Build your own antenna! There's a ton of ham radio literature about log periodic antennas. No doubt it'd be a challenge, but I bet you could get it done under 100 USD. I specifically recommend the ARRL Antenna Handbook.
the difficult part is not building the antenna, but measuring how its gain changes with the angle and frequency, which you need for the EMC measurements. For measuring your antenna, you need one of the echo-less chambers.
@@pyrokinetikrlz Right... was thinking can be done cost effectively other than tuning and standardizing with calibration certification or whatever the electronics test equipment industry uses.
Dave, I think you 'described' it without 'explaining' it. I always figured that a layman's explanation for the difference between near-field and far-field signals was that near-fields are the portion of the field that is close enough (at the speed of light,) to collapse back onto its source as the signal's source polarity reverses, while far-fields are the portion of the field that is not close enough to get back before the reversal, and thus is 'pushed away', and must continue on as free-space radiation, an effect that essentially turns an antenna into an electromagnetic-field 'pump'. This explains why longwave near-fields can extend a mile from an antenna, while microwave near-fields are centimeters or less.
Good one Dave. Being a physicist, I was gonna call you out on the H-field vs B-field thing but you included it in passing (which is plenty good enough!) 10:06 - nicely done!
Zeta Crucis That section of physics is one i utterly failed at. Any chance of telling me what the function of this microwave part is? ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-DZLMiUEVKNM.html
Hi Dave! Go for Schwarzbech, good antennas, and fair prices! But it depends if you want an antenna just to cover 30 to 250 or 300 MHz. In this case go for a biconical antenna. If you want to cover the band from 30 MHz to 1GHz with a single antenna, go for a so called hybrid antenna, the union of a biconical and log antenna, otherwise called bi-log. Happy EMCing!
It would be iteresting to understand the difference between near field (NFC) and far field devices. As you said in one of your videos, near field devices work only in close proximity and in magnetic fields more like transformator with air core. I like your tutorial type videos.
Howdy. Nice presentation. Some comments. Near field. The E and H components are 90 degrees out of phase. Almost all field energy collapses back into the antenna periodically. Far field. The E and H components are in phase. The energy has broken free and travels forward. The intermediate field or the Fresnel field. The phase shift is between 90 degrees and 0 degrees. Some of the energy collepses and the rest propagates. What is the mechanism that splits the components into collapsing parts and propagating parts ? Collapsing takes time. Field lines closer to the antenna have time to collapse. Field lines farther away do not have time to collapse completely they will encounter the build up of the next half wave field at the antenna. The ends of the field lines disconnect from the antenna and interconnect to form loops that travel forwads. Regards.
@@750kv8 Thank you very much sir / ma'am. Check this wonderful clip: ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-ZaXm6wau-jc.html I remeber from my college, eq. in Finland, days no teacher would explain what stuff is all about or what is the usefulness of stuff. Instead I just got a battery of math and theoretical physics stuffed down my throat. And was left bewildered. For myself I derive pleasure from dwelling on stuff until I am able to convey (in my mind) the point in clear language. Regards.
Thank you so much. I've seen textbooks that depict the E and M waves in phase, and other books that show them 90 degrees out of phase. I've asked physicists about that and never got a clear answer. I always thought it would make sense for them to be 90 degrees out of phase because then the energy of one field is at peak when the other is at zero, so the energy is converted from one to the other. But it also makes sense for them to be in phase at the antenna. You just wrapped it up, so thank you.
You were asking about lower freq antennas. Mil461 describes designs for biconical antennas that cover the lower end of the radiated field. Not overly difficult to make. I don’t believe they detail it also requires a 4:1 balun. The bigger wide range hyperlogs are practically a log periodic crossed with a biconical of course.
This "E field" probe is actually an electric potential (voltage) probe. Electric potential is a scalar and has no direction. In isotropic materials B and H are always proportional to each other. In fact it is quite difficult to find or to engineer materials having tensor magnetic susceptibility; materials having tensor electric susceptibility are numerous. In free space E=cB and B=mu H, but in materials mu is larger than 4pi X 10^-7. So whether we call the magnetic component of EM waves B or H is purely theoretical other than the proportionality constant with E.
@@redknight344 Well, you wouldn't say a light meter is really a voltage probe because it converts light into voltage, or that a bathroom scale really measures distance because it converts weight into linear deflection.
If you look at the datasheets of those expensive antennas you'll apreciate how much variation is on parameters like the antenna factor, antenna gain, directivity and so on. The reason for those prices, apart from high quality materials that imporve the performance, are the calibration data that you get with them in order to be able to calculate with accuracy radiated and received power (Basically same concept as you explained in your video about why Fluke meteres are so expensive). So basically you're buying parameters to fill in your Friis ecuation, and certifications needed if you want to be a EMC certification lab. I think you can build your own antenna but you'll end up with the same problem of having several cheap antennas that cover all the frequencies range, the only advantage I see is that you won't need to switch from one antenna to other while sweeping the entire band.
When testing in the chamber if we got within 3db of the limit or so, we would have to take the kit outside and retest at the points that got too close to the limit line. As reflections off the chamber walls could be making our test sample appear better than it is. This means you may o ly need to a few test frequencies out side. We used to get problems outside as the test site was 6 miles for a small city with its airport so quite a lot of rf noise. Also on damp days electric fences can be heard clicking away over a very wide spectrum. So even looking at a single frequency outside could take up quite a long time. And ho boy their time is a lot of money.
I am totally not a EM radiation hobbyist ... I have never heard abaut near field zone, where separate electric and magnetic fields are converging in to a point where they become an electromagnetic field/ radio wave. Very interesting info. Could you make more video tutorials about EM fields and radio wave basics? Maybe viewers could build up some kind of cool and usefull radio devices after these tutorials...
Dave, you have said in the past you're not a radio guy, but you should look into an antenna called the Quad Loop. Basically, a simple square(or any shape, really) antenna with a variable capacitor inserted at the feedpoint. I have built several of these(including the caps), and while the dimensions and cap values are important, they are not critical to function, as it's a tuneable antenna. Bonus: it has a fairly narrow(bilateral - dependent upon shape) gain pattern, so isolation is easier. One I manufactured for 20 meters(14MHz) measures only about 2/3 of a meter in its largest dimension. Getting down to 30MHz or low-VHF should not be a problem at all, and it would be even smaller. I used a simple SDR as a frequency domain analyzer to verify my tuning range, but I'm pretty sure you have one of those laying around the shop. As you tune the cap, you'll see the spike in the signals move around the frequency domain. Barring that, a coil of wire of the correct length or one wound around the correct ferrite rod will get you reception in high HF/low VHF. Or even lower - this is what most AM receivers use and I have seen those measuring only inches long; albeit with sometimes strange gain patterns, but it you put on in a Faraday Bucket(a cage with an open side), you may get good results. PS, I'm trying to get down to the site for the Solar Road demo in Baltimore and get some pics for you. I try to avoid that place, as it's all a "bad neighborhood"...
Great video Dave! Nice overview and good range of information. I've wondered how to make the complete test setup in the most cost effective way since I've done the SDLC work for lab test equipment and apps in pharma other than the RFI/EMI/EMC testing per the whatever the range of FCC and more standards are required in the World, CE/UL/etc.?. I still need to read into the range of standards.
I had to learn the basics about Nearfield effects of RF transmitters for MRI Safety. It's absurd that a Magnetic Resonance Imaging scanner, designed for imaging humans, can cause RF burns. If a patient has any part of their body within 5mm of the inside bore of the MRI scanner, nearfield effects are possible. This would be like building a car with holes on the floor, then blaming the driver then they lose a foot by putting it through the hole. Hoping to understand better from this video :)
Why do many other videos put lambda in the denominator? Distance can't be the reciprocal of a distance unit, but it's not a one-time typo, when a lot of lecturers say the same thing.
Back in the bad old days in Silicon Valley, I recall that VDE certification and testing was performed in an empty steel warehouse (probably originally intended for agricultural produce). It was an expensive process.
