Author Topic: Transformerless power supplies circuit, cheatsheet and questions  (Read 8593 times)

0 Members and 1 Guest are viewing this topic.

Offline wasteTopic starter

  • Contributor
  • Posts: 32
  • Country: de

I have already built a dozen Transformerless power supplies with X2 rated capacitors and I 've been very happy about them (mostly around LEDs). They are cheap, easy to build, pretty nice in low amperage needs.

The negative side is that everything should be enclosed as the whole circuit is potentially hot and not isolated.


For the 230V people out there my little chart and cheatsheet goes as follows.

for every 100nF you get around 7mA  (10mA peak),
so 150nF would be around 10mA (15mA peak)
220nF around  15mA (22mA peak)
330nF around 25mA (33mA peak)
470nF around 33mA (47mA peak)

the inrush current resistor would be 1/2watt and I use the following values:
2.2K for 100-150nF,
1K for 220-330nF and
470Ω for 470nF

It's pretty nice to have a 680KΩ or 1MΩ resistor parallel to the X2 capacitor as it discharges the capacitor when plugged out (and prevents your fingers from getting shocked).

Another nice addition to have is a smoothing capacitor. In my LTspice simulations I get the following results for 6LEDs in series (normal 20mA-30mA MAX LEDs)

220nF X2 capacitor

  6μF smoothing capacitor (3.5mA min 20mA max)
22μF smoothing capacitor (8mA min 16mA max)
50μF smoothing capacitor (10.5mA min 14mA max)

330nF X2 capacitor

33μF smoothing capacitor (13mA min 23.5mA max)
50μF smoothing capacitor (14.5mA min 22mA max)


The circuit from LTspice and the graph is attached.


Now my question part for the more experienced here. In the past 2 months I had 2 of them I gave to friends as gifts failing. It's a night lamp made of concrete and acylic and as everything is safely enclosed in concrete and acrilyc, its also a pain in the ass to repair :)))))

Inspecting the failing circuits I found in both of them a failed electrolytic capacitor and a burned LED :) As the LED's are in series one dead and everything stops working. As I wrote in the beggining that I love this circuit because of its simplicity and durability. I have some of them running for years 24/7 with no problems whatsoever.

So my next step is use the oscilloscope to see if the LTspice simulation actually makes sense in real life.

But as I treat the whole circuit as potentially live, I want to be sensible and safe in my approach. I don't have a differential probe but Im in the process of isolating my USB osciloscope and I always use a hands off approach. I setup everything before a turn the circuit on take the measurment and then off again.

In my bench I can be sure which wire is the Hot in the circuit and I attach the probe there, as it is a high impendence path. I also use AC coupling. But what stopped me from testing was that I could measure 60-70VAC between the aligator on the probe and the DC negative on my circuit. Not being able to explain it I turned for your help :)

My 3 thoughts of how to approach this issue are.

a) finish isolating the USB and powersupply of the oscilloscope
b) build a lamp in series on the hot wire for extra security
c) build a small isolation transformer using two 30Watt transformers I have around back to back.

 

Offline OwO

  • Super Contributor
  • ***
  • Posts: 1250
  • Country: cn
  • RF Engineer.
Re: Transformerless power supplies circuit, cheatsheet and questions
« Reply #1 on: February 05, 2019, 02:13:05 pm »
Because you have a full bridge rectifier AC ground is not DC ground. There is some AC potential difference between the two. Try grounding the AC supply in your simulation.
Email: OwOwOwOwO123@outlook.com
 

Offline Cliff Matthews

  • Supporter
  • ****
  • Posts: 1911
  • Country: ca
    • General Repair and Support
Re: Transformerless power supplies circuit, cheatsheet and questions
« Reply #2 on: February 05, 2019, 02:20:47 pm »
Nice guide. I don't usually go much above 50V but it'd be nice to see handy numbers for 120V people.
 

Offline wasteTopic starter

  • Contributor
  • Posts: 32
  • Country: de
Re: Transformerless power supplies circuit, cheatsheet and questions
« Reply #3 on: February 05, 2019, 03:03:49 pm »
Nice guide. I don't usually go much above 50V but it'd be nice to see handy numbers for 120V people.

