Bandwidth in electronics

[vigsgb]

As I understand it bandwidth is a frequency range and the bigger the range the more data can fit on it?   I thought I knew what bandwidth was until I watched the op amp video and started looking around.  Now I question if I even know what it is or what it means technically with electronics.    I hear things such as a 13 kHz bandwidth is not good but what does that mean in relation to an op amp?    Is it in relation to the input?  Anyone have a good explanation or video tutorial or something that explains this?

[Dave]

The bandwidth figure (frequency) is specified where the output power of the circuit drops to 1/2 of its nominal value (-3dB). If we talk in terms of amplitude, its 1/Sqrt(2) of nominal.

[mrkev]

The thing is that bandwidth in communications is completely different thing that bandwith of OPA. Every OPA has limitations and bandwidth is about limitation in frequency range. Easy way how to look at it is that it has some speed (V/us) when you have quick signals, it's not enough and the gain therefore drops. Rest said Dave pretty well.

[Nerull]

Bandwidth is the 'width' of a frequency range. What exactly that refers to depends on the context - usually either the baseband bandwith or the passband bandwidth. Baseband bandwidth is specified from 0Hz to an upper limit equal to the specified bandwidth. Passband bandwidth is specified from a lower bound to an upper bound.

Oscilloscopes, for example, are limited by signal attenuation at higher frequencies, and speced by the highest useful frequency - defined as a -3dB drop in signal strength - they can operate on. This is their baseband bandwidth.

A radio transmission, on the other hand, will have a specified bandwidth that the signal occupies - say 20kHz - and also a carrier frequency - say 435Mhz - where the signal is centered.

Op-amps, like oscilloscopes, have a baseband bandwidth where they are useful. As you increase the frequency, and op-amp loses its ability to follow a signal, and the gain drops. The bandwidth is given as the point where the signal out of the opamp is -3dB lower than it would be from an 'ideal' op-amp.

There is also a completely different meaning of bandwidth in computer networking, where it refers to bit-rate. It's important to understand that these are all different things and not try to confuse them.

[Rudane]

I agree with you, it's confusing. The term bandwidth shouldn't be used in digital systems. It should be called bit rate, but for some reason we chose bandwidth. We say we transmit 200 kilobits per second and call it a bandwidth. It doesn't make sense, because it's not a bandwidth at all, it's the bit rate. The true bandwidth is like the one Dave talked about, and even though there is a relationship between bit rate and bandwidth I wish we didn't interchange them.

[vigsgb]

My background is in software development and the networking definition of bandwidth was what I was thinking about while watching the video.

Thanks for all the replies thus far.

[Rick Law]

The term "band width" like many other terms, when it somehow gets into the general public's lingo, it adopted an entirely different meaning.  Even as far back as 10-15 years ago, it was not uncommon to hear an exec ask: "Does your team has the band width to deal with the issue?" which to him (or her) means resources or ability to handle tasks.

This is no different than many other typical day-to-day words and phrases.  Even beauty is in the eyes of the beholder.

A statement is best interpreted by using the frame of reference of the speaker (user of the term) and not the listener.

[vk6zgo]

Many forms of modulation yield an "apparent bandwidth" greater than that which the transmission medium is capable of.
As a basic case,Quadrature Amplitude Modulation,by separating two modulating signals in phase,can "share"the existing amplitude bandwidth of a system.

(In this world,"you get nothing for nothing,& bugger all for sixpence",so the system now has stricter requirements as to phase response.)

Some forms of modulation use multiple phase shifts,multiple levels,& other tricks to squeeze a larger bit rate into a narrower bandwidth.

Computer people are "laypersons" as far as Electronics are concerned,& as the above techniques have blurred the relationship between  bandwidth & "bit rate",have adopted what is a wholly inappropriate term .

[miceuz]

I'm no EE pro, just a computer guy, but I'd say frequency bandwdth and bitrate are closely related. To transmit a square-like waveform, one needs some specific bandwidth defined by the minimum rise time one needs to achieve the specific bit rate. Perfectly square waveworms require infinite bandwidth, finite bandwidth puts the upper bound on how fast the digital level can change, thus limiting the bitrate.

I don't recall exactly, but I think, one of the signalpath videos talks about that.

