SMPS Isolated power stage driver using a pulse transformer
HTML-код
- Опубликовано: 8 сен 2024
- #70 In this electronics tutorial I look at how power stages, either simple or with multiple transistors can be driven using a pulse transformer. This is a technique usually applied in isolated switch mode power supplies (smps) but it can be used in any circumstance depending on the needs.
I first look at some general operating principles behind the transformer and then try things out on a practical circuit.
Links to documents used in the video:
www.ti.com/lit/...
www.power.com/...
i.stack.imgur....
Building the Constant Current Load: • DIY - Constant current...
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 you can at: / feszelectronics
There are less than 5 channels I watch and enjoy on RUclips, and this is one of them. I have over 700 subscribed channels.
Interesting, you don't see pulse transformers that often anymore. If I am not mistaken, used a lot in the 80's/90's.
In fact, I don't know any new design with one in it to be very honest. Basically you can already provide enough isolation with just the output transformer as well as the optocoupler.
A lot of modern gate drivers are even isolated and can handle gate side of 600V and up!
Nice video as usual!
Hello @Pier Muijs ! I used to see these quite commonly in PC supplies as long as the power transistors where bipolar and there was a lot of current needed to drive them; what I've seen lately is that the IC's have the drivers internally and are capable of much higher voltage ratings than in the past.
I think this method might still be used in high power circuits like welding machines to drive IGBT's.
Anyway, for hobby purposes its quite simple to apply and drive half bridge or even full bridge circuits so it could turn out usefull.
Hi,
Here are some interrogations after viewing your video. I do not know if you are wrong, but i'm looking for knowing what is right ! :) And before doing comments, once again, I thank you for your videos ! It is the best channel video about electronics that I ever seen.
I'm surprised to see that you have to reduce the magnetizing inductance for having a rectangular waveform at the output... When you write the differential equation of an RL circuit for having a rectangular waveform, it shows that the magnetizing inductance have to be high and the resistor to be low, if I did not do mistakes. Nevertheless it may come from the fact you have a coupling factor of 0.99, so you have some leakage inductance in your transformer. At the beginning you have 500uH so a leakage inductance of about 10 uH (Lleak = L*(1-k²) and then you have an inductance of 50 uH so you have a leakage inductance of about 1uH . Leakage inductance and intrawinding capacitance have an impact on how the waveform rises and falls (Christopher Basso and M. Dixon have written documentation about leakage inductance in transformer if you are interested in) . I will try to do some simulations to see what is really true.
I did not want to send the message ... But I wanted to go to the line so I press enter x) ... and my comment was posted... "Flute". Tell me if you know the answer :) I m really insterested in ! I will continue to search from my side !
Hello @Denis Leo ! You are right, by setting an ideal coupling factor of 1 the output waveform looks much better. I set the value to 0.99 to illustrate a more real life situation - its easier to reduce inductance than to make the real transformer closer to ideal. I wasn't really aware of this, I just set the coupling to 0.99 out of habit - this tends to make simulations more realistic in general. Also in these simulations I did not add any parasitic capacitance.
Please share the link on transformers! I am quite curious.
@@FesZElectronics It is a real pleasure to share you documents ! (An eternety later I founded the links !) Here are the links : cbasso.pagesperso-orange.fr/Downloads/Papers/Flyback%20Converter%20with%20Leakage.pdf
Again I press enter .... The second link : www.smps.us/slup199.pdf . If you have any question. Do no hesitate It would be a pleasure to try to answer it !
Thank you for the links! Just for info, you can edit your own comments at a later time; so its not over once you posted it - this should be valid in general on any youtube channel. There is an "action menu" symbol with 3 vertical dots.
love your videos keep it up..
This video is very good for my study. Thanks!
Love your videos. BTW, your sharp looks does make sure that the audience pay attention to the videos.
Ghay
Hello! Good video! I thought about how to build a simple HI-side MOSFET driver from pulse transformer - but when a pulse duty is quite high - voltage on MOSFET becomes very high, and vice versa.
Like your work and wondering if you are available via Upwork or Freelancer to do some freelance work reviewing a SMPS design?
