Help me to solve AC current leakage out of the SMPS.

[papaisou11]

Greetings to all.

Lets discuss about my problem -

I am making a power supply to power Audio instrument having 12V 1.25A rating across the output.

Here is my circuit -

Now when I connect the AC it is working as expected but the problem is I am feeling AC shock of low voltage across the output. When I conenct a neon tester it is glowing on both the outputs.

What I have done, I have connected a Filter (3.3uH Choke) directly with the secondary RTN and the Earthing like this way -

Sorry for my poor image.

The Neon doesnot glow, and I do not feel any electrical shock like sensation in the above case.

I want to use this smps in locations where the Earthing will be a problem. So, is there any way to eliminate this issue without earthing. (I want to have earthing option as well).

Thanks a lot to all.

 

[Easy peasy]

remove c4

 

[papaisou11]

Yeah, but I hope it is done for EMI supression related issue. Is there any better option without removing this one? Or changng the EMI circuit that produce same effect without the C4?

Thanks a lot.

 

[mtwieg]

A 0.15nF should not cause enough leakage to be felt, or light up a neon lamp. Such capacitors usually go up to 2.2nF, even in medical equipment. Are you sure it's not faulty, or a higher capacitance? Have you measured the leakage current?

 

[KlausST]

Hi,

Connect C4 to secondary amd earth GND ...and connect a 1M in parallel.

Klaus

 

[cupoftea]

I believe C4 is needed for EMI reasons.
You should not be getting shocked with your top schem....you must have too much pri-sec capacitance in your txfmr...or in some way youve wound it wrong.

--- Updated Jun 26, 2021 ---

did you use triple insulated wire?
If not , did you use enough tape, and tubing, and margin layering?
How doe the txfmr behave on your hipot test?

 

[Easy peasy]

what is the capacitance of the transfomer pri to sec ? this should be easy to measure if you disconnect one leg of C4.

--- Updated Jun 27, 2021 ---

Also double check the value of C4 you have ACTUALLY used in the circuit, if it is 15nF - there's your answer ......!

 

[scopeprobe]

Greetings to all.

Lets discuss about my problem -

I am making a power supply to power Audio instrument having 12V 1.25A rating across the output.

Here is my circuit -

View attachment 170416

Now when I connect the AC it is working as expected but the problem is I am feeling AC shock of low voltage across the output. When I conenct a neon tester it is glowing on both the outputs.

What I have done, I have connected a Filter (3.3uH Choke) directly with the secondary RTN and the Earthing like this way -

View attachment 170417

Sorry for my poor image.

The Neon doesnot glow, and I do not feel any electrical shock like sensation in the above case.

I want to use this smps in locations where the Earthing will be a problem. So, is there any way to eliminate this issue without earthing. (I want to have earthing option as well).

Thanks a lot to all.

Try removing C4 and put a Y Capacitor from each side of C2 to Earth to act as a common mode filter which may help with your EMC issue. A CM Choke will give limited effect without the Y capacitors. Try also putting a Y capacitor from the secondary 0V to Earth to couple noise away too.

If you have a common mode output currect issue maybe consider putting CM filter on the secondary also

 

[FvM]

As stated by the OP, there's no earth ground (protective earth) available. That's the standard situation for most wall plug and many other small SMSP.

These devices have a certain leakage current through Y-capacitors. Removing Y-capacitors is no option if you want to comply with EMC standards. I agree with mtwieg that about 10 µA leakage of a 150 pF capacitors are hardly perceivable. Usual Y-capacitor level is rather 500 to 1000 pF. Self-evident that it must be safety rated.

 

[scopeprobe]

As stated by the OP, there's no earth ground (protective earth) available. That's the standard situation for most wall plug and many other small SMSP.

These devices have a certain leakage current through Y-capacitors. Removing Y-capacitors is no option if you want to comply with EMC standards. I agree with mtwieg that about 10 µA leakage of a 150 pF capacitors are hardly perceivable. Usual Y-capacitor level is rather 500 to 1000 pF. Self-evident that it must be safety rated.

