Good evening from Greece, really all the videos I have watched are amazing, your work is great. you are a master of presentation and I am sure you are also a good person, good luck in whatever you do in your life.
I would expect the „integrated high ohmic resistor“ is a current source as a resistor is difficult to implement in silicon. One additional factor in the regulation error of this topology is that the ohmic losses in the secondary winding cannot be eliminated correctly. So the output voltage will be „softer“
Difficullt or not, this is integrated in quite a lot of SMPS controllers Here is an example that has an integrated 'start up cell' which acts basically like an internal high ohm resistor to generated the initial Vcc. Of course like you mention the actual implementation could be done in some other silicon manner pdf1.alldatasheet.com/datasheet-pdf/view/312676/INFINEON/3BR0665JZ.html
You are right on both counts. Regulation can be acceptable, especially for something like this circuit where the required precision regulation is done by the down-stream battery charging circuit. Some of the control ICs have moderately elaborate circuits to sample the feedback voltage at the optimum time. If the converter is operating in discontinuous current mode (in a flyback what you're really talking about with DCM/CCM is whether or not the inductor is fully discharged each cycle since current flow in both primary and secondary is always discontinuous). If you can sample the feedback winding voltage just as the inductor is nearing discharge, the current in the secondary is minimal so you get a pretty accurate reflection of the voltage on the output filter capacitor. I haven't looked at this stuff in years, but when such ICs were starting to show up the manufacturers were a bit secretive about just how they did this.
Hi Rick. Very well explained. Just you should know that you have explained that case in video series for SMPSUs. I don't remember exactly which one, but it was not new to me. Still it might be also useful as a separate short video instead of the other long one where you cover several cases. God job.
Yeah I have explained how the AUX winding powers the PWM controller in at least a couple other videos. Never before though have I shown on video how a PSU can also regulate it's output voltage by monitoring the voltage on the AUX winding, therefore using considerably less components.
Very good video... The interesting thing is that lightning symbol under every 2nd port might was supposed to present "fast charging", obviously not really but to deceive whoever buys this
Heya, make sence as there is no optoisolator it had to be the transformer with a extra winding still learning from you but understanding more and more thanks for your teaching
Ah, the venerable SMPS Power Supply - found right next to LCD displays, PIN numbers and DMM multimeters. 🤪 Anyway, minor niggles aside, great video once again. 😀
Iv seen this a feed back type a few times in cheap chargers on "diode gone wild's" channel.. he has a load of video where he opens these chargers. Iv learnt a lot off that channel..
If you wanted a bit better regulation, on the top side feedback resistor could you use lets say a 10V zener diode? that would make the feedback voltage drop at a 1:1 ratio instead of a 3:1 ratio
The regular smps with opto has a secondary winding for powering the chip. The regulation on these cheap ones is outside the spec for USB power (4.8-5.2 iirc). They will easily drop to 4.7V before they react to compensate the load. So, no good
Since supplies like this are almost universally used for battery charging, being outside of USB spec really isn't too important. The charger circuit is what does the precision control. It isn't a matter of "before they react." The issue is simply that the accuracy of the control isn't great because of imperfect tracking of the output and the feedback winding. That said, dynamic performance some switchers _is_ pretty dismal.
An advantage is cost saving and easy design of your psu If you control the primary side it´s called primary side feedback. Optocoupler with TL431 gives you a higher precision voltage regulation on the secondary side because you monitor the secondary voltage directly but also adds more parts and complexity to the circuit. The disadvantage of primary side feedback is of course lack of precision of voltage regulation because you only control the transformer voltage ratio not the real secondary voltage.
Out of curiosity, when working, what is the frequency of the feedback circuit opto-isolator? For instance, how many times is it turning on and off the opto to maintain the voltage output?
