Virtual 240Vac Bipolar Active Zener to Extend the SMPS High Voltage Limit

[KerimF]

Hi,

The presented circuit (attached as PDF) will be connected in series to the protected SMPS (AC to DC Switch-Mode Power Supply). It is surprising that it will act as if it is a bipolar hi-voltage zener connected in parallel to it. It works from about 100 to 400 Vac. Below 240Vac (the adjusted hi-limit, as Vpeak=340V), the circuit (in series) looks as a short- circuit between mains supply and the SMPS.

I am afraid that unless it is possible to model the AC input of an SMPS properly, a simulator (as LTspice) won’t give the real trace of the voltage on the SMPS input (50Hz sinewave clipped at the adjusted limit voltage in each half cycle). The real trace (when mains voltage is above the voltage limit) can be seen on an oscilloscope.

I shared this topology here with the hope that someone may know a simpler way to achieve this function.

Thank you.

Kerim

Edited:
A simple overvoltage protection is presented on post #11

[coromonadalix]

you want to limit the input voltage of the smps  outside the main ac voltage 

personally i don't understand where your going,   you can use  scr controlled rectifiers or even triac controlled rectifier bridges

in no way your irfp460 circuit is bipolar, since you have a protection diode in the mosfet ??


or you want to limit the inrush current ??
https://www.radiolocman.com/shem/schematics.html?di=183767

[KerimF]

you want to limit the input voltage of the smps  outside the main ac voltage 

personally i don't understand where your going,   you can use  scr controlled rectifiers or even triac controlled rectifier bridges

in no way your irfp460 circuit is bipolar, since you have a protection diode in the mosfet ??


or you want to limit the inrush current ??
https://www.radiolocman.com/shem/schematics.html?di=183767

Sorry for not being clearer.

Let us suppose a conventional (Chinese) SMPS may be supplied, by mistake or due to an electric fault, by a relatively high voltage (say >240V up to 400V). This circuit, if connected in series, clips the AC voltage at its input (say at +/- 340V, the peak of 240Vrms). This is equivalent to a virtual bipolar hi-voltage zener connected in parallel at its input.

The added 4-diode bridge lets the MOSFET acts as a bipolar device.

A triac cannot clip the top of an AC supply voltage when it exceeds a certain limit (as the peak of 240Vrms). it can vary its RMS value only, not its peak (unless the phase delay of the trigger pulse is made greater than 90 degrees).

This circuit is not about protecting from inrush current but from overvoltage.

[Kim Christensen]

Sounds like you need fuse in series and a MOV across the AC input of the SMPS.

[DavidAlfa]

Search "ac crowbar protection circuit".

[KerimF]

Sounds like you need fuse in series and a MOV across the AC input of the SMPS.

Good solution, but the purpose of this circuit is to keep the SMPS running (its DC supply not interrupted) even if the mains voltage becomes relatively very high (up to 400Vrms).

[coromonadalix]

some smps can handle some spikes at 400v   if you have some crap    not sure   loll 

some will work from 85-100 vac up to 265 vac


you have power bars  or line filters who will kill some of theses spikes


sound like you want some line conditioner to avoid those surges, an active solution you could say

some solution
https://www.littelfuse.com/~/media/electronics/application_notes/littelfuse_high_power_semiconductor_crowbar_protector_for_ac_power_line_application_note.pdf.pdf

[KerimF]

Search "ac crowbar protection circuit".

Thank you for the hint.
But I wonder if such a circuit will keep the SMPS running and supplying its DC voltage to a controller board while the mains voltage is high (say 380Vac).

Please note that the circuit, presented earlier, is not just a protecting one but regulating as well.

[DavidAlfa]

No, it's just a protection, blowing the fuse instead burning everything.
Neither does your original circuit, trying to clamp mains can only end in fire, the power from mains is virtually unlimited, so if it supplied 400V and you tried to clamp it down to 200V, it would need hundreds of amps.

