EMI filter for 24VAC

[boylesg]

Can some one give me an idea if an EMI filter normally used for 240VAC would work for 24VAC?

And if not how would I go about designing one for 24VAC?

 

[D.A.(Tony)Stewart]

You don't start until you define some measurable requirements.

 

[Easy peasy]

Yes it would - but unless you have a scan result of the filter performance - you are shooting in the dark.

 

[D.A.(Tony)Stewart]

Filters work on impedance ratios as do transformers except 24024 is 10:1 for V but 100:1 for Z. That means the impedances will be expected to be higher so beware of current rating.

 

[betwixt]

... and any over voltage protection that clamps above say 275V AC is unlikely to be of much use at 24V but if it is employed, you can change to lower clamping voltage devices.

Brian.

 

[boylesg]

Is there an online calculator that any of you can recommend? To give me start in this new endeavor for me.

 

[boylesg]

... and any over voltage protection that clamps above say 275V AC is unlikely to be of much use at 24V but if it is employed, you can change to lower clamping voltage devices.

Brian.

So these things are common mode filters right?

With a separate inductor for each AC line wound on to the same core.

I have seen these on tv circuits boards.

The circuit schematics are simple enough. But knowing the correct values is the hard part.

OK, so if these things generally expect 240VAC then, the inductance of those two coils will be set accordingly.

If I want to use 24VAC then I assume I will have to reduce the inductance of the coils and therefore the impedance.

And I have a spread sheet calculator to calculate the correct inductance value for 24VAC, given 2A max for this particular board. Or whatever current I want.

Am I OK so far guys?

So either I buy one of these boards, pull the inductors apart and re-wind them. Or I just make one of these from scratch.

Am I still on the right track?

As for the voltage rating for the caps it doesn't really matter in my case. As long as it is 50V or more.

The values will be the same I am guessing?

EMI Filter Module AC Power Interference Sound Enhancer Socket Reduce Noise | eBay

Find many great new & used options and get the best deals for EMI Filter Module AC Power Interference Sound Enhancer Socket Reduce Noise at the best online prices at eBay!

www.ebay.com.au

--- Updated Oct 6, 2021 ---

The schematic for the above board will simply be this:

 

[KlausST]

Hi,

such a filter has two functions:
* PASS signals. This is 24V AC for your connected device to operate.
--> We don´t know how much current your device draws nor what voltage drop is allowed.
* ATTENUATE signals. High frequency. We don´t know which frequency range of noise you expect (or the frequency range that you want to be suppressed). And we don´t know the HF volotage levels. Neither the expected noise level nor how much you want it to be suppressed.

***
Let´s compare the filter with a bungee rope.
* on the one hand the rope needs to be strong enough to carry your weight,
* on the other hand it needs to be elastic enough to softly decelerate the speed ... without hurting you.
So it depends on a lot of things if the bungee jump is "successful", like: Jumper´s weight, jumping height, rope length, rope elasticity, maximum rope force rating....

Klaus

 

[boylesg]

Hi,

such a filter has two functions:
* PASS signals. This is 24V AC for your connected device to operate.
--> We don´t know how much current your device draws nor what voltage drop is allowed.
* ATTENUATE signals. High frequency. We don´t know which frequency range of noise you expect (or the frequency range that you want to be suppressed). And we don´t know the HF volotage levels. Neither the expected noise level nor how much you want it to be suppressed.

***
Let´s compare the filter with a bungee rope.
* on the one hand the rope needs to be strong enough to carry your weight,
* on the other hand it needs to be elastic enough to softly decelerate the speed ... without hurting you.
So it depends on a lot of things if the bungee jump is "successful", like: Jumper´s weight, jumping height, rope length, rope elasticity, maximum rope force rating....

Klaus

The digital electronics is going to draw up to 1A perhaps. Or not much more than that.

A 20 x 4 LCD screen, WiFi, and relay board.

Also Arduino Mega, data logger shield, HC-05 - all low current stuff.

Let's say 1.5A at 5V = 7W

May be 1A max at 24VAC (something like that).

The 24VAC solenoid valves draw approximately 250mA when on.

--- Updated Oct 6, 2021 ---

The digital electronics is going to draw up to 1A perhaps. Or not much more than that.

