ZN414 TRF radio

[GK]

I grew up reading about ZN414 radio circuits in electronics magazines, but never built one. On a order of unrelated stuff I had going a while ago with Wiltronics, in a nostalgia trip, I plonked on a few ZN414 in their current guise (TA7642) along with that specific model of AM broadcast band receiver tuning capacitor and ferrite rod antenna which have been part of the jelly bean parts stock of our popular kitset suppliers for longer than I have been alive.

So now I am finally going to build that ZN414 radio. Here is where I am at after bread boarding and testing the prototype. A proper PCB layout is up next and the finished design will get assembled into a proper "professional" enclosure.

Some notes on the apparent complexity:

1) I want this to be a portable (battery powered) radio with the ability to drive a speaker at a decent volume.
2) I want low quiescent current drain and long battery life.
3) I want that low quiescent current drain without horrible crossover distortion.
4) The design is to use the same garden variety parts that I could have bought with my pocket money 30 years ago.

There are plenty of designs out there that augment the ZN414 with a discrete transistor power amplifier to drive a loudspeaker at acceptable volume. Most of these designs, if intended to have low quiescent current consumption, are quite terrible. They generally have essentially class C power output stages and need a supply voltage of at least 9V to deliver an acceptable output to a speaker. The "class C" output stages are due to the general difficulty of properly and stably biasing a complementary emitter follower output stage just a bit beyond the very verge of conduction. So as to keep the quiescent power consumption as low as practical, a typical design would "bias" the power output transistors below the threshold of conduction in the quiescent state and rely on global negative feedback around the amplifier to clean things up. The problem with this under bias is that it generates a dead band in the power output stages transfer characteristic that no amount of negative feedback can fix as the amplifier can't instantaneously slew through the dead-band. Anyone with a 'scope who has tried to use something like an LM324 for audio above a few 100 Hz knows of this problem.

The result of all of this is low quiescent current consumption, but terrible amounts of crossover distortion and, generally, gravelly and not very pleasing El-cheapo walkie talkie-like sound quality. Aside from their abundant shortcomings, well designed TRF receivers and in particular the ZN141, are well known to produce decent quality demodulated audio. It would be a pity to mate any contemporary effort with the ZN1414 with a crappy audio power amplifier.

After some study and experimenting, in order to drive a speaker to an adequate volume even on faint and distant stations, I decided than an audio stage gain of at least 200 (46dB) is required. There are designs out there which achieve this much audio gain in something like a single 4-transistor, speaker-driving power amplifier. However the requirement to operate all bipolar transistors at low currents (resulting in low transconductance) and the fact that a non-Darlington complementary emitter follower power output stage doesn't present a very high impedance to the collector of the preceding driving stage, typically results in an amplifier with barely enough loop gain to stabilize the DC operating point, let alone enough to provide any worthwhile distortion reduction. To mitigate this problem, I split the total audio gain between the power amplifier and an additional (single transistor) pre-amplifier stage.

With respect to 2) above, when driving a loudspeaker at decent volume something like a 9V battery, which was typical in the day, isn't going to give a great deal of battery life. I don't fancy the idea of lugging a car battery around with me either, so I figured that a 6V lantern battery would be a good compromise for the power source. However, this makes the power amplifier design even harder. I tried a few of the established low-power circuits designed to run on 9V or more and found that in general, into a 4 ohm speaker, they typically struggle to deliver even 1V peak before pooping out. The problem is in delivering enough base current drive to the output transistors from a preceding stage having a very low standing current.

Complementary Darlington outputs for high current gain are ruled out due to the two additional Vbe drops subtracting from my measly 6V rail. In order to get decent current drive from a voltage amplifier stage not having an excessive standing current, I went for a full complementary, push-pull amplifier design. The end result is the ability to deliver 3.4V pp of audio into a 4 ohm load and 4V pp into an 8 ohm load without any visible crossover distortion throughout the entire audio band. Scope photos are attached below. This level of drive voltage really does result in quite a high level of volume of clean audio from a speaker with decent efficiency.

I solved the power transistor thermal stabilization and biasing problem with the complementary quasi-current mirror connected power transistor pairs. In the final build these four TO-126 transistors will be thermally bonded and stacked with single 3mm bolt through the lot of them. Though even on the breadboard, without any close thermal coupling, the thermal stability of the output stage quiescent current, at a little under 3mA, really is extremely stable. This is basically due to a combination of relatively large emitter degeneration and the very low power supply voltage. With only ~3V across each power output transistor, you just aren't going to get any major heating and thermal runaway.

The quiescent current for the complete receiver is 7 - 8mA. That should give many hours of operation from a 6V lantern battery.






Unloaded clipping:



7.5 ohm load:



4 ohm load:



 
EDIT: fixed a couple of errors in the schematic.

 

[Audioguru]

An AM radio sounds awful. The poor frequency response cuts frequencies above about 2.5kHz so three octaves of sounds are cut. The clicks and pops of AM are also awful. FM radio reception is MUCH better.

