Please help dumb digital guy w/ FSK - RF Cafe Forums
Because of the high maintenance needed to monitor and filter spammers from the RF Cafe Forums, I decided that it would
be best to just archive the pages to make all the good information posted in the past available for review. It is unfortunate
that the scumbags of the world ruin an otherwise useful venue for people wanting to exchanged useful ideas and views.
It seems that the more formal social media like Facebook pretty much dominate this kind of venue anymore anyway, so if
you would like to post something on RF Cafe's
page, please do.
Below are all of the forum threads, including all
the responses to the original posts.
Post subject: Please help dumb digital guy w/ FSK Posted: Sun Jan 14, 2007 10:24 pm
Joined: Sun Jan 14, 2007 9:52 pm
I am trying to make the receiver portion of a
Universal Access Transceiver (UAT). This is an FSK modulated data-link for aircraft. The specifics are as follows:
Frequency 978 MHz, +/- 20 PPM
Delta f/2 +312.5 kHz
-Delta f/2 -312.5 kHz
Modulation Type Continuous
Modulation Rate 1.041667 MBPS
Sensitivity -93 dBm
I know a lot of RF guys and all of them
say “you can probably get a chip that does all that”. Or another quote is “They make a chip that does that”. Who
is “They”? I have found several FSK/ASK chips that work up to about 920 MHz. I think these are designed for
Keyless Entry, Car Alarms, Remote Controls and that sort of thing. It looks to me like I will have to make a
circuit out of discrete components. I don’t think this is a big deal but I am not sure how to get started. I think
I can handle the gain and filtering portion and I can for sure handle the bit stream once it gets turned into ones
and zeros, but I need help with the demodulation. Does anyone have any specific suggestions on how to build this
type of circuit? Or can someone point me to a book? I am not too interested in communications theory as I am
building a radio to a standard. I would really like a book with circuit ideas.
From the information I have
been able to figure out on the web I have two ideas how to do the demodulation. I am guessing I am way off base so
any constructive comments are welcome.
Filter, amplify, down convert, split then pass each
leg through a filter/rectifier and apply each leg to the input of a comparator.
Second idea: (and this is
my lame idea)
Filter, amplify, down convert, then square up the signal w/ a comparator. Send the signal into a
fast counter and use a CPLD to keep reading the counter to determine the frequency.
Thank you in advance,
Post subject: FSKPosted: Mon Jan 15, 2007 1:47 am
Joined: Wed Feb 22, 2006 3:51 pm
While you might be able to get a chip to do what
you want to do, I suspect the market isn't large enough for a specialized part.
You might be able to
re-purpose chips for the cellular telephone market, but that's not at all a sure bet. And "consumer" devices have
a deservedly bad reputation for being discontinued during the production life of industrial/commercial equipment.
Still, you might browse through Qualcomm's offerings.
All of which puts you squarely in the "do it
Generally, the "best" way to demodulate FSK is with a phase-locked loop (PLL). For weak
signals, this is best done by a receiver with the following stages:
Filter - amplifier -downconverter (mixer +
oscillator) - amplifier - pll - comparator.
Filter - amplify - quadrature down-convert to
DC, apply the two signals to an ADC, and demodulate using DSP.
(This isn't as unthinkable as all that -
some of TI's TMS320C2000 series DSP's have multiple high-speed 10-bit ADC's built in, and still don't cost much.)
The front-end filters are critical. In most aircraft, there are several transmitters: voice, transponder,
radar, etc. The signal from those must be reduced to the point where the FSK receiver low-noise amplifier doesn't
overload. BUT, any loss in that filter adds dB-for-dB to the noise figure: bad news. Given that this isn't a
unique system, some manufacturer of surface-acoustic-wave (SAW) filters might actually have one as a stock item.
LNA's are widely available as IC's from companies like RF Micro Devices, Mini-Circuits Labs, and others; and
mixers/IF amplifiers are not hard to get. (Motorola used to make some nice ones - but they were discontinued when
Mot spun out On Semi, then Freescale Semi.
You might even get a 74HC4046 to do the PLL job (ouch - I'm
showing my age! Pay no attention to that man behind the curtain...)
Bear in mind that in the 900 MHz
region, "breadboarding" is an art requiring minimal lead length, a good ground plane, and just plain luck, if you
don't have experience. Many people just lay out a pcb, and plan on doing at least one revision.
By the way,
the methods you mention are classics: the first one is generally referred to as a "discriminator" in the
literature - of which there's a lot before PLL's took over. The second one is also a standard - implemented in
analog (yes, an "analog counter"!), it's used in high-end audio. So your intuition isn't bad.
Me, if I were
going to do the project, I'd take the DSP route - but that has its own challenges - particularly code efficiency.
Post subject: Posted: Tue Jan 16, 2007 1:50 am
Joined: Sun Jan 14, 2007 9:52 pm
Thank you for the input. What
a great reply! Great news, I like the "do it yourself" category the best. I have about a million more questions
for you. I wish I worked with you.
I downloaded a data sheet for the 74HC4046. I will read it and try to
figure it out.
I though about trying to use a DSP but I thinking that with a bit rate of over 1 MBPS. I
would have to sample at over two million samples per second. Is this correct? I am very comfortable writing low
level micro code so if can be done this way that would be good.
Also I would like to thank you for you
advice on the SAW filter. I will look into it.
Post subject: FSKPosted: Tue Jan 16, 2007 3:09 am
Joined: Wed Feb 22, 2006
Well, maybe a TI '6000 series DSP would be better
If your speed were
lower, Analog Devices has a transceiver chip - but it only goes up to 384 kbits/second - not enough. (ADF7025, I
think - but you should check their website rather than depend on my ability to remember part numbers
What kind of design constraints do you have? Size, power, cost? If cost, is it
"cost-per-unit" or "cost-to-first-unit"? (That is, parts cost or development cost).
"You can have it good,
fast, and cheap - pick 2..."
I'll try to write more later - but it's already too late.
Post subject: Posted: Tue Jan 16, 2007 12:14 pm
Joined: Sun Jan 14, 2007 9:52 pm
My target is “good and cheap”. I am paying for development
out of pocket so I would like to keep that cost down although I realize nothing is free, especially good tools.
And I would like to keep "cost-per-unit" down too so that I can keep the price down. Avionics are relatively
expensive so I think I would like to keep parts down in the $100 or $200 range. The good news is I do have time.