If you need it to keep operating under overvoltage condition, then properly design it so it can work at 400V.

[KerimF]

some smps can handle some spikes at 400v   if you have some crap    not sure   loll 

some will work from 85-100 vac up to 265 vac


you have power bars  or line filters who will kill some of theses spikes


sound like you want some line conditioner to avoid those surges, an active solution you could say

some solution
https://www.littelfuse.com/~/media/electronics/application_notes/littelfuse_high_power_semiconductor_crowbar_protector_for_ac_power_line_application_note.pdf.pdf

I needed to design the OP circuit because I produce mains stabilizers for houses and offices whose input could be 290V down to 110V (50Hz). Their controllers are supplied by conventional (economical) AC-DC SMPS. And I liked that the 3+3-digit LED display of their controller shows the input voltage even if it becomes as high as 400V (extreme possible limit!). I did it by regulating the peak of the AC voltage supplying the SMPS.

[KerimF]

I corrected the title of the topic.
Based on what I heard so far, I realized that the word 'protect' in the original title was misleading. Sorry for any inconvenience.
 

[KerimF]

By the way, in my low-cost mains stabilizers, I simply protect the SMPS by a simple circuit (attached as PDF). It just cuts the mains supply when the mains voltage reaches a high voltage (as 250Vac, for example). It is much like a mechanical breaker but for overvoltage.

If its 24Vdc relay is somehow sensitive (it turns on at a DC voltage around 15V or below), a 1/2W resistor (not shown on the schematic) is added in parallel to 470uF (CP1). Its value (usually between 3K3 and 10K) needs to be found empirically so that the 5K trimmer (RV1) covers the range around the desired AC voltage limit. For instance, the DC voltage on the relay coil will be around 24V @380Vrms (50Hz).

[KerimF]

Neither does your original circuit, trying to clamp mains can only end in fire, the power from mains is virtually unlimited, so if it supplied 400V and you tried to clamp it down to 200V, it would need hundreds of amps.

You are totally right in case a circuit is connected in parallel. But my original circuit is connected in series.

After all, it seems that what I did is somehow a novel topology (based on the well-known saying: Necessity is the mother of... ). I wonder if someone else needed to also design such an active regulator (made for small conventional SMPS, not heavy loads).

So, if someone will have time (and the necessity) to build and test it, he will be surprised (as it happened to me while watching its regulation on the oscilloscope screen for the first time) how a series circuit could act as a virtual zener connected in parallel which also clamps the top of the high voltage sinewave.

[soldar]

Neither does your original circuit, trying to clamp mains can only end in fire, the power from mains is virtually unlimited, so if it supplied 400V and you tried to clamp it down to 200V, it would need hundreds of amps.

You are totally right in case a circuit is connected in parallel. But my original circuit is connected in series.

Then it is not acting as a zener. You are using that word very loosely and confusingly.

[KerimF]

Neither does your original circuit, trying to clamp mains can only end in fire, the power from mains is virtually unlimited, so if it supplied 400V and you tried to clamp it down to 200V, it would need hundreds of amps.

You are totally right in case a circuit is connected in parallel. But my original circuit is connected in series.

Then it is not acting as a zener. You are using that word very loosely and confusingly.

I understand your confusion.

I usually add LTspice files to simulate the circuit I present.

Unfortunately, here there is no good model to emulate the input of a conventional SMPS. Do you think it could be done?
So, for the time being in the least, the only way to see how this real circuit (in series) acts if it were a bipolar zener connected in parallel is by building it then testing it (with an SMPS as a load in series) by using a scope to see the positive and negative flat tops of the regulated AC voltage when the high limit is exceeded. Otherwise, it is very hard for someone to get how it works.

 

[T3sl4co1l]

I don't quite understand "220V" being wired to a CD4093 logic gate. 

