Can I use a Directional Detector as an ASK-Demodulator?

[Dr. von Rosenstein]

first: sorry for my bad english!

hi,
i need an ASK-demodulator. carrier frequency: 24 GHz. data frequency: some kHz. i wonder if a directional detector (for example ) can assume this function. responds such a device quickly enough?

thank you for all answers!

 

[biff44]

It should work. But this chip is made as a power detector, not a fast envelope detector. The only 2 concerns are:
1) you need around +5 dBm power hiting it to get a good output voltage. that is a little high for a detector! A simple diode would work as well with a -15 dBm signal. the problem is there is an internal directional coupler that lowers the RF power hitting the actual detector
2) those idiots writing the data sheet did not put down a response time or 3 dB bandwidth for the video output. If your modulation rate is too high, this detector might not work...or worse might work for some data bit patterns but not others. You will have to try it yourself with an oscilloscope to watch the leading/falling edges of the video output. You might need to play with R2,R3,C4 values for it to work.

 

[Dr. von Rosenstein]

thank you, that helps me!
i want to build a simple receiver to test a transmitter (connected using SMA-cable) without great effort. in a later stage i build the detector using schottky-diodes. but in this first stage i thought i use such a directional detector, because of the simple PCB compared to that one with a "full wave voltage doubler diode rectifier demodulator".
because of the connection using SMA-cable i think the needed power is no problem. only the response time i'm worrying about.
i intend to use a hysteresis comparator to convert the Vdet from the detector into a digital-signal.

 

[FvM]

I think, a single schottky diode what do for the basic receiver, needing no bias voltage at all.

The Avagotech chip isn't bad, but actually the solution for a different problem.

 

[Dr. von Rosenstein]

another reason why i thought this detector is a good idea is that it has the 50ohm input-impedance needed by the SMA-cable. at the output i thought to use a simple SMA-terminator. if i build the demodulator using a diode, i have to ensure the 50ohm myself.

 

[biff44]

yes, matching to a schottky diode detector is always a challenge, since the impedance is high.
but in practice, most do not try. They simply put a 3 dB pad or shunt 50 ohm chip resistor in front of the detector and call the combination "matched".

For 24 Ghz, you would put series cap (maybe 1 pF) at rf input, shunt 50 ohms chip resistor, series schottky diode, shunt 10 pF cap, then video out. Keep everything really close together.

 

[FvM]

Depends on exact device parameters, but the impedance of typical microwave detector diode will be high compared to 50 ohm at low or zero bias, so you can still use a termination. Or accept the mismatch.

P.S.: I see, biff44 has already answered the question with detailed suggestions.

 

[Dr. von Rosenstein]

hm, know i am undecided.

 

[jiripolivka]

hm, know i am undecided.

Everything depends on the top modulation frequency you need to detect. Most detectors on the market are designed to have the video bandwidth of ~10 kHz and their video output has often a large shunt capacitor inside.
Terminating the video output in 50 OHms allows to obtain a better frequency response but then the sensitivity is reduced.
The best way is to make your own detector.You can use HSMS series Schottky detector diodes or a microwave mixer diode. It can be mounted in a waveguide or in a coaxial or stripline. Matching both sides in 50 Ohms allows to extend the video bandwidth up to several GHz if needed.
Or take any detector at hand and take data on its video bandwidth by modulating the 24 GHz signal source. Use 50 Ohm video termination and try to increase this value to see the video output.

 

[Dr. von Rosenstein]

the word "video" you used i read several times during my research (-impedance, -resistance, -bandwidth, -output etc.). i'm not sure what this means. in no book i've found an explanation. in the application note of the directional detector () on page 4 is a passage "video bandwidth". is this the maximum frequency of the modulation i am searching for? 30 MHz? is "video" similar to "data"?

 

[biff44]

not sure where the term comes from, maybe back from analog TV days. But basically the output of an RF diode always has some series resistance and some shunt capacitance. So even though the diode itself might be quick enough to rectify GHz rate signals, the DC output can not change very quickly because of that RC time constant (and what is loading the diode, like load resistance/capacitance). A big part of the series resistance is the diodes own "dynamic resistance", which you can measure by the slope of the I-V curve. Often the rising edge of a detector output is quick, but the falling edge (where the diode turns off) is slow unless the diode is loaded with 50 ohms.

To determine the "video bandwidth" one AM modulates an RF signal, applies it to the detector, and sweeps the modulation rate from perhaps 100 Hz to 1 MHz. At some point the sine wave video output drops by 3 dB, and that is what you call the "video bandwidth". Very useful when you have high data rate digital AM modulation....if the video bandwidth is not wide enough, the modulation bits average out to a flat DC voltage and can no longer be detected. i.e. you diode envelope dector becomes a DC power rectifier instead.

 

[Dr. von Rosenstein]

which of these resistances is the "video resistance/impedance" i have read about? i found values of 8k-10k ohm video resistance.

the video bandwidth is based on a sine modulation? ok. this means for a square-wave modulated signal that the video bandwidth should be much higher than the modulation frequency to achieve a tolerably steep edge? then i think the 30 MHz should be enough for a 100 kHz ASK-modulation. what do you think?