Maker Pro
Maker Pro

PLL as modulator og demodulator

R

rasmusms

Jan 1, 1970
0
I would like to use one PLL for both FSK modulator and demodulator
because of price. I probably gonna switch between a tvxco and FSK input
to "sig in" on a 4046. I gonna modulate the FSK on after the Loop
filter. My consern is about the loop filter. When I am using txvco as
center carier frequency I would like a small BW for loop filter (to
stay looked on carrier frequency), but when pll is used as demodulator
the loop filter should inclued the two FSK frequency?. Or am I wrong?
I would like to use the txvco as center carieer because of large
temperatur swing (-30 - 80) degrees celsius.
 
B

Ben Jackson

Jan 1, 1970
0
I would like to use one PLL for both FSK modulator and demodulator
because of price. I probably gonna switch between a tvxco and FSK input
to "sig in" on a 4046. I gonna modulate the FSK on after the Loop
filter. My consern is about the loop filter.

Variable bandwidth loop filters are also used to get faster lock time
(and better frequency agility) while still having good stability once
lock is achived. If you google around you'll find some appnotes with
techniques for accomplishing that.

You might want to consider other PLLs like the MC145170. It has serial
programming of the divider/counter values which saves on external
parts, but it does lack an integrated VCO.
 
R

rasmusms

Jan 1, 1970
0
My center carrier is about 1-3Mhz

Do you know who makes the MC145170? I can't find it on
fairchild/motorola homepage anymore.
It might be a alternativ.

Do you have a example on an appnote?
 
A

Andrew Holme

Jan 1, 1970
0
rasmusms said:
I would like to use one PLL for both FSK modulator and demodulator
because of price. I probably gonna switch between a tvxco and FSK input
to "sig in" on a 4046. I gonna modulate the FSK on after the Loop
filter. My consern is about the loop filter. When I am using txvco as
center carier frequency I would like a small BW for loop filter (to
stay looked on carrier frequency), but when pll is used as demodulator
the loop filter should inclued the two FSK frequency?. Or am I wrong?
I would like to use the txvco as center carieer because of large
temperatur swing (-30 - 80) degrees celsius.

For a demodulator, the loop bandwidth must be appreciably wider than
the data rate. The actual mark and space frequencies themselves are
irrelevent; it is how fast you switch between them that matters. For a
modulator, the loop gain must roll-off well below the switching speed,
or else the loop will fight against the modulation. Injecting after
the loop filter doesn't avoid this. You may be able to control the
bandwidth by varying the loop gain. You will have to do the PLL
stability analysis.
 
A

Andrew Holme

Jan 1, 1970
0
Andrew said:
For a demodulator, the loop bandwidth must be appreciably wider than
the data rate. The actual mark and space frequencies themselves are
irrelevent; it is how fast you switch between them that matters. For a
modulator, the loop gain must roll-off well below the switching speed,
or else the loop will fight against the modulation. Injecting after
the loop filter doesn't avoid this. You may be able to control the
bandwidth by varying the loop gain. You will have to do the PLL
stability analysis.

Another thought: if you leave the modulator idling (at mark or space)
for too long, it will return to the carrier centre frequency.
 
W

Winfield Hill

Jan 1, 1970
0
rasmusms wrote...
Do you know who makes the MC145170? I can't find it on
fairchild/motorola homepage anymore.

The modern pin-compatible version is the mc145170-2, which
had a new 32-page rev-5 Freescale datasheet issued last year
http://www.freescale.com/webapp/sps/site/prod_summary.jsp?code=MC145170_2
The official p/n was the mc145170P2, D2, DT2, etc.

Freescale (the Motorola RF semi spinoff) has dropped the part
(along with many other "legacy" ICs), perhaps to avoid making
RoHS versions, but some are available from surplus distributors.

Lansdale Semiconductor is offering a version, the ML145170.
Get a 13MB (!) datasheet on their webpage, www.lansdale.com
Do you have a example on an appnote?

Two classic appnotes were the an1207 and the an1671.
 
R

rasmusms

Jan 1, 1970
0
Is it possible to avoid varying the loop gain by detecting on the
output from the phase comparator and not after the loop. So the loop
stays constant looked on carrier frequency ( rescieve signal has both
carrier and one of the FSK frequencies)
 
A

Andrew Holme

Jan 1, 1970
0
rasmusms said:
Is it possible to avoid varying the loop gain by detecting on the
output from the phase comparator and not after the loop. So the loop
stays constant looked on carrier frequency ( rescieve signal has both
carrier and one of the FSK frequencies)

I don't understand what you mean about the receive signal having both
carrier and one FSK frequency. That is not how FSK normally works.

We're trying to modulate and demodulate FSK using the same PLL without
varying the loop gain (or bandwidth) on switching between send and receive.
I'll assume the loop bandwidth is narrower than the data rate, to prevent
the loop fighting against the modulation on send. So, the loop will
filter-out the modulation on receive; and the phase detector output will be
a complex waveform with a repetition frequency equal to half the FSK mark /
space frequency shift. Depending on the type of PFD used, it might be
possible to determine if the received tone was higher or lower than the
VCO; but it would probably be easier to change the loop bandwidth. Since
you also need to switch the reference input when changing between send and
receive, might it not be simpler to use two PLLs?
 
Top