E fields are divergent, they radiate outwards from a source, so its like having a net. H fields are curl, they effect orthagonal to current vector, this is why orientation matters. Its kinda like measuring the rotation on intimately tiny hoolahoops
Not to be pedantic or anything :) but your hand drawing of E and H fields would be propagating in the other direction. The direction is given by the Poynting vector S = E x H. The animation which you placed over it (very cool by the way) had the direction correct.
Hi Dave, Would you please do a video about how a near field becomes a far field -- specifically the details of what is needed to "spit out" the energy rather than have it fold back in. (And maybe a word of how quanta and classical physics really throw a bum-wrench in the understanding of wireless transmission.) I feel like I should understand those particular things thoroughly, but I don't. I only fake it and have always let the equations do the work for me in the past. It would be really nice to have a better conceptual model of radio transmission. It's all about that mysterious place in the in-between --- the near-to-far transition. In any case, a Merry Christmas! to you and yours and may 2020 bring good fortune to you all.
You explained that the E field and the H field move 90 degrees relative to each other and later in the video that the H field is a round loop surrounding the conductor. For the E field I understand well perpendicular to the conductor, but how can the H field be at 90 degrees from E if it is a round loop as per your animation?
I was wondering this too. I guess the probe is single-ended (i.e. not differential), and it's just measuring the field with respect to ground, not the gradient. ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-S27zk_Gj9nU.html seems to confirm this.
Welcome to EMC hell !!, wait till you package it in your designed box, send it to the very expensive test house and it leaks like a siv, both radiated and conducted. keep up the excellent vids
Hello We can see the peaks with near field H probe on our board. How we can ensure that this peaks will effects our product compliance testing (CISPER22)which is tested at far field(>3m)?
Bit of a weird suggestion, but would it help if you immersed the antenna in a tank of distilled water? Seems to have a fairly consistent permittivity of about 80, and a relatively low loss factor at 100mhz (~5e-3) The impedance mismatch would lead to a ~4.4 dB loss due to reflection, but the wavelength in the water Is Lower by a factor 9 which would bring it close to the 30mhz you needed? Even tap wateright work, should have a skin depth measured in meters, which seems workable Don't the antennas in mobile phones to a similar thing to reduce the needed length, but with better dielectrics than water? Just a random thought, not an EE
Have you looked into magnetic loop antennas for 30MHz? I don’t think they’re commonly used for EMC hence nothing really being available, but they’re as easy to build as a mag NF probe. Not sure if they can result in an FCC/IC/etc compliant measurement though.
I'm sure you were saying something for the first 5 or 6 minutes of this video, but I was too distracted by that 4 CPU socket motherboard on the shelf behind you.
The "far field probe" is basically a regular antenna. I doubt that you can buy a cheap log-periodic antenna for HF (3-30 MHz). Firstly, because of the wavelength the antenna will be huge. Secondly, because these antennas are manufactured for amateur radio operators. This market is really crazy, a piece of wire with a ferrite core can easily cost 100-200 USD. If you need an HF antenna the cheapest and simplest option is to make one. There are many articles online that explain how to make a dipole, two element yagi and wideband variations of these antennas. From personal experience it takes one Saturday to make a dipole or two element yagi if you never tried this before. However you will need some free space, about 30ft of 22 AWG wire, a 10m fishing rod (two for yagi) and an antenna analyzer / SWR meter. The FAA-450 from Electroparts is a good product in terms of price/quality and it covers 0.5-450 Mhz. Also it's an open hardware project by EU1KY amateur radio operator.
The obvious answer is that they were calculated using the speed of light. But I have another doubt. The far field starts quite close according to Dave explanation. From the Antenna, HF and propagation knowledge one can get, that far field region depends on frequency and the size of the antenna, which is not mentioned here. And as already said - starts very close
Does Australia allow you to do your own certifications? It used to be allowable in the UK if done by a 'competent' person and using calibrated equipment and done to the correct standards. You still had to hire calibrated equipment and the process is very time consuming, so in the end might not be much savings as compared to having it done by a professional test house. I'm going back to the mid nineties here - I worked somewhere where they decided to self certify a custom built power supply for use in their own equipment. Due to EMC rules, it became very costly for the small independent electronics manufacturer to supply ready built equipment as everything needed to have EMC compliance in addition to complying with electrical safety standards.
