Dual Gate MOSFET Direct Conversion Receiver
Based on inputs I have received either in direct comments on the blog or from email messages, there seems to be a desire to better know the "why" of various choices and actions. In this particular post we will look at the Dual Gate MOSFET being used in a Direct Conversion Receiver (DCR). We will also use an LT Spice simulation to demonstrate the "why".
In the Part 5 we described the Direct Conversion signal frequencies and the resultant outputs. Essentially to receive an 800 Hz CW signal on the 7.030 MHz QRP frequency we would need to supply a local oscillator signal at either 7.0292 MHz or 7.0308. Again the Direct Conversion Receiver is simple to build and is quite sensitive; but does not give single signal reception. You will receive the same signal at TWO places on the dial!
Below is a schematic representation of a Direct Conversion "detector" using JFET's configured as a Dual Gate MOSFET. We have chosen two JFET's for this evaluation and includes the very popular 2N3819 and the 2N4393. The same circuit was used to evaluate both devices and the only difference is the device used in the test bed circuit. The incoming signal has been set at 0.3 Microvolts and the LO is 1.414 volts. The output scan is from 10 Hz to 50 kHz. [Note these photos are GIFs as I was severely criticized by one blog reader for using JPEG's the reasoning for which is still not clear to me but to stop receiving emails --they are GIF's.]
You only need to connect the audio amplifier to this circuit, connect an antenna to "Gate 1" and your favorite Local Oscillator to "Gate 2" and you are in business. Keep in mind the LO signal frequency must be in the same range as the incoming signal.
As homebrewer's the usual process is to dig into the "junque box" and find a couple of devices and heat up the iron AND then wonder why the receiver seems dead. Well this is where my term "noodling" comes into play. Turning on the soldering iron should be the very last step. Evaluating what you are doing and how you are doing it IS the first step!
If you start with the circuit above utilizing the LT Spice simulation you would see below the following expected response. In the first case we use the ever popular 2N3819 and in the second case is the 2N4393. The output curves are very similar BUT the 2N4393 has about a 6dB greater gain than the 2N3819. Will they both work --yes but the 2N4393 is more sensitive to weak signals.
Keep in mind that you need a real antenna to hear signals. Throwing a 10 foot piece of wire on the ground is a compromise (a poor compromise) in comparison to a 40 Meter Dipole at 30 feet. Hooking your Direct Conversion Receiver to a rain gutter may provide some signals but is not as desirable as a real antenna. Put you time and resources into an antenna and then you can remove that as a variable in the reason "why the receiver seems deaf".
Having a signal from a Local Oscillator is one thing but having a signal of one volt or better (1.414 Volts Peak to Peak is 7 dBm) is needed to make this play. Some of the LO devices (AD9850, Si5351) are output frequency sensitive. Thus at lower frequencies there is plenty of output; but at higher frequencies the voltage output drops off and thus may not be sufficient for the mixing process. Thus an outboard amplifier may be needed between the LO and the detector. But unless you measure the output over the frequency ranges you may not realize that insufficient LO drive is causing marginal performance.
In the next post I will have an actual test hardware amplifier that is based on Parts 5 & 6.Stay tuned!