I hear fractals are used for designing cell phone antennas to maximize receptive power in a minimum amount of space. So why can't you just use a length of wire wound in a spiral arrangement to occupy the same space with the same length of wire and do the same thing? Why the complex pattern?
Spirals work and are old hat in microwaves:
Fractal antennas were discussed here a few years back. The gist is that they're inherently broadband, more so than a spiral, random length, or meandering antennas. They can also be nearly omnidirectional.
Log periodic arrays are directional fractals.
Mark L. Fergerson
Think of an antenna as a device that matches the typically 50 ohm output of a transmitter (or input of a receiver), with the impedance of free space (377 ohms). The basic antenna requirements.
To achieve these goals, antenna modeling programs (NEC2, NEC4, etc) perform iterative (trial and error) calculations on the impedance (VSWR), gain, directional pattern, bandwidth, and small physical size. Each of these requirements involve a trade-off with the others. For example, I could optimize the antenna to give me the maximum gain within a given physical size as found in a cell phone, at the expense of gain and bandwidth. In simple terms: "Gain, bandwidth, small size... pick any two"
In a cell phone, small size is the key requirement because external projecting antennas are considered a customer repellent by the aesthetics committee. Wide bandwidth is a necessity due to the increasing number of bands (operating frequency ranges) allocated for cellular communications. It would not do to have individual antennas for each band as that would waste too much space inside the phone. So, gain and antenna pattern are what is compromised which manifests itself as lousy range and dropped calls.
Fractals allegedly produce the highest gain possible within the confines of a limited area. Adjusting the shape of the antenna, to optimize the antenna pattern adds another layer of iteration as trial, error, tweak, and recalculate are used to produce the best antenna possible within the confines of the cell phone. Although fractal antennas allegedly produce optimum results, they're not the only workable solution. For example, meandering line antennas and patch antennas work well and are what is found in most cell phones. The antenna pattern may look strange if not weird, but it's the result of a tremendous amount of computer modeling used to produce an optimum compromise between SAR (specific absorption rate or how much RF is absorbed by your head), pattern, gain, bandwidth, size, shape, hand position, body detuning, etc. In other words, a fractal antenna is just one possible solution for the antenna problem, and isn't necessarily the best when you consider all the compromises required.
Dragged to an extreme, Apple uses the metal frame of its iPhone 4 and later for various antennas. The antenna is essentially an untuned monopole with little optimization beyond SAR safety requirements. While the function of this type of untuned antenna is inferior to a highly optimized patch or fractal antenna, Apple has successfully demonstrated that the buying public is willing to tolerate a slight loss in performance in trade for award winning styling. Oh well.
To answer your question, a spiral antenna would certainly function. It has great bandwidth, small size, but lousy gain. It's also highly direction, which is not a good thing in a cell phone or smart phone, which needs to be fairly orientation insensitive.
-- Jeff Liebermann jeffl@cruzio.com 150 Felker St #D http://www.LearnByDestroying.com Santa Cruz CA 95060 http://802.11junk.com Skype: JeffLiebermann AE6KS 831-336-2558
Fractals allow a lot of room in trading the "area" of the antenna against the "length" of a "piece of wire" from which you would construct an antenna. I expect it just provides more possible solutions.
-- Les Cargill
Excellent post, just one tweak: it's actually half (~188 ohms).
Consider the inverse of a conical dipole: a cone is a cone, you're just looking from the inside rather than the outside. Cones are self-dual structures (at least, 90 degree ones), so offer very wide bandwidth, and have a simple structure. Spirals (suitably proportioned) are self-dual. Fractal spirals are logarithmic rather than Archimedian, which gives flatter frequency response.
For a self-similar structure, the impedance will also be self-similar, therefore the "positive" version has the same impedance above zero, that its inverse has below Zo. 1 / (2) = (1/2), so, it's half of Zo, of course. :)
Tim
-- Seven Transistor Labs, LLC Electrical Engineering Consultation and Contract Design Website: http://seventransistorlabs.com
Doesn't a spiral also have to be 3D as opposed to the flat fractal?
There are planar spiral antennas with a reflector behind them. There are conical antennas, and there are helical antennas. So, all forms are apparently workable for specific purposes. I think the spiral and helical are narrow-band, and the conical is a wideband form.
Jon
Nope. There are other shapes and configurations: What you're thinking of is a conical spiral: which offers more gain and directionality than a flat spiral. It's really a helical antenna with a logarithmic geometry.
Both types offer the advantage of broad operating bandwidth, but at the expense of gain and size. For example, the approximately 6ft long dual polarized conical spiral antenna in this example: goes from 350 to 2250 MHz (almost 3 octaves), but only has a gain of
6dBi. If a different antenna of similar size (well, of similar volume), such as a reflector type antenna, it probably would have much less bandwidth and far more gain.Chorus: "Gain, bandwidth, small size... pick two".
Putting a flat spiral antenna into a smartphone is possible. The circular polarization gives the benefits of being insensitive to orientation and reduced fading, at the price of a 3dB polarization loss. The necessary bandwidth, currently from 700 to 2700 MHz would work nicely with a spiral. However, the current fashion of using a fat dipole or multiple IFA (inverted F shaped antenna), coupled with elaborate matching schemes or possibly ATU (MEMS automagic antenna tuner), seems to be adequate. Or better yet, using some metal structural elements of the smartphone as an antenna, as in the back of the HTC One and the frame of the iPhone. Some detail: At this time, it's rather difficult to even find the antenna(s) inside the phone. Following the coax cables is an obvious clue, but even so, antennas are rather obscure when using the FCC ID photos and iFixit disassembly photos. I have yet to see anything that looks like a classic fractal antenna. To be fair, I recently haven't seen anything that looks like any type of recognizable classical antenna geometry.
There is also competition for real estate on the back of the phone between NFC (near field communications), wireless battery charging, and various antennas. The result is a general decrease in antenna sizes. If one could cram a fractal antenna into the remaining space, it would need to be rather small and fit at the bottom of the phone, the farthest away from the head, to meet SAR requirements.
-- Jeff Liebermann jeffl@cruzio.com 150 Felker St #D http://www.LearnByDestroying.com Santa Cruz CA 95060 http://802.11junk.com Skype: JeffLiebermann AE6KS 831-336-2558
They figured the overwhelming majority of iPhone users are not using the phone to make voice calls 95% of the time, but simply as an Internet-enabled pocket computer. And as the overwhelming majority of their customers (like most people) are technologically illiterate, if calls are dropping they figure it will be the cell provider that gets the blame, anyway.
Doesn't the antenna get used for Internet access?
-- Rick C
Internet/text is a lot more forgiving of dropped packets than is voice. Neither is a big problem with the phone end, though. The problem with dropped connections is usually something between the phone and the cell, like a big pile of dirt.
I think you might mean 95% of the bytes transferred, and not 95% of the (air) time. The problem is that the overly compressed and unintelligible garble that represents the cellular vendors idea of voice only takes a few kilobytes of bandwidth to function, while data gobbles far more bandwidth and megabytes of bandwidth.
However, much of this "cellular" traffic is being offloaded onto Wi-Fi and home nanocellular or femtocellular radios: "Mobile offload exceeded cellular traffic for the first time in 2015. Fifty-one percent of total mobile data traffic was offloaded onto the fixed network through Wi-Fi or femtocell in 2015. In total, 3.9 exabytes of mobile data traffic were offloaded onto the fixed network each month." Translation... those handsets with really crappy wi-fi performance are going to see customer complaints. Also, indoor cellular coverage, and handset reliability is so bad, that many users and offices are buying internet connected nanocellular or femtocellular base stations in order to get functional cellular service.
In terms of quality time that users spend with their smartphones, you're correct. It's basically a pocket computer spending much of its time doing email, facebook, twitter, or other messaging systems. Voice is so insignificant these days, that the latest greatest is just a service that rides on top of LTE (VoLTE) data. However over half the bytes moved these days is video. From the above Cisco URL: "Mobile video traffic accounted for 55 percent of total mobile data traffic in 2015. Mobile video traffic now accounts for more than half of all mobile data traffic."
Perception is everything, and most users would never blame the expensive device in their hands. Better to blame the service provider. Actually, the service providers welcome such criticism as it give them an opportunity to upsell the complainer to a later model phone:
So, will a stronger signal and better SNR (signal to noise ratio) do anything useful for all this? Yep. Speed and SNR are directly related. The better the SNR, and by implication the stronger the signal, the faster the bytes will move through the system. When the signals get weak or noisy, the system slows down in order to improve the SNR. This slowdown consumes more air time to move the same amount of bytes. What you pay for with both voice and data is air time and is what the cellular vendors are actually selling. If everyone suddenly began tolerating crappy radios and antennas for some reason, that slowed traffic by half, we would have a cellular system with half the number of users (assuming the system is operating near capacity which most are). Half the users, means half the revenue to the cellular provider, which would certainly be considered unacceptable. Being able to cram more users onto their systems is why cellular providers have an interest in providing the best possible handsets and antennas. Anything that reduces air time is a win for the provider. If it means we have to live with unintelligible garble for audio, no problem.
-- Jeff Liebermann jeffl@cruzio.com 150 Felker St #D http://www.LearnByDestroying.com Santa Cruz CA 95060 http://802.11junk.com Skype: JeffLiebermann AE6KS 831-336-2558
Thanks. Hopefully that partly compensates for my previous screwups.
Ok, Babinet's principle: which makes the characteristic impedance of a dual log periodic spiral antenna equal to 188 ohms. (Equation #4)
I use the model where half the power is dissipated in the source impedance of the antenna, and the other half is delivered to mythical free space at 377 ohms. I think (not sure) that the results are the same as using 188 ohms for the free space impedance with a symmetrical self-similar structure.
-- Jeff Liebermann jeffl@cruzio.com 150 Felker St #D http://www.LearnByDestroying.com Santa Cruz CA 95060 http://802.11junk.com Skype: JeffLiebermann AE6KS 831-336-2558
Actually, it just would mean twice the number of cells. The basic and funda mental idea behind cellular radio is that you divide up the spectrum spatia lly until the cell tower are sufficiently closely spaced to be able match t he number of customers you've got
Which is why they'd build more in-fill cell towers.
There's always error-detecting and -correcting code. As krw says, data can work with retransmitting corrupted packets (errpr-detecting codes), but err or-correcting coding can work with voice - up to some given error rate, whe n you get too many uncorrectable errors.
-- Bill Sloman, Sydney
I was thinking fractals don't have any resonant frequency due to the complexity and are good for broad band applications. Anything of a reqular pattern woud be resonat somewhere and have some gain. It appears cell phone frequencies are allocated to the 800-900 MHz band and the 1800 Mhz band. Seems like it would be better to put all the cell phone frequencies in one band around 1800Mhz. That would only require a 10% bandwidth to accommodate
900 channels at 200K each and a narrow band antenna could be used.
See page 12 (first page of article) and notice what happens when you change fractal geometry. Lots of resonances. 0db return loss means a very high VSWR. -10dB return loss = 1.7:1 VSWR, which is usable. So, the fractal antennas in the article will only work where the various graphs are below about -10dB return loss. Broadband? Methinks not.
Nope. Resonance has nothing to do with that gain of the antenna. More generally, most antennas (especially broadband antennas) do not operate near resonance. Resonance is defined as the frequency where the inductive and capacitive reactances are equal. There's no guarantee that this point on the frequency will produce a 50 ohm feed impedance, maximum gain, or even a recognizable peak.
Nope. That's the US "cellular" and "PCS" bands. They're different in Europe. Add in the 3G, 4G, GSM, UMTS, LTE, and other bands, and the operating frequencies looks like Swiss cheese.