I hope everything is okay out there, because of the fire. This is not good at all. Hoping everything will go out fine, I just saw that video with a koala and a fireman in the news. Love, Markus
4:25 You asked, these are also far-field antennas: www.saelig.com/product/M00154004.htm (1/8.33th of the price of the log antenna). Is this what you're looking for? Edit: they have an even smaller version nowadays: www.saelig.com/tem-cells/tbtc0.htm to test cellphone-sized devices. Edit2: random quote: "The wave travelling through the cell has essentially a free-space impedance (377 W) (sic), thus providing a close approximation to a far-field plane propagating in free-space."
Dave! Big fan here greetings from CA! Funny I was just watching your mini series on DIY NFP last week and built my own. Thought I'd let you know it worked GREAT! Quick question though, where did you get your enclosure? Does a shielded enclosure work better for these applications? Many thanks friend and keep up the great work!
Thanks for the explanation! So I guess, as an extreme example, when I build and run a ZVS flyback (which I have), most of the low voltage side circuit, with the MOSFETs and caps and chokes etc. mostly emit H-field, and the HV output of the flyback emits mostly E-field, when it's not arcing, just... held up into air like that... And I guess the output emits mixed fields, both E-field and H-field when arcing.
@@openyourmind7288 Why? Maybe with an ESD electric field tester perhaps if you wanted to measure a surface spot? The electric field has to be defined over a given surface area. You have to know what, how and why you want to do this first before answering that question.
It could be an interesting video of you trying to DIY a log periodic + bowtie antenna... I bet there are some tutorials out there, maybe The Signal Path can help also
If you test in an OATS facility, then you have to deal with all of the TV and radio stations in the area, as well as whatever smaller transmitters are near (probably wifi and the like). Dave was talking about emissions testing, but there is also the issue of immunity testing, where RF energy is directed at the device under test. In an OATS, this can get you in trouble with the FCC or the local equivalent.
@@SkyhawkSteve Does immunity testing matter for most consumer grade equipment? I'm thinking the FCC class B notice you often see, with the wording "This device must accept any interference received, including interference that may cause undesired operation."
@@threeMetreJim I've never designed consumer electronics, so I'm not sure about the specifics. FCC part 15 is the basic spec on emissions, which is that the device isn't allowed to interfere in the standard radio and TV bands, etc. (if I recall correctly, etc) Dealing with EMC regulations isn't for the faint of heart. For a small hobbyist selling a few items, it's not a concern. If you are selling a lot of stuff, you might be liable for any problems caused by your product or may have to provide replacements or refunds. This is when it's good to consult with an EMC test house and get their guidance. regarding immunity testing specifically.... that probably depends more on the type of product. If it's a toy, then it is no big deal if it stops working when someone uses a cell phone nearby. If there are safety implications, then you'll want to do whatever the industry standard testing is. Even concerns about customer satisfaction might be a good reason to do immunity testing... no one wants to deal with thousands of customers seeking refunds or replacements.
That's E not S. E is vertical if the radio wave is "vertically polarized" which depends on antenna orientation, since the convention is to refer to polarization in terms of the electric field. If a waveguide is used then it also depends on its orientation (which can be twisted), and as it happens the magnetic field follows the wider side of a waveguide, so it is vertically polarized when the opening looks like it is horizontal.
You have VNAs and all that stuff, just build one yourself and show us the process of tweaking the design... btw it would be fun to have a robot that runs evolutionary algorithms to bend and stack together some antennas...
