Do you have an EMC probe in your toolkit? Probably not, unless you’re in the business of electromagnetic compatibility testing or getting a product ready for the regulatory compliance process. Usually such probes are used in anechoic chambers and connected to sophisticated gear like spectrum analyzers – expensive stuff. But there are ways to probe the electromagnetic mysteries of your projects on the cheap, as this DIY EMC testing setup proves.
As with many projects, [dimtass]’ build was inspired by a video over on EEVblog, where [Dave] made a simple EMC probe from a length of semi-rigid coax cable. At $10, it’s a cheap solution, but lacking a spectrum analyzer like the one that [Dave] plugged his cheap probe into, [dimtass] went a different way. With the homemade probe plugged into an RTL-SDR dongle and SDR# running on a PC, [dimtass] was able to get a decent approximation of a spectrum analyzer, at least when tested against a 10-MHz oven-controlled crystal oscillator. It’s not the same thing as a dedicated spectrum analyzer – limited bandwidth, higher noise, and not calibrated – but it works well enough, and as [dimtass] points out, infinitely hackable through the SDR# API. The probe even works decently when plugged right into a DSO with the FFT function running.
Again, neither of these setups is a substitute for proper EMC testing, but it’ll probably do for the home gamer. If you want to check out the lengths the pros go through to make sure their products don’t spew signals, check out [Jenny]’s overview of the EMC testing process.
[via RTL-SDR.com]
Would this method using the Spectrum Analyzer software for the SDRplay be much of an improvement? Seems better lower noise floor and frequency range capabilities at the least and a little more bandwidth. Neat.
Not sure what happened to my last post, but here it goes:
One may estimate the ppm drift after warming a 820T based dongle with gqrx for 20 minutes.
https://github.com/steve-m/kalibrate-rtl
(we wrote some simple scripts that automate this process for students)
This RTL-SDR scan tool does work for EMC projects:
https://github.com/EarToEarOak/RTLSDR-Scanner
Follow the calibration setup instructions for getting the best results by profiling your hardware. It does several sampling report modes, but I tend to prefer max-peaks over .25 sec per sample range when tracking an issue.
Mine was repeatable within ~2dB, accurate within a constant 5dB offset from the lab chamber straight out of the box (fixed gain, no external LNA or AGC), and less then 2.4kHz off marker for most of the spectrum. Our test unit had a regular xtal with a 17ppm correction, and not that better TCXO version in this article.
For around $10, the RTL dongles provide excellent value… and I still use it before exposing equipment to unknown RF sources. ;-)
73
Very nice tips, thanks.
The spectrum analyzer functions in DSOs can do a very good job of checking SMPS noise. But don’t forget an E field probe. A bit of heat shrink to prevent shorts on the tip of a scope probe with the grabber and ground wire removed is quite amazing. I’ve not yet tried a small toroid cut in half, but I expect it will also prove exceedingly useful at isolating noise sources to particular traces.
Thats a lot of coax without being amplified at the end for something like an EM probe
You can get dirt cheap 5mhz – 3ghs amplifiers off ebay and the RTL-SDR Blog SDR can be configured with a bias T
but i do have to question the usefulness of an EM probe that can only go down to 27mhz
it would be interesting to break out my HackRF and pull an amp from my pile of RF stuffs and see if i can make a good EM probing setup but id have nothing to compare it to.
All your questionsions are answered in the blog post.
There you’ll see that the preferred setup is to use the usb extension and the probe connected straight on the dongle, instead of using the antenna extension. Also the first setup worked a bit better for me because I’ve used the dongle as a handle.
The RTL-SDR V3 can go lower than 27MHz and down to KHz range without any hardware modification, by just switch the ADC input from the tuner to the antenna. This is called direct sampling mode and for the V3 is just a click on the sdrsharp settings.
if i remember right the direct sampling mode in the R820T2 down converters is a very lossy harmonic based method and would be filled with reflections and is something that should be used with external filters while also not retaining 50 ohms
seeing how it would probably create crazy images from the midpoint (14mhz i think) you could have switching EMC and clocking EMC all jumbled together
i guess it would be a fun tool to poke around and maybe get a rough idea of where to probe if you have crosstalk on a long trace on a large board but using this for pre-regulation stuff to me sounds like a bad idea
I did the same thing, very useful tool to have, works like a charm. Though I would really advise to cover the probe in something – epoxy, lacquer, duct tape, anything will do. It allows safer handling – you can accidentally touch the device/board you are probing without shorting anything. I learned this the expensive way :-(
I’ve used a black rubber coating paint for that (there is a photo in the post) and definitely it’s not meant to be used without coating, otherwise as you said bad things will happen…
https://github.com/AD-Vega/rtl-power-fftw is nice for getting power spectrum, but it’s still pretty slow for trying to interactively find a noise source.
Maybe it’s me, but I’m having a little bit of hard time finding semi-rigid cables such as those. I need it for antenna tuning on a board I made, to hook up a VNA. Anyone have a link? Prefer eg Mouser or digikey since it’d be much faster. Yet, I seem to only find semi-rigid cable on very large spools, not small 15-30 cm pre-made ones with SMA connectors. Thanks!
For short lengths miniature cables will support their own weight just fine. I have used SMA pigtails like this https://www.nooelec.com/store/male-mcx-to-male-sma-pigtail-rg316-0-5-length.html and cut off the end that I don’t need.
When I first saw this in the tipline I ordered a semi-rigid 30-cm cable off eBay – https://www.ebay.com/itm/223304942483
It’s one of those “slow boat from China” things, might get here sometime in May. Or June. But it was only $3, and I don’t have the time to play with it right now, so I’m fine with that.
And like magic, the slow boat arrived! Got the package via USPS yesterday, mere minutes after I posted this message.
Now all I need is a decent SDR dongle.
https://www.digikey.com/product-detail/en/taoglas-limited/CAB.058/931-1391-ND/5287277
I know what an EMC is but could the writers take a moment and at least write it out once? We all see lots of letters that mean lots of things and if your only job is to write a summary paragraph or two from existing text and documents, it seems like an easy enough thing to do. It also helps in feeds which will get you more clicks and dough. No-brainer.
You’re right, it is pretty easy to do. That’s why I defined EMC in the second sentence:
“Do you have an EMC probe in your toolkit? Probably not, unless you’re in the business of electromagnetic compatibility testing or getting a product ready for the regulatory compliance process.”
I prefer this style to the parenthetical explanation of acronyms and initialisms. I think it’s more subtle and gives the reader more credit.