Better names I guess would be "220V" --> "HOT", "NTR" --> "NEUTRAL", "SMPS" --> "LOAD"; "220V" net bridged to a "VDD" or something like that, and "GND" can be a high side VEE, or floating GND or something like that, maybe it's fine to leave as "GND" as such but just as long as it's clear.  A connected drawing would be easier to make sense of than net labels, making it clear that the logic side is "hanging below" one AC line, and there is a series bridge element connecting the load.

So I guess the point is a dropper to limit load voltage?  But this is exactly inverse to a zener shunt, not equivalent to it!

What current range?  What fault current capacity?  Not much I guess given the 1N4007s, and the poor IRFP460 will expire rather quickly under any kind of startup conditions, but maybe it'll last long enough to start a small (10s of W?) SMPS, I don't know.


The problem with putting absolutely any kind of semiconductor directly on the mains, is transients.  You need at least a 1.2kV rating, with generous MOVs provided; at 400V, you're probably asking closer to 2.5kV.  Even with SiC MOSFETs and a generous budget, this is a huge ask, just in switching capacity, let alone in linear operation.  SiC do at least switch quickly, which suggests a more promising route: make an AC buck converter with current and voltage limiting behavior; the dropped power still must be dissipated, but it can be clamped by diodes and dumped into a generously sized resistor if nothing else.

Such a gross error (literally wired to the wrong power source) is best solved by a similarly gross corrective method: relay switching, or fusing.  Perhaps this device is promising in certain settings, but without those circumstances set out, without any motivation, without any specification -- it just sounds like a poor attempt to deal with something that should never, ever be a "nuisance" anywhere on the planet.

Tim

[soldar]

I understand your confusion.

I usually add LTspice files to simulate the circuit I present.

Unfortunately, here there is no good model to emulate the input of a conventional SMPS. Do you think it could be done?
So, for the time being in the least, the only way to see how this real circuit (in series) acts if it were a bipolar zener connected in parallel is by building it then testing it (with an SMPS as a load in series) by using a scope to see the positive and negative flat tops of the regulated AC voltage when the high limit is exceeded. Otherwise, it is very hard for someone to get how it works.

Again, if you are trying to do what I think you are trying to do, you are expressing it in a very confusing way which I would interpret as you not being entirely clear and being beyond your limits.

You say you want like a zener in parallel except it would not be in parallel and not be a zener.  That makes no sense.

It seems to me you just want to limit the voltage. The fact that it goes to a small wall wart SMPS is not really that important. It is probably just a load with a rectifier bridge and capacitor. You are probably focusing on unimportant points.

One way to stabilize voltage is with ferroresonant transformer
https://www.sunpower-uk.com/glossary/what-is-a-ferroresonant-power-supply/
https://voltage-disturbance.com/power-engineering/ferroresonant-transformer/
https: // www. youtube. com/watch?v=T5Je6l-OaoQ

[KerimF]

I don't quite understand "220V" being wired to a CD4093 logic gate. 

To let the SMPS be connected to neutral (not floating), I decided to let my circuit be the floating one. So, its ground is connected to the hot line.
You are right. The word '220V' is misleading here. But my assistants whose English is rather very weak are used to hear it (220V) as being 'the hot line'.   
 

Better names I guess would be "220V" --> "HOT", "NTR" --> "NEUTRAL", "SMPS" --> "LOAD"; "220V" net bridged to a "VDD" or something like that, and "GND" can be a high side VEE, or floating GND or something like that, maybe it's fine to leave as "GND" as such but just as long as it's clear.  A connected drawing would be easier to make sense of than net labels, making it clear that the logic side is "hanging below" one AC line, and there is a series bridge element connecting the load.

All you said makes sense. But when I drew it, after many other design failures and for in-house use only (that is just for me), I didn't notice that its schematic became clear to me only Please note that the regulating circuit is done for SMPS specifically (not any load). Therefore, no one here will likely be interested in keeping a conventional 220V SMPS running normally at high voltages. Usuallly, a simple passive protecting circuit (as the one at post #11, for example) could be enough in most applications.