I also don't usually go over 50V , the 6LEDs that are in my circuit total around 20V, and my biggest nightlam has 11LEDs in series which is around 33V :)


Because you have a full bridge rectifier AC ground is not DC ground. There is some AC potential difference between the two. Try grounding the AC supply in your simulation.

It's not in my simulation I see the potential, it's on the circuit in my multimeter when checking between the - of the LEDs and the aligator of the probe  :)
That's why I haven't tried using the probe yet :)
 

Offline soldar

  • Super Contributor
  • ***
  • Posts: 3595
  • Country: es
Re: Transformerless power supplies circuit, cheatsheet and questions
« Reply #4 on: February 05, 2019, 07:57:21 pm »
The only way I see for that damage to happen is that there was a spike in voltage and current and for that to happen R2 and C1 are suspect. Are they OK? How about the four diodes of the bridge? It may be that a high voltage transient caused a spike in voltage.

R2 is not strictly necessary and may be a cause of failure so you might consider removing it

C1 should be of adequate voltage. for 230 volt circuits I like to put 630 volt caps.

After the bridge you could place a zener as extra protection.

I cannot think how else it could have failed.

In any case, I like to make things easily repairable for this very reason.
All my posts are made with 100% recycled electrons and bare traces of grey matter.
 

Offline mariush

  • Super Contributor
  • ***
  • Posts: 5170
  • Country: ro
  • .
Re: Transformerless power supplies circuit, cheatsheet and questions
« Reply #5 on: February 05, 2019, 08:14:32 pm »

You should check out DesignNote 001c : Transformerless Power Supply Design

Link : http://www.designercircuits.com/DesignNote1c.pdf

Starting from page 6, you have formulas you can use to determine the values for all parts.


For something more basic, even though the formulas are not quite 100% correct, you may want to check out  Microchip's appnote :

AN954 : Transformerless Power Supplies: Resistive and Capacitive

Link : http://ww1.microchip.com/downloads/en/appnotes/00954a.pdf

Lots of formulas and explanations ... here's a "cheat sheet" :

Note that the example and data is for 110v AC , so with 230v you'd use a C1 capacitor rated for more than 250vA AC.  And the example uses only half wave rectification.

Also, output capacitance... depends on how much ripple you want... a good approximation can be done with formula  C (farads) = Current (Amps ) / [2 x Mains Frequency x (Vdc peak - Vdc minimum desired) ]


« Last Edit: February 05, 2019, 08:27:48 pm by mariush »
 
The following users thanked this post: Cliff Matthews

Online Benta

  • Super Contributor
  • ***
  • Posts: 6420
  • Country: de
Re: Transformerless power supplies circuit, cheatsheet and questions
« Reply #6 on: February 05, 2019, 08:40:06 pm »
I've also built a lot of these capacitor-drop supplies, and a couple of points stand out.
For 470 nF X2 drop-caps I use 330 ohms wirewound for R1 These are very robust when it comes to current surges, which you'll certainly have here when switching on. I never had any problems.
R2 needs to be a high voltage type, I normally use Vishay VR37, 470 kohms.
I also always use a 1 W Zener after the bridge to have a somewhat constant DC voltage.

Now to your failure mode: what happens is, that one of your LEDs fails (probably due to a high current spike when switching on). When this happens, you have no voltage regulation any more, and the voltage over your filter cap rises until destruction.
BTW: where's the current limiting for your LEDs? I see no resistor.

To the 120 V, 60 Hz people: resize the cap as: (230/120) / (50/60) * C
Leave the rest as it is.
« Last Edit: February 05, 2019, 09:02:52 pm by Benta »
 

Offline AnalogSteph

  • Newbie
  • Posts: 3
  • Country: de
    • My Web 1.0 Home
Re: Transformerless power supplies circuit, cheatsheet and questions
« Reply #7 on: February 06, 2019, 12:35:13 am »
Now to your failure mode: what happens is, that one of your LEDs fails (probably due to a high current spike when switching on). When this happens, you have no voltage regulation any more, and the voltage over your filter cap rises until destruction.
I concur.

The crux with dropper supplies is that they are inherently a highpass. Any voltage spikes will be attenuated much less than the 50/60 Hz mains and may cause damage. Worst case, turning on or off exactly when line voltage peaks.

