There's a saying in EMI that any conductive feature longer than a quarter wavelength can be an antenna. All it takes is one piece of metal a quarter wavelength long and another piece of metal to be the “ground” or the other half of the dipole. The metal will radiate all frequencies above this point, the saying goes.
The positive thing about this is you can generally use a larger antenna for shorter wavelengths. An antenna that resonates on the 10m band will usually work decent on the 2m band. At my home I have a 5-meter-long vertical designed to be a half-wave antenna for 10m. Despite it being poorly matched for 2m, I can still use it for that band. When I measure it with a network analyzer, about half the power I send to it at 2m is reflected back. Some of the power is lost both ways in the coax, so maybe only 25% of the power gets to the antenna. All received and transmitted signals will be 6dB weaker than they would be with a good antenna. If two stations with such poor antennas try to communicate, they lose 12dB of link budget. That means they could cut their output power by over 15 times if they both used good antennas. That sounds horrible, but each 6dB is only one S-unit. If one side switched to a well-matched antenna, that side would only send and receive one unit stronger on a signal meter.
I tried to take advantage of this a few weeks ago when I was at a park about 30 miles outside of town. I wanted to listen to NOAA weather radio, so I connected my Uniden BC95XLT scanner to a Cobra HGA-1000 mag mount antenna on my car roof. The antenna is designed for the 11 meter CB band. It's one of those designed to look like an 80s-era mobile phone antenna. There were few cell towers in those days, so you needed a decent antenna to access the AMPS cellular network. A mobile phone and antenna was a status symbol. 11m CB was a free-for-all and not a status symbol. The look-alike antenna is much shorter than a quarter wavelength. A built-in loading coil matches the antenna to the 50-ohm feedline.
The NOAA weather stations are around 162MHz, just under 2 meters, so I figured the 11m antenna would work fine. Surprisingly, the rubber duck that came with the BC95XLT scanner worked much better even in the passenger compartment than the 11m antenna centered on the roof.
When I got back to town I threw both antennas on the vector network analyzer (VNA). As expected the CB antenna looked fine in the CB band, which is centered on 27.185MHz:
Here is the Smith Chart plot of S11 of the CB antenna:
Even though the BC95XLT scanner covers the CB band and works pretty well there, the duck antenna that comes with it is god-awful for VHF-low (26-54MHz).
The duck on the scanner just looks like an open at low frequencies on the Smith Chart.
The CB antenna on the car stays at a constant level of mismatch, with an SWR around 6.5:1 but makes a loop into near perfect match in the CB band.
When I look at the rubber duck across the entire band (a zoomed out version of the 20MHz-120MHz plot above), it looks decent across the entire coverage of the scanner, except for VHF-low.
The CB antenna shows modest reflections at higher frequencies, but this may be because the feedline and the matching network are lossy at high frequencies. Despite this plot, the CB antenna clearly does not work well for VHF:
After I took this data, I tried again to listen to weather stations with both antennas. The duck that came with the scanner clearly does better at 162MHz than the antenna on the car.
This case shows that just because the power isn't being reflected doesn't mean the antenna is working well. I suspect the loading coil on the CB antenna absorbs power at VHF, making it look better on the SWR meter.
This case also shows that the notion that antennas will work at okay somewhat on bands well above their design frequency is not always true.
