FSK technical question
 

Is anyone familiar with the hardware details of the FSK process?
Specifically what components are involved in transforming a frequency
into a bit?

Thanks!
Dave

David Harper ) writes:
Is anyone familiar with the hardware details of the FSK process?
Specifically what components are involved in transforming a frequency
into a bit?

Thanks!
Dave


It depends on how it's done.

The earliest schemes were switching a capacitor in and out of the circuit
of a VFO. I think they may have used relays originally, but once
semiconductor diodes became common, they were used to do the switching.
Apply a voltage, the diode conducts, and thus connects the capacitor into
the circuit, lowering the frequency.

When SSB came along, a common means was to feed an audio oscillator into
the mic input, and that audio oscillator was shifted between two frequencies
in the same way as above. Since a single tone in an SSB transmitter is
the same as a cw transmitter (the transmitter translates the tone to
a radio frequency), then shifting the audio oscillator shifts the transmitter
between two distinct frequencies. This method was handy since it required
no modification of the transmitter, and you got the same shift across
the band. The previous scheme could not give the same shift as the
vfo shifted frequency, since it was a fixed capacitor in parallel with
the variable capacitor, so the fixed capacitor would give more shift as
the frequency went up (since it was a larger percentage of the capacitance
of the variable capacitor). The downside is that if the SSB transmitter
wasn't adjusted for good carrier supression and unwanted sideband supression,
or the audio oscillator did not put out a pure enough sinewave, then one
would get more than a single frequency out of the transmitter.

As an aside, on VHF AFSK (the shifting of an audio oscillator fed into
the mic input of the transmitter) was pretty much the only scheme used
for RTTY. Here, the result was not FSK, since you were using the regular
modulation, AM or FM, of the transmitter. And the audio oscillator
did not need to be as pure, since it didn't affect the spectrum of the
transmitter.

I actually don't know what the common scheme for FSK is these days.
If the transmitter is using DSP, the fsk can be done digitally.
I assume that DDS-based VFOs in current rigs may actually be able
to be reprogrammed at a fast enough rate that one just keeps loading
the two frequencies into the synthesizer as the shifting is needed,
but I've not kept track of recent rigs to know if that's what's being
used.

Michael VE2BVW

David Harper ) writes:
Is anyone familiar with the hardware details of the FSK process?
Specifically what components are involved in transforming a frequency
into a bit?

Thanks!
Dave


It depends on how it's done.

The earliest schemes were switching a capacitor in and out of the circuit
of a VFO. I think they may have used relays originally, but once
semiconductor diodes became common, they were used to do the switching.
Apply a voltage, the diode conducts, and thus connects the capacitor into
the circuit, lowering the frequency.

When SSB came along, a common means was to feed an audio oscillator into
the mic input, and that audio oscillator was shifted between two frequencies
in the same way as above. Since a single tone in an SSB transmitter is
the same as a cw transmitter (the transmitter translates the tone to
a radio frequency), then shifting the audio oscillator shifts the transmitter
between two distinct frequencies. This method was handy since it required
no modification of the transmitter, and you got the same shift across
the band. The previous scheme could not give the same shift as the
vfo shifted frequency, since it was a fixed capacitor in parallel with
the variable capacitor, so the fixed capacitor would give more shift as
the frequency went up (since it was a larger percentage of the capacitance
of the variable capacitor). The downside is that if the SSB transmitter
wasn't adjusted for good carrier supression and unwanted sideband supression,
or the audio oscillator did not put out a pure enough sinewave, then one
would get more than a single frequency out of the transmitter.

As an aside, on VHF AFSK (the shifting of an audio oscillator fed into
the mic input of the transmitter) was pretty much the only scheme used
for RTTY. Here, the result was not FSK, since you were using the regular
modulation, AM or FM, of the transmitter. And the audio oscillator
did not need to be as pure, since it didn't affect the spectrum of the
transmitter.

I actually don't know what the common scheme for FSK is these days.
If the transmitter is using DSP, the fsk can be done digitally.
I assume that DDS-based VFOs in current rigs may actually be able
to be reprogrammed at a fast enough rate that one just keeps loading
the two frequencies into the synthesizer as the shifting is needed,
but I've not kept track of recent rigs to know if that's what's being
used.