And yes, running LEDs on what's still mostly a voltage source isn't exactly recommended. I would consider implementing a shunt current limiter with one extra resistor and transistor. The resistor would be in series with the LEDs and sized such that R ~= 0.65 V / I_led,max, and the transistor quite arguably should be medium power with plenty of Vceo to spare. This should also work as a current regulator, but obviously dumping part of your current is not going to help efficiency for obvious reasons, and I would tweak the circuit for a decent balance.
 

Offline not1xor1

  • Frequent Contributor
  • **
  • Posts: 716
  • Country: it
Re: Transformerless power supplies circuit, cheatsheet and questions
« Reply #8 on: February 06, 2019, 08:19:01 am »
a couple of notes:
-1) it is better to use 2-3 resistors in series for discharging the input capacitor, you do not need a power resistor and low power ones have a limited range for the maximum voltage
-2) for high current PSUs (i.e. 100mA or more) an active circuit for limiting input surge current allows to get a more efficient PSU
 

Offline StillTrying

  • Super Contributor
  • ***
  • Posts: 2850
  • Country: se
  • Country: Broken Britain
Re: Transformerless power supplies circuit, cheatsheet and questions
« Reply #9 on: February 06, 2019, 08:52:56 am »
Inspecting the failing circuits I found in both of them a failed electrolytic capacitor and a burned LED :)

Start the Vin Sine source with a delay of 90 Deg. to check for start up surges.
On the odd occasion when you get an intermittent switch on or switch off, I think the peaks could be over 1A, mostly going into the 33u cap.
« Last Edit: February 06, 2019, 08:56:32 am by StillTrying »
.  That took much longer than I thought it would.
 

Offline soldar

  • Super Contributor
  • ***
  • Posts: 3595
  • Country: es
Re: Transformerless power supplies circuit, cheatsheet and questions
« Reply #10 on: February 06, 2019, 09:23:21 am »
Max forward voltage for the LEDs is about 4.0 V. You could put a 5V zener in parallel with each LED and that way if one LED dies the rest will keep working.

That circuit is putting about 20~22 mA through the LEDs which is a bit too close to the absolute max of 25 mA, especially if it is enclosed and cannot dissipate heat well. I would feel more comfortable bringing the current down to about 13~15 mA by decreasing C1 to about .22~.25 uF. That is the first thing I would do. Decrease current and increase refrigeration.

As has been said, when one LED fails, there goes the capacitor (which is not really needed).

Contrary to what has been said, this type of power supply approximates an AC current source, not voltage. Short the LEDs and you get a certain current through the short. Now place the LEDs and the current is almost the same while the voltage rises. Open the circuit and the voltage keeps rising.
« Last Edit: February 06, 2019, 09:54:19 am by soldar »
All my posts are made with 100% recycled electrons and bare traces of grey matter.
 

Offline wasteTopic starter

  • Contributor
  • Posts: 32
  • Country: de
Re: Transformerless power supplies circuit, cheatsheet and questions
« Reply #11 on: February 06, 2019, 11:40:00 am »
Thanks for all the interesting comments, I will try to adress and keep asking if you may :)

I've also built a lot of these capacitor-drop supplies, and a couple of points stand out.
For 470 nF X2 drop-caps I use 330 ohms wirewound for R1 These are very robust when it comes to current surges, which you'll certainly have here when switching on. I never had any problems
I also always use a 1 W Zener after the bridge to have a somewhat constant DC voltage.

Now to your failure mode: what happens is, that one of your LEDs fails (probably due to a high current spike when switching on). When this happens, you have no voltage regulation any more, and the voltage over your filter cap rises until destruction.
BTW: where's the current limiting for your LEDs? I see no resistor.

Can you elaborate in what way the wirewound is more robust? It attenuates the surge better for the rest of the elements, or it just more robust for itself. I never got a failing resistor. I think the failure mode you described is what most people agree on :)

You don't need a resistor for the LEDs, as the X2 Capacitor (C1) acts as the current limiting factor. Capacitive droppers to my understanding are current limiting devices. You can add up as many LED's as you want in series, it will still give the same amperage.



a couple of notes:
-1) it is better to use 2-3 resistors in series for discharging the input capacitor, you do not need a power resistor and low power ones have a limited range for the maximum voltage

1) about the low powered resistors (ex 1/4watt) . What do you mean they have a limited range for maximum voltage? that they fail easier (acting as fuse) or attenuating the surges better?



