Active rectifiers (2/2)
Вставка
- Опубліковано 22 лип 2024
- #158 In this video I continue working on active diodes by performing some tests. I start of with the regular bridges built with PN diodes and then Schottky diodes and then move on to a couple active bridges - first a simple one, with both diodes and transistors and then a full on active bridge.
Related content:
Active rectifiers part1: • Active rectifiers (1/2)
Diy current Probe: • DIY Current Probe - Re...
Active diode schematics:
drive.google.com/file/d/1cD5r...
drive.google.com/file/d/1-rjl...
Used datasheets:
www.onsemi.com/pdf/datasheet/...
www.vishay.com/docs/88712/s2a...
www.ti.com/lit/ds/symlink/tl0...
Special Thanks to all my supporters on Patreon! Especially Ralf B., Paul Pr. Richard, Jonathan Alvarado and Tosta!
If you liked this video be sure to check out my other videos and you can also subscribe to be up to date with all the new ones!
If you want to support the creation of more and better videos please consider checking out: / feszelectronics - Наука та технологія
So comprehensive yet intuitive!
I imagine the experiments on real boards are time consuming, but totally worth it.
Thanks a lot for your lectures, you are one of the best teachers on UA-cam for advanced topics. 👏👏👏
Fun fact: you can use a schottky diode as a temperature sensor/overtemperature protection. When I upgraded my bug zapper, I used a mosfet with a schottky diode on the same chip. A NPN shuts the gate off if enough leakage current is passed to it through the schottky. As the gate is fed by analogue means, this reduces duty cycle gradually as the chip gets hotter.
Love your channel. Lot of practical stuff.
Your being very thorough has generated another subscription. Nice work.
My favorite electronics channel has a new upload! Thank you for your awesome content.
Again, it's so helpful that you go into a lot of practical detail! I've been playing with ideal diode ICs that control a MOSFET lately, as an efficient anti-backfeed measure for battery charging. Their internal circuits look pretty similar to this; an op-amp using negative feedback to ensure a small positive voltage drop across drain-source.
I'm happy you enjoyed it! Normally the commercial products will not have the issues you get with a discrete implementation. Also, some ideal diode controllers come with a built in charge pump to be able to drive high side Nchannel transistors.
Don't practically all mosfets have reverse diode that will backfeed to the input if the load is capacitive?
My deep gratitude and appreciation for your excellent electronics lessons. Greetings.
Excellent lecture and very practical demo. thank you .
The reverse recovery time for the Schottky diode is very fast for a line frequency bridge driven by a transformer front end. A normal designed minimum PIV for a bridge diode is 2x the peak voltage at high line. So, you would measure the reverse current at 50%, it should never reach 100%, although the characteristic is fairly constant except at 100%. The temperature rise in a good thermal design will perhaps at most have a 20 degree C rise, although with an excellent heat path that includes a large SMT diode package like a D2PAK with thermal vias and heatsink, and the fact that the current is not constant and the bridge legs alternate (with the layout being interleaved) a heat rise of 10 degrees C is not uncommon, and heat sinking is usually needed for other components like a linear regulator (which may have multiple pass transistors) and also cooling for SMT capacitors. So, when the design reaches a certain power dissipation the design concept actually starts with where is the heat generated how to move it to ambient. This then gives direction on if the heat dissipation in the bridge is a concern, which it usually is not for up to about 50 Watts output power with a transformer front end. Also, the transformer is operating in a nonlinear fashion it is only supplying current when the bridge is conducting, which tends to distort the input sine wave line waveform.
The challenge with an active bridge with a transformer front end is what happens during turn on at high line or low line. The waveforms for current and voltage are very dynamic due to the input power switch (which will bounce) and the input line filter (usually needed to pass conducted EMC and will add a little inductance to the line) the fact that the filter capacitor(s) are discharged and you have transformer magnetization current (and other transformer effects), the characteristic of the regulator upon start up and the load characteristics on start up. Here a simple robust bridge is really a blessing, especially if you use large Schottky diodes with a high surge rating. The use of a Schottky bridge I find very appealing, however it is many times hard to justify the cost unless the equipment is critical or apart of instrumentation which needs to be reliable. So, you will many times see a basic diode bridge specified even if it generates more heat.
