EEVblog Electronics Community Forum
Electronics => RF, Microwave, Ham Radio => Topic started by: mike449 on January 29, 2024, 06:10:05 pm
-
It is easy to measure the differential noise on the output of a power supply. It is also relatively easy to filter out.
But the real problem is power supplies that present an EMI source between the mains side and the output side, which is common mode on the output. Especially if a longer than usual wire is coupled to the output side, creating a big dipole antenna radiating over wide frequency range. An example is a router with Ethernet cables, or a landline phone, or supply for a remote antenna preamp feeding it over a long cable.
I know it is a good idea to use old transformer based wall warts, but these have their own issues. So I want to be able to measure the noise generated between the mains and the output for any PS that comes my way, to see it it is suitable for use around my receivers, of needs common mode filtering to be added.
Is there a way to do this in a hobby setup?
-
A spectrum analyzer is the best tool for that.. The power supplies for ham radio use (for example) get sometimes a knob for tuning their switching frequency (and its harmonics) off your RX band of interest..
You may try with a cheapo SDR receiver as well - you will see the harmonics in the waterfall then.
-
I understand that I can look at the spectrum. The question is about the connection method. I am scared to hook up something to the mains (even the neutral). There should be some way to couple at RF without galvanic connection.
-
Below is a document on setting up for mains testing for EMI.
https://www.digikey.com/en/blog/an-introduction-to-emi-and-emc-measurement-of-power-supplies-part-1?utm_adgroup=&utm_source=google&utm_medium=cpc&utm_campaign=Pmax_Shopping_DK%2B%20Supplier_GEM%20Suppliers&utm_term=&utm_content=&utm_id=go_cmp-20940607809_adg-_ad-__dev-c_ext-_prd-_sig-CjwKCAiAtt2tBhBDEiwALZuhANgtpSmbtpLXFfneivxh0Yl4G2U3TcW7ZExegCxXQefPYWNAMuW4iBoC_54QAvD_BwE&gad_source=1&gclid=CjwKCAiAtt2tBhBDEiwALZuhANgtpSmbtpLXFfneivxh0Yl4G2U3TcW7ZExegCxXQefPYWNAMuW4iBoC_54QAvD_BwE (https://www.digikey.com/en/blog/an-introduction-to-emi-and-emc-measurement-of-power-supplies-part-1?utm_adgroup=&utm_source=google&utm_medium=cpc&utm_campaign=Pmax_Shopping_DK%2B%20Supplier_GEM%20Suppliers&utm_term=&utm_content=&utm_id=go_cmp-20940607809_adg-_ad-__dev-c_ext-_prd-_sig-CjwKCAiAtt2tBhBDEiwALZuhANgtpSmbtpLXFfneivxh0Yl4G2U3TcW7ZExegCxXQefPYWNAMuW4iBoC_54QAvD_BwE&gad_source=1&gclid=CjwKCAiAtt2tBhBDEiwALZuhANgtpSmbtpLXFfneivxh0Yl4G2U3TcW7ZExegCxXQefPYWNAMuW4iBoC_54QAvD_BwE)
-
Use a loop on your SA.
Connect the GND lead to the tip of a scope probe and connect it to your SA.
Put it near your wall wort.
You should see the spectrum emitted.
-
Use a loop on your SA.
Connect the GND lead to the tip of a scope probe and connect it to your SA.
Yep, that is how it works. Play with the scope's probe 1/10 setting, use a longer wire for lower frequencies (or make several turns)..
-
but the question was how to measure and not to estimate some undefined signals
-
You can get a good idea of unwanted signal on the line by wrapping a wire around the output several times and connecting Your SA to the wire wrap..You can also connect your SA directly to the output with a load with appropriate probe and get a magnitude.
The output of any SMPS interference will change with load however,
I try to separate Line Noise from Radiated Noise that comes from the inside of the SMPS and is radiated directly from there and cannot be filtered by output filter.. SMPS have input and output filters. You can have very little line noise and still have lots of interference.
Sometimes shielding helps.
All SMPS have noise. the SMPS for HAM radios are designed so that the noise does not interfere with the radio it was designed for. .Just because you see noise does not necessarily mean it will interfere with your radio.
-
On the theme of probes:
https://www.youtube.com/watch?v=RH9zHsdk5iE (https://www.youtube.com/watch?v=RH9zHsdk5iE)
-
You want "accurate" radiated emissions measurements eh?
Step 1: Build an anechoic chamber...
Step 2: Buy 100s of thousands of dollars of RF equipment... (or just merely many 10s of thousands of dollars of used equipment and then pay to have it recalibrated...)