So I guess the point is a dropper to limit load voltage?  But this is exactly inverse to a zener shunt, not equivalent to it!

Yes, it is a voltage dropper in series but, as I explained many times earlier, its end effect is of a bipolar zener (the positive and negative tops of the voltage sinewave are clamped, so two flat tops could be seen, exactly what a bipolar zener is supposed to do).
Yes, it is not a zener. It just regulates the AC high voltage as if it were a bipolar zener. This is why I added the word 'Virtual' in the OP title.

What current range?  What fault current capacity?  Not much I guess given the 1N4007s, and the poor IRFP460 will expire rather quickly under any kind of startup conditions, but maybe it'll last long enough to start a small (10s of W?) SMPS, I don't know.

It is not a general-purpose regulator. It was done for small SMPS supplies only.

The problem with putting absolutely any kind of semiconductor directly on the mains, is transients.  You need at least a 1.2kV rating, with generous MOVs provided; at 400V, you're probably asking closer to 2.5kV.  Even with SiC MOSFETs and a generous budget, this is a huge ask, just in switching capacity, let alone in linear operation.  SiC do at least switch quickly, which suggests a more promising route: make an AC buck converter with current and voltage limiting behavior; the dropped power still must be dissipated, but it can be clamped by diodes and dumped into a generously sized resistor if nothing else.

Your remarks are interesting. If I was born and live elsewhere, I would likely need to observe them. So far, the circuit, as it is, seems working fine in my surrounding (local consumers). 

Such a gross error (literally wired to the wrong power source) is best solved by a similarly gross corrective method: relay switching, or fusing.  Perhaps this device is promising in certain settings, but without those circumstances set out, without any motivation, without any specification -- it just sounds like a poor attempt to deal with something that should never, ever be a "nuisance" anywhere on the planet.

I agree with you. It was not done/designed for a NASA program
About nuisance, I am afraid that, if we are realistic, there are always new devices which are created and made, almost daily, just for it

[coromonadalix]

well  instead of creating it,  just buy such device ...  you have active line stabilizers / protectors  etc ...  they are certified and tested ....

You don't provide enough knowledge,   nor protections,  nor insurances this project will work as intended, 

For myself i would not use such a DIY design if you will use it say for professional use or protect some smps used somewhere  etc ...

the implications ... the safety  etc ...

too much misleading information, bad wording, sentences corrections, etc ....    this thread should not be pursued ...

no one will take the risk helping you like this,  and in case of something happens ....

my 2 cents,  sorry

[KerimF]

I understand your confusion.

I usually add LTspice files to simulate the circuit I present.

Unfortunately, here there is no good model to emulate the input of a conventional SMPS. Do you think it could be done?
So, for the time being in the least, the only way to see how this real circuit (in series) acts if it were a bipolar zener connected in parallel is by building it then testing it (with an SMPS as a load in series) by using a scope to see the positive and negative flat tops of the regulated AC voltage when the high limit is exceeded. Otherwise, it is very hard for someone to get how it works.

Again, if you are trying to do what I think you are trying to do, you are expressing it in a very confusing way which I would interpret as you not being entirely clear and being beyond your limits.

You say you want like a zener in parallel except it would not be in parallel and not be a zener.  That makes no sense.

Sorry, I don't try doing anything. It happened that after the circuit was designed and done, it works on the regulated sinewave voltage as if it were a parallel bipolar zener while it is actually connected in series. Isn't this weird. Yes, it is. I personally didn't expect this result. But a fact is a fact. I wouldn't believe it if I didn't test it and saw how the regulated top was clamped as a virtual bipolar zener in parallel is supposed to do.

It seems to me you just want to limit the voltage. The fact that it goes to a small wall wart SMPS is not really that important. It is probably just a load with a rectifier bridge and capacitor. You are probably focusing on unimportant points.