Nice explanation 👌
DC restorer on secondary to remedy the duty ratio/voltage issue.
The hand-on-chin 'thinker' meme is a joke now, but you wear it well because you clearly are a thinker!
Good review but I don’t recommend the single pulse to generator two through the transformer as you said to much issues I recommend two signal to get the same at the secondary side and yes use a pnp for fat switching off
Why don't we see inductive spiking in from coil L1 when voltage is reaching zero?
Now i know how to make a h-bridge class d amp
Can you pleases explain how i would design a audio or measurement transformer for analog signals with low distortion.
How gets the Initial Power into the Secondary side in the first time? Is it like a self oscilating?
can you make a video about ratio detector and simulation?
Do you believe in Cold electricity, the stuff Peter lindeman cites in that video from the 80 s ?
@FesZ Electronics i need sugestion i am working on building a synchronous buck converter that works using tl494 it only have one output for driving high side(p-channel) mosfet and a free wheeling diode if i plan to use a mosfet for this than please guide me how can i drive the n channel (low side ) FET so that it also have some dead time in it also what IC did you used in this video for pratical demonstration ?
Hey FesZ electronic your content is very unique and interesting. You've inspired me to create a RUclips channel on electronics and I plan to upload cool stuff like this too. I would appreciate some support/feedback just to get a good start
I guess the most important thing is do videos about things that YOU consider interesting, don't try to copy someone else. And of course have fun while making them, if you are enjoying yourself, the viewers will also enjoy the content.
Good luck!
@@FesZElectronics do things that you like that will refresh you not tired
Hi. Might be related ... or unrelated questions. What is keeping the secondary (-) node, in an isolated power supply, from floating to ... infinity, or at least until the internal components break down? I understand that it's not uncommon for the secondary (-) be quite a bit higher above chassis/earth ground potential ... but what keeps it from going into hundreds of volts?
In theory I guess nothing. Depending on the transformer and construction you can have guaranteed isolation voltages of ~1.5kV; Also the various safety capacitors, like Y type will be rated to multiple hundred volts. At some point though if a large voltage difference appears you will have some breakdown in components; but its not uncommon to have a capacitor in parallel with a large resistor (400k-1meg) between ground and chassis. This resistor will remove any voltage build up.
18:16 Re interference with radio broadcasters: If the entire device is shielded, then the individual inductors don't have to be shielded, correct? In that case, what about interference with adjacent circuitry within the same device?
This again depends on the application - shielding is usually used for immunity in rare cases - ex. sensitive radio reception equipment. You can also resolve issues by keeping physicals distance between the noisy and the sensitive part of a circuit. If an inductor is causing issues, you could try a more compact core (ex IHLP from Vishay) or one that has a built in shield (www.vishay.com/docs/48365/_did-you-know_ihle_ms7521.pdf) . If non of these work, then I would look into a dedicated shield.
I would recommend you take apart a similar product already on the market (or just google a teardown of it) and take particular notice of this - is there any shielding, or is there a special protected area. A shield is something that does stick out and is usually visible. If the product is similar to yours then it ran into the same issues.
Hi Mr. FesZ, I have couple of question :
1.) by providing isolation to control IC by pulse xformer, then how the control IC is powered at the first time it powered? since the auxiliary windings doesn't have a voltage at that initial time
2.) do we need more winding to drive the high side mosfet? since we know that the voltage at the gate should be Vth+Main DC Voltage
Thank you
Hello Garin,
Hope this helps:
1) There are multiple ways to power the controll IC; one common method is to supply it both from the input high voltage trough a linear regulator and secondly from the aux winding when a voltage is present. Since the linear regulator is used for a very short time, the high inefficiency is not an issue.
2) No need for more turns; the high side mosfet needs a high gate voltage only in reference to GND, but the secondary of the gate transformer (the winding that drives the high side) is not connected to GND, only to the G and S of the transistor.
@@FesZElectronics About answer number 2, thank you, it help me so much!
About answer number 1, I use aux winding to powers the IC. But the situation is when first time power supply is pluged in, the aux winding has no voltage since the primary has no voltage too, and the primary has no voltage because of at that time the IC is not yet powered. The IC is not powered because of aux has no voltage. So who does the job to initiates the oscillating voltage at the primary side?
