance
Nice video! thanks a lot. Hopefully you can make a video about pi filters using Chokes vs pi Filters using Ferrite beads. :D
....and pi filter with plain inductor :P
Thank you for the suggestion! Before I get to do that though, if your main interest is in using the pi filter for switching power supplies, I highly recommend - www.analog.com/media/en/technical-documentation/application-notes/AN-1368.pdf
If nothing else, this might clarify some topics regarding the ferrite bead.
@@FesZElectronics I will take a look. I am more interesting pi filters at transmission lines, I have seen sometimes chokes and others beads. I will take a look in the AN. Thx
It's great. Would you mind to record something for real begginers? Something like analysis of common emiter amp? How dc and ac circuit flow, why do behave this way, etc? Greetings from Poland
Look at this...
ua-cam.com/video/_ikRwYd4ytw/v-deo.html
take a look at gyrator circuits!
What is the used simulation software?
In all my videos I focus on using LTspice - a free circuit simulation SW.
Nice video again!
But why not using a RCRC filter to this type? That way you even filter a lot more!
I actually find it strange that not a lot of people seem to use a rcrc filter with these kind of circuits.
Another useful addition is to mention that for quite some circuits (mostly audio), noise and ripple is far more important than perfect voltage regulation. So in this case a simple zener will do the job.
In the end you'll have a circuit that has much better PSRR than a standard 78xx regulator (and probably also better than a lm317).
I like your videos, they have a good pace and are well put together 👍🏻
Well I covered that(RCRC) in the beginning, I just refereed to it as a "second order RC filter", and afterwards I did not include it in the schematic to keep things simple. I fully agree, its more efficient than the simple RC filter and depending on the frequency you are trying to filter and where the corner frequency is, it can be even better to have 2x(R/2+C/2) rather that R+C so the size of the 2 options is comparable.