Labs 7-9: Heterodyne radio
Introduction
For reference, here, here, and here are the lab manuals that were followed.
While the basic radio is great, it doesn't select stations at all. This means that two stations on adjacent bands would likely play over each other.
There may be some freaks out there who want to listen to two talk shows at once, but it isn't me so filtering is an important step
Note: Due to how horrible our antenna is, we will likely only pick up the one strong station so this doesn't really matter
but the theory is of interest for real radios with decent antennas
Lab 7: Bandpass Filters
Filtering for a particular station is a simply a matter of filtering for the carrier wave plus and minus 5kHz.
This can be acheived by a bandpass filter but problematically, this would have very poor tuning capabilities, so the construction
would dictate the station received not the user. The next two labs fix this by changing the carrier wave down to a fixed frequency
regardless of the station, so for now just know that we want to filter for around 200kHz plus and minus 5kHz. This is also advantageous
as it reduces the necessary quality factor of the filter.
The bandpass constructer here is 1 stage passive filter. Won't be the highest quality but is extremely cheap. Passive bandpass filters
take the form of the picture below:
The filter was constructed with two different values for the resistors and tested.
The reults gathered are below and agree closely with theory likely only deviating due to parts being different.
| Lf1 | 10 uH | 10 uH |
|---|---|---|
| Cf1 | 0.069 uF | 0.069 uF |
| RL,RS | 100 Ω | 1k Ω |
| Fres | 202 kHz | 208 kHz |
| Flow | 181 kHz | 197 kHz |
| FHigh | 241 kHz | 221 kHz |
| BW | 25.5 kHz | 2.7 kHz |
| Q | 7.88 | 75.52 |
Lab 8: Mixer
By exploiting some high order terms in the functionality of a bjt, two waves can be subtracted to move the center frequency to difference of two waves.
Using this in a radio means that a local oscillator can be combined with the incoming signal to move the signal down to their difference. So if we want to received
1230 kHz, we set the LO to 1022 kHz and the mixer will move the intelligence down to 208 kHz, exactly the center frequency of our bandpass filter.
Making the LO is the focus of lab 8, here we just build the mixer. Another channel on the fgen acted as the LO for this lab.
The circuit of a mixer can be seen here:
Building and testing this circuit yields the output below, with a clear peak at the IF of 200 kHz but also a huge number of other peak:
To fix this abundance of peaks, the filter from Lab 7 was added yielding this:
As can be seen, the IF is now only one a few signals. With the only one larger being at the LO, which is not of concern as it has no audio signal to distort the desired signal.
An easy success.
Lab 9: The LO
While it seems like the two labs above could produce a working het radio, they would be tethered to a function generator which isn't particularly useful.
The oscillator design to be used here is a hartley oscillator which you can read about here if you care.
Just know it causes rail to rail swings of an output via a transistor and a bunch of passives. The circuit below is the schematic followed.
It generates a pretty awful sine wave as can be seen here:
This will have high frequncy components that will cause the mixer to do strange things but no quick solutions were found in lab that made it a decent sine wave across the full range.
Our oscillator puts out a signal anywhere between 1 MHz and 1.55 MHz depending on the tuning of the capacitor. This is adequate for the stations we want to receive.
Conclusions and Reflections
We put the full thing together and it just didnt have the gain.
Adding an additional RF amp after these het stages rectified that but it was still extremely finnicky
to tune to a station and caused the overall system to be rather sensitive to random problems.
The biggest culprit is likely the single stage passive filter both attenuating the signal too much and not attenuating the noise enough.
Fixing this would be a matter of switching to an active filter or using a ton of additional stages.
This won't see any further developments as we have to move on to the final project but it was a good introduction to how real radios work.