[GK]

An AM radio sounds awful. The poor frequency response cuts frequencies above about 2.5kHz so three octaves of sounds are cut.

Errm, the ZN414 implements a TRF receiver with a single tuned stage. The actual audio bandwidth is determined by the Q of the tuned L/C "tank" circuit, and the selectivity actually varies from one end of the AM broadcast band to the other, but the audio bandwidth is for the most part much higher than 2.5kHz.

[vk6zgo]

An AM radio sounds awful. The poor frequency response cuts frequencies above about 2.5kHz so three octaves of sounds are cut. The clicks and pops of AM are also awful. FM radio reception is MUCH better.

This is only because of the abysmal quality of modern AM radio designs.
At least in Australia, AM Broadcast transmitters are better than 2dB at out to 9 kHz (it used to 10kHz, but they reduced the "channel spacing").

Well designed Superhet receivers can reproduce pretty close to that limit.

Unfortunately, the "AM" section of most "AM/FM" receivers is an afterthought, with similar specs to the
"pocket" transistor sets of old.
Within the service area of a MF AM transmitter, signal/ noise ratio is not very much different from that of FM.

Back in the heyday of TRFs the term was normally used for radios with several tuned circuits, like an antenna tuned circuit, an RF amplifier, then another tuned circuit, & so on.

Things like the ZN414 were regarded as "a single tuned detector with several stage of audio amplification".

[GK]

Some revision an additional complexity. I added a pair of current sources so that the tail currents of the long-tail pairs are no longer a strong function of the power supply (battery) voltage. Basically before doing this crossover distortion started to become faintly evident on the oscilloscope trace when the supply voltage was reduced down to 5V - the tail currents of the long tail pairs define the standing currents throughout the entire amplifier. Now this doesn't happen until the supply is reduced down to around 4V or less.

Probably getting a little overkill compared to a typical ZN414 radio circuit, but all just for fun. I think I might refine the bias regulator for the ZN141 as well, so that the 1.5V provided is better maintained down to a supply voltage of 4V. 

[floobydust]

It looks like a very good discrete audio power amplifier. I wonder if you could get more gain out of it and eliminate the transistor stage after the ZN414.

If you look at historical pocket radio builds, they use an output transformer with Class AB driver.
Modern stuff uses chip amp, such as the LM386 which sounds terrible, I hate that IC and it's noise and distortion as a >40year old design. It's almost an AM radio by itself.

To get the voltage efficiency up, some chip amps use bootstrapping to get a bit more swing making up for losses on the Darlington top. But they're still 10% THD and a big compromise on PNP laterals compared to using discretes.

I made a ZN414/TA7642 radio to experiment with lightning detection and as an AM radio it's got poor selectivity and picks up many stations overlapping. The 0.1uF cap at the ZN414 output does roll off highs quite a bit, you can use a lower value. I'm using a red LED as a zener for the 1.6V supply.

[GK]

I get 3 stations in with decent separation as they're not particularity close to each other, but one transmitter is local and the RF gain needs to wound right down and the tuning off center otherwise that station just overloads the ZN414.  This chip has a pretty limited and arguably almost useless AGC range on its own accord; more so if there are any kind of strong station nearby.

And yeah, the generally specified 100nF corners with ~680 ohms pretty low 

About this project I'm suddenly rather underwhelmed and am starting to think; meh.  For the most part the ZN414 makes a pretty crap radio. It might be interesting to take the thing out camping and string out a long wire antenna to see how many distant country stations can be pulled in, but around the house I think I'm over the nostalgia trip already. The neighbors WiFi (if  not something else) seems to add an unpleasant low-level growling to any AM reception around here and it seems whenever I try to tune into something in the evenings the most I seem to get is theologians and assorted religious types just talking out of their arses.

Back onto other less frivolous things now I think.

[HB9EVI]

An AM radio sounds awful. The poor frequency response cuts frequencies above about 2.5kHz so three octaves of sounds are cut. The clicks and pops of AM are also awful. FM radio reception is MUCH better.

I don't agree to that statement. When testing any kind of LW/MW/SW antennas I really stuck sometimes on the BC bands listening to the program for quite a while. Certainly AM-BC is much more audiophile than listening to the hamradio transmissions with their 300-2k7 bandwidth.
Of course with a lousy S/N ratio AM is barely pleasant, but that's a matter of reception conditions and not AM by itself.

I still have a ZN414 from my teen days experiments. I wouldn't bother anymore building something with it but it's still a funny experiment

[TurboTom]

I remember having built a few 27MHz AM R/C receivers in my teen years, utilizing the ZN414 as the 455kHz IF amp / demodulator. These units performed pretty nicely with good sensitivity and selectivity. At the price of an OEM R/C receiver back then these DIY versions were quite economical as well (which was (-is?) always a concern for an teen  ). And due to the simple design and low parts count, they were quite small despite the home-brew PCB technology... Unfortunately, none has survived the decades since, so no photos to prove.