230 Volts LED

[RODALCO]

I did read in one of the older posts a request for a simple circuit for a 230 Volts LED.
Well here it is:

(background info)
Here in the local powerboard this circuit has been used for about 15 years with no problems for potential indicators in CT meters.
With the new high efficiency LED's only about 3 mA is required for a good light output hence not much energy is wasted in the series resistor.
The last 5 years we use 2 * 33 or 39 k Ohm 1 Watt resistors in series and a 1N4007 diode to get the running current to around 2 to 3 mA.
A 1N914 or 1N4448 is used in antiparallel with the LED.

Previously we used 47 k Ohm 4 Watt resistor in series with the older type of LED's.
Capacitors of 22 and 47 nF 1 kV were tried too, but spikes on the mains sometimes caused the LED's to fail. For reliability on utility meters to check for blown PT fuses the resistive method was preferred.


**broken link removed**

Blue LED with 2 * 39 k Ohms , Red LED with 2 * 33 k Ohms.
These are standard 5000 milli candela LED's.

**broken link removed**

Blue and Red LED from different angle.

Of course all mains exposed wires require proper insulation heatshrinks fitted around the exposed wiring ends for safety.

 

[fever]

coool dude.
can u upload the schematic so that we can understand its working in a better way.
and u said that u used them extensively.so can u tell which part getiting heatup,is it giving fumes etc....

 

[fever]

current 3ma?????
i think it shld be 25 or 30 mA?

 

[eblc1388]

If an anti parallel diode is already connected across the LED, then the 1N4007 diode serves no useful purposes.

Alternatively, there is no point to fit an anti-parallel diode if a 1N4007 is used in series as the current is limited to uni-directional by the 1N4007.

 

[RODALCO]

Schematic is as you can see the parts

{Phase}{39k Ohm}{1N4007}{LED} then {1N914}{39k Ohm}{Neutral}

I always fit an antiparrallel diode in case the 1N4007 didn't get used for whatever reason.
Also the 1N4007 is rated for 1 kV while the LED is only rated at 5 or 6 Volts reversed polarity.

I fit the 1N4007 to reduce the power in the series resistors to 0.707 * power.

20 mA at 230 Volts is 4.6 Watts in heat dissipated in the resistors, for an optical indicator less than 1 Watt is better, and with the high efficiency LED's you don't notice much difference in light output between 3 and 20 mA.

These LED's I have used in many 1000's of CT watthourmeters and lately fitted in substations for optical indications for streetlight and hotwater circuits.
The key here is longevity and reliability.

 

[RODALCO]

**broken link removed**

Here is a panel from the Henderson Valley NZ substation for Hotwater and Streetlight controls where these 3 mA LED's have been fitted about 2 years ago.
The two blue LED's off are because of blown fuses on the streetlight pilotwires and currently controlled via an outside lightcell till the cables are repaired.

 

[AllVol]

Amazing, isn't it, how we sometimes quickly deny 15 years of reliability?

 

[fever]

the zener diode(1N914)is connected across the LED right?

**broken link removed**
AM i right?

 

[Dr.EM]

The 1N914 is a silicon signal diode, not a zener.

 

[eblc1388]

Amazing, isn't it, how we sometimes quickly deny 15 years of reliability?

Open discussion is how I learned from others and I have been proved wrong many times.

I have made two statements in the above postings. If you have better ideas or disagree to one or both of them, please shows your reasoning so we can all learn as well.

 

[Roff]

Open discussion is how I learned from others and I have been proved wrong many times.

I have made two statements in the above postings. If you have better ideas or disagree to one or both of them, please shows your reasoning so we can all learn as well.

Well, if one wanted to nit-pick, one could point out that the 1N4007 saves power in either case. Probably not a big reliability issue.

With the 1N1007, but without the antiparallel diode, the LED will break down due to capacitive current during each cycle. However - the reverse current is dominated by the spike that occurs when the mains voltage goes through zero, and 1N4007 zero bias capacitance is about 20pF, so this will yield a current spike of only about 4uA for a few microseconds each cycle - not enough to damage the LED, I would think, unless there is some sort of degradation due to reverse current over and above power dissipation,which will be negligible.

 

[eblc1388]

Well, if one wanted to nit-pick, one could point out that the 1N4007 saves power in either case. Probably not a big reliability issue.

If saving power is a concern, then one would normally use a capacitor instead. In this case, a series 1N4007 diode would not work.

With the 1N1007, but without the antiparallel diode, the LED will break down due to capacitive current during each cycle. However - the reverse current is dominated by the spike that occurs when the mains voltage goes through zero, and 1N4007 zero bias capacitance is about 20pF, so this will yield a current spike of only about 4uA for a few microseconds each cycle - not enough to damage the LED, I would think, unless there is some sort of degradation due to reverse current over and above power dissipation,which will be negligible.

The 1N4007 is in series with the LED, which itself is also a diode. So there are two capacitors in series. If there is any reverse current spike via the capacitance of the 1N4007, it might as well go through the LED also as capacitance current in stead of breaking the junction down and flow as leakage current.

However, I agree the situation becomes more involved after the reversed voltage exceeds the breakdown voltage of LED junction. One then has to "simulate" the circuit in order to find out what happens next.

 

[fever]

what do u guys think at this one
http://www.marcspages.co.uk/tech/6103.htm

 

[RODALCO]

The reliabilty issue is here the most important.
We had in 1990 around 200 kWh meters with potential indicators with 22 nF and 47 nF 1 kV capacitors.
Several LED's failed or the 1N914 failed either due to inrush current or power spikes.
So the resistive circuit was preferred because over around 2000 Ct meters we have had NO faillures of the LED's and most of these converted meters are still in service.
I make a photo of the very simple schematic because I don't know how to put it on this forum via XL.
The power usage is NOT an issue here but for the sake of soldering in a 5 cts antiparrallel diode across the LED, an extremely reliable mains indicator is available for under 1 NZ$ or 0.63 US$.

fever cool circuit, have tried that too but it is more work.
The 50 Hz flicker across the LED is hardly noticable and not an issue as far as i'm concerned with a mains indicator.

as Ron H explains that is exacltly why the 1 N914 or any other type of diode can be fitted across the LED.
The 1N4007 saves a little bit of power, hence less heat in the series resistors.

 

[eblc1388]

The reliabilty issue is here the most important.
We had in 1990 around 200 kWh meters with potential indicators with 22 nF and 47 nF 1 kV capacitors.
Several LED's failed or the 1N914 failed either due to inrush current or power spikes.

Is there a resistor in series with the capacitor?

Failure is inevitable and expected if there is no current limiting resistor in series but just one single capacitor being used. One would still need to use a resistor of several K in series with the capacitor to prevent failure.

 

[RODALCO]

We used a 470 Ohms ¼ Watt resistor in series with the Capacitor, this one sometimes ruptured by the inrush current. Didn't try the 1 Watt Resistor which may perhaps work ok.

here is the schema, photo taken of hand drawn schema
where

U is 230 Volts 50 Hz
R1 and R2 are 33 or 39 k Ohms 1 Watt resistors
D1 is 1N4007 (1kV 1Amp diode)
D2 is 1N914, 1N4448, any rectifier diode of 50 mA or more will do here.
D3 is high efficiency LED ( 3000 to 5000 mcad or more ) any colour.

**broken link removed**

 

[fever]

thanks dude.
i'll definetly try this one....

 

[fever]

last one..
can i keep this in a closed space?
imean if it get(resistor) heatup forming fumes,then it might create problem

 

[RODALCO]

Heat up of series resistors is negligible as long you stick to the values quoted in diagramme.

 

[Hero999]

D2 doesn't do anything, because D1 is already blocking the negitive cycles.