Apologies he showed a connection to earth in his schematic so I assumed it was possible. I agree it would provide some leakage current.

Alternatively he could screen the windings of the transformer to reduce the primary to secondary capacitance and eliminate the issue at source

 

[Easy peasy]

pri and sec screens would be a good idea.

 

[cupoftea]

i believe Easy Peasy coined it in #7.....the wrong value y cap has been used...or its faulty and partially shorting etc.

 

[KlausST]

Hi,

What I miss is the PCB layout, it may cause resistive as well as capacitive leakage.

With the given informations I see two general options.
* reduce the leakage current (omit/reduce C4, optimized T1, modified EMI circuitry)
* give the leakage current a path to earth.

Klaus

 

[Easy peasy]

Also C4 must be Y2 or Y1 rated - if it is only 50VDC - you have a real problem ...

 

[papaisou11]

Wow... Lots of advices. Please give me time to reply one by one. Sorry I was inactive yesterday.

--- Updated Jun 28, 2021 ---

A 0.15nF should not cause enough leakage to be felt, or light up a neon lamp. Such capacitors usually go up to 2.2nF, even in medical equipment. Are you sure it's not faulty, or a higher capacitance? Have you measured the leakage current?

Hi... I have not used the 0.15nF, I have ordered few and let all of you know.
I have used 1nF 2kV Y cap, generic one, no partnumber since it is locally available. I have 1nF and 2.2nF both are showing same result. Changed several times, no effect what so ever.

Let me check removing the cap and what is the effect.

--- Updated Jun 28, 2021 ---

Hi,

Connect C4 to secondary amd earth GND ...and connect a 1M in parallel.

Klaus

I will let you know after trying this method. I am slow understanding technical things. What will the 1M in parallel will do? Thanks for your advices Klaus. You always helped me to solve problem. Thanks again.

--- Updated Jun 28, 2021 ---

I believe C4 is needed for EMI reasons.
You should not be getting shocked with your top schem....you must have too much pri-sec capacitance in your txfmr...or in some way youve wound it wrong.

--- Updated Jun 26, 2021 ---

did you use triple insulated wire?
If not , did you use enough tape, and tubing, and margin layering?
How doe the txfmr behave on your hipot test?

I have attached the transformer construction file. I have asked a transformer vendor to provide me the transformer as per spec. They told me it is done exactly waht is given in the spec. I have no idea about hipot test. Google gave me the answer. I suppose that it is done by the transformer manufacturer since they told me that the specification is asper given in the document.

--- Updated Jun 28, 2021 ---

what is the capacitance of the transfomer pri to sec ? this should be easy to measure if you disconnect one leg of C4.

--- Updated Jun 27, 2021 ---

Also double check the value of C4 you have ACTUALLY used in the circuit, if it is 15nF - there's your answer ......!

To measure the transformer capacitance I need LCR meter right? I have one but it is not with me and due to covid restriction it is quite impossible to get it right now. I am looking into this. I undertand that without getting proper instruments it is marely waste of time and no meaning at all. Lets see if I could manage the LCR meter and check this.

The C4 is 2.2nF or 1nF. I have checked in both cases, but sadly no improvements. Also, what is the best way to measure this AC leakage current to provide all of you a concrete number for undertanding this much better?

--- Updated Jun 28, 2021 ---

Try removing C4 and put a Y Capacitor from each side of C2 to Earth to act as a common mode filter which may help with your EMC issue. A CM Choke will give limited effect without the Y capacitors. Try also putting a Y capacitor from the secondary 0V to Earth to couple noise away too.

If you have a common mode output currect issue maybe consider putting CM filter on the secondary also

Yeah this is a good option, I will try this as well. But I have made the PCB, I need to solder this on the PCB. Let you know.

 

[KlausST]

Hi,

From your post we know that there is voltage and current from primary secondary (edited) to earth GND.
We don´t know if it´s purely AC, purely DC or a combination.

From your schemati we see possible paths for current form mains in to secondary side. These are:
* capacitive (purely AC) coupled via C4
* erroneously via T1
* erroneously via U2
* erroneously via PCB layout (dirt, humidity, wrong creepage distance, wrong clearance...)

As written you may reduce the unwanted currents or give those currents a path to flow (in a way that theycan´t flow through your body anymore)

C4 in your circuit clearly causes current one does not want. (besides the wanted EMI suppression)
Thus omitting C4 in your circuit reduces current.
Additonally connecting C4 to earth GND generates a path EARTH.
.. but this only is for AC currents.

By adding a 1M to EARTH you enable DC currents a path to earth. So static charge will be dicharged.

****

I really want to see your PCB layout. It may give much more information than you can guess now.

Klaus

 

[papaisou11]

As stated by the OP, there's no earth ground (protective earth) available. That's the standard situation for most wall plug and many other small SMSP.

These devices have a certain leakage current through Y-capacitors. Removing Y-capacitors is no option if you want to comply with EMC standards. I agree with mtwieg that about 10 µA leakage of a 150 pF capacitors are hardly perceivable. Usual Y-capacitor level is rather 500 to 1000 pF. Self-evident that it must be safety rated.

In INDIA there are lots of SMPS available with two plug pins. Unfortunately, while earhing is a standard safety practice, It is hardly available in every here and there places in INDIA.

I have checked JioFiber Router (A Router for 4G internet in INDIA) that uses 12V 2A three pin wall socket adapter. When I connected the adapter on wall outlet where Earthing was not available, there are no problems at all, I checked with the same neon tester. I could not brake it open since I only have one and it is the providers property.

My problem doesnot lies in situations where the Earthing is available. I am bothering about the situation where the Earthing is not available. Believe me 80% of people in INDIA do not use earthing in there home or offices and unfortunately whenever I need to play the guitar, I need to connect this power supply and almost all known places of that does not have earthing.

 

[KlausST]

Hi,

you are free to build your own "earthing" by connecting the secondary to a water pipe or anything else that carries EARTH potential.....
May not worke everywhere, maybe not the best solution ... but still worth a try.

Klaus

 

[papaisou11]

Hi,

From your post we know that there is voltage and current from primary to earth GND.
We don´t know if it´s purely AC, purely DC or a combination.

From your schemati we see possible paths for current form mains in to secondary side. These are:
* capacitive (purely AC) coupled via C4
* erroneously via T1
* erroneously via U2
* erroneously via PCB layout (dirt, humidity, wrong creepage distance, wrong clearance...)

As written you may reduce the unwanted currents or give those currents a path to flow (in a way that theycan´t flow through your body anymore)

C4 in your circuit clearly causes current one does not want. (besides the wanted EMI suppression)
Thus omitting C4 in your circuit reduces current.
Additonally connecting C4 to earth GND generates a path EARTH.
.. but this only is for AC currents.

By adding a 1M to EARTH you enable DC currents a path to earth. So static charge will be dicharged.

****

I really want to see your PCB layout. It may give much more information than you can guess now.

Klaus

Obviuosly, Without getting information about the PCB it is quite difficult for everyone.

PCB is new. Soldered and tested. No resudal of solder flux, properly cleaned.

--- Updated Jun 28, 2021 ---

Hi,

you are free to build your own "earthing" by connecting the secondary to a water pipe or anything else that carries EARTH potential.....
May not worke everywhere, maybe not the best solution ... but still worth a try.

Klaus

Such a great trick, never thought before. Worth a try.

 

[mtwieg]

Hi... I have not used the 0.15nF, I have ordered few and let all of you know.
I have used 1nF 2kV Y cap, generic one, no partnumber since it is locally available. I have 1nF and 2.2nF both are showing same result. Changed several times, no effect what so ever.

Are you saying that you feel a shock even when C4 is removed? If so, then something is very wrong here and I wouldn't use the supply until it's figured out.