Ummm... 10% bandwidth would be 180 MHz. Here's ROUGHLY what is currently allocated and in use in the USA: Band Use MHZ BW
710-716MHz 740-746MHz AT&T current 4G/LTE 12 746-757MHz 776-787MHz Verizon current 4G/LTE 22 806-866/869MHz LTE 4G 144 824-896MHz Traditional Cellular (part of above) 1710-1755, 2110-2155MHz AWS Band (T-Mobile 3G/4G) 90 1850-1990MHz Traditional PCS 140 1850-1990MHz Sprint 4G/LTE (overlaps above) 2496-2690MHz LTE data 194 ============================================================= Total 602 MhzSo, assuming I didn't screw up the frequencies and math, you will need at least 602 MHz of bandwidth in order to handle the EXISTING traffic[1]. Also note that the frequencies are somewhat different in other parts of the planet. Of course, everything will need to be coordinated through an assortment of acronym infested national and international regulatory bureaucracies. See "Spectrum Harmonization" for clues of frustration to come:
From a long term standpoint, you're correct that a spectrum shuffle is required. However, considering the huge amount of equipment that would be obsoleted and need to be replaced, I suggest that the exercise might not be worth the cost.
[1] If you believe this report, traffic and bandwidth requirements are gonna grow rapidly. HTML and PDF formats: While the numbers are probably optimistic marketing hype for selling Cisco hardware and services, the general trends are in the right direction and the historical and current numbers are staggering.-- Jeff Liebermann jeffl@cruzio.com 150 Felker St #D http://www.LearnByDestroying.com Santa Cruz CA 95060 http://802.11junk.com Skype: JeffLiebermann AE6KS 831-336-2558
In urban areas the 1800 MHz band works great, with lot of reflections from buildings and a lot of paying customers, so a large number of base stations can be employed. Having a large number of base stations makes it possible to use small cell sizes, reducing the handset transmit power demand and also making it possible to reuse the same frequency every few hundred meters, having a huge total transfer capacity within a city.
In rural areas, lower frequencies propagate better, making it possible to use large cell sizes and hence fewer cell towers. In rural areas there are also fewer customers paying for the infrastructure.
For these reasons, multiple frequency bands are needed in the same mobile device.
Adding cell sites these daze is more a matter of politics than engineering. As soon as a vendor proposes a new cell site, the tin foil hat crowd arrives to testify that the cell site will turn their brains to mush. Never mind that with most, this is a pre-existing condition. When they're done, the aesthetics committee arrives to insure that the tower is properly camouflaged. I've been through a few of such planning department hearings locally, and I must say that they provide me with substantial entertainment value and science fiction material.
Small cells somewhat solve most of these problems. They belch less RF so the tin hats are more tolerant. They are smaller and less obtrusive. They're also cheaper. Planning boards are more tolerant of small cells. Unfortunately, there are problems. They have lousy range, limited backhaul capacity, and tend to have limited features found in tower based cell sites (such as MIMO for LTE). However, in my never humble opinion, the worst part is that the lack of range, combined with the tendency to install them along major highways, offers good highway mobile coverage, but little or nothing further than maybe 1/2 mile away from the highway.
The small cells that are not along major traffic routes, tend to be installed near recently constructed buildings, where foil backed insulation, low-E glass, and tons of rebar attenuate cellular signals. As more people are cutting the cord and using wireless as their only telephone service, this is becoming more important.
Yep. Frequency reuse is the name of the game. The more channels that can be re-used means more users per square mile. However, that requires careful antenna pattern calculations and adjustments as well as careful channel selection. It also requires a working handoff mechanism for moving handsets and spare channels at all sites to take care of handoffs without dropping calls. Some cellular providers do a better job of this than others.
Data can be delayed for many seconds or even minutes with streaming by buffering in the handset before something is noticed by the user. Voice is delayed at least 50 msec due to compression/decompression delay. If it gets over about 250 msec, some systems switch from full duplex to half duplex, so it's less obnoxious. (I'm not sure of the exact delays and too lazy to look them up).
Gotta run...
-- Jeff Liebermann jeffl@cruzio.com 150 Felker St #D http://www.LearnByDestroying.com Santa Cruz CA 95060 http://802.11junk.com Skype: JeffLiebermann AE6KS 831-336-2558
Remember back when international phone calls got bounced up to one synchronous satellite, across to another, and down again?
Delay management technology got quite good, quite quickly, and the users learned not to interrupt one another too often ...
-- Bill Sloman, Sydney
ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.