Start the Vin Sine source with a delay of 90 Deg. to check for start up surges.
On the odd occasion when you get an intermittent switch on or switch off, I think the peaks could be over 1A, mostly going into the 33u cap.

Would be using 5-6μF ceramic caps, or higher voltage rated electrolytic caps help prevent this problem??




Max forward voltage for the LEDs is about 4.0 V. You could put a 5V zener in parallel with each LED and that way if one LED dies the rest will keep working.
That circuit is putting about 20~22 mA through the LEDs which is a bit too close to the absolute max of 25 mA, especially if it is enclosed and cannot dissipate heat well. I would feel more comfortable bringing the current down to about 13~15 mA by decreasing C1 to about .22~.25 uF. That is the first thing I would do. Decrease current and increase refrigeration.

As has been said, when one LED fails, there goes the capacitor (which is not really needed).

puting a zener diode in parallel with each LED kind of defeats the simplicity of this project. Most of the times I use 3825LED's in 12V strips which are already connected in parallel and then you can connect them in series of 3s together (see attached pictures).
Using lower current is a solid and good idea, I usually don't go over .22nF and the smoothing capacitor decreases the max current even more apart from minimizing flickering. Maybe even that is too much though and should have gone with 150nF X2 Capacitor.


The only way I see for that damage to happen is that there was a spike in voltage and current and for that to happen R2 and C1 are suspect. Are they OK? How about the four diodes of the bridge? It may be that a high voltage transient caused a spike in voltage.
R2 is not strictly necessary and may be a cause of failure so you might consider removing it
C1 should be of adequate voltage. for 230 volt circuits I like to put 630 volt caps.
After the bridge you could place a zener as extra protection.
I cannot think how else it could have failed.
In any case, I like to make things easily repairable for this very reason.

adding a zener in parallel I think it's not going to work with LED's so well as the zener should be really precise to the Vf of the LEDs . 1-2 volts higher and the LED s are toast anyhow :)
Everything else apart from the smoothing capacitor is working nicely. R2 is "needed" if you dont want to get a little sting every time you take the plug out and touch the metal, I dont want to scare my friends with my present :)



Do you have any suggestions about how to probe correctly?

Thanks again a bunch for all the help and ideas
« Last Edit: February 06, 2019, 11:45:32 am by waste »
 

Offline not1xor1

  • Frequent Contributor
  • **
  • Posts: 716
  • Country: it
Re: Transformerless power supplies circuit, cheatsheet and questions
« Reply #12 on: February 06, 2019, 01:19:48 pm »
a couple of notes:
-1) it is better to use 2-3 resistors in series for discharging the input capacitor, you do not need a power resistor and low power ones have a limited range for the maximum voltage

1) about the low powered resistors (ex 1/4watt) . What do you mean they have a limited range for maximum voltage? that they fail easier (acting as fuse) or attenuating the surges better?

The datasheet specifies the maximum working voltage.
With higher voltage there is loss of insulation so you risk to burn all your leds.
Probably that is affected by environmental conditions (temperature/pressure/humidity) too.
 

Online Zero999

  • Super Contributor
  • ***
  • Posts: 20357
  • Country: gb
  • 0999
Re: Transformerless power supplies circuit, cheatsheet and questions
« Reply #13 on: February 06, 2019, 01:54:42 pm »
A transformerless power supply is a current source if the LED forward voltage is low, compared to the mains.

My advice is to avoid them where at all possible. Use a switched mode power supply instead. The only reason to use a transformerless PSU is for cost, which perhaps makes sense at low power levels, but at more than about 50mA, a switched mode becomes more attractive, because decent, large capacitors aren't cheap.
 
The following users thanked this post: Yansi

Online Benta

  • Super Contributor
  • ***
  • Posts: 6420
  • Country: de
Re: Transformerless power supplies circuit, cheatsheet and questions
« Reply #14 on: February 06, 2019, 03:00:34 pm »
"Can you elaborate in what way the wirewound is more robust? It attenuates the surge better for the rest of the elements, or it just more robust for itself."

