5V Zener Regulated 5V from 240V mains, Why Blow Up?

[Mechatrommer]

i assembled the following circuit and blew up the traces, any good reason why? other than possibly my careless mistakes such as short circuited traces, leakage etc, but i checked for short between mains etc, everything were fine. from circuit point of view, i cant see a reason why it should blow up, even if damaged zener. i have 180Kohm resistor to 5V zener, that should limit the current to 330V/180Kohm = 1.8mA tops... although 0.6W flowing through the resistor, the resistor is doing fine (not blown), its the zener that died. simulation indicates everything works the way they should... but, i need confirmation before reassembling the same circuit and potentially blowing another zener...

[IanMacdonald]

Resistors do have voltage limits irrespective of dissipation, and 330v is a bit much for some of the smaller styles.

Also worth noting that you will get a  fair a mount of ripple with this arrangement since C1 can only charge to a 'hard limit' of the zener voltage. If a problem,  the best workaround is to use a higher voltage zener and postregulate.

[Mechatrommer]

Resistors do have voltage limits irrespective of dissipation, and 330v is a bit much for some of the smaller styles.

i'm using jelly bean china 1/4W through hole resistor. is there any issue?

Also worth noting that you will get a  fair a mount of ripple with this arrangement since C1 can only charge to a 'hard limit' of the zener voltage. If a problem,  the best workaround is to use a higher voltage zener and postregulate.

this is 2nd concern after i get this circuit working (not blown), with blown circuit, i dont have enough time to probe for undesirable ripple. in fact i'm not worry too much with ripple, the 5V will be used for powering non-critical non-precisian part, such as just powering a mcu...

[amyk]

I'd recommend using a capacitive dropper instead of a resistive one.

[Benta]

I'd recommend stopping this approach, otherwise you'll earn a Darwin Award.

[janoc]

This is 2nd concern after i get this circuit working (not blown), with blown circuit, i dont have enough time to probe for undesirable ripple. in fact i'm not worry too much with ripple, the 5V will be used for powering non-critical non-precisian part, such as just powering a mcu...

  Good luck debugging that. Excessive ripple will cause all sorts of issues with spurious signals.

[cdev]

This is a question, not an answer..

Since we're on the subject of zener diodes used for voltage regulation..

OTOH it is reasonable (right?) to use this voltage drop across a zener diode approach to get 3.3v or 5v at very low ma off of a regulated 12v or 9v rail, without needing another full blown voltage regulator, right?

How hot will it get? or would it? Would it get hotter than a linear voltage regulator or not? Both would emit heat, which one would emit less into a closed space..

Remember in my circuit, we're just taking about low voltage DC.  dropping 12v or 9v down to 5 or 3.3v

[Karlo_Moharic]

just use a regulator like everybody  else

[max666]

To elaborate on Benta's post. Your circuit is mains referenced, Mechatrommer. One careless swap of the mains leads or a breakdown of the likely underspeced resistor and you could easily kill yourself.

[wraper]

What was connected to this circuit? It should't blow up traces by itself. My guess is there was something connected which have current path to earth. I hope you did not attach an oscilloscope to it.

[wraper]

i'm using jelly bean china 1/4W through hole resistor. is there any issue?

Although it shouldn't blow up normally, they are rated for lower voltage. Higher power resistors usually have higher max voltage rating, should check a datasheet for particular resistor.

[SeanB]

Typically these use 2 zener diodes, the first being a higher voltage ( around 12V) with the mains smoothing capacitor across it, and then a second resistor to the 5V zener, so that it is fed with something approaching DC, which it then regulates down to 5V. Can then be used to supply a low current ( under 1mA with your values, and will still have some residual ripple on it along with a rather long ( as in around 10 mains cycles) start up voltage build up as the capacitor charges.

[Mechatrommer]

What was connected to this circuit? It should't blow up traces by itself. My guess is there was something connected which have current path to earth. I hope you did not attach an oscilloscope to it.

see attached, it was floating in the air held by the helping hand by the mains cord... nothing connected, just the mains cord to the circuit. i was just testing for circuit stability when connected to the mains, about a second later, boom! btw, i know how to probe this safely dont worry people, i got shocks since i'm very little i'm very well aware of the danger to the point of all my kids are too scary when come close to a naked wire sigh ... the circuit was not probed yet, as i said, only stability test...

attached is the circuit i'm working on. nevermind other components they are not populated yet... only fuse, D1, R3, D2 (zener), and C2 and the mains cord (2 wires cord) are populated. do i have with copper pour clearance issue? i dont thing so, since the blast points are along the path of D1,R3,D2 (see 4th picture, i reattached the broken traces ready for the next blast? but waiting for the counsel verdict here). only D2 zener is dead, D1 and R3 still ok... fwiw...

