How much IF gain do I need?
You can 1st make link budget, then you know the lowest receiving power level, its should be about 10dB greater than sensitivity. Assume your sensitivity is -90dBm, and say you should indicate RSSI for -110dBm, and say RF RX gain is 25dB, and the lowest level for decoder is -20dBm, so your IF gain is -20-(-110)-25dB+Margin, say margin is 5dB, IF gain is about 70dB.
Hello,
Assuming noise figure of 6 dB and RF bandwidth of about 2.4 kHz, noise will be about -134 dBm (that is 43nV in 50 Ohms).
Now you need to know the input range for your final IF to RF mixer (demodulator) with LF amplification.
Just some numbers: Imagine your maximum input voltage for the detector is 50 mV and you want the noise about 20 dB below maximum input voltage. In that case the IF noise level (at the entrance of the detector) should be about 5mV.
This means that the overall voltage gain from RF to detector input must be 5mV/44nV = 114k (101dB).
For good dynamic range, you want the gain from RF to first narrow band IF filter as low as possible. The minimum gain requriement for the RF amplifier + Mixer + filter is determined by the noise contribution in the complete chain. Let us assume gain from RF input to IF filter output of 10 dB. In that case tthe total IF gain should be in the 100 dB range. When your demodulator can accept less input signal (without becoming too noisy), you can reduce the IF gain).
When you use a double super scheme, your overall IF gain also should include the loss of the 2nd IF filter and gain of the 2nd mixer.
Depends on the sensitivity of the demodulator. Noise figures for various demodulators range from about 6 db to 20 db. The worse the noise figure the more gain required which takes away from the dynamic range of the demoduator.
The gain distribution, along with the selectivity distribution determines intermod performance.
I prefer to work in noise temperature which more obviously shows the noise floor levels at each stage. You can model each stage as an idea NF of 0 db device with an input summation of device's real noise temp (NF and noise bandwidth derived) summed with noise power output from prior stages.
Noise temp also makes you more aware of narrow band takeover. This is desired comm channel RF bandwidth in comparison to the potentially wider noise bandwidth of backend stages. Working solely in NF does not take into account that demod noise bandwidth might be significantly wider then the intended comm bandwidth. This wider noise bandwidth in backend stages can degrade overall signal to noise ratio if there is insignificant gain from the frontend stages that sets the comm channel bandwidth.
If I used op amps to get the IF gain and used resisters for impedance matching (like a 50 or 75 ohm in series with op amp output and the same value across the successive stage's input), would the net gain be 1/2 what it would be without the resisters?
Hello:
Yes, this is true, but why you need to add the resistors? For these frequencies and all stages close together, you can connect the outputs to inputs directly. To get the published shape of the filter, you need to drive from and load with the published resistor values.
IF you expect out of band instability, you may add some series resistance.
Other thing, why using opamps? If you use 455 kHz IF, general purpose transistors like BC547/847 can be used.
So, am I right to say that I would need about 10 or 11 stages of amplification if I could get 10dB out of each at the IF which is 3MHz. Should I use a different IF than the 3MHz that I planned? Other circuits are designed around this. Maybe it was a mistake.
Your calculation for the number of stages is correct
Selection for your IF will depend mostly on the filters you can get. Assuming SSB voice, you require about 2...2.4 kHz BW. The number of "poles" determines how steep the filter curve drops outside the passband.
In case of a single conversion setup, a high IF frequency will make it easier to get sufficient image rejection, but increases design complexity. Are you using a regular mixer for demodulation, or a (good) image reject demodulator (where you can select upper or lower side band)?
Is this receiver for single frequency use? If so, an image reject filter at RF is relatively easy (combining parallel resonance with series (notch) resonance). Will you use it with a small indoor antenna or large outdoor antenna? In case of large outdoor antenna, your receiver will be subjected to large RF signals. This requires a mixer with good dynamic range, in combination with good RF input filtering.
Regarding the IF amplifier. It sounds strange to me that each stage just does 10 dB. At these frequencies >20 dB/stage should be possible without getting instability.
With regards to the overall gain (in my example about 100 dB), very good layout and inter-stage power supply decoupling is required to avoid instability. IF this is your first RF project, actually building >100 dB at 3 MHz can be a challenge.
As you speak of SSB only, you may do your demodulation at lower voltage level. In that case you can save on IF gain and audio gain is easier to make.
When you finish your budget calculations, you may put a block diagram with all gain and voltage levels here. It may make discussion easier.
Can you recommend an op amp with which I can get > 10dB?
Hello,
When you want, for example, a voltage gain of 10 at 3 MHz (that is 20 dB), you need a gain-bandwidth product >> 30 MHz. If you would like to stay with opamps, try to find some opamps with gain bandwidth product >> 30 MHz. Nowadays there are lots of fast low voltage opamps, but I don't what sources you have access to.
www.national.com has many high speed opamps to give you an idea what is possible and how they specify them.