Help circuit compares TWO FREQS, one fix one sweep, trigger

[gmcjetpilot]

The idea is compare a FIXED RF signal (any where from 455Khz to 12 Mhz at least, more range the better) with a sweeping Frequency (RF signal of fixed voltage). The idea is when they are the same, it triggers a pulse to the Z intensity input of the scope.

**broken link removed**

 

[biff44]

Re: Help circuit compares TWO FREQS, one fix one sweep, trig

Looks neat! How does it work?

 

[keith1200rs]

If you mix (multiply) two frequencies together you get the sum and difference frequencies. When they are exactly equal you will get a DC value equivalent to the cosine of the phase difference (I think).

Depending on how close you want the frequency match, you could detect the DC, or more likely detect a very low frequency which is close enough (say, less than a few hundred Hertz).

Keith

 

[gmcjetpilot]

Re: Help circuit compares TWO FREQS, one fix one sweep, trig

If you mix (multiply) two frequencies together you get the sum and difference frequencies. When they are exactly equal you will get a DC value equivalent to the cosine of the phase difference (I think).

Depending on how close you want the frequency match, you could detect the DC, or more likely detect a very low frequency which is close enough (say, less than a few hundred Hertz). Keith

Do you have any part and circuit ideas?
Here are the design parameters.

Sweep - Rate about 60 Hz
Sweep - Width (Frequency) about 300 Khz to 1 Mhz. No more than 3 MHz
Fixed Freq - Would be close to center of Sweep width, or symmetrical sweep.
Signals are unmodulated and have control of attenuation of both


A few hundred Hz? Not bad. I would like to keep it in the 100 hz ball park. Giving it some "calibration" that might be improved with external checks, if the circuit had some tune-ability, such as hold off and trigger duration.

 

[keith1200rs]

I have limited access to data at the moment but try the MC1496. The matching frequency will be determined by the low pass filter you use and can be as low as you like.

Keith

Added after 2 hours 7 minutes:

By the way, your detection frequency will be limited by your sweep speed. In other words, you cannot detect with a low frequency difference with a fast sweep speed. The lower frequency difference you want to detect the slower the sweep speed required.

Keith

 

[gmcjetpilot]

Re: Help circuit compares TWO FREQS, one fix one sweep, trig

By the way, your detection frequency will be limited by your sweep speed. In other words, you cannot detect with a low frequency difference with a fast sweep speed. The lower frequency difference you want to detect the slower the sweep speed required. Keith

Thanks Keith. I can sweep as slow as needed, like 1 Hz, but 60 Hz is normal. An slow sweep is not viewable on the oscilloscope. I can handle a blinking display. I have an analog Tek 2225 scope, not a DSO.

Analog wise when two freq mix simply a fixed Freq (call it a "Marker") and sweeping Freq, they heterodyne. When the sweep goes past the fixed Freq, you get a BEAT Freq or harmonic. On a scope it looks like a big wavy line. It is very visible. The PROBLEM is these markers are TOO big for low freq narrow band width sweep of RF circuits.

If you turn up the amplitude of the fix the harmonic marker it gets bigger but also wider. That is a problem EVEN on the lowest power you can see the "Marker"; it is too wide for low band width traces. The marker is 20-40Khz wide. This is OK when sweeping a wide band width. However at lower Freq like 455 Khz and 10 Khz band width, a FAT marker is no good. It overwhelms the trace and obliterates it. It is useless. This is where a synthetic marker of a pulse of finite width (time) of sweep is needed. One way is use the Z intensity of the scope to make a marker. The circuit just needs to respond to the peak of the harmonic beat freq. Now this is complex I guess. It also might have the WIDE issue (in time), but I think with the right filters and circuit logic it can get with in 100 or 200 Hz..... that is better than a 40 Khz wide blob marker, especially if it is a sharp defined mark. This will aboslutly tell me what freq I am at along the trace. This indication is used to align the RF or IF section of a radio, for example.

I found some RF comparator IC's that compare power, phase shift and other such good things. These are the two. Not sure I could use them but they tend to go for $17-$28 each!

MAX2016
**broken link removed**

AD8302
http://www.analog.com/static/imported-files/data_sheets/ad8302.pdf

Even if they could be used or something like it, I am running out of technical design horse power to integrate the whole design.

 

[keith1200rs]

If you mix without low pass filtering I think you will find it more difficult to interpret the signal on an oscilloscope.

What you seem to be trying to make is a spectrum analyser!

I would caution against using 2.5GHz devices when you only want 3MHz. You may get all sorts of layout issues, the chips are expensive and really overkill. You would need to ensure a 50 ohm system and I am not sure how they would help your problem anyway. I have used the AD8302 for phase measurement at up to 1GHz.

I think you need the following:

1. mixer (multiplier/double balanced mixer/Gilbert Cell).
2. low pass filter (a few hundred Hertz depending on sweep rate)
3. peak/envelope detector

There may be other ways of achieving what you want. I assume you are trying to simply match two frequencies?

Keith.

 

[biff44]

Re: Help circuit compares TWO FREQS, one fix one sweep, trig

A lot of fm radio chips use gates or flip flops as frequency discriminators. It compares rising edges of the two frequencies, and can tell which one is higher or lower. Since it is a quick response, it seems like a good thing to pursue.

Look at when one of the rising edges goes from leading to lagging and that was your frequency crossing point.

Trying to do the same thing with a mixer and a zero beat note is not easy to do when one of the signals is sweeping, as it takes a long time to figure out it is really zero hz beat note.

If I were going to look at the beat note out of a mixer, I would use an I/Q mixer, and instead look at when the frequency went from positive to negative. In other words, the frequency is 1 khz, 100 hz, zero hz, -100 hz, -1khz. It might be easy to detect the sign change.

Another thing to investigate would be how rubidium oscillators work. The rubidium gas absorbs a specific microwave frequency, and the light output dips when you are right on frequency. I do not remember the details, but the use the derivative of the light output to frequency lock the crystal to the rubidium cell. They use a differentiating circuit. Maybe you should also be looking at the derivative of the mixer or frequency discrimanator output.
To be clearer, the light detector is not-monotonic. In other words, as the frequency gets closer to the rubidium resonance, the light output gets smaller....smaller...smallest (right at resonance), then bigger...bigger...bigger. By differentiating the light output, it is easy to detect when the resonance happened.

 

[gmcjetpilot]

Re: Help circuit compares TWO FREQS, one fix one sweep, trig

Thanks GOOD replies, appreciate it.