They are much more robust when it comes to surge overloads.
Think about it: if the circuit is switched on at max. AC voltage (~320 V), we're talking about almost 1 A in 330 ohms, as all caps are discharged.
This works out as a peak power surge (very short, admittedly) of ~300 W. A wirewound will withstand this, a film resistor will not in the long run.
If the resistor is 680 ohms, we're still talking ~150 W.
 
The following users thanked this post: waste

Offline mikeselectricstuff

  • Super Contributor
  • ***
  • Posts: 14117
  • Country: gb
    • Mike's Electric Stuff
Re: Transformerless power supplies circuit, cheatsheet and questions
« Reply #15 on: February 06, 2019, 03:06:06 pm »
If going for higher powers, check out the Power integrations Linkswitch series devices - costs may be comparable to larger capacitive dropppers and certainly smaller 
Youtube channel:Taking wierd stuff apart. Very apart.
Mike's Electric Stuff: High voltage, vintage electronics etc.
Day Job: Mostly LEDs
 

Offline wasteTopic starter

  • Contributor
  • Posts: 32
  • Country: de
Re: Transformerless power supplies circuit, cheatsheet and questions
« Reply #16 on: February 06, 2019, 06:39:07 pm »
A transformerless power supply is a current source if the LED forward voltage is low, compared to the mains.

My advice is to avoid them where at all possible. Use a switched mode power supply instead. The only reason to use a transformerless PSU is for cost, which perhaps makes sense at low power levels, but at more than about 50mA, a switched mode becomes more attractive, because decent, large capacitors aren't cheap.

I agree with you that for higher wattage, there are more attractive alternatives, but it's not only the cost that makes them attractive compared to the alternatives.

If you go for a 5V switching power supply then you will have to either use a transistor to current limit your leds, or something like the AMC7235 chip, which usually comes in 350mA increments. Don't get me wrong I love how cheap and simple is the AMC7235, but it's not exactly efficient in 5V (it's mostly designed for li-ion batteries) and then you are constrained to 350mA which for many applications where you want ambient light, is too much.


Then you can go for a switching current limited board like the chinese ones built around the BP2836 and it's many siblings that go around. They are cheap and give okeish stable output, but even the smallest ones are quite powerfull for ambient light and they are not exactly reliable .

Additionally both solutions are quite bulky compared to a capacitive dropper and not so versatile. With the capacitive dropper you just add LED's in series and you are done.

To sum it up, I agree that for higher wattage/current needs, other solutions are better and more suitable, but I came to love those capacitive droppers when I started using them 2 years ago (before I was kind of timid going around the 220V), that's why I want to make sure they will stay as reliable as my first designs that go on and on and on :)



The datasheet specifies the maximum working voltage.
With higher voltage there is loss of insulation so you risk to burn all your leds.
Probably that is affected by environmental conditions (temperature/pressure/humidity) too.


So using a 1/2watt or a 1watt resistor makes sense as it can take more abuse from higher voltages, the issue can be with the 1/4 1MΩ resistor that discharges the capacitor. Did I get it correctly?
 

Offline StillTrying

  • Super Contributor
  • ***
  • Posts: 2850
  • Country: se
  • Country: Broken Britain
Re: Transformerless power supplies circuit, cheatsheet and questions
« Reply #17 on: February 06, 2019, 07:22:37 pm »
I doubt applying some 320V square waves is a very accurate way to simulate a bouncing or arcing switch, but I've simulated it that way anyway, :) using the values as in the 1st post.

Green and Grey is with the 33u smoothing cap disconnected, Purple and Yellow is with it connected.
At least in this simulation a way to protect the LEDs from switch-on current spikes Yellow, is to double the 33u value to 68u Blue. YMMV.
« Last Edit: February 08, 2019, 07:10:14 pm by StillTrying »
.  That took much longer than I thought it would.
 
The following users thanked this post: waste

Online Benta

  • Super Contributor
  • ***
  • Posts: 6420
  • Country: de
Re: Transformerless power supplies circuit, cheatsheet and questions
« Reply #18 on: February 06, 2019, 07:35:50 pm »
"So using a 1/2watt or a 1watt resistor makes sense as it can take more abuse from higher voltages, the issue can be with the 1/4 1MΩ resistor that discharges the capacitor. Did I get it correctly?"

