How to get more bandwidth?

What carrier frequency ? in interesting way is to convert the audio to a carrier of about 40khz then use a filter to get rid of the unwanted sideband, before final SSB upconverting.

Colin =^.^=

Hi Colin, Thanks for the replay,

The care frequency is 40kHz What do you mean by " convert the audio to a carrier"? If you are talking about the first way of generating SSB (filtering method), so it seems it wont work with the bandwidth which I need too.

It is relatively easy to design a filter with any desirable bandwidth.

Luhan

Are you talking about filtering method or phasing method(hilbert filter)?

For the filtering method, the filter must be a high pass filter and could reject all frequencies lower than 40kHz and passes all frequencies higher than that ( 40 to 60kHz)

Hello Colin,

Or use ye olde Hilbert for the lower audio range, then a filter for the upper and join them. The different group delays might be a challenge though.

In message , dated Tue, 29 Aug 2006, Luhan writes

the

phase

to

bandwidth

phasing

but

music

or

For the filtering method its a bandpass filter 40-60khz wich at that range isnt at all demanding to design or build.

Colin =^.^=

Ultra-high-speed sampling scope followed by an FFT?

Easy in a simulator, well he only said easy to design, didnt say anything about implementing.

but maybe a saphire resonator ?

Colin =^.^=

In message , dated Tue, 29 Aug 2006, colin writes

Maybe, or a VERY small tuning fork. (;-) I'd settle for downshift -> filter -> upshift.

The Hilbert must be connected to the input of modulator so how to use a filter while the output of hilbert is not more than 3kHz?

You can't design if the end-goal is not in sight.

You've said nothing about the purpose of this, and you talk about "high bandwidth" but then want bandwidth that will convey music.

You've also said nothing about what this is all about. You can get away with a lot with voice-only SSB, because mistuning will result in a funny sounding voice (which may not even be noticed unless you are familiar with the voice). But music counts on very precise tones, and if you mistune an SSB signal carrying music, it will sound awful.

If this is a closed system, ie the "transmitter" and "receiver" are directly connected and you can actually use the same carrier generator on both ends, then mistuning won't be an issue. But once you start using a separate carrier at the "transmitter" and at the "receiver" the issue will arise.

Note that with DSB with suppressed carrier, you can determine where the carrier should be reinserted because you have two sidebands to derive that center point from. But there's nothing there with SSB, unless you start sending the carrier along (be it at full strength or reduced strength).

FOr these reasons, SSB has rarely been used to transmit music. A few shortwave stations have dabbled in it, but if it was so viable it would have become the norm long ago.

Obviously, commercial shortwave broadcasters that have played with SSB would be a source of information on how they deal with the required larger bandwidth.

Michael

Hello Adam,

You would need two mixers (multipliers). But again, this is a rather esoteric scheme and would need close attention to group delay and phase matching. It might not be practical and others have pointed out better options.

For example, you could do the whole modulation at very low frequencies where it is easier to build steep bandpasses. Or use crystal filters. With crystal filters I usually start by looking at what's available and what it would cost. Let's say you find a nice 15kHz wide crystal filter at 10.7MHz that has a sufficient shape factor and thus enough suppression of the unwanted sideband. The you just taylor you oscillator to ride on one of the slopes. Take a look at crystal IF filters for aircraft VHF AM radios.

Something else that would be easier on the budget: There are fancy FFT packages from the ham radio community that can turn your PC and its soundcard into an SSB receiver. Some also work the other way around, audio in on one channel, carrier in the 20-30kHz range out the other. I don't know how far you could push the total audio bandwith here considering that the sample rate of most cards is below 100kHz.

Hello John,

And if it's for production where you don't want to deal with 5% cap tolerances I'd do the middle part in a DSP.

In message , dated Tue, 29 Aug 2006, Michael Black writes

It isn't in fact so very difficult to make a 90 degree phase shifter with 20 kHz bandwidth, but the OP has pre-emptively ruled that out.

Hello Michael,

Politicians can :-)

With some "modern" music it won't make much of a difference because it already sounds awful ;-)

Main reason: Lack of receivers. Those who rely on shortwave often don't have the means to buy anything beyond a $10 transistor radio. Once. Or somehow make uncle Leroy's old tube set work again. This is also which DAB will IMHO be a fruitless effort. Keep it simple is the name of the game there.

But one has to keep in mind that a commercial station manager wouldn't even flinch if the new SSB modulator came in a few thousand above budget.

I've seen this done, a combination of a standard filter system (balanced modulator+bandpass filter) augmented with a phasing-type sideband rejector to zap the close-in unwanted sideband, so that the filter skirts don't have to be insanely steep. But mixing and balancing all that stuff gets nasty.

One could do the Hilbert digitally, in an FPGA, and get the 90 degree phase shift over a huge frequency range. So then go ahead and do the entire SSB generation in the FPGA and then just drive a DAC.

John

Joel Kolstad a écrit :

You'll need at least 1E9 samples.

Hello John,

It does get nasty. Sometimes we have to deal with this kind of stuff in ultrasound. Everything has to line up with the delay line granularity and that can be as low as 10nsec.

Or pour it all into a Blackfin. But that's like giving up driving a stick-shift. Maybe after my 80th :-)