Added to the mix we now have SiC diodes which share current. Which is not immediately viewed as beneficial in bridge design unless you need to spread the heat out across the PCB. You could perhaps interleave three bridges across the board which would have a very large thermal footprint of relatively low temperature, although at a very high cost. However, this may be beneficial in an off-line front end where active bridges are not really viable due to high voltage requirements, and the SiC recovery time (almost zero) and leakage current are very small. SiC diodes PIV options are 650 and 1200 volts, which makes them good off-line candidates if the design can justify the cost.
As you can see much of the design is all about heat when dealing with power supply front ends. Where it is being generated and how it is moved. If you went to a power supply conference where you could see the inside of the power supplies while they are operating you could learn a lot about the designs just by using a thermal camera as your only measurement.
The challenge in measuring power factor is in having a controllable spectrally pure AC power source, like a programmable AC power supply, which is usually very expensive. It does however provide you with a very pure sinewave (much more than the facility AC line input) and as such provides a very good PF measurement.
It is important to mention to use not only a comparator with low offset voltage but also with a rail to rail input. Using a dual comparator you can control both upper rail MOSFETs to avoid the fifth MOSFET. The gate capacity and gate charge of the MOSFETs should be not too high in order to make them fast enough, especially for frequencies considerably higher than 50 Hz. It is a trade-off with the pull down resistor at the gates.
The higher the voltages the lower is the contribution of electronic diodes to efficiency. If you rectify 100 V a full wave rectifier has allready an efficieny of about 97%. The gain of efficiency due to electronic rectification in this case is questionably because also the channel on-resistance of higher voltage MOSFETs is higher.
Most interesting and well researched, thank you.
Another fantastic video, thank you! :)
Very helpful!
you can use the REF button on the oscilloscope to mark the reference waveform to compare with. It is the button next to MATH button.
1st time viewer... A "bit" over my head but I have a family sized bottle of Excedrin so let's do this! I'l keep checking back in. Cheers from So. CA. USA, 3rd House On the Right.
Fantastic analysis and demo!
I’m wondering how your proposed active bridge implementation compares to Jung’s super regulator design? Cheers!
Interesting video.
The TL061 is fairly old, slow (1MHz) and has poor Vos (6mV) compared to modern parts.
A good reminder to consider the effect of inductance when circuits turn on and off, including cable inductance.
You are right, the TL061 isn't the newest part, but I had it around, and the problems it has are interesting to highlight various limitations to look out for :D also it offers the feature that the inputs can go up to the positive rail, which was important for this application.
Tsv7722 is part from ST , 22Mhz opamps.
Great video, I learned alot thanks
Nice video, keep it up!
Does power factor correction necessarily cause losses because of the indance?
I future video about PFC topologies would be very nice
Edit: oh you already have a video on that
👍👍
my lawnmower has 1 diode
is it time for an upgrade to a recitifer so I can add more lights
I've got question/challege for circuit of choice, want to draw as much power as possible out of bicycle dynamo 6v/3w AC
How would you approach that? Is there useful IC that would work well?
I have 2 ideas, either yolo hand wound ferrite ring transformer to ~22V, where diode rectification is not as punishing, then use out of the box step down board
Or design AC step up boost circuit
In perfect world there would be IC with step up/step down, with high efficiency boosting or stepping down voltage to required level, but I doubt there is one for so low voltage applications, final voltages I aim for is 19/5/4.2/1.41V
Excellent!
We should the call active rec. a synchronous rectifier. It's goal is less loss via the low resistance of the FET switching and logic driven.
thanks for easy and lovely learing
Very interesting topic and well presented. As someone previously unaware of active rectification, and never having needed to be concerned about the voltage drop when using regular diodes, I am left wondering in what applications one would seek to use either Schottky diodes or make the leap to active rectification.
Usually, if the voltages are relatively low (
Loved this, but I never get any sleep after your videos. So many questions. Anyway I have a high frequency application (200 kHz, 12v input, 10 amps output), I surprisingly got some amazing efficiencies (at least on LTspice) with careful selection of shottky diodes, but I am keen on trying to get an active rectifier design to work. I have looked, but all silicon mosfets I have tried fairly consistently seem to crap out (large current spikes appearing, ruining efficiency) at somewhere in the high hundreds to low thousands of hertz, even when I try practically every mosfet there is in the hybrid circuit. I am assuming this is because of slow on and off times, or high gate charges, so I'm going to try simulating GaNs tonight to see what happens, using a high resistance voltage divider to feed the gates, hopefully without too much of a power hit.
Sorry if this is so elementary, (I'm not that bright) but I noticed your statement "These can block negative voltages but will not block reverse current" at around 9:40. Is this statement related to why we need to use the P channel mosfets for the high side? My question about the full bridge circuit: is there a way to use 4 N-channel mosfets, without using a p-channel device? (I cant get models for P-channel GaNs).
Did you get your answers I am interesting to know about it because I want to do a rectifier at 10 MHZ
Hi FesZ, when I simulated the final design, in ltspice, I found that the AC input could not be below 14V-15V without significant distortion and higher voltage drops compared to the input voltage, on the output wave forms. Why is this?
It might be transistor dependent - you need to have enough voltage to get the mosfet conducting - the voltage needs to be above the Vgsth to actually turn the transistor on.
@@FesZElectronics Thanks I think the mosfet with the OP amp was not turning on all the way, it's Vgs was way too low.
@@cat-ie6yp I guess another thing to look out for is the voltage range both input and output of the op-amp - I specifically used the TL062 since inputs inputs can work with up to V+; not all opamps can do this; similarly with the output, can it actually turn on and off the transistor - is there a sufficient voltage range?
@@FesZElectronics Yes that's a good suggestion, additionally it may be easier to use an Over-the-top comparator to do the trick, but may be more expensive than a general purpose op-amp
They actually make special active diode controllers that fix a lot of these issues automatically.
Indeed, you can skip over all of these problems with a proper controller, but usually those components are expensive so they are not used unless in high quality equipment.
@@FesZElectronics Yes but the controllers only cost ~$6 and make all the problems go away. The controllers for the secondary side cost only ~50c.
Interesting. Also regarding that inductor would a choke work better then 2 inductors? maybe a zener across the output could have kept the voltage more stable but that is some more leakage.
Regarding the op amp wouldn't a chopper op amp be better in this type of application? Since it is self tuning so no manual adjustment is needed.
A and regarding that big inductor, would making it resonate with the cap at 50Hz make the circuit more efficient? (might be reinventing the wheel here).
6:25 That's unexpected! Does it apply to any other types of components?
It is true for bipolar transistors: when they are heated up, collector-emitter saturation voltage drops to some extent. And it's opposite for MOSFETs: their on-resistance increases when heated. But that is actually very good effect (and bad for bipolars) as several MOSFETs can be connected in parallel and get automatically balanced: if one had lower on-resistance and began to draw most current through itself, it heats more and increases its resistance. So in the end current is distributed evenly which is good for heat management. But bipolars can't be connected in parallel as well: when one of them gets hotter, it draws more current through itself, so it gets even hotter and probably burns up. What's even more "funny", simple bipolar transistor can be thought of as a lot of little transistors connected in parallel and the same mechanism applies here in certain conditions. That's how secondary breakdown is formed. That's region of high enough voltages and rather small currents. As MOSFETS are self-balancing in this way, they don't have secondary breakdown at all!
You'll be surprised at what changes with temperature, nearly every part will shift somewhat...
I planned on playing with inductors today. Maybe 25Tube knew before me. 🤔
It's shocking that high power and voltage active rectifiers still aren't available.
High voltage active rectifiers have no sense, due to low profit.
About high current, You are writing this message from device, having at least 5pcs of active rectifiers inside.
;-)
Hey Sussy cat here , sorry forget to comment , i was studing for exams SO :D