Another transformerless power supply design question!

[MrAl]

I was under the impression that the series resistor was in case the capacitor goes S/C, and act's as a kind of fuse.

Nigel, I'm glad you brought this up.
I almost forgot to mention that the resistor should be a flameproof type
unless we use a fuse in series too, a low amp fuse that is.
A standard resistor could start a fire.

Poodle, my web site is being taken down because the stupid company
that hosted it decided to stop doing that kind of web stuff. I'll see
if i can post my schematic here anyway...

**broken link removed**

The zener voltage depends on how many LEDs are connected in series.
The current through the LEDs is (Vz-Vleds)/100.
This is a full wave type and provides more light per farad than a half wave type.

Note also that this unit uses a zener to help limit the surge through the LED
when first plugged in or for random line transients.

One thing i almost forgot to mention is that a 1Meg resistor (1/4 watt) should
be placed across the cap to make sure it discharges in a reasonable time after
the unit is unplugged.
It's funny however that without this resistor i was able to unplug the unit
from the wall and place my index finger across the two 120vac plug prongs
and get the LED to light up for a couple seconds I didnt feel any
shock at all for some reason.

 

[MrAl]

Hi again,


Here's a formula that can be used to calculate the current through
the series resistor (the resistor connected to the 0.6uf cap in the
schematic in my previous post) and the power in it during normal
operation. This formula can also be used for other lights like this
by changing vLED in the formula to whatever the LED voltage is,
or to what the zener (if any) voltage is. In my schematic, vLED
must be made equal to the zener voltage but many lights like this
do not use a zener so that would be made simply the total LED
voltage. If more than one LED in series, then vLED is made equal to
the total series voltage of all the LEDs.

I=(2*pi*C*F*(V-vLED))/sqrt(4*pi^2*C^2*F^2*R^2+1)
P=I^2*R

where

I is the rms current flow through the series resistor
pi=3.14159265
C is the capacitance of the series capacitor, in farads
F is the line frequency in hertz
V is the rms voltage of the line (typically 120v)
vLED is the LED voltage or the zener voltage in power supplies
R is the series resistor (100 ohm connected to cap in the schematic)
P is the average power in watts
This calculation is good for a full wave version, but also close for
a half wave version for calculating power in the series resistor when
vLED is made equal to zero.

EXAMPLE:

For an example, say we use one white LED directly on the output of
the bridge rectifier, this makes vLED equal to 3.5 volts. Lets say
we make the series resistor 100 ohms, and the line frequency is 60Hz
and the line voltage is 120vrms, and C is 0.68uf, this makes I equal to:

I=(2*pi*C*F*(V-vLED))/sqrt(4*pi^2*C^2*F^2*R^2+1)
I=0.029855 amps.

Now to calculate the power:

P=I^2*R
P=0.08913

which is close to 100mw in the resistor. This means a 1/4 watt
(0.250 watts) resistor would work ok here.

It's interesting that with these same parameters except make the
resistor 200 ohms, we get 0.178 watts, which is getting a little
high for a 1/4 watt resistor, but going to 300 ohms we get a power
of 0.266 watts, which is clearly too high for a 0.25 watt resistor.

 

[poodlenuggets]

I almost forgot to mention that the resistor should be a flameproof type
unless we use a fuse in series too, a low amp fuse that is.
A standard resistor could start a fire.

I've actually never heard of a flameproof resistor - is there a specific type of resistor, like metal film, that are flameproof, or do you have to check the datasheet for each one? I searched for "flameproof" on Digikey and came up with a small, not very useful selection of resistors.

 

[ericgibbs]

I've actually never heard of a flameproof resistor - is there a specific type of resistor, like metal film, that are flameproof, or do you have to check the datasheet for each one? I searched for "flameproof" on Digikey and came up with a small, not very useful selection of resistors.

hi,
Look here:
How Electronic Things Work... and ... - Google Book Search

 

[Optikon]

I've actually never heard of a flameproof resistor - is there a specific type of resistor, like metal film, that are flameproof, or do you have to check the datasheet for each one? I searched for "flameproof" on Digikey and came up with a small, not very useful selection of resistors.

Why are you bothering with safety things like flameproof resistors? You have already resigned yourself to a very foolish design and if you are going to do it right, then do it right!

 

[Optikon]

I was under the impression that the series resistor was in case the capacitor goes S/C, and act's as a kind of fuse.

The notion of a resistor acting like a fuse requires a type that will fail reliably as an open circuit and not arc-over. This is not a trivial part. IF resistors worked wella s fuses, there wouldnt be fuses. I'll never understand why someone would ever want to replace a fuse with a resistor hoping it will act like a fuse (especially in a CAT II+ installation)

 

[Optikon]

In that case for the cost of less than 1p for the resistor, I would always fit it.

I would recommend a good quality capacitor rated at least for the local mains supply.

A professional would never do this.

 

[ericgibbs]

A professional would never do this.

I dont understand your post.

Look up flame proof resistors also the UL regulations.

 

[poodlenuggets]

Why are you bothering with safety things like flameproof resistors? You have already resigned yourself to a very foolish design and if you are going to do it right, then do it right!

I'm happy to hear any criticism, but please explain it so I can learn something new. I abandoned this design for a long time because I was concerned with what I saw as the main safety issue - it's not isolated from the mains.

My mind was changed when I found that 2 of our competitors use this design. They saved a lot of money and a lot of board space with it, and they both have UL approval. At that point, I thought it would be foolish of me to ignore it. I came to this forum to make sure I design our PSU as safely as possible.