A 20 x 4 LCD screen, WiFi, and relay board.

Also Arduino Mega, data logger shield, HC-05 - all low current stuff.

The 24VAC solenoid valves draw approximately 250mA when on.

As for frequency - well all the AC line filters in TVs and laptop power supplies must aim to block a similar range of frequencies likely to come in through the wall sockets.

I was pondering this while having a shower and I think I get why I am misunderstanding this whole thing.

I am thinking about the inductors and caps in isolation rather than what type of electronics construct they are forming.

The mylar caps and the inductors have got to be forming a low pass filter right?

They are just dual low pass filters for each AC line.

That have low impedance for 50Hz coming in from the wall socket.

But high impedance for frequencies above that, so that they are greatly attenuated.

So what I need to do is find a low pass filter calculator and calculate the values for low impedance at 50Hz.

Now have I got the right idea?

 

[KlausST]

Hi,

you said you want to use this filter on 24V AC.
Now you talk about LCD screen, WiFi, Arduino, relay, solenoids...
--> I doubt that devices run from 24V AC.
So you hide some information.

Then you talk about valves and a relay board.
--> We don´t know how many valves and how many relays there are.. and how many are active at the same time.

Indeed it´s less important for us to know all this. It´s more important that you know how all this is related.

Especially for you, I recommend to use a sheet of paper and a pencil and make a sketch how all is connected and at which point you expect which current, noise and so on.

For sure you are free to show us your sketch.

Klaus

 

[boylesg]

Here is my schematic:

 

[scopeprobe]

Is there an online calculator that any of you can recommend? To give me start in this new endeavor for me.

Wurth Electronik has a EMI filter design tool on their weksite. You may need to register first.

https://redexpert.we-online.com/redexpert/

 

[boylesg]

Wurth Electronik has a EMI filter design tool on their weksite. You may need to register first.

I found this to save me the hassle of registering:
electronicbase.net/low-pass-filter-calculator/

LC low pass filter calculator.

How far above 50Hz (mains AC) would you recommend going?

 

[scopeprobe]

I found this to save me the hassle of registering:
electronicbase.net/low-pass-filter-calculator/

LC low pass filter calculator.

How far above 50Hz (mains AC) would you recommend going?

I don't think the filters on that site will work as a Line Filter as they assume no load impedance so are likely to interact with your circuit. For an EMI filter you have to specify your impedance of your load, know the frequencies your trying to filter, and the type of noise otherwise the effects will be undersirable to say the least. 50Hz is the line frequency which far away from digital/switching circuitry noise. You really need to know what frequencies are generating to understand what needs filtering. Looking at your circuit the biggest source of noise will be the DC/DC conveter which will have high energy dv/dt nodes and probaby operates @ >100Khz. Noise will likely be at the switching frequency upwards so as far as filtering in concerned knowledge of this and the amound of noise its generating is mandatory for EMI. As a start if i were guessing i'd assume differential emissions will be the target for the filter you could try a filter pitched at 20db Attenuation 100khz with a 2A load as a start. Its highly like the converter is higher than this but the attenuation will be greater at higher frequencies. If any issues are common mode then this won't help.

L = 120uH
C = 470nF

As its an AC circuit you'll get about 100mW Reactive Watts. Outside of this you really need to supply data sorry.

 

[KlausST]

Here is my schematic:

what about the other information?

Klaus

 

[dick_freebird]

Now that you have mentioned solenoid valves, presumably driven by in-box sources and switches, I think you want to think more about what-all will be happening EMI-wise on the right hand side of that filter. Ground and power spikes etc. need a short wide path back home, current loops separated and explicitly set, and s on.

 

[boylesg]

what about the other information?

Klaus

I don't understand Klaus. What other information specifically.

 

[KlausST]

Hi,

You ask for a filter, but your sketch is missing the filter at all.
You ask for a filter (which is a solution), but you did not mention the problem you want to solve with the filter.
Is it for passing an EMV test? Or do you want to solve a malfunction problem? Something else?

I asked for the expectable currents and frequencies.

Currents:
Now there is a DC/DC converter. You did not mention it before.
Depending where you want to place the filter, the "2A" may be valid or not. The filter needs to be calculated for the current that runs through it.