The optocoupler operates in "linear" mode thought it isn't necessary very linear. The description of how a TL431 and optocoupler in this video, and a few others I've seen, is not accurate. The whole TL431/optocoupler circuit is linear. It does not switch. The TL431 is actually used as a combination of a reference voltage and a "voltage error amplifier." In the circuit as shown at about 5:25 in this video, if the output voltage rises above the setpoint the anode current of the TL431 increases, as does that in the IRED of the optocoupler. This increased current causes corresponding increase in current in the transistor of the optocoupler. This current is used to "turn down" the power delivery circuit, typically by reducing the duty cycle. If the output voltage decreases, the opposite effect occurs. What is "always" ignored in descriptions of these circuits is the very important "frequency compensation." Effectively the error amplifier is run "open loop" at DC - its gain is the inherent gain of the amplifier used at DC and is typically very high - in the range of thousands to tens of thousands. That makes the regulation at DC very good. But you can't do that as frequency rises. What you need to do is make the gain "roll off" with rising frequency. Ultimately you want the gain to pass through "unity" (one) on its way done with adequate "phase margin." If you don't do that the whole thing can become a power oscilliator - you get a phase shift around the loop of 360 degrees with a gain of one. If you make sure the gain goes to less than one with a phase shift of less than 360 degrees, maybe something like 315 degrees so you have a phase margin of 45 degrees, the circuit will be stable and still able to respond quickly to changes in load current or line voltage. To make this happen you use a capacitor for feedback in the error amplifier. As the frequency rises the impedance of the cap drops and the gain of the amp is reduced. It can be more complex than this, depending on the exact circuit topology, but the general objectives, high DC gain and preserving phase margin at unity gain, are fundamental. Anyway, this means that the feedback path does not switch in normal operation. Ordinary optocouplers aren't highly linear. The "current transfer ratio" may vary quite a bit over the working range and with temperature. You do have to consider that in designing, but the non-linearity is "cancelled" by the error amplifier.
@@d614gakadoug9 Thanks for the explanation. I was wondering why the components didn't seem to be special due to what I perceived a high frequency they would have to operate at, given the original explanation of operation.
In that case why does it say 5V 3.5A on the side of the PSU, which rates the PSU at 18W max? Where are you seeing the other 30W comes from? And I don't believe the 5V 3.5A is correct either. The QC5783 datasheet (and this is the chip that drives the PSU) says it is : *Power management IC 5V 2.1A 12V 1A Mobile phone charger adapter IC PWM power converter* www.aliexpress.com/i/1005003681351155.html So that's 12W at 12V and 10.5W at 5V max that this PSU can supply actually, yeah? Hmmm in that case then, and I do rather believe the datasheet is correct, where did the 5V 3.5A figure come from?
I just saying how i use to read when the Amp times voltage doesnt match up with the wattage.. I have seen it a couple of times on both usb chargers and power supply's in general but In this case there is something wery wrong, i can see that.. Great video by the way
@@MrDacorp For sure there is something very wrong with the rating of this PSU. When we first looked at it on the live stream(@theelectronicschannel 17:00 London time every second Sunday, we connected a TS80 soldering iron to it, and it was drawing a max of 10W when we tried soldering a heavy groundplane
Yeah the reference would be 2.5V, because the TL431 is a 2.5V 'zener' but the output voltage of the PSU would be 5V due to the ratio of the two resistors. I'm sorry if I didn't make that clear, it was a bad choice of words on my part.
I think I can explain that. Both secondaries are effectively in parallel with each other, though they don't electrically connect to each other. The primary is driven at high frequency by the PWM mosfet. If you put a load on any one of the secondaries, not only do you reduce the voltage on that secondary, you reduce the voltage on all of the secondaries. This is because the energy that induces voltage into the secondaries is all coming from a single source, the primary. When the PWM then drives the mosfet harder to increase the output voltage on the loaded secondary it also increases the output voltage of the other secondaries. Another way to think of it. We all know (hopefully) that the output voltage of two secondaries is proportional to the number of turns of wire in each. So in a given transformer, if one secondary has 100 turns and is generating 12V and another secondary has 200 turns, the second secondary (lol) would generate 24V yeah? The only way this formula can remain true, is if you load one secondary thus reducing the output voltage to 10V, the output on the other secondary must drop to 20V at the same time. 🙂 Does that make sense, otherwise I'll have to think of another way to explain it.
@@LearnElectronicsRepair the "other way to think about it" helped me greatly. If I reduce the voltage on one winding, all other windings must reduce voltage to keep the same ratio. Thanks Richard!
These cheap Chinese products are very cleverly designed for cost efficiency but i still never use them because I know if anything you should never cheap out on power supply that pull out such high voltages in a compact size, specially mobile chargers. But sometimes they work really good being cheap.
@aftab277 Make up your mind. Which is it "i still never use them". or. "they work really good". How can you know they are good, if you never use them ?
It is interesting but it strikes me as a dangerous way of doing it. If there is a load on the VCC from the startup capacitor leaking or a faulty controller chip then the voltage will drop. The controller will drive the transformer harder to bring it back up and in the process exceed 5V on the main output, but it will not know because it is not directly sampling it. Nobody but the cheapest Chinese manufacturers would run a power supply that didn't directly sample the output.