[Rerouter]

  In order to send out a square 1/0 logic level square wave the system would need a bandwidth higher than the bit rate or it would resemble a triangle wave

not to mention your bit rate can be much higher than your signalling frequency, such as the case of phase or amplitude modulation,

while the complexity increases on the end points there isn't much stopping someone from implementing a 16 state analog signal with a 16 state phase encoding on a clock rate of 10Khz as opposed to a digital one at 2.56Mhz, (bit rate is 256 times higher than bandwidth)

[ivaylo]

The thing is that bandwidth in communications is completely different thing that bandwith of OPA.

Err, not that different... See if these examples work for you.

Think how FM packs two channels of high quality sound only because of the wider bandwidth - http://www.diffen.com/difference/AM_vs_FM . Now, there is some trickery how that is done but nothing is stopping you from expanding the bandwidth further and packing even more communication channels.

Like these machines (which I had the "good fortune" of learning 25 years ago ) - http://militaryforces.ru/weapon-4-83-551.html . I am not finding the exact bandwidth number, it's a few MHz, but see how that one can do 6 voice channels. And this one - http://militaryforces.ru/weapon-4-83-550.html 21 voice channels. Yup, you guessed it - higher bandwidth...

And when it comes to digital, don't those "more bits" come with faster rise times, i.e. higher analog bandwidth?

[T3sl4co1l]

What's missing:

Analog bandwidth (in the information-theoretic sense) is frequency response times SNR.

Digital bandwidth is baud rate times bits per symbol.

10Mb ethernet for example is Manchester encoded, so the bitrate is actually half the transition rate (20MHz).  The SNR can be very poor indeed, as long as the signal levels can be discriminated.

100Mb is multilevel; I forget exactly; I think it's four level -- 2 bits per symbol -- with some scrambling, line encoding (8b/10b or whatever), DC offset compensation and some frequency compensation (I think?).  (Don't take my word for it, look it up!)

Gigabit ethernet is heavily encoded (error correction and scrambling and everything else), heavily multilevel (4096 levels -- 12 bits per symbol -- comes to mind, but I may just be thinking of something else, and it may be adaptive), and has frequency response compensation besides.

All three of these could potentially have the same analog bandwidth -- they can all run on Cat6 cable and 8P8C connectors.  But in the presence of a noisy transmission medium of limited bandwidth (typical of a very long cable run, the noise arising from thermal noise and unbalanced ambient interference, and the bandwidth being limited by dispersion in the cable), the actual limits will become apparent.

When talking about bandwidth outside of communications, analog bandwidth is always the range of frequencies for which a circuit passes a signal, "pass" being defined as a certain threshold (usually -3dB, but occasionally -6, -1 or -20dB or more where strong rejection is required, or even just -1dB or even tighter, when precision is required).  The SNR may be large or small; it doesn't matter because it's not part of that consideration.

Tim

[cyr]

while the complexity increases on the end points there isn't much stopping someone from implementing a 16 state analog signal with a 16 state phase encoding on a clock rate of 10Khz as opposed to a digital one at 2.56Mhz, (bit rate is 256 times higher than bandwidth)

What stops you in the end is noise, if you have more possible symbols they become harder to tell apart.

http://en.wikipedia.org/wiki/Shannon%E2%80%93Hartley_theorem

Analog bandwidth and maximum bitrate if certainly closely related, it's just not the only parameter to consider.

[mrkev]

The thing is that bandwidth in communications is completely different thing that bandwith of OPA.

Err, not that different... See if these examples work for you.

Think how FM packs two channels of high quality sound only because of the wider bandwidth - http://www.diffen.com/difference/AM_vs_FM . Now, there is some trickery how that is done but nothing is stopping you from expanding the bandwidth further and packing even more communication channels.

Like these machines (which I had the "good fortune" of learning 25 years ago ) - http://militaryforces.ru/weapon-4-83-551.html . I am not finding the exact bandwidth number, it's a few MHz, but see how that one can do 6 voice channels. And this one - http://militaryforces.ru/weapon-4-83-550.html 21 voice channels. Yup, you guessed it - higher bandwidth...

And when it comes to digital, don't those "more bits" come with faster rise times, i.e. higher analog bandwidth?

You missed my point... BW in communications is about how much of freq. spectrum is occupited by the signal, it has nothing to do with physical limitations of device... So yes, they are very different...