Limit the gain of grounded gate amplifier
by adding an input potentiometer (but what about input impedance variation?) or by somehow affecting the gate of the fet?
Hello,
If you want to reduce the gain, you may increase the value of the 180 Ohms resistor and put a resistor from the filter to ground to make sure that the filter sees the same impedance.
when you only increase the 180 Ohms resistor (between filter and fet), you will change the filter curve.
Note that the actual input impedance of the amplfiier is about 180 OHms + 1/gm (of the fet).
To ensure that the I/O impedance stays the same, wouldn't it be nicer to just have a 50 ohm input attenuator like this Spectrum analyser input attenuator instead of affecting the amplifier stage?
Most simply by reducing Id and gm. Increase the 180 ohm source resistance to an arbitrary value. Applying a negative gate voltage (with respective bypass capacitor) will achieve the same.
So a switched input attenuator like those used on spectrum analyzers like this Spectrum analyser input attenuator won't be better in terms of not affecting the amplifier/filter behavior?
Neazoi: you can add (home made) 50 Ohms attenuators, but by decreasing Gm and/or increasing the 180 Ohms resistor, you will get less degradation of noise performance. If noise isn't an issue, the input attenuators will be fine.
By reducing Id alone, your large signal behavior will degrade.
The ultimate goal would be to attenuate more the powerful signals whereas at the same time attenuate less (or not at all) the weak signals. something like a log amplifier does. could such a behavior be accomplished with this simple circuit?
This is what most reveivers with AGC do. You can review some popular circuits how they achieve it. As WimRFP mentioned, you get a trade-off between large signal capability and noise. Pin diode attenuators with a parallel branch as first element have possibly better large signal behaviour. Measurement instruments often use relay switched input attenuators.
Thank you all for your replies.
Another solution for an input attenuator apart from the switched one, could be this one RF-Workbench-2 the Serebriakova Attenuator. This does not provide an accurate 50 ohm output impedance during all attenuation values, but it is simple to make.
But after all, do one really should care too much about accurate 50 ohm I/O in an attenuator, since the purpose is to attenuate and this can in some extent achieved by mismatching...?
With a gain controlled amplifier, you attenuate or amplify all signals, both the strong and the weak. You can't attenuate powerful singals without attenuating the weak ones. The only solution for that is using filters.
If you want to limit the output by using clipping (for example with anti-parallel diodes), you will create lots of "new" signals (harmonics and intermodulation products). This is not what you want for a receiver input stage. Limiting should only be done to protect input circuits from excessive input that may damage the circuitry.
Interesting! I bet these clipping diodes behave like mixers too. I think one should consider to use the potentiometer attenuator for simple qrp RX equipment, like small pre-amplifiers followed by regenerative detectors, because of it's simplicity, even at the cost of greater noise addition.
I do not know if relatively small impedance mismatches on the output of the attenuator would cause a problem to the following 50 ohm filter though.
If it is just to attenuate a narrow band (AM, FM SSB, etc) signal, the mismatch isn't a problem, so you can use a potentiometer circuit. In applications where the shape of the filter curve is of importance, deviation from the design impedance will change the filter curve.
For your application, circuit noise isn't an issue, as you only attenuate when the signal is (too) strong (that means your circuit noise is small w.r.t. the signal power) .