That won't happen, if there is excessive load or a short on the output then the AUX winding will not generate enough voltage to keep the PWM controller running so the PSU will simply shut down. Or do the chirp chirp thing as it repeatedly tries to restart and fails
The more resourcefull engineers are employed by companies who want to make compromises so their products are as simple and cheap as possible.And of questionable safety as well.these chapeo power adapters are not recommended at all,that is if you care about your phone,tablet etc.
How is it clear that whoever designed the psu doesn’t know Ohm’s Law? I’d presume that they do know it because they’d have to choose components based on calculations or it wouldn’t work properly. What I think is a possibility is that it had nothing to do with the engineer, I think it could have been their marketing department making fictitious claims to increase sales.
You got it right, then corrected yourself to be wrong for the rest of the video. Why did you not unrap the transformer to demonstrate the physical isolation of the "AUX" winding. You are unbelievable...
Very interesting. Thanks Richard.
Good evening from Greece, really all the videos I have watched are amazing, your work is great. you are a master of presentation and I am sure you are also a good person, good luck in whatever you do in your life.
Another excellent video. Thank you ☺
Very interesting. Thanks.
I would expect the „integrated high ohmic resistor“ is a current source as a resistor is difficult to implement in silicon.
One additional factor in the regulation error of this topology is that the ohmic losses in the secondary winding cannot be eliminated correctly.
So the output voltage will be „softer“
Difficullt or not, this is integrated in quite a lot of SMPS controllers
Here is an example that has an integrated 'start up cell' which acts basically like an internal high ohm resistor to generated the initial Vcc. Of course like you mention the actual implementation could be done in some other silicon manner
pdf1.alldatasheet.com/datasheet-pdf/view/312676/INFINEON/3BR0665JZ.html
You are right on both counts.
Regulation can be acceptable, especially for something like this circuit where the required precision regulation is done by the down-stream battery charging circuit.
Some of the control ICs have moderately elaborate circuits to sample the feedback voltage at the optimum time. If the converter is operating in discontinuous current mode (in a flyback what you're really talking about with DCM/CCM is whether or not the inductor is fully discharged each cycle since current flow in both primary and secondary is always discontinuous). If you can sample the feedback winding voltage just as the inductor is nearing discharge, the current in the secondary is minimal so you get a pretty accurate reflection of the voltage on the output filter capacitor. I haven't looked at this stuff in years, but when such ICs were starting to show up the manufacturers were a bit secretive about just how they did this.
Hi Rick. Very well explained. Just you should know that you have explained that case in video series for SMPSUs. I don't remember exactly which one, but it was not new to me. Still it might be also useful as a separate short video instead of the other long one where you cover several cases. God job.
Yeah I have explained how the AUX winding powers the PWM controller in at least a couple other videos. Never before though have I shown on video how a PSU can also regulate it's output voltage by monitoring the voltage on the AUX winding, therefore using considerably less components.
Very good video...
The interesting thing is that lightning symbol under every 2nd port might was supposed to present "fast charging", obviously not really but to deceive whoever buys this
Heya, make sence as there is no optoisolator it had to be the transformer with a extra winding still learning from you but understanding more and more thanks for your teaching
Thanks for another good one!
Great explanation, thank you
Ah, the venerable SMPS Power Supply - found right next to LCD displays, PIN numbers and DMM multimeters. 🤪
Anyway, minor niggles aside, great video once again. 😀
Iv seen this a feed back type a few times in cheap chargers on "diode gone wild's" channel.. he has a load of video where he opens these chargers. Iv learnt a lot off that channel..
If you wanted a bit better regulation, on the top side feedback resistor could you use lets say a 10V zener diode? that would make the feedback voltage drop at a 1:1 ratio instead of a 3:1 ratio
Well explained.
the world's best teacher thanks
well explained Richard another excellent video
Your are doing great videos I always appreciate the way you explain during your videos. I have learned a lot from you keep it up. You are awesome.
Another super video - thanks again!
The regular smps with opto has a secondary winding for powering the chip. The regulation on these cheap ones is outside the spec for USB power (4.8-5.2 iirc). They will easily drop to 4.7V before they react to compensate the load. So, no good
Since supplies like this are almost universally used for battery charging, being outside of USB spec really isn't too important. The charger circuit is what does the precision control.
It isn't a matter of "before they react." The issue is simply that the accuracy of the control isn't great because of imperfect tracking of the output and the feedback winding. That said, dynamic performance some switchers _is_ pretty dismal.
Great lecture🎉🎉
Thank you for the video.
@ 7:06 OMG it is gay a rainbow 🤣
Trans = transformer, perhaps?
@@I_Don_t_want_a_handle PMSL!
What is the advantage and disadvantage of this circuit compared to an optocoupler?
An advantage is cost saving and easy design of your psu If you control the primary side it´s called primary side feedback. Optocoupler with TL431 gives you a higher precision voltage regulation on the secondary side because you monitor the secondary voltage directly but also adds more parts and complexity to the circuit. The disadvantage of primary side feedback is of course lack of precision of voltage regulation because you only control the transformer voltage ratio not the real secondary voltage.
@@Telectronics Yeah I agree this is for cost saving
Thanks @@Telectronics
You are welcome !@@tonysheerness2427
Out of curiosity, when working, what is the frequency of the feedback circuit opto-isolator? For instance, how many times is it turning on and off the opto to maintain the voltage output?
0hz.
it's dc.
dunno if you watched previous video on tl431?
it tells the primary when output voltage is reached, it has no refresh rate.
The optocoupler operates in "linear" mode thought it isn't necessary very linear. The description of how a TL431 and optocoupler in this video, and a few others I've seen, is not accurate.
The whole TL431/optocoupler circuit is linear. It does not switch. The TL431 is actually used as a combination of a reference voltage and a "voltage error amplifier."
In the circuit as shown at about 5:25 in this video, if the output voltage rises above the setpoint the anode current of the TL431 increases, as does that in the IRED of the optocoupler. This increased current causes corresponding increase in current in the transistor of the optocoupler. This current is used to "turn down" the power delivery circuit, typically by reducing the duty cycle. If the output voltage decreases, the opposite effect occurs.
What is "always" ignored in descriptions of these circuits is the very important "frequency compensation." Effectively the error amplifier is run "open loop" at DC - its gain is the inherent gain of the amplifier used at DC and is typically very high - in the range of thousands to tens of thousands. That makes the regulation at DC very good. But you can't do that as frequency rises. What you need to do is make the gain "roll off" with rising frequency. Ultimately you want the gain to pass through "unity" (one) on its way done with adequate "phase margin." If you don't do that the whole thing can become a power oscilliator - you get a phase shift around the loop of 360 degrees with a gain of one. If you make sure the gain goes to less than one with a phase shift of less than 360 degrees, maybe something like 315 degrees so you have a phase margin of 45 degrees, the circuit will be stable and still able to respond quickly to changes in load current or line voltage. To make this happen you use a capacitor for feedback in the error amplifier. As the frequency rises the impedance of the cap drops and the gain of the amp is reduced. It can be more complex than this, depending on the exact circuit topology, but the general objectives, high DC gain and preserving phase margin at unity gain, are fundamental.
Anyway, this means that the feedback path does not switch in normal operation.
Ordinary optocouplers aren't highly linear. The "current transfer ratio" may vary quite a bit over the working range and with temperature. You do have to consider that in designing, but the non-linearity is "cancelled" by the error amplifier.
@@d614gakadoug9 Thanks for the explanation. I was wondering why the components didn't seem to be special due to what I perceived a high frequency they would have to operate at, given the original explanation of operation.
Are we using un rectified feedback?
awesome!
5V/3.5A,there is 6 ports but not 3.5A on all of them at the same time, you could pull max 48W all together... Thats how i would read that
In that case why does it say 5V 3.5A on the side of the PSU, which rates the PSU at 18W max? Where are you seeing the other 30W comes from?
And I don't believe the 5V 3.5A is correct either. The QC5783 datasheet (and this is the chip that drives the PSU) says it is :
*Power management IC 5V 2.1A 12V 1A Mobile phone charger adapter IC PWM power converter* www.aliexpress.com/i/1005003681351155.html
So that's 12W at 12V and 10.5W at 5V max that this PSU can supply actually, yeah? Hmmm in that case then, and I do rather believe the datasheet is correct, where did the 5V 3.5A figure come from?
I just saying how i use to read when the Amp times voltage doesnt match up with the wattage.. I have seen it a couple of times on both usb chargers and power supply's in general but In this case there is something wery wrong, i can see that..
Great video by the way
@@MrDacorp For sure there is something very wrong with the rating of this PSU. When we first looked at it on the live stream(@theelectronicschannel 17:00 London time every second Sunday, we connected a TS80 soldering iron to it, and it was drawing a max of 10W when we tried soldering a heavy groundplane
At 4:00 wouldn't the reference voltage be 2.5V dropped between the two 10k resistors
Yeah the reference would be 2.5V, because the TL431 is a 2.5V 'zener' but the output voltage of the PSU would be 5V due to the ratio of the two resistors. I'm sorry if I didn't make that clear, it was a bad choice of words on my part.
I don't really understand how a voltage drop on the secondary winding causes a voltage drop on the auxiliary winding
I think I can explain that. Both secondaries are effectively in parallel with each other, though they don't electrically connect to each other. The primary is driven at high frequency by the PWM mosfet. If you put a load on any one of the secondaries, not only do you reduce the voltage on that secondary, you reduce the voltage on all of the secondaries. This is because the energy that induces voltage into the secondaries is all coming from a single source, the primary. When the PWM then drives the mosfet harder to increase the output voltage on the loaded secondary it also increases the output voltage of the other secondaries.
Another way to think of it. We all know (hopefully) that the output voltage of two secondaries is proportional to the number of turns of wire in each. So in a given transformer, if one secondary has 100 turns and is generating 12V and another secondary has 200 turns, the second secondary (lol) would generate 24V yeah? The only way this formula can remain true, is if you load one secondary thus reducing the output voltage to 10V, the output on the other secondary must drop to 20V at the same time. 🙂
Does that make sense, otherwise I'll have to think of another way to explain it.
@@LearnElectronicsRepair the "other way to think about it" helped me greatly. If I reduce the voltage on one winding, all other windings must reduce voltage to keep the same ratio. Thanks Richard!
i love it
Well, it doesn't say 48 watts but 48w, model number maybe. I know it's misleading!
Yeah I guess you are right, 48W could stand for anything 😆
👌👌👍👍
These cheap Chinese products are very cleverly designed for cost efficiency but i still never use them because I know if anything you should never cheap out on power supply that pull out such high voltages in a compact size, specially mobile chargers. But sometimes they work really good being cheap.
@aftab277
Make up your mind. Which is it "i still never use them".
or.
"they work really good".
How can you know they are good, if you never use them ?
U r perfect
I have an old AT&T phone charger that stopped working and I took it apart and it had a 2032 battery in it.
It was a very confusing thing to see.
It is interesting but it strikes me as a dangerous way of doing it. If there is a load on the VCC from the startup capacitor leaking or a faulty controller chip then the voltage will drop. The controller will drive the transformer harder to bring it back up and in the process exceed 5V on the main output, but it will not know because it is not directly sampling it. Nobody but the cheapest Chinese manufacturers would run a power supply that didn't directly sample the output.
That won't happen, if there is excessive load or a short on the output then the AUX winding will not generate enough voltage to keep the PWM controller running so the PSU will simply shut down. Or do the chirp chirp thing as it repeatedly tries to restart and fails
LGBT power supply. I like that somebody has courage to make jokes about that.
LOL Not being brave, if something is funny then it's funny, simple as.
I almost spit my drink when he said that, luckily this isn't twitter xD
48W my ass. Those are chinese watts. The secondary side rectification looks like a MOSFET or synchronous rectifier chip.
It's a rainbow Charger.. that's why basic "FACTS" like ohms law is wrong on this... basic "FACTS" are their kryptonite..
(Now don't ban me TH-cam..)
Me ban anyone? Not gonna happen 😉
The more resourcefull engineers are employed by companies who want to make compromises so their products are as simple and cheap as possible.And of questionable safety as well.these chapeo power adapters are not recommended at all,that is if you care about your phone,tablet etc.
How is it clear that whoever designed the psu doesn’t know Ohm’s Law? I’d presume that they do know it because they’d have to choose components based on calculations or it wouldn’t work properly.
What I think is a possibility is that it had nothing to do with the engineer, I think it could have been their marketing department making fictitious claims to increase sales.
You got it right, then corrected yourself to be wrong for the rest of the video.
Why did you not unrap the transformer to demonstrate the physical isolation of the "AUX" winding.
You are unbelievable...
What?
LGBT power Supply 😂😂😂
It took you 12 minutes to get to the point. This is not a channel for me. Unsubscribed.
Fair enough, but often it takes context to explain something.
You should ask ugreen to sponsor the channel, they do really good product 😉