 

[ericgibbs]

I'm happy to hear any criticism, but please explain it so I can learn something new. I abandoned this design for a long time because I was concerned with what I saw as the main safety issue - it's not isolated from the mains.

My mind was changed when I found that 2 of our competitors use this design. They saved a lot of money and a lot of board space with it, and they both have UL approval. At that point, I thought it would be foolish of me to ignore it. I came to this forum to make sure I design our PSU as safely as possible.

hi poodlenuggets,
Most, if not all domestic PIR units use the transformerless mains supply.

Provided the electronics are fitted into an totally enclosed housing they work just fine.

My personal concern is where I think a 'hobbyist' is planning to use one, obviously you know what you are doing and whats required by UL.

Good Luck with your project.

 

[poodlenuggets]

hi poodlenuggets,
Most, if not all domestic PIR units use the transformerless mains supply.

Provided the electronics are fitted into an totally enclosed housing they work just fine.

My personal concern is where I think a 'hobbyist' is planning to use one, obviously you know what you are doing and whats required by UL.

Good Luck with your project.

Thanks! And that's actually a great tip - when you say PIR units, you mean passive infrared detectors, right? I should pick a couple up to see more examples. As MrAl said, they're also in LED nightlights - are there any other devices that commonly use this type of power supply? Perhaps any consumer electronics device that uses low power.

 

[Optikon]

I'm happy to hear any criticism, but please explain it so I can learn something new. I abandoned this design for a long time because I was concerned with what I saw as the main safety issue - it's not isolated from the mains.

My mind was changed when I found that 2 of our competitors use this design. They saved a lot of money and a lot of board space with it, and they both have UL approval. At that point, I thought it would be foolish of me to ignore it. I came to this forum to make sure I design our PSU as safely as possible.

Keep in mind, your two competitors can still be sued into the ground if their designs kill somebody. UL doesnt mean squat if the design is inherently dangerous and shortcuts can be pointed out (no transformer etc..).

After reading your initial posts, I saw no indication that this is what you were after. For all I know, you are some cheapskate tinkerer who will get killed trying to save some $$$ at the expense of safety. So you are making a "me-too" product?


Your product will be tested to regulatory standards like mains surges and EFT etc.. those kinds of things right? You came to this forum (mostly noobies these days) looking for safety-design advice? Doesn't your company have an expert or committee? My advice to you is hire a consultant.

 

[MrAl]

Hi again,

As Optikon pointed out, UL is not the end all of end's all. For example, as far
as i know they are still passing 16 gauge cords on 1500 watt heaters, and while
this works when the unit is new, any aging effects quickly diminish the cord to
a smoking, melting pile of junk. I recently had to replace a 120vac plug for
someone with just such a heater, and the cause was that the 16 gauge cord
wore such that some of the strands internal to the wire were broken. This caused
heat to melt the 120v plug that came with it, but it could have been worse.
In other words, while the 16 gauge cord works there is no margin of safety
for when the unit ages and gets used by humans in a way that they typically
do things.

There is the other side of the coin too though. When 120vac (or other line voltage)
is used to power a device there is immediately a shock concern for the user in
that if it is used wrong it could cause a fatal shock. For example, dropping it
into water like a bathtub could be a big problem. Yet, we dont have to water
proof every device that connects to the 120vac line to make it safe.

In the case of many devices they are already connected to the 120vac line
anyway, so if they are used wrong they will cause a fatal shock, even if
design perfectly and with an internal double isolated isolation transformer.
The key idea is whatever we do we want to make sure that the user can
not come into contact with 120vac during 'normal' and even a bit of 'abusive'
use. In the case of the off line power supply (being discussed in this thread)
this means the user can not be allowed to come into contact with any part
of the conductive part of the circuit, while with the isolated power supply
circuit we only have to make sure they cant come into contact with the
primary side. Still, there are often other concerns too, such as signal
noise injection by a human finger for example. We dont want a good audio
amp connections exposed where the user would be able to touch them
with normal use so they end up injecting lots of hum into the system.
So even with isolation sometimes we are paying attention to what can
be and what can not be touched, and taking the appropriate steps to
take care of this.
With the off line power supply (no transformer) if we take the right steps
to ensure safe operation we can end up with a very successful design
just as with an isolated power supply design.
I had designed several devices including night lights, but also
speed controls ranging from a few amps up to 25 amps AC, and they
all run off of an off line power supply. They are completely safe to
use barring crushing the whole case and exposing live terminals.
Many times the cases are made of durable plastic too, so even if
crushed they may not expose live terminals.

The trick with off line power supplies then is to make sure the user
can never come into contact with the circuit, and that means any
part of that circuit. This might exclude power supplies made for
testing other devices because we often use aligator clips and such
for connections, and that means most likely at some point someone
will touch the metal on one and get a shock.

Just to note, even though i have built many off line power supply driven
devices in the past, all of my test power supplies are run from wall warts
or other types of isolating transformers. Many wall warts are especially
safe because they are double isolated, meaning if one thing goes wrong
with the wall wart the second means of isolation may still keep the line
ac from contacting the output leads.

Good luck...

P.S.
Another example of an off line powered device is a lamp dimmer.

 

[Hero999]

hero,
Look at this extract from a microchip application note:
Posted earlier.

Connecting the neutral to ground through a fuse?

That sounds like rubbish to me.

I don't know about the US electrical code but the IEE regulations state that the neutral should never be connected to ground by anything other than a filtering capacitor.

Sorry I still don't believe it, there must be an error in the datasheet. I agree that there should be a resistor in parallel with the capacitor and in series with the capacitor. I just don't think the series resistor needs to be such a high value, as long as it limits the surge enough to prevent the fuse from blowing or the components being damaged then it doesn't matter.