I love your videoe a lot. for this SMPS, the switch transistors I was trying to use Si4800DY, fast enough. the low side gate drive signal is good, but high side's gate drive signal is very bad. could you share the switch transistors spice model your used please ? many thanks.
I used the following model: www.onsemi.com/pub/Collateral/NTD18N06L.REV0.LIB
Usually most serious manufacturers provide a simulation model for their transistors, at least for the newer ones..
Hello, very interesting video, thanks ! i have a question if you don't mind: can i use a SMPS Pulse Transformer to convert 220V ac to 12Vac at 50Hz ? thanks
I'm not sure ... with a fixed high frequency at 50%, and without capacitors on the input, the circuit will modulate (amplitude modulation) at 100hz (after the bridge rectifier), filtering the output will give you 12v 100hz, similar to rectified ac voltage.
it's only my guess, I'm not sure if what I told is what will happen...
(I think with two transformers and each supplied with only one side of the mains (single diode rectifier) and joining the two (filtered) outputs may be possible to get what you want)
only thoughts...
(sorry for my bad english)
@@VictorC173 Thank you very much for the answer.
The short answer is no. SMPS transformers are designed to work at high frequency typically 50khz, 100khz etc. Mains transformer operates at 50/60 Hz. That is why SMPS transfomers are small compared to mains transformer. Connecting SMPS transformers directly to mains will blow up.
18:16 How to prevent antenna?
First thing to keep in mind is the circuit size - make it small - have the isolation transformer as close to the power transistors as possible. If pricing allows, add a shield. Working at lower switching frequencies will also reduce emissions, compared to working at higher frequencies.
@@FesZElectronics Fantastic info! What's a safe frequency? Would emissions under 1 Mhz be safe without shielding?
Well it depends on the standards you are trying to comply with. Around the 1MHz point the most sensitive limit is related to the Medium-Wave AM radio band 520-1700KHz or so but for certain applications you are not measuring Radiated emissions below 30MHz. Its common to use either ~450KHz or ~2.1MHz as center switching frequency if the AM band is an issue. 450KHz has the disadvantage that the first 2 harmonics are still in the AM band and you need larger components (compared to 2MHz) and for the 2MHz the disadvantage involves increased switching losses and more high frequency noise (from ringing).
If shielding is necessary or not, relates to the exact converter and emission specs.
@@FesZElectronics Superb detail. Therefor, would somewhere around 250 kHz avoid all these issues?
I have read that the reduced inductor size from increasing frequency offers diminishing returns as frequency increases. And that around 250 kHz is the inflection point. So for that reason, plus RF, could 250 kHz be a good target?
"At some point, an additional increase in switching frequency does not significantly decrease the inductor volume. As the inductor becomes smaller, the packaging and solder pads take up an increasing percentage of overall volume, thereby diminishing the reduction in inductor volume with increasing frequency."
www.ti.com/lit/an/slyt117/slyt117.pdf
"Due to diminishing
returns of reduced inductor size at frequencies greater than 250 kHz..."
scholarworks.uark.edu/cgi/viewcontent.cgi?article=1043&context=eleguht
@@FesZElectronics "for certain applications you are not measuring Radiated emissions below 30MHz"
Which applications? Are you referring to FCC certification?
what do you used software name?
For circuit simulation - LTspice. It's a free software; I have quite a few videos on how to use it, if you are interested.
@@FesZElectronics thank you so much. i will see it.
Hello, I watched your video, and I think you are not driving this circuit properly. This transformer shoud be driven with symmetrical push-pull signal, but you are using non symmetrical signal. Something like this ruclips.net/video/pffOJdCQ7kw/видео.html
Well, "properly" is a relative term. Although perfect symmetry and an H-bridge driver are the ideal use case I wanted to show that you can get away with simpler circuits (using a single ended driver and capacitor) and also other than 50% duty and the concept of the isolation transformer will still work. Depending on the desired circuit complexity and efficiency you can chose one method or the other.
Ads on this video were terrible!