An SMPS supply (24V/2A for example), working from 100Vac up to about 240Vac, may not be important to most engineers. On my side, I use it to supply most of the controller boards in my products.

One way to stabilize voltage is with ferroresonant transformer
https://www.sunpower-uk.com/glossary/what-is-a-ferroresonant-power-supply/
https://voltage-disturbance.com/power-engineering/ferroresonant-transformer/
https: // www. youtube. com/watch?v=T5Je6l-OaoQ

Thank you for your suggested good solutions. In these days (since after year 2011), it becomes hard to me, where I live, to get ferro transformers. So, I used to avoid using them in my new designs. I just use the components that could be available locally, in every period of time (for this, I had to also update my designs. one a while).

[KerimF]

well  instead of creating it,  just buy such device ...  you have active line stabilizers / protectors  etc ...  they are certified and tested ....

You don't provide enough knowledge,   nor protections,  nor insurances this project will work as intended, 

For myself i would not use such a DIY design if you will use it say for professional use or protect some smps used somewhere  etc ...

the implications ... the safety  etc ...

too much misleading information, bad wording, sentences corrections, etc ....    this thread should not be pursued ...

no one will take the risk helping you like this,  and in case of something happens ....

my 2 cents,  sorry

All you said makes sense.

It just happens that we have no choice but to live in two very different worlds while we are on the same planet.

In brief, what seems good or bad to you, or what could be available to you or not... may not be so to me. Why? I am afraid it is a big topic and could be discussed rationally among 2 or 3 persons at most... surely not in a forum.

Anyway, thank you for your sincerity.

[soldar]

An SMPS supply (24V/2A for example), working from 100Vac up to about 240Vac, may not be important to most engineers. On my side, I use it to supply most of the controller boards in my products.

I mean it is not really relevant to the issue of limiting maximum voltage.

If you are trying to supply power to a device that will take between 100 and 240 V and you want to prevent the mains 230V going over voltage then you can put a transformer (or autotransformer) that will reduce the voltage from 230V nominal to, say, 160V. That way you are sure to never go over voltage.

[temperance]

Is there a valid reason for you to not just develop your own SMPS capable of doing exactly what you need? Switcher with integrated 700V rated MOSFET's are available for peanuts.

[T3sl4co1l]

Is there a valid reason for you to not just develop your own SMPS capable of doing exactly what you need? Switcher with integrated 700V rated MOSFET's are available for peanuts.

I suspect the answer is again availability.  OP doesn't give background (and, perhaps, has many reasons not to), but if we were to read into their country flag a little, we might guess they aren't in a very familiar (to western minds), or orderly, or even peaceful, environment.

Reading into that flag a bit further still, we might also expect reasons why OP is reluctant to give more details -- whether because middle-east anything is a "hot button" political topic (which, to be fair, a pretty sizable number of people here are capable of having a peaceful and nuanced conversation over; but due to the few who cannot, we generally shy away from such topics entirely), or even for reasons of personal security (whether giving more detail would create personal risk, e.g. deanonymizing, or indeed be seen as criticizing the local powers-that-be and thus a direct threat).

And, for the record, and by all means, do read this as a benefit of a doubt -- you have all my sympathy, being in such a contested region; my above reply is simply what it is, a response to the information as given, without trying to make much assumption about it.

Tim

[soldar]

Is there a valid reason for you to not just develop your own SMPS capable of doing exactly what you need? Switcher with integrated 700V rated MOSFET's are available for peanuts.

Besides the point of limited availability I would recommend adapting the input voltage to the device rather than trying to design and build a new device. It is much simpler to lower the voltage than it is to design and build a SMPS. Pretty much anyone can use a transformer to lower voltage but it takes a very specialized and experienced engineer to design a good SMPS. Not to mention the availability of special ferrite transformers, etc.