The trouble with dipoles is that you have to adjust the length for each frequency if you want to get calibrated results. This quickly gets very, very tedious if you are testing multiple frequency ranges
30-300 MHz log periodic? It's not going to be cheap, because it'll be quite large. The bowtie is a great antenna for 300 MHz, because it's relatively small, and it's bandwidth is almost 100 MHz. It allows the log perodic section to start just below 400 MHz, which makes it smaller, too. At 30 MHz, the bowtie's bandwidth drops to 10 MHz, and it would be something near 24 ft wide. It saves you about a foot each way (width and length) on the log-periodic-only (30-300 MHz) version (vs. a 30-40 MHz bowtie plus 40-300 MHz log-periodic). Not enough space savings to justify the extra materials and work, really, and either way - Huge. It's not hard to make your own log-periodic antenna. There are a few log-periodic design programs around. Here's one site that gives all the details and a calculator: hamwaves.com/lpda/en/index.html, though you'd have to use larger square tubing and 1/4" aluminum tubing for those lengths. If that's just too huge, you could try a Fan Dipole. They're very popular with Amateur Radio people who just can't afford one of those huge beam antennas, and you can make it out of wire and roll it up for storage when you don't need it, or just leave it flat on the wall.
So is it correct to say that an SPI signal will mainly produce E field (high and fast voltage variations but low current) while the switching node in a converter will rather produce H field (big currents involved)?
Yes except that it is big /changes/ in current that would produce the alternating H field that is relevant here. A large constant current would produce a constant magnetic field like a magnet, and wouldn't be measurable with that probe, and wouldn't form a radio wave when combined with the E field beyond the far field distance. For more on magnetics, and measuring constant magnetic fields with a different kind of probe, see the AppliedScience video on the BH Curve.
The switching node is actually high E field due to high dv/dt, while your input/output loop (depending on buck/boost) is high dI/dt and therefor high H field.
Hi Dave, I have always been suspicious about the drawings or animations that show the E and B parts of the EM fields in phase with each other. Indeed Faraday and Ampere equations that are part of Maxwell's equations precisely state that there should be a quarter wavelength shift between E and B when these contain only a single frequency. But I could never find *any* drawing actually showing the right representation. Do you have any clue? I may misunderstand something.
If you do the Maxwell business then that can popout for various polarisations. But for linear polarisation the phase angles are the same. And yes, I don't think I've ever seen it pictured that way either though. Maybe because it more confusing drawn that way? Don't know.
In free space E and H are always in phase. It is curl of E that is equal to the partial time derivative of B and in absence of lossy materials (i.e. real wave impedance) the E and H field will be in phase. If E and H were to be out of phase by 90º the time average of Poynting vecotor (E*conj(H))would be 0 and there would be no transmission of power
@@TeslaLegend I completely forgot the affect of the curl operator. You are right about the average power too. That's one more question answered I don't need to deal with. Thanks a lot!
I advise you to watch the following: ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-nFtNCPUMoYA.html I must warn you though it's a bit heavy on math, but well explained nevertheless.
I seen a guy in a video(not sure who it was) with one of those probes but instead of a ring at the top he had made his own with a English two pence piece!! (copper coin) he was using it to find stuff in a car engine!! cool...
I've never seen proper response curves for cheap log periodic TV antennas on eBay, nor someone take a VNA to one, but some claim to go down to 40-50MHz for FM radio. Might be some gaps between 230-400 MHz though? Other than that, DIY?
The cheap antenna are easy to figure out. Usually take little more than basic geometry to compare the expected waveform to what the antenna is configured to be sensitive to. Since your talking cheap, you can safely assume there's only bare minimum signal processing if any and rely on physical characteristics.
Testing is for compliance with regulations that forbid radio interference. If a device emits excessive unwanted EM waves then it would interfere with other radio signals.
New EMC antennas are not cheap, the cheapest new one I found is www.rowielektronik.de/emv-emi-antennen/hta-2003-20-mhz-3000-mhz.html It comes with individual calibration data (antenna factor table). Of course you can try to buy used one on e-bay. Sometimes you can find bi-log for 200....500$.
You'll want to build a biconical (modified cage dipole) like this: www.com-power.com/products/antennas/biconical-antennas/ab-900a Covers 25 to 300 MHz. The standard I used over the last 20 years for MIL-STD 461 testing has been the A.H. Systems SAS-540.