No. The correct way is reading the resistor data sheet. That's responsible design methodology.

Example: the Vishay VR37 high voltage resistors I mentioned previously are metal glaze high voltage resistors that will withstand pulse voltages up to 3.5 kV. Just to demonstrate: the tolerance colour band is yellow instead of gold to achieve this.

Standard metal film resistors will accept 300 V peak, 75 V continuous voltage. Not what you want.

EDIT: the resistors I use are actually VR25, apologies.

« Last Edit: February 06, 2019, 09:33:53 pm by Benta »
 

Offline not1xor1

  • Frequent Contributor
  • **
  • Posts: 716
  • Country: it
Re: Transformerless power supplies circuit, cheatsheet and questions
« Reply #19 on: February 07, 2019, 06:32:55 am »
A transformerless power supply is a current source if the LED forward voltage is low, compared to the mains.

My advice is to avoid them where at all possible. Use a switched mode power supply instead. The only reason to use a transformerless PSU is for cost, which perhaps makes sense at low power levels, but at more than about 50mA, a switched mode becomes more attractive, because decent, large capacitors aren't cheap.

a 1µF 275VAC X2 capacitor is around 25-30 euro cents (a few months ago I bought 10 for 2.51€ from TME)
the main drawback is that unless you use some active circuit (e.g. a high voltage mosfet current limit) the surge current limit resistance resistor would add to the overall power loss and greatly affect efficiency
« Last Edit: February 08, 2019, 06:17:37 am by not1xor1 »
 

Offline not1xor1

  • Frequent Contributor
  • **
  • Posts: 716
  • Country: it
Re: Transformerless power supplies circuit, cheatsheet and questions
« Reply #20 on: February 07, 2019, 06:39:37 am »
The datasheet specifies the maximum working voltage.
With higher voltage there is loss of insulation so you risk to burn all your leds.
Probably that is affected by environmental conditions (temperature/pressure/humidity) too.


So using a 1/2watt or a 1watt resistor makes sense as it can take more abuse from higher voltages, the issue can be with the 1/4 1MΩ resistor that discharges the capacitor. Did I get it correctly?

that was just an example...
other manufacturers may produce resistors able to withstand higher or lower voltages
anyway in most cases a 1W resistor might work
« Last Edit: February 07, 2019, 06:42:23 am by not1xor1 »
 

Offline mariush

  • Super Contributor
  • ***
  • Posts: 5170
  • Country: ro
  • .
Re: Transformerless power supplies circuit, cheatsheet and questions
« Reply #21 on: February 07, 2019, 11:43:23 am »
As long as we're at the subject of non-isolated power supplies it's worth mentioning that there are also linear regulators that can work with up to 440v input voltage.

For example LR8K4 can output up to 20mA, provided you don't exceed the power dissipation: https://uk.farnell.com/microchip/lr8k4-g/linear-volt-reg-0-02a-440v-to/dp/2448524

So for example, let's say for a micro that uses only 1-2mA, you could simply have a diode for half wave rectification and let's say a 1uF capacitor, or the minimum to keep the input voltage above around 17v (the regulator needs 12v above output voltage for regulation)

 

Offline mikeselectricstuff

  • Super Contributor
  • ***
  • Posts: 14117
  • Country: gb
    • Mike's Electric Stuff
Re: Transformerless power supplies circuit, cheatsheet and questions
« Reply #22 on: February 07, 2019, 12:08:53 pm »
As long as we're at the subject of non-isolated power supplies it's worth mentioning that there are also linear regulators that can work with up to 440v input voltage.

For example LR8K4 can output up to 20mA, provided you don't exceed the power dissipation: https://uk.farnell.com/microchip/lr8k4-g/linear-volt-reg-0-02a-440v-to/dp/2448524

So for example, let's say for a micro that uses only 1-2mA, you could simply have a diode for half wave rectification and let's say a 1uF capacitor, or the minimum to keep the input voltage above around 17v (the regulator needs 12v above output voltage for regulation)
Not sure I can see much benefit of a HV regulator over a simple zener.
Youtube channel:Taking wierd stuff apart. Very apart.
Mike's Electric Stuff: High voltage, vintage electronics etc.
Day Job: Mostly LEDs
 

Online Gyro

  • Super Contributor
  • ***
  • Posts: 10173
  • Country: gb
Re: Transformerless power supplies circuit, cheatsheet and questions
« Reply #23 on: February 07, 2019, 12:32:56 pm »
"Can you elaborate in what way the wirewound is more robust? It attenuates the surge better for the rest of the elements, or it just more robust for itself."