[Mechatrommer]

just use a regulator like everybody  else

with too many components involved if i'm going the normal regulator route... then the china wins.. see below..

Typically these use 2 zener diodes, the first being a higher voltage ( around 12V) with the mains smoothing capacitor across it, and then a second resistor to the 5V zener, so that it is fed with something approaching DC, which it then regulates down to 5V. Can then be used to supply a low current ( under 1mA with your values, and will still have some residual ripple on it along with a rather long ( as in around 10 mains cycles) start up voltage build up as the capacitor charges.

i've seen china circuit using only one zener succesfully, this is my inspiration of very few components hence cheap china style... see attached of the other board i teared apart few weeks ago, its a drybox machine.. i havent pull it out of the casing so i cant trace the circuit, only the top board 1 zener visible (through hole technology) so i suspect nothing smd or whatsoever is underside... edit: err you are right 2 zeners there, i will consider this, but still i cant explain the blow up...

[AlfBaz]

Two words
Creepage and clearance

[Cerebus]

OTOH it is reasonable (right?) to use this voltage drop across a zener diode approach to get 3.3v or 5v at very low ma off of a regulated 12v or 9v rail, without needing another full blown voltage regulator, right?

Yup, nothing wrong with it as long as you don't need a precise voltage rail. It would be quite reasonable to power a single 5V op amp like this in a circuit that's otherwise all 12V.

How hot will it get? or would it? Would it get hotter than a linear voltage regulator or not? Both would emit heat, which one would emit less into a closed space..

Remember in my circuit, we're just taking about low voltage DC.  dropping 12v or 9v down to 5 or 3.3v

It'll get a little hotter than the best choice of regulator, a little less hot than a poorly chosen regulator. But we're talking about maybe chucking 5mA through a zener, 5mA to a load, dropping 10mA from a 12V rail to a 5V one, so (25mW+70mW) =  95 mW dissipation + another 25mW in the load. That'll warm up a 500mW BZX84 series diode by 12.5C above ambient and an RC55 series 1/4 watt resistor by 7.7C.

[rstofer]

Did the Helping Hand short anything?

[AlfBaz]

Wow, I'll try again

Mecha, get rid of the ground pour around the 240VAC mains part of the cct

[cdev]

Thank you. Think I will use some kind of partition to separate my GPSDO from its controller and two voltage regulators.

[32]

OTOH it is reasonable (right?) to use this voltage drop across a zener diode approach to get 3.3v or 5v at very low ma off of a regulated 12v or 9v rail, without needing another full blown voltage regulator, right?

Yup, nothing wrong with it as long as you don't need a precise voltage rail. It would be quite reasonable to power a single 5V op amp like this in a circuit that's otherwise all 12V.

How hot will it get? or would it? Would it get hotter than a linear voltage regulator or not? Both would emit heat, which one would emit less into a closed space..

Remember in my circuit, we're just taking about low voltage DC.  dropping 12v or 9v down to 5 or 3.3v

It'll get a little hotter than the best choice of regulator, a little less hot than a poorly chosen regulator. But we're talking about maybe chucking 5mA through a zener, 5mA to a load, dropping 10mA from a 12V rail to a 5V one, so (25mW+70mW) =  95 mW dissipation + another 25mW in the load. That'll warm up a 500mW BZX84 series diode by 12.5C above ambient and an RC55 series 1/4 watt resistor by 7.7C.
[/quote]

[Brumby]

Two words
Creepage and clearance

Wow, I'll try again

Mecha, get rid of the ground pour around the 240VAC mains part of the cct

My first recommendation also.  Sorry, but this just screamed at me!  It's asking for trouble.

Second is to take that 180K resistor and use at least 4 resistors in series (47K) - but check the datasheet for voltage limits on the resistors to work out the right number.  I couldn't do a single resistor like that.

[AlfBaz]

this just screamed at me!  It's asking for trouble.

One of my first home made pcb's used 240VAC and I did the obligatory ground pour over the whole board. Didn't see anything wrong at all on the screen. It wasn't until I went to solder the mains in I thought to myself WTF! and out came the dremel pcb modifier bit

[AlfBaz]

Second is to take that 180K resistor and use at least 4 resistors in series (47K) - but check the datasheet for voltage limits on the resistors to work out the right number.  I couldn't do a single resistor like that.

Yes, first time I saw that was on an industrial 600kW VVVF drive. If I remember correctly they used what looked like 4 standard 1% 1/4W resistors in series to feed an opamp with its 415VAC output

[wasyoungonce]

Please why are you not using an isolation step down transformer for safety.   Never run directly from mains.

Another fail is mains neutral in Australia is not earth....most house development us MEN (Main Earth Neutral) where earth is linked to neutral at one point in the power box.

You can control mains using a low pwr cct, its much safer so please consider this as direct mains circuits are very particular in components ratings.

[Mechatrommer]

Did the Helping Hand short anything?

its holding the mains cord. The circuit is floating on the air by the cord's stiffness.. if helping hand was to short anything,it will be the cord's core copper strand..

Wow, I'll try again

Mecha, get rid of the ground pour around the 240VAC mains part of the cct

its not a ground pour.. its the floating, or no net pour.. just to save my ferric chloride,and possibly as a place to put additional component if modification is needed. Do you mind sharing your experience?

[Brumby]

Whether it's ground or not is basically irrelevant.  It's a lot of metal around mains that doesn't need to be there.

Don't get cheap in addressing the needs of adequate:

Two words
Creepage and clearance

[amyk]

i've seen china circuit using only one zener succesfully, this is my inspiration of very few components hence cheap china style... see attached of the other board i teared apart few weeks ago, its a drybox machine.. i havent pull it out of the casing so i cant trace the circuit, only the top board 1 zener visible (through hole technology) so i suspect nothing smd or whatsoever is underside... edit: err you are right 2 zeners there, i will consider this, but still i cant explain the blow up...

That picture you posted is a capacitive dropper. See the big blue cap?

[wraper]

Did the Helping Hand short anything?

its holding the mains cord. The circuit is floating on the air by the cord's stiffness.. if helping hand was to short anything,it will be the cord's core copper strand..

Wow, I'll try again

Mecha, get rid of the ground pour around the 240VAC mains part of the cct

its not a ground pour.. its the floating, or no net pour.. just to save my ferric chloride,and possibly as a place to put additional component if modification is needed. Do you mind sharing your experience?

You have almost no isolation between that pour and mains voltage, no wonder it blew up. It actually shorted even before the fuse. Hint: did you ever see such pours in commercially made mains powered things?

[Loboscope]

its not a ground pour.. its the floating, or no net pour.. just to save my ferric chloride, ...

Don´t save your ferric chloride, save your health (if not your life!).
There must always be enough space between the traces conducting mains and the other traces/pour. I would not hesitate to add some extra gaps like the safety-gaps found in DMM´s.

But like others told you already I would also prefer to use l little transformer. Tiny, cheap and healthful  .

[Mechatrommer]

Thanks guys, i'm going as amyk suggested.. googling 'capacitive dropper' led me to interesting info esp one of the old thread over here with link to microchip app note.. it turned out what i'm trying to invent is a transformerless psu resistive dropper.one prominent drawback is excessive power waste in resistive dropper (as i already noted) , so not good. Rather than try to rectify whats wrong and potentially blowing another zener, but still stuck with less efficient circuit.. i'm drawing another pcb for capacitive dropper. My test last night showed the 180kohm resistor i was using is ntc, ie lower resistor when get heated, so it can be the worsening of the earlier problem, beside creepage and clearance that i will also take into account in the new pcb.. i'll post for advice before making pcb when i complete drawing the new circuit..

[kalel]

I heard about capacitive droppers from BigClive's LED videos. Maybe you could take a look at some if you're interested, but... it's not for powering MCUs (maybe in some case).


It seems that some lamps even use a capacitive dropper for power bank-like USB output! So, while charging, the USB is live.


This is the problem of resistive and capacitive droppers (from what I understand, I am by no means experienced or knowledgeable in the field), that they are not isolated from the mains, so there is a safety risk for products that are not fully enclosed. For a lamp in a fully covered case, I guess this is not an issue.

[Mechatrommer]

Non isolated circuit is a big no no when user interfaces or have access to it.. my project is fully encapsulated,except a potentiometer that user will operate by the plastic knob. The potentiometer exterior metals are not in contact with any node in the circuit,thats why i think its ok for my purpose.. with usb charger, there is possibility of exposure such as on iphone usb cord charger plug..

[kalel]

Quote from: Mechatrommer on Today at 11:22:59 AM

Non isolated circuit is a big no no when user interfaces or have access to it.. my project is fully encapsulated,except a potentiometer that user will operate by the plastic knob. The potentiometer exterior metals are not in contact with any node in the circuit,thats why i think its ok for my purpose.. with usb charger, there is possibility of exposure such as on iphone usb cord charger plug..

Yes, for USB charger it is not acceptable.

Just a thought, not a suggestion, but probably the cheapest solution from China is not (too) much more expensive if you could work with 5v from a cheap USB charger or module (charger is enclosed). Of course I can't really recommend them either - obviously it's not high quality or proven to be safe.

[Mechatrommer]

anyway transformerless psu only for low power in mA range. modern usb charger can charge up to 1-5A. transformerless psu will become less economic in that range, hence we commonly see smps transformer psu for them, i havent seen non-isolated battery/usb charger.

[thermistor-guy]

I have seen capacitive dropper circuits implemented in mains electricity meters, to power a small MCU inside the meter.

The circuit used a self-healing metallized film capacitor. When the capacitor get hits with brief high-voltage transients on the mains, the capacitor can break down for a short time; metal film near the breakdown location vaporizes; but the fault clears. The capacitor keeps working, but its capacitance now is slightly less.

After a few years of exposure to such transients, the capacitor's capacitance value can drop to a point where it no longer provides enough current for the circuit it powers. That's what happened to this particular batch of meters - after a few years of service, the MCU wouldn't operate reliably.

If you feed mains power to the zener using a resistor, you still have to keep these transients in mind, and design accordingly. Some resistors will seem to go short-circuit during kV level transients, due to the insulation on the resistor body breaking down. Under these conditions, the nominal high value of resistance is no longer in effect. Your circuit no longer has a simple resistor.

The approach I have seen in some telecommunications power supplies is to use two long (eg. 1/2 inch) power resistors in series for the resistive feed. The resistors were chosen not for the power rating but for the body length. The increased body length helped the resistors avoid breakdown under commonly occurring transient conditions (and also reduced sneak paths due to stray capacitance). Other posters have mentioned multiple resistors in series.

240 V mains isn't just 240 V. It has all sorts of transients on it, like lightning strikes and near-strikes, and switching transients from heavy machinery located in your area. If you want your circuit to last a few years, think about what voltage and energy spikes you want it to withstand. Keep in mind that a capacitive dropper may have a wear-out mechanism, as described.

[ludzinc]

 But with your home etched PCB I'd bet you have small copper deposits between traces.

At low voltages, it's not an issue.

At Mains voltages, it's conductive enough to be an issue. 

Use more clearance.  Don't want to etch it off?  Peel it off then!

[Mechatrommer]

this is the capacitive dropper circuit i'm currently working on (attached).. reference: http://ww1.microchip.com/downloads/en/AppNotes/00954A.pdf i just swap few components to the top side, easier for me to visualize gnd node mentally...

[Brumby]

Whatever solution you go for - whether resistive, capacitive or a toroidal transformer --- as others have said ... MAKE SURE you get decent clearances around the mains input!

[wraper]

this is the capacitive dropper circuit i'm currently working on (attached).. reference: http://ww1.microchip.com/downloads/en/AppNotes/00954A.pdf i just swap few components to the top side, easier for me to visualize gnd node mentally...

That circuit does not have bleeding resistor (likely present somewhere else in the circuit or has other means of discharging). When you remove it from mains socket and touch the plug, it will hit you with electricity stored in capacitor.

[Ice-Tea]

Seems folks missed the problem in the OP? 1/4W resistor blowing something like 300mW?

[woody]

I used to muck around with this problem a lot. How to get 5V or 3.3V from mains in a simple and safe way. My latest answer to that:

http://nl.farnell.com/vigortronix/vtx-214-001-305/power-supply-ac-dc-5v-0-2a/dp/2775086

Small, simple, safe and cheap. Wide VACin.  No worries if my widget uses 1, 5, 10 or 200 mA. And if I ever design something that needs to be rock-bottom cheap because I need to make a million of 'm I'll pay someone else to design a safe power supply 

[Mechatrommer]

MAKE SURE you get decent clearances around the mains input!

yup, i heard this since the beginning, i will implement this in the next pcb, thanks.

this is the capacitive dropper circuit i'm currently working on (attached).. reference: http://ww1.microchip.com/downloads/en/AppNotes/00954A.pdf i just swap few components to the top side, easier for me to visualize gnd node mentally...

That circuit does not have bleeding resistor (likely present somewhere else in the circuit or has other means of discharging). When you remove it from mains socket and touch the plug, it will hit you with electricity stored in capacitor.

thanks for reminder, i saw it last night on the other website i just forgotten it when drawing schematics, i will add the bleeding resistor.

Seems folks missed the problem in the OP? 1/4W resistor blowing something like 300mW?

yes this is my highly suspicion, as the resistor heated up, it went lower resistance, more and more current flowing some sort of thermal runaway until the things are blown. beside the other possibility of too narrow clearance and the floating copper pour got shorted due to unclean creepage.

[woody]

Seems folks missed the problem in the OP? 1/4W resistor blowing something like 300mW?

yes this is my highly suspicion, as the resistor heated up, it went lower resistance, more and more current flowing some sort of thermal runaway until the things are blown. beside the other possibility of too narrow clearance and the floating copper pour got shorted due to unclean creepage.

I might be completely wrong here but then I'll learn something. Is it not so that the .25W resistor behind the diode blows .5* ((240V/220k)^2 *220k) = .13W?

Apart from that, a moment ago I fried a 160k resistor between the leads of my meter, and it went from 159.3k cold to 158.0k when the paint came off so I do not see the lower resistance thing causing this.

Looking at the state of the PCB I would guess that a short is the most likely reason for this to happen. Or else the fact that where most .25W resistors are 250V maybe this Chinese one was a bit lower.

[LukeW]

Read Microchip AN954 on Transformerless Power Supplies.

http://ww1.microchip.com/downloads/en/AppNotes/00954A.pdf

You can only get very small load current... you will get a lot of power dissipation in the resistor and it will rapidly become impractical as the load current goes up.

You may also consider an offline buck regulator IC such as Power Integrations LinkSwitch-TN series. This won't give you galvanic isolation, just like a capacitive dropper does not, but you'll get less heat, more efficiency and more output current.

http://www.es.co.th/Schemetic/PDF/LNK304-306.PDF

[Ice-Tea]

You are right and wrong: wrong because it's 180k, not 220k. So, I calculated 0,32W. However, I was wrong also: it's not a full wave rectifier, so only half that. A 250mW resistor should be fine (but quite hot) unless it's a terrible Hung Wong Do resistor

[woody]

Ah, I see. I used the 220k resistor as shown in C1 but the schematic in the OP uses the 180k  you mentioned.

[orolo]

I know this is slightly off-topic, but after experimenting for a while with capacitive supplies and the such, I really don't find any redeeming value in them except for maybe weight. Consider a tiny epoxy encapsulated transformer like this (first search result). It's dirt cheap (0.9-1.6$), offers much more current than a capacitive dropper, provides galvanic isolation, much lesser ripple and is roughly equivalent in  power waste. Even if the transformer costs more than the dropping capacitor, the safety (no double encapsulation) and lifetime more than offset the additional cost.

Is there any advantage in capacitive droppers excepting some weight, a little volume, and maybe the chance to use the cheapest mains rated capacitors around?

Edit: to clarify, I mean for low voltage regulation. It can be argued that, for long LED strings and such, a capacitive dropper is a sensible option.

[xani]

The only advantage is slightly lower price (and size I guess) which is why it is so often used in cheap chinese stuff. I see no reason to bother with it for one-off hobbyist projects or small runs

Edit: to clarify, I mean for low voltage regulation. It can be argued that, for long LED strings and such, a capacitive dropper is a sensible option.

IIRC from bigclive videos some even just put resistor in series, but those had enough LEDs in series that resistor only dropped few volts