Step 3: ...
Step 4: Profit!!
I kid. But even if you just want "approximate" radiated emissions measurements, you are probably looking at a legit pre-compliance setup which will still cost you many thousands of dollars of stuff (TEM cell, spectrum analyzer, etc) and will provide "some" results that you are then likely unable to independently verify if they are actually measuring what you think they are measuring.
You can also take your wall wart to a certified test lab. That will only cost you about $5k and an afternoon.
-
Below is a document on setting up for mains testing for EMI.
https://www.digikey.com/en/blog/an-introduction-to-emi-and-emc-measurement-of-power-supplies-part-1?utm_adgroup=&utm_source=google&utm_medium=cpc&utm_campaign=Pmax_Shopping_DK%2B%20Supplier_GEM%20Suppliers&utm_term=&utm_content=&utm_id=go_cmp-20940607809_adg-_ad-__dev-c_ext-_prd-_sig-CjwKCAiAtt2tBhBDEiwALZuhANgtpSmbtpLXFfneivxh0Yl4G2U3TcW7ZExegCxXQefPYWNAMuW4iBoC_54QAvD_BwE&gad_source=1&gclid=CjwKCAiAtt2tBhBDEiwALZuhANgtpSmbtpLXFfneivxh0Yl4G2U3TcW7ZExegCxXQefPYWNAMuW4iBoC_54QAvD_BwE (https://www.digikey.com/en/blog/an-introduction-to-emi-and-emc-measurement-of-power-supplies-part-1?utm_adgroup=&utm_source=google&utm_medium=cpc&utm_campaign=Pmax_Shopping_DK%2B%20Supplier_GEM%20Suppliers&utm_term=&utm_content=&utm_id=go_cmp-20940607809_adg-_ad-__dev-c_ext-_prd-_sig-CjwKCAiAtt2tBhBDEiwALZuhANgtpSmbtpLXFfneivxh0Yl4G2U3TcW7ZExegCxXQefPYWNAMuW4iBoC_54QAvD_BwE&gad_source=1&gclid=CjwKCAiAtt2tBhBDEiwALZuhANgtpSmbtpLXFfneivxh0Yl4G2U3TcW7ZExegCxXQefPYWNAMuW4iBoC_54QAvD_BwE)
Wow - I bet those 1kW resistors aren't cheap! :)
(https://www.digikey.co.uk/-/media/Images/Blogs/2022/February/An%20Introduction%20to%20EMI%20and%20EMC%20Measurement%20of%20Power%20Supplies%20Part%201/an-intro-to-emi-and-emc-measurement-of-power-supplies-pt1-img2.jpg?la=en&ts=5aee3674-fe9b-41a4-8f76-87c149a5fc9a)
-
If you need a 5uH LISN: https://www.eevblog.com/forum/projects/5uh-lisn-for-spectrum-analyzer-emcemi-work/msg3789893/#msg3789893 (https://www.eevblog.com/forum/projects/5uh-lisn-for-spectrum-analyzer-emcemi-work/msg3789893/#msg3789893)
...
Wow - I bet those 1kW resistors aren't cheap!
...
https://www.mouser.com/c/passive-components/resistors/?power%20rating=1%20kW (https://www.mouser.com/c/passive-components/resistors/?power%20rating=1%20kW)
-
Wow - I bet those 1kW resistors aren't cheap!
Just in case someone is reading this post literally, I'm pretty sure the "W" symbol is some weird font/unicode incompatibility, despite the "ohm" symbol being correctly displayed in the red text.
In other words, the resistors are 1 kilohm/1,000 ohm, not 1 kilowatt/1,000 watt.
-
Just in case someone is reading this post literally, I'm pretty sure the "W" symbol is some weird font/unicode incompatibility, despite the "ohm" symbol being correctly displayed in the red text.
In other words, the resistors are 1 kilohm/1,000 ohm, not 1 kilowatt/1,000 watt.
OK then, post edited to add a smiley to show I wasn't being serious... ::)
-
Wallace's suggestion to wrap some sensing wire around the output cable going to the load seems to work.
Not really a "measurement", but I can see the difference between a known good and a known bad wall warts, and changing EMI spectrum with changing load. I also see the change on HF when I add a 3m piece of wire to the ground on the load end. This tells me I am really picking up the common mode. Differential ripple is not that different between the 2 supplies (except the bad one is more erratic in frequency), but there is at least 30dB difference in the coupled common mode noise.
This method is good enough for a quick check.