2007-07-17
The "Noisy Regen"
This circuit is essentially the same topology as the "noisy oscillator" in EMRFD (page 4.13, diagram 4.21). I've just used PNP transistors (2N3906) so I can "ground" the tuning capacitor rotor and still use a "conventional" -ve potential ground plane (making it easier to integrate with common-grounded test equipment).
Although running at a very low current (which gives the transistors a reasonably high Re) the tight coupling of the resonator with the transistors severely degrades its Q, especially as the regeneration is turned up high. It actually performs fairly well when the reaction coefficient is just less than 1, but once the -ve resistance is increased until it oscillates is becomes one of the noisest oscillators I've ever seen. Even when just regenerating it is pretty noisy and broad.
OK, so it sucks electrically, but it does have some utility. It is extremely easy to get going, virtually any inductor that isn't completely hopeless will work in the circuit. It pulls very, very little current and works from LF to VHF. It needs only just over a volt to oscillate on the lower bands, and 3 volts will work into the low VHF region. The only frequency determining components are the L and the C, although the transistor capacitances are very significantly coupled to the resonator, limiting its MUF and complicating estimating the appropriate values for a particular frequency - two seconds of empirical capacitor swapping will get you in the right ballpark. The regeneration control modulates the transistor capacitances as well, pulling the frequency a bit, but this isn't unusual for a regenerative receiver and the selectivity is so bad anyway it doesn't matter much above MF.
About 2 minutes after I assembled it on the solderless breadboard I was listening to Radio Australia on 6.025 MHz with just a short clip lead for an antenna (coupled directly to the tank at the collector though a 2.2 pF cap - the tank itself was simply a 10 uH Ohmite RFC and my low-C jig). It truly is an instant gratification receiver, and can be used with a crystal earpiece resulting in a circuit that pulls less than 300 uA, meaning no off-switch is required for a pair of AA batteries. I added a stage of audio amplification, using an 2N3904 with the familiar self-biasing (1M feedback, 4k7 collector load). Even without this, the circuit can drive the old Archer "Mini Amplified Speaker" Tandy used to sell, but the extra gain is helpful for weaker stations.
Other Uses
The circuit can be setup to super-regenerate by changing the emitter circuit to contain an RFC and a capacitance across the emitter resistor. In this mode it seems the circuit would be an excellent telemetry or control receiver; micropower, sensitive, and not too tuning critical. It would be fairly easy to integrate this circuit with a microcontroller for remote-control use. My surplus of 27.195 MHz crystals suggests some direction for future experiments with this topology.
Another use might be a LF receiver for a lightning detector. Many moons ago I built a version of Charles' lightning detector (a great little project, build one, they are fun). One thing that bugs me about it though is the need for an antenna. I think this little cross-coupled regen circuit would be perfect (being micropower) to build a lightning detector with a loopstick magnetic antenna rather than an electric whip.
Building a more Permanent Version
The "Noisy Regen" might not be much of a radio, but I couldn't help myself, I just had to give it a permanent place amongst my homebrew receivers. The circuit was rebuilt on PCB, this time with a 65 pF trimmer for tuning and a toroidal inductor wound on a T37-2 core (an additional 200 pF of fixed inductance was needed to put the receiver on the 49 metre band - this gives you a good idea of just how much the transistor capacitance is coupled to the tank, the inductor measures around 2.3 uH, which means it must see an additional 50+ pF of capacitance elsewhere to resonate near 6 MHz with the mid-trimmer capacitance).
A simple three transistor audio amplifier was thrown together to drive 32 Ohm headphones. Complete the radio pulls about 3.5 mA and delivers enough audio for comfortable listening at home. (Omitted from this diagram is a 1 nF capacitor across the 2M2 feedback resistor in the first amplifier, this rolls off the AF response and kills the HF noise.)
The PCB was cut and filed down to be a perfect friction fit in the top of a Jaycar potting case. The case itself houses half a 9V battery, the power switch and a stereo 3.5 mm headphone socket. To create the battery I dismantled a cheap 9V unit and snapped the wax-potted cell stack in half. (Some batteries have 6 sub-AAA cells instead of a linear stack of "pellet" cells, but even sub-AAA cells were too large to fit comfortably in the case.) This gives a 4.5 volt battery of moderate internal resistance. Actually the higher source resistance meant I had to raise the decoupling cap in the audio amplifier to 470 uF to prevent howling at max gain. My general purpose bench supply and AA batteries did not need this much capacitance for stability. A 10 nF cap from the emitter follower base to ground might help a little too, but I'm happy with the final result.
The antenna connection is simply a PC pin, which a clip lead is connected to, leading to a random wire for quite reasonable reception of the stronger stations. The radio tunes approximately 5.8-6.7 MHz which covers all of 49 metres, but it might be helpful to pull it a little lower to spread out the channels a little.
The little unit is quite cute. It is not a huge performer as receivers go, and you have to adjust it with a screwdriver, but it was a nice project none the less. This weekend it let me listen to Saturday Night Country on Radio Australia, but I also heard Radio New Zealand and a bunch of other stations. Sure, I can use any number of other radios I have, some with the comfort of AGC and other features, but there is nothing quite like using a radio you designed and built yourself, even one as backward and hacky as this one.
Suggested Improvements
- Add a 220-1K resistor and large (at least 10 uF) capacitor before the 10K volume control pot. This will stop AF noise on the power supply rail getting into the AF amplifier and causing howling at minimum volume with bad layouts and/or weak batteries.
- Adding a similar decoupling network to the intermediate BC549 and the detector stage can't hurt either if you are having stability problems.
- Consider building something less primitive than the emitter follower audio output stage!
- Add band-switching. (41 and 31 metres would be nice too)
- Build it bigger! (Although its small size is a large part of its charm)
Update: 2007-08-25
I use this little radio quite a lot, so much that I've flattened the battery three times now. The process of replacing the battery has grown tiresome; it involves hacking open a particular brand of 9 Volt battery and fixing up the anode contact with some care. I decided to install an external power jack (mono 3.5 mm socket) and run with conventional AA batteries in a 4-cell holder.
The "complete" radio isn't a self-contained 2 cubic inches any more, but I still get great joy out of using it.
7 comments.
Attachments
| title | type | size |
|---|---|---|
| Circuit Source | application/postscript | 14.563 kbytes |