Frequencies:
What "frequency" do you think is the problem?
* any frequency that is generated by the "digital electronics"?
* any frequency that is generated by the DCDC converter?
* any frequency (switching noise) that is generated by the relays and solenoids (sparks)?
* now that you showed a rectifier ... are you aware that a rectifier generates noise?
And are you concerned about "common mode noise" or "differential mode noise"

****
I fully agree with post#2.
And in post#6 you ask about a calculator. But every calculator needs values to be input.

So my idea is:
* if you want properly calculated values, you need to give input values first.
* if you don't have these values, one can not calculate filter part values. Neither one of us, nor a calculator. Then it's "guessing".
For me "guessing" is the same as "trial and error". While the one is a theoretical approach, the other is a practical approach.

Btw: someone recommended L and C. I can not say whether the values are useful or not. But generally I don't like undamped LC filters, because they generate resonances. Resonance may cause (uncontrolled) high resonance frequence amplitude. Be sure the resonance is damped somehow.

Klaus

 

[boylesg]

I don't see how it can be anything to do with my board.

Hi,

You ask for a filter, but your sketch is missing the filter at all.
You ask for a filter (which is a solution), but you did not mention the problem you want to solve with the filter.
Is it for passing an EMV test? Or do you want to solve a malfunction problem? Something else?

I asked for the expectable currents and frequencies.

Currents:
Now there is a DC/DC converter. You did not mention it before.
Depending where you want to place the filter, the "2A" may be valid or not. The filter needs to be calculated for the current that runs through it.

Frequencies:
What "frequency" do you think is the problem?
* any frequency that is generated by the "digital electronics"?
* any frequency that is generated by the DCDC converter?
* any frequency (switching noise) that is generated by the relays and solenoids (sparks)?
* now that you showed a rectifier ... are you aware that a rectifier generates noise?
And are you concerned about "common mode noise" or "differential mode noise"

****
I fully agree with post#2.
And in post#6 you ask about a calculator. But every calculator needs values to be input.

So my idea is:
* if you want properly calculated values, you need to give input values first.
* if you don't have these values, one can not calculate filter part values. Neither one of us, nor a calculator. Then it's "guessing".
For me "guessing" is the same as "trial and error". While the one is a theoretical approach, the other is a practical approach.

Btw: someone recommended L and C. I can not say whether the values are useful or not. But generally I don't like undamped LC filters, because they generate resonances. Resonance may cause (uncontrolled) high resonance frequence amplitude. Be sure the resonance is damped somehow.

Klaus

I don't see how it can be anything to do with anything on my board Klaus.

I have tried two new boards with all new components and the same thing has happened. Either the SD card reader fails and cannot initialise the SD card. Or the ESP13 shield fails in that it cannot be initialised and my software can't communicate with it.

This has only happened consistently at one location.

I have the same hardware in 3 other locations, and here at my home. And I never have this problem.

So it has got to be something external to my board surely?

But I have no clue where to start, other than it is particularly strong EMI interference coming in through the wall socket.

And I lack the necessary filter in my hardware to stop it.

Where do you even start with something like this?

--- Updated Oct 10, 2021 ---

Consumer electronics never seem to have this problem, no matter where they are plugged in.

So surely there is a a fairly standard filter system that manufacturers employ on the AC power supply inputs of their consumer electronics?

 

[D.A.(Tony)Stewart]

I repeat what I said on comment #2

You don't start until you define some measurable requirements.

Noise is bidirectional and can conduct and radiate therefore couple by proximity due to capacitance due to fast switching regulators or flyback dV/dt or inductive from dI/dt currents. The most common newbie problem is two different systems separately powered by DCDC converters connected by a data cable with ground circulating currents affecting communication on the cable. This happens frequently.

So start with a block diagram showing modules of all interference signals vs power flow from AC power & PE ground to DC power & 0V return with physical lengths of cables and orientation. Worst case is parallel nearby for crosstalk and best case is shielded twist pair at right angles (90deg) w.r.t each other.

So where is your block diagram and some measurement of results . e.g. ripple, SNR, ok /not ok

A simple but accurate block diagram of signal and power is a good start with path of interference