Michael VE2BVW

"Michael Black" wrote in message
...

I actually don't know what the common scheme for FSK is these days.
If the transmitter is using DSP, the fsk can be done digitally.
I assume that DDS-based VFOs in current rigs may actually be able
to be reprogrammed at a fast enough rate that one just keeps loading
the two frequencies into the synthesizer as the shifting is needed,
but I've not kept track of recent rigs to know if that's what's being
used.

This is pretty reasonable at low bit rates. But some more recent DDS chips,
for example, TI's TRF4900, actually have a data pin, and the two frequencies
are programmable, so the chip has the two divisors onboard making it a
simple matter to switch between divisors.

...

"Michael Black" wrote in message
...

I actually don't know what the common scheme for FSK is these days.
If the transmitter is using DSP, the fsk can be done digitally.
I assume that DDS-based VFOs in current rigs may actually be able
to be reprogrammed at a fast enough rate that one just keeps loading
the two frequencies into the synthesizer as the shifting is needed,
but I've not kept track of recent rigs to know if that's what's being
used.

This is pretty reasonable at low bit rates. But some more recent DDS chips,
for example, TI's TRF4900, actually have a data pin, and the two frequencies
are programmable, so the chip has the two divisors onboard making it a
simple matter to switch between divisors.

...

David Harper wrote:



Is anyone familiar with the hardware details of the FSK process?
Specifically what components are involved in transforming a frequency
into a bit?

Thanks!
Dave

I think you already have a couple of messages on how the frequency can be
generated from a bit.

It appears to me you are asking for the opposite.

Basically, the modem runs the receive frequencies through narrow filters. If
the frequency received matches the passband of the filter, the audio tone
gets through the filter. The tone can then drive something as simple as a
transistor amplifier or darlington pair from one state to another (e.g.
from Vc=5v to Vc=0.5v). These outputs can then drive gates in the right
combination to get your serial output.

What has changed over the past two decades is how the filters have been
implemented and how the serial signal is generated.

What used to be narrow filters built with big inductors and critical
capacitors became filters implemented using op amps which, in turn, became
filters using dsp techniques.

Serial signal generation started out with straight off and on pulses to rtty
ksrs, moved to simple uarts built with logic gates, which became uarts on
ic's.

The basics are the same, however. Narrow filters for the tones and logic to
generate the serial signal.

73,

tim a0bwr

David Harper wrote:



Is anyone familiar with the hardware details of the FSK process?
Specifically what components are involved in transforming a frequency
into a bit?

Thanks!
Dave

I think you already have a couple of messages on how the frequency can be
generated from a bit.

It appears to me you are asking for the opposite.

Basically, the modem runs the receive frequencies through narrow filters. If
the frequency received matches the passband of the filter, the audio tone
gets through the filter. The tone can then drive something as simple as a
transistor amplifier or darlington pair from one state to another (e.g.
from Vc=5v to Vc=0.5v). These outputs can then drive gates in the right
combination to get your serial output.

What has changed over the past two decades is how the filters have been
implemented and how the serial signal is generated.

What used to be narrow filters built with big inductors and critical
capacitors became filters implemented using op amps which, in turn, became
filters using dsp techniques.

Serial signal generation started out with straight off and on pulses to rtty
ksrs, moved to simple uarts built with logic gates, which became uarts on
ic's.

The basics are the same, however. Narrow filters for the tones and logic to
generate the serial signal.

73,

tim a0bwr

tim gorman ) writes:
David Harper wrote:



Is anyone familiar with the hardware details of the FSK process?
Specifically what components are involved in transforming a frequency
into a bit?

Thanks!
Dave

I think you already have a couple of messages on how the frequency can be
generated from a bit.

It appears to me you are asking for the opposite.

You're right, I completely missed that.

Michael VE2BVW

tim gorman ) writes:
David Harper wrote:



Is anyone familiar with the hardware details of the FSK process?
Specifically what components are involved in transforming a frequency
into a bit?

Thanks!
Dave

I think you already have a couple of messages on how the frequency can be
generated from a bit.

It appears to me you are asking for the opposite.

You're right, I completely missed that.

Michael VE2BVW