FET Regenerative Receiver

This fully portable receiver has superhet sensitivity and selectivity. Speaker is 4".

This MW receiver is based on the ETI-062 "Simple AM Tuner" which appeared in ETI for March 1977. It was also published in ETI's "Project Electronics", a compilation of various ETI projects. The now defunct Dick Smith Electronics sold kits for each of the projects, but in 1980 DSE published their own project book "Dick Smith's Fun Way into Electronics" and Project Electronics faded into history.
For non Australasian readers, ETI was "Electronics Today International", an Australian based electronics magazine, which in the 1970's and 80's was the main competitor to "Electronics Australia".

ETI did not specify any particular method of construction or use of an enclosure.

My first experience with the ETI-062 goes way back to 1978. Unfortunately, with very little knowledge of electronics at the time, I was not successful with the kit.
Around 1986, frustrated with ZN414, and not knowing how to improve its performance, I decided to re-visit the ETI-062. I simply built the circuit up on a breadboard and was quite astounded at its performance. [An improved circuit configuration was developed for the ZN414 which resulted in considerable improvement. See here.]
It didn't take long to notice the amazing sensitivity and selectivity, which was on par with a superhet. It was quite a simple matter to listen to 3XY from Melbourne in Sydney at night with just the ferrite loopstick aerial. I ordered the printed circuit board from RCS radio to make a more proper construction of the circuit. This was simply mounted on a piece of wood, with an aluminium front panel to support the speaker, dial, and other controls. A two transistor amplifier was used to drive the speaker.
The ultimate goal was to make a portable receiver out of the ETI-062, but that took another 30 years to get around to!

The ETI-062 Circuit.

The design of the circuit is unlike anything that I have seen published anywhere else. Although the article did not give the author's name, it appears credit must go to Barry Wilkinson, one of the project designers for ETI. It is the unconventional design which is behind the unusually good performance. It is certainly not "just another regenerative receiver"!

Unlike most solid state regenerative receivers, this one does not need a low impedance winding on the aerial coil to feed the first transistor. This is because the first transistor, Q1,  is a FET. Not only that, the FET is connected as a source follower, with R1 as its load. The result is an extremely high input impedance seen by the aerial coil. The Q of the tuned circuit remains high, and good selectivity is retained. Q2 is a buffer connected as an emitter follower, which reduces loading on the FET. The output of Q2, which is the amplified, but undetected AM signal, proceeds two ways. One is to the detector, Q3, via an RF gain control. This transistor has its bias set so it operates in class B, and thus detects the AM signal. At the collector of Q3 is the recovered audio signal. C5 filters residual RF.
Q4 provides further audio gain prior to feeding to an external audio amplifier. C7 provides final filtering to any RF that may be present.

The second signal path from the output of Q2 is to the regeneration circuit. This consists of two turns of wire around the ferrite loopstick, and a 22K pot to adjust the amount of out-of-phase signal fed into this coil. In strong signal areas, ETI suggested the regeneration circuitry was not required, and this was found to be true. In this case it becomes a local station only receiver.

ETI's suggestion was to feed the audio into their ETI-061 amplifier, a simple four transistor complimentary symmetry design. I'm not a fan of class B amplification, so I used a two transistor class A design instead. On its own, the ETI-062 will drive headphones, despite the poor impedance matching. Obviously, the higher the impedance of the phones, the louder the sound will be. I found 8 ohm phones to be tolerable in a strong signal area.

Construction of a Portable Receiver.
The receiver was built on an aluminium chassis, attached to a Marvi-Plate steel front panel. This assembly slid into a plastic case with a handle and internal battery.

Rear view of chassis. The PCB is between the tuning condenser and the chassis. On the right side of the chassis is the audio amplifier and output transformer.

The tuned circuit consists of a ferrite loopstick with pre-wound coil, once sold by Jaycar. The tuning condenser is an obsolete MSP (AWA) type which is adjusted by a miniature vernier dial on the front panel.

Closer view of the RF section. The audio output transistor is visible in its heatsink at the rear of the chassis.

Underneath the vernier dial are the three other controls; regeneration, RF gain, and on/off volume. The pots also secure the chassis to the front panel. On the other side of the chassis is the aerial coil, audio amplifier, and 4" speaker.

Close-up of the audio amplifier circuitry.

The loopstick was mounted on brackets sharing the upper speaker bolts. 3/8" rubber grommets secure the rod in the brackets. Note that the brackets are notched at the grommet holes to eliminate the shorted turn that would otherwise be formed.
The audio amplifier was built on tagstrips and the transistors mounted in sockets.

8 x AA cells provide power for the receiver.

I considered several options for powering the receiver. However, I had to keep in mind that once the chassis was in the enclosure, there wasn't actually a lot of room left. I had a 14.4V Li-Ion battery which would have been an ideal choice otherwise. This would necessitate the use of a special charger circuit to balance the charge for each cell, but these are available cheaply on eBay pre-assembled - which they really need to be, because of the surface mount components.
Ni-Cd's are of course an option and easy to charge. As the audio amplifier needs 12V, this would require 10 such cells. The off putting feature however, is they're liable to leak. Batteries are by their nature chemically unstable. In my years of servicing, batteries would have to be the most problematic component, before modern electrolytic capacitors - another chemically unstable component. Left unattended, most batteries will release their corrosive electrolyte. At best it goes no further than the terminals. At worst, the corrosion travels along the connecting wires to the PCB or switch. Additional to this is the incidental corrosion of nearby metallic parts.

Out of the lesser of evils, I chose AA alkaline cells. These are cheap if bought from the right place. In view of the convenience, this the choice I took. Unfortunately, many types of alkaline cell start leaking early on, well before their capacity has been used up. This seems to be a recent phenomenon too. I don't ever recall Duracells or Energizers leaking back in the 1990's. So far, I have found Varta to be the only brand I've not seen leak. And they're cheap from Bunnings in bulk packs. So cheap in fact, that I made a 45V B battery with them for this receiver.

Two four-cell AA holders were mounted on the back inside the case. As the plastic here is thick enough, I was able to use short 2.5mm screws tapped into the plastic to secure the holders. Hot melt glue was used to secure the wiring.

Circuit of the new receiver.

Due to components I had available, a slight departure was made with some capacitor values in the ETI-062 circuit. C4 and C6 are now 4.7uF. C8 was changed to 220uF. These values are non critical and have no effect on performance. The ETI-062 is fed from the 12V supply via a 390 ohm decoupling resistor.

Audio Amplifier.
Turning now to the audio amplifier, two PNP germanium transistors are used. These, and the output transformer were obtained from the remains of an Astor G5G portable radiogram.
The circuit is one I've used many times before. However, because of the PNP transistors and the positive supply, the amplifier circuit is connected upside down so the transistors receive the correct polarity.
The first transistor is a 2N408 and this is direct coupled to the output transistor, an AT128. This is just Astor's in-house version of the AC128. Output transistor current is stabilised by the 47R emitter resistor. Current through this provides bias to the 2N408. If the AT128 emitter current increases, the 2N408 is biassed harder on, removing bias current from the AT128, and so its current drops. It is an effective stabilising circuit which allows for quite some variation in transistor characteristics. Output transistor current is 30mA to suit the output transformer.

The output transformer has an impedance ratio of 375R to 4R. In its previous life, it operated into a 3.5R speaker, but the half ohm difference is trivial. Because of its large core size, bass response is very good. The speaker is a 4" 4 ohm type. Included in the speaker circuit is a stereo 3.5mm socket for headphone use. Both left and right channels are connected in parallel.

Because of the very high audio gain of this circuit; Q3, Q4, the 2N408, and AT128; instability will occur at anything more than a moderate volume control setting. Output of the ETI-062 is not quite enough to drive the AT128 to full output on a weak station, so the 2N408 is required. However, gain is then too high. To bring the gain down, the 2N408 emitter circuit has a 220R resistor to provide degeneration.
Voltage gain from the 2N408 base to the speaker voice coil is 13.6 times. Sensitivity is 87mV P-P. Output power before clipping is 44mW. That may sound like an absurdly low power, but with a good sized baffled speaker it easily provides room filling volume. It is well to remember that most domestic audio equipment is normally operated at less than 100mW, despite, say, the typical 3W rating of a mantel radio.

The first thing that strikes the user is the selectivity. With critical adjustment of the regeneration control, the bandwidth can be reduced to the point where the higher frequency audio components are lost. Selectivity is knife-edge, and with the regeneration adjusted correctly, the vernier dial is really needed. Needless to say, stations on nearby frequencies do not cause interference.
Because of the incredibly high Q of the tuned circuit, anything brought near the ferrite loopstick will detune the receiver. Once the receiver is placed on a table and tuned in, this is not a problem, but if it is then lifted by the handle, the tuning will shift slightly.

The regeneration control is really the only deficiency with this receiver. Unfortunately, it suffers backlash. That is to say, if the control is taken just a little to far, the receiver bursts into oscillation and the control has to be turned back considerably to stop it. For distant station reception, tuning the receiver requires steady adjustment of the regeneration control while the tuning is adjusted, not letting it go over the threshold.

For all the years I've used this receiver circuit, I have not once found the RF gain control necessary. Conceivably, the detector could be overloaded in areas of high signal strength, which is why it was included. The idea of using the RF gain control as the volume control, allowing the omission of the normal volume control is interesting in that one control could be eliminated. It does actually work in that the volume can indeed be controlled. But, a few problems make it impractical to use it this way. Firstly, the audio amplifier is operating at full gain all the time, the result of which is some noise audible at "low volume" settings. There isn't the audio amplitude sufficient to mask it. Secondly, there is DC flowing through the RF gain control pot which results in a small scratching sound as the pot is rotated. And, thirdly, there is a slight interaction between the RF gain control and regeneration circuit, which becomes evident when the regeneration is critically set.

Both sensitivity and selectivity are the same as a superhet with a ferrite loopstick. In some instances it's better. This receiver is quite suitable for DX reception.

Update 27/02/23: It was discovered that at anything above low volume settings would result in supersonic oscillation in the audio amplifier. While not audible, it caused an increase in battery drain, as well as reducing sound quality. The amplifier was modified thus: