MC1496 as FM Demodulator

[kroto]

Can I use MC1496 as a FM demodulator?
I have 2 pcs of this IC on my bench, used for nothin' ( a wrong purchased part).
the FM tuner is sanyo LA1260, could I make a stereo FM tuner from LA1260 and MC1496?

oh, the FM tuner already have a LA3361 (Sanyo PLL Multiplex stereo demodulator), but the sound (output audio) doesn't stereo at all, so I think those MC1496 could be used as the Demodulator.

any input? schema? suggestion?

datasheet :
MC1496 : http://www.datasheetcatalog.org/data...8l8pw8d97y.pdf
LA3361 : http://www.datasheetcatalog.org/data...f_e/LA3361.pdf
LA1260 : http://www.datasheetcatalog.org/data...f_e/LA1260.pdf


regards,

[DF96]

The 1496 could be used as part of an FM discriminator, or part of a stereo decoder - these are two separate stages in a stereo receiver. Much better to use a dedicated chip, unless you want to make your own circuit. For example, what about IF limiting - dedicated FM chips do this but 1496 will need you to do it separately.

Much better to repair the radio, not redesign it.

[firechief]

The data sheet does in fact reveal the idea of using this part as an FM detector "An FM detector may be constructed by using the phase
detector principle. A tuned circuit is added at one of the
inputs to cause the two input signals to vary in phase as a
function of frequency. The MC1496 will then provide an
output which is a function of the input signal frequency."
But even with two of them you would not be able to decode FM stereo. I dosent work that way.

[kroto]

@DF96 : yes, I think I'll repair that tuner later (with advice). it just that the schema doesn't the same as in the datasheet, and I didn't quite understand about radio design...
@firechief : yeah, thats why I make this post, I don't know how to make them usefull
a few month ago my brother buy a wrong number, it should be MC1495, not MC1496.

thanks guys...

[stratus46]

Originally Posted by kroto

Can I use MC1496 as a FM demodulator?
I have 2 pcs of this IC on my bench, used for nothin' ( a wrong purchased part).
the FM tuner is sanyo LA1260, could I make a stereo FM tuner from LA1260 and MC1496?

oh, the FM tuner already have a LA3361 (Sanyo PLL Multiplex stereo demodulator), but the sound (output audio) doesn't stereo at all, so I think those MC1496 could be used as the Demodulator.

any input? schema? suggestion?

datasheet :
MC1496 : http://www.datasheetcatalog.org/data...8l8pw8d97y.pdf
LA3361 : http://www.datasheetcatalog.org/data...f_e/LA3361.pdf
LA1260 : http://www.datasheetcatalog.org/data...f_e/LA1260.pdf


regards,

FWIW when I worked in broadcast TV the precision audio demodulator they used did a second conversion to get to around 700 KHz IF which detected 0 crossing and then triggered a monostable set to 50% duty cycle with no modulation. You then integrate the pulse stream to recover the audio. The detection 'curve' is a 'Z' that's dead linear as long as there is no overmodulation.

If FM radio was worth a nickel (severely over processed in LA) I'd consider modifying a tuner but after you're all done there's still nothing to hear here.

G²

[wrenchone]

Well, yeah, it's modulation vs. signal radius. The signal slightly in the red (engaging the compressor) hits the fringe areas better. Here at KFJC (kfjc.org), it's still human hands on the console, so the individual can choose how hard he wants to push the signal, if he notices... You have to have a compressor, so that you don't bash all hard-edged into the bandwidth limits for FM broadcasting. Usually what is used is something like the Orban Optimod (or, no doubt, its modern digital counterpart), a multi-band compressor. How one engages that compressor is up to the individual station. I shudder to think what some of the commercial stations do . After all, "louder is better".

[oshifis]

Originally Posted by stratus46

FWIW when I worked in broadcast TV the precision audio demodulator they used did a second conversion to get to around 700 KHz IF which detected 0 crossing and then triggered a monostable set to 50% duty cycle with no modulation. You then integrate the pulse stream to recover the audio. The detection 'curve' is a 'Z' that's dead linear as long as there is no overmodulation.

I am thinking of building an FM demodulator based on this principle for quite long time. It is called scaler detector, as I remember. It could be quite simple: a 10 MHz crystal, a mixer chip, amplifier/limiter, a monostable and a low-pass filter. Do you have a circuit diagram? I am concerned about the noise of this solution, but perhaps it could be realized with discrete components, so the noise could be less than with using TTLs.

[davidsrsb]

It is very hard to make the monostable trigger at the same threshold as the clock frequency varies, causing jitter. This leads to some very strange high order distortion.

I had a tuner (Sansui?) years ago that worked with this "pulse counting" detector and it did not sound very good.

The standard double tuned coil phase detector works pretty well i practice

[stratus46]

Originally Posted by davidsrsb

It is very hard to make the monostable trigger at the same threshold as the clock frequency varies, causing jitter. This leads to some very strange high order distortion.

I had a tuner (Sansui?) years ago that worked with this "pulse counting" detector and it did not sound very good.

The standard double tuned coil phase detector works pretty well i practice

In broadcast TV it's of paramount importance to lock to the leading edge of the analog sync pulses - even in digital high definition. They solved that problem many years ago by using a tracking precision slicer (comparator) to always trigger at the 50% point of the sync. Since TV has known repetitive pulses they take a sample at back porch (blanking level) and sync tip. Sum the two analog values with matched resistors to get the slice reference point. Even if the DC level of the video is shifting with program material, the tracking slicer takes care of it. FYI they need to achieve a spread of a few nanoseconds.

For the FM demod you could so a similar trick by looking at the output of the limiter and sampling the max and min values and do the resistor sum for the slicer reference but in reality, how is it so different from the ratio detector? It too would have issues with jitter that shows up as hiss in the output. As signal strength increases the noise goes down as the ambiguity goes down. I would expect the 'digital' demodulator to behave similarly with varying signal strength.

G²