They are much more robust when it comes to surge overloads.
Think about it: if the circuit is switched on at max. AC voltage (~320 V), we're talking about almost 1 A in 330 ohms, as all caps are discharged.
This works out as a peak power surge (very short, admittedly) of ~300 W. A wirewound will withstand this, a film resistor will not in the long run.
If the resistor is 680 ohms, we're still talking ~150 W.

I would go the other way and use an appropriately rated fusible resistor. This sort of circuit tends to be difficult to fuse due to the low current - you don't want a very hot resistor for any significant time. Decent commercial circuits use flameproof fusible resistors to prevent the fire risk.
« Last Edit: February 07, 2019, 12:52:09 pm by Gyro »
Best Regards, Chris
 

Offline wasteTopic starter

  • Contributor
  • Posts: 32
  • Country: de
Re: Transformerless power supplies circuit, cheatsheet and questions
« Reply #24 on: February 07, 2019, 10:30:20 pm »
I doubt applying some 320V square waves is a very accurate way to simulate a bouncing or arcing switch, but I've simulated it that way anyway, :) using the values as in the 1st post.
Green and Grey is with the 33u smoothing cap disconnected, Purple and Yellow is with it connected.
At least in this simulation a way to protect the LEDs from switch-on current spikes Yellow, is to double the 33u value to 68u Blue. YMMV.

thanks for the graphs, so bigger capacitors with higher voltage rating are actually quite helpful. I'm actually building a small isolation transformer using 2 transformers back2back so I can test the circuit live and measure the transient volts (not the amperage though my clamp meter doesn't have a max function).

how did you simulate the spikes? Do I get it correctly that square wave is kind of a worse scenario than sine wave? (Ive read that somewhere about generators)


So using a 1/2watt or a 1watt resistor makes sense as it can take more abuse from higher voltages, the issue can be with the 1/4 1MΩ resistor that discharges the capacitor. Did I get it correctly?

that was just an example...
other manufacturers may produce resistors able to withstand higher or lower voltages
anyway in most cases a 1W resistor might work
[/quote]

The same goes for higher rated resistors :)


As long as we're at the subject of non-isolated power supplies it's worth mentioning that there are also linear regulators that can work with up to 440v input voltage.
For example LR8K4 can output up to 20mA, provided you don't exceed the power dissipation: https://uk.farnell.com/microchip/lr8k4-g/linear-volt-reg-0-02a-440v-to/dp/2448524
So for example, let's say for a micro that uses only 1-2mA, you could simply have a diode for half wave rectification and let's say a 1uF capacitor, or the minimum to keep the input voltage above around 17v (the regulator needs 12v above output voltage for regulation)
That would be nice for voltage regulated applications (like the micro). Capacitive droppers are current regulated, which is exactly why I find them so sexy for the LEDs.


Not sure I can see much benefit of a HV regulator over a simple zener.

Given the fact that LEDs are quite constrained in the Voltage category a zener would have to be very close to their top rating (for the 6 LEDs  example around 19-20Volts), to actually protect them from the surges. Am I interpreting you correctly?


I would go the other way and use an appropriately rated fusible resistor. This sort of circuit tends to be difficult to fuse due to the low current - you don't want a very hot resistor for any significant time. Decent commercial circuits use flameproof fusible resistors to prevent the fire risk.

The flameproofness of my design is in the enclosure. No air and concrete all around the circuit :)
My main concern wasn't the resistor, but also the X2 capacitor which can fail open or closed if I am not mistaken.
using higher rated resistors will make them also run cooler, isn't that true? I usually go for at least double the rating, so instead of 1/4 on the R1 I use 1watt resistors


again thanks for all the input
« Last Edit: February 07, 2019, 10:54:43 pm by waste »
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf