a 74xx-defined radio

Video Appetizer


I built a shortwave radio receiver from scratch using only cheap and easily available components, i.e. standard transistors, op-amps and 74xx logic chips. No typical radio parts – no coils, no variable capacitors, no exotic diodes. This project is easy to build and gives you a hands-on experience with radio technology which you won’t get from a fully integrated SDR.

Here is the schematic.

Edit: If you don’t actually want to build a radio, but you want to have an FX pedal to make sth sound like an old radio, do the following: Build the radio as shown, but exchange C13 for 1nF. Feed your guitar/synth/audio source into the antenna input and turn the tuning knob until you “receive” your instrument sounding pleasingly LoFi-ish.


A friend of mine had an idea: He wanted to build a guitar FX pedal that is essentially a short wave radio receiver. It should use the guitar cable as an antenna and it should be very straightforward to build. No weird coils, no exotic rotary capacitors, no Russian voodoo detector diodes. And it would be okay (or even desirable) if it’d be LoFi. He had some weird sound experiments in mind, probably in combination with synthesizers and funny effects.

So he asked me if I had an opinion on how to build such a thing. I answered that I am not the “radio type” of tinkerer. I had never built anything resembling a ham radio and had never played with a software defined radio receiver. But I said that “at least I think I understand the general working principle of AM radio“.

So, long story short, my friend had me “nerd-sniped” and I found myself doing some LTspice simulations … modulating/multiplying sine waves, etc. Let’s sum up the rules (“dogmas”) again that I set for myself:

  1. No coils and no transformers -> No tuned LC circuits
  2. No (mechanically) variable capacitors -> No tuned LC circuits
  3. No special/exotic diodes
  4. Must receive short wave broadcasts
  5. KISS (Keep It Simple, Stupid)

So essentially to build a radio from scratch with no radio parts … but at the same time as simple as possible.

The Heterodyne – An old but smart concept

So I consulted the Wikipedia page about short wave broadcast and I learned that short wave is defined as everything ranging from 3 to 30 MHz. And I figured that, due to my self-proclaimed ban of LC-circuits, I probably want to build a heterodyne receiver.

Superheterodyne receiver block diagram 2

(Picture embedded from Wikipedia)

I’m not going to explain the heterodyne receiver in full glory here, there are others who already have done a good job doing just that. Let’s just highlight the most important idea: A tunable oscillator (the “Local Oscillator”) is tuned to a frequency not identical but very close to the carrier frequency of the radio station that we want to tune in to. The received carrier and the local oscillator waveforms are fed into an “RF Mixer”, a device that essentially performs an analog multiplication. The two slightly out of tune waves will “beat against each other” in the mix, with the beat frequency being exactly the frequency difference of the input signals.

For example: We have a radio station broadcasting at 10 MHz and we tune our local oscillator to 10.1 MHz, then the signal after the mixer will have a signal component oscillating at 10.1 MHz – 10 MHz = 100 kHz. The catch: The 100 kHz signal still has the same envelope function as the the 10 MHz AM station, i.e. the same audio information that was modulated onto the carrier. We basically only “tune down” the carrier wave, without changing the information that is carried. The beat frequency is often called IF – intermediate frequency.

Why do we do that? Because the lower the frequency of the signals, the easier it becomes to amplify and filter them! So in order to listen to our station we have to have a bandpass filter and filter out and amplify only the intermediate signal at (in our example) 100 kHz. Be aware that all the other stations have also been mixed with the local oscillator and have landed at some other intermediate frequency (that we don’t want to listen to). Because we already have a tunable local oscillator that determines which station lands where after the mixing, the IF band pass filter can be set to a fixed frequency (here 100 kHz).

Now all we have to do is rectify and low-pass filter the selected IF signal to recover the envelope function (i.e. the AM audio information).

The most humble mixer

Okay, so the heart of a heterodyne receiver is a RF mixer. I did some googling and found the classical implementations:

  • There is the famous diode ring mixer, which is technically identical to the famous ring modulator sound effect you might know from pop music. It relies on signal transformers to generate galvanically insulated input and output signals. It did not fit my design dogmas.
  • There is the Gilbert Cell mixer, a clever discrete analog circuit comprising six transistors to implement an analog multiplier. The Gilbert Cell, too, has differential signal inputs and outputs. Although six transistors look like an okay amount of complexity, the prospect of biasing all of them properly turned me off a bit. There is a very nice integrated Gilbert Cell IC, the NE612/SA612, which is sort of straightforward to use, but it is not really a convenient off-the-shelf component, that you find in every hobbyist distributor’s online shop. Also it is rather priceous (around 2€) and I was only able to get an SMD package.

I was reading some more about RF mixer theory and I came across this document from analog devices which seems to be a course about RF/IF circuits. In the theoretical introduction to RF mixers they discussed an ideal (switching) RF mixer:

An ideal switching mixer

I have never viewed it like this before, but it sort of makes sense: Make two copies of the RF input signal with opposite polarities and switch between these two at a fixed frequency f_LO. This is effectively the same as multiplying (mixing) the RF input signal with a square wave of frequency f_LO. Well, mixing with a square wave is not exactly the same as mixing with a pure sine wave (because you also mix with the higher order harmonics of the local oscillator frequency) but for building our heterodyne short wave receiver it does not make that big of a difference.

In the analog devices course they went on and discussed the diode ring and the Gilbert Cell mixer as examples of real-world mixers, which gave me the impression that using an actual switch as a mixer is impractical or yields poor results. Also the usual sources of RF circuit knowledge on the web don’t seem to mention switching mixers that often.

But I wanted to see it with my own eyes. At least I wanted to build a crappy mixer to see that the principle works. Yes, using an actual mechanical switch is very much impractical, if we are talking about reversing the polarity of an RF signal several million times a second. But what about a good old analog multiplexer? For example the humble 74HC4051?

The 74HC4051 has 8 analog inputs, A0-A7, and one analog output A (or 1 in and 8 out, the switch works both ways). Internally the ‘HC4051 acts just like a single-pole 8 throw switch. The switch position is determined by the three logical select inputs S0, S1 and S2.

So in order to use it as a mixer, we first use our good old friend, the 2N3904 npn transistor, wired as a phase splitter (like an emitter follower, but you get an additional inverted copy for free). This will give us a positive and a negative (inverted) copy of the RF input. This is fed to the analog inputs A0 and A1 of the multiplexer. The local oscillator waveform has to be provided as a logical TTL or CMOS signal and is fed into S0. S1 and S2 are wired to GND, this way we will only switch between A0 and A1. The IF output (A) is then connected to the base of another 2N3904 wired as an emitter follower, just to buffer the signal in order to drive the next stage.

I tested it with a two channel signal generator and an oscilloscope. The above device behaves exactly like an ideal switching mixer! In the below example I mixed a 30 MHz sine wave at the RF input with a 31 MHz square wave at the LO input. The result is this spiky beast of a waveform. But unmistakably the dominating low frequency component is a sine wave at 1 MHz. Hooray! It might not be the best mixer ever. But it was cheap and straightforward and seems to work up to 50 MHz until you begin to see noticeable injection loss. Over the entire short wave band (3-30 MHz) the mixer works with a gain of more or less exactly 1. I am impressed.

A caveat: Please use the 74HC4051 chip and not its older and slower cousin the MOS4051.

A simple RF local oscillator

Okay, we have our mixer. Now we need a local oscillator that provides a stable rhythm by which to flick the switch. Desired frequency range: The entire shortwave band of 3-30 MHz. Luckily the toolbox of 74xx chips provides exactly what we are looking for! It is called 74HC4046 and it is a Phase Locked Loop (PLL) IC. A PLL always comprises a voltage controlled oscillator (VCO) in combination with a phase detector which compares the phase of the VCO with the phase of the oscillation that the PLL is supposed to synchronize with. But let’s not talk about PLLs. We are only interested in the first half of the chip – the VCO.

Elliot Williams from Hackaday.com has written a very nice article introducing the ‘4046 as a fun an versatile source of logic noise synthesizer sounds: https://hackaday.com/2015/08/07/logic-noise-4046-voltage-controlled-oscillator-part-one/

This basically explains everything about the ‘4046 you need to know. We’ll be using it in a similar way, but not in the audio frequency range, but at much higher frequencies. This is achieved by changing the external resistor and capacitor to smaller values. By trial and error I ended up with 10k and 47p. Tuning the oscillator (and thus selecting the radio station) is done via two regular potentiometers. One for coarse and one for fine tuning. Making the second poti a fine tuner is achieved simply by connecting it though a ten times larger resistor than the coarse tuner. This way its influence on the VCO control voltage is also ten times smaller.

A caveat: Please do use an actual 74HC(T)4046 and not it’s older and slower cousin the MOS4046 (aka HEF4046, CD4046). I got good results with a Texas Instruments CD74HCT4046 (with a T). Funny enough I got bad results (too slow, too low frequency) with a Texas Instruments CD74HC4046 (without the T), even though the datasheets claimed both chips have similar performance. If possible, check the frequency range of the oscillator output with an oscilloscope or with a frequency counter. You don’t actually have to have an oscilloscope that can measure all the way up to 30 MHz. You could feed the output of the VCO into the clock input (CP) of a 74HC4024 (7-stage binary counter) and study the output of its Q6 output. It gives you a square wave with 1/128 the frequency of the VCO (30 MHz/128 = 234.375 kHz). This is something you can study with a very cheap oscilloscope. Or you could just count oscillation cycles with an Arduino. If this is still too fast for your measurement device, then just cascade two 74HC4024 and divide the frequency by some more factors of two.

IF filter and amplifier

Mixer – check, local oscillator – check, now we need a narrow filter! Remember, we want to pick out one narrow frequency band which we want amplified, all other mixing products which don’t land near our target frequency (here 100 kHz) shall be suppressed. Once again: no tuned LC circuits! So we go for a classic op amp band pass:

An LTspice simulation helps us selecting the right values to get our filter response right. In the end we want a very high gain and a small resonance maximum. But not infinitely small: The bandwidth of the pass band should be at least around 5 kHz, otherwise you don’t have enough bandwidth to fit in a significant part of the audio spectrum.

Note that the virtual GND of the OpAmp is at half the VCC = 2.5V. This reference voltage is created with a resistive divider and stabilized by a 100n capacitor. The 2.5V reference is also used for the demodulator OpAmp.

Maybe this is not the most ideal IF filter, but it does the job 🙂


Okay. After the IF filter we have a single 100 kHz sine wave with audio information modulated onto it. To get to the audio we “only” have to rectify the waveform and smoothen it a bit with a low-pass filter.

A lot of classical AM demodulator circuits that I found on the web use a germanium or Schottky diode as a half-wave rectifier (to cut away the negative half of the IF wave). They perform better than your regular silicon diodes, because they have a lower threshold voltage and thus can rectify smaller input signals. But what if we don’t have and don’t want to buy special magic radio diodes?

Fear not, there is a nice OpAmp circuit that goes by the name of “active half wave rectifier” and makes the use of special diodes obsolete. Here we use negative feedback to actively mitigate the finite threshold of the rectifier diode (D2). The second diode (D1) is only there to limit the negative output swing during the negative (removed) half wave. By choosing the feedback resistor ten times larger than the input resistor (R14/R13=10) we also get a gain factor of 10 for free.

R15 and C10 form a low-pass filter that reduces the depth of the high frequency ripples (stuff that you don’t hear anyway). In the end we have a volume poti and an AUX-level audio output jack.


Here is the complete schematic.

As you see and hear it works … and it makes some nice cliché AM radio noises. As an antenna I used a thin wire, circa 6 meters long, fixed with thumbtacks to the wall.

If your reception is poor – do not despair. Sometimes the air is full of radio stations some hours later. Shortwave reception depends massively on the time of day (because sun and atmosphere and stuff).

You can cross check your results with an interactive WebSDR, like this well-crafted Web-App for a receiver in the Netherlands: http://websdr.ewi.utwente.nl:8901/ … there are others, maybe you find one near your area.

Superheterodyne receiver block diagram 2

(Picture embedded from Wikipedia)

One more thing: In the initial heterodyne block diagram (which I borrowed from Wikipedia) you see an RF filter and and RF amplifier. If you payed close attention you’ll have noticed that I have built no such things.

I noticed that the receiver worked just fine when I attach the antenna directly to the mixer. The necessary signal gain comes solely from the active IF filter (in theory x100) and the active rectifier (x10). If we amplify only the frequency that we already tuned in to, then we don’t need an RF filter anyways. This way we can keep it nice and simple. After all this is just for fun.

Hope you succeed in building your own receiver. Please do send me a video of your receiver in action!

Bonus section

If you read this far, then you already have your basic receiver. But you can make it slightly better 🙂

In the following video you see the shortwave receiver with both “upgrades”, the Arduino controlled local oscillator and an additional RF filter+amplifier. The mini oscilloscope displays the amplified and filtered IF waveform before the demodulator.

Optional Upgrade I: A better local oscillator

The most wacky part of the whole system is the local oscillator. If you want precision tuning and excellent stability you can buy a ready made HF clock generator. There is an inexpensive and very versatile Arduino breakout board out there that’s based on the si5351 chip.

Adafruit Si5351 Clock Generator Breakout

Here is a link to the original module from Adafruit: https://learn.adafruit.com/adafruit-si5351-clock-generator-breakout

It can synthesize you anything between 8 kHz and 160 MHz with quartz perfection. It has an i2c interface and can be programmed with any arduino. There exists a control library that works straight out of the box: https://github.com/adafruit/Adafruit_Si5351_Library – or just search for “si5351” in your Arduino library manager. Just connect one of the clock outputs to the S0 input of the 74HC4051-mixer and you’re good to go.

There exist China copies of this board that you can get for less than 5€ including shipping – and they work with the Adafruit library just as fine.

Optional Upgrade II: RF input amplifier

I just told you that the radio works perfectly fine without the RF filter and amplifier before the mixer. And it does.

But: Maybe you understand more about antennas than me and, say, you have built a very compact antenna that has very cool properties but picks up less RF than my 6 m wire. So you wish you can get a little more RF signal gain from the electronics.

Okay, so here is a simple RF filter and amplifier to have some more flexibility regarding input gain. The first block is a (quick and dirty) second-order high-pass filter that just suppresses everything below roughly 1 MHz. I expect the antenna to pick up first of all 50 Hz power line hum and switching power supply interference in the dozens to hundreds of kilohertz range. Feeding this dirt into the RF amplifier might lead to clipping and distortion of the amplified waveform (destroying information). We don’t want that, so we filter it out before the amplifier.

The second block – The RF amplifier, makes use of a cheap HF transistor – the BFR92 – to boost our RF signal by a factor of 50. I could only find it in an SMD package. But fear not, with a bit of fiddling you can solder it on a regular protoboard. The output of the RF amp can be connected to the antenna input of the RF mixer module.

A logarithmic potentiometer between filter and amplifier lets you adjust the gain of the RF input circuit.

Here is the schematic as PDF.

Visitor Showcase

As I mentioned in the introduction, the whole project started with the FX pedal idea of my friend [Sten]. Meanwhile he also tried and succeeded at building his shortwave FX pedal. Here is a demo of the beauty. Since it’s designed to generate radio NOISE and is fed into a guitar amplifier … expect the sound to be noisy. In this case it is a FEATURE! 😀 Remember the antenna IS the guitar cable (and the guitar).

Video by Sten
Video by Sten

Here is another build from our friend Vito from Italy – This looks like a REAL radio build!

If you have build a 74xx shortwave receiver, too, please do send me a picture or youtube link. I’d be glad to showcase it here.


79 responses to “a 74xx-defined radio

  1. Pingback: Un receptor superheterodino con un giro 74xx - Govannom·

  2. Pingback: A Superheterodyne Receiver With A 74xx Twist | Business, Energy, Science and Technology News·

  3. Pingback: A Superheterodyne Receiver With A 74xx Twist - Latest Hacking NEWS - Lazy Hackers LLP·

  4. Pingback: A Superheterodyne Receiver with a 74xx Twist – Ham Kar Chan·

  5. Pingback: A Superheterodyne Receiver with a 74xx Twist – News Bazzar·

  6. > If we amplify only the frequency that we already tuned in to, then we don’t need an RF filter anyways.

    Two questions: can the local oscillator bleed back to the antenna without a filter? If the input from the antenna is not filtered, can you drive the mixer to saturation with spurious out-of-band signals and transients?

    • First question: can the oscillator bleed back into the antenna? – I didn’t check for this but I find it very unlikely. The voltage signal from the antenna is buffered by an emitter follower. This is a very good approximation of a signal “one-way street”. Second question: In principle the mixer can be saturated by interference. But it does not seem to happen. The mixer can easily process RF input signals up to circa 2V RMS or so, and the mixer itself has a gain of slightly less than 1. Gotta have some damn nasty interference to saturate it.

      • The reason why it might bleed back is because the loading on that emitter follower circuit changes at the time of switching, which propagates backwards through the BJT base current. Although diminished by the reversed gain factor of the transistor, as more or less base current is suddenly needed to maintain the voltage at the output, something must travel back and the local oscillator signal, or its higher harmonics, may come out of the mixer input. The question is, how much?

      • Also, don’t forget the BJT’s miller capacitances (Cce, Cbe), which are basically pass-through for the high harmonics of a square wave. While the emitter follower may block signals from going backwards at the switching frequency, you can still cause nasty stuff to bleed through at higher frequencies.

      • You are right. That is a possibility. However, there is one more barrier to cross. The LO-signal also has to go through the analog multiplexer “the wrong way”. In the end it seems you made me curious enough so i might actually measure this ….

      • Unless there is an active amplifier in the analog multiplexer, the load at the output changes the load at the input.

  7. Pingback: A Superheterodyne Receiver With A 74xx Twist | Sverige Energy·

  8. Pingback: A Superheterodyne Receiver with a 74xx Twist – AnonBoard – AnonBoard·

  9. Pingback: A Superheterodyne Receiver with a 74xx Twist – AnonBoard – AnonBoard – AnonBoard·

  10. Pingback: A Superheterodyne Receiver With A 74xx Twist | Hackaday·

  11. Great post! I am thinking of building something similar to this but on some cheap PCB’s and build them up at home! It’ll be a great start for some RF engineering. Thankyou!

  12. Hi, what components set the frequency range of this receiver? Can we set it up for MW or LW operation? If yes, what component values would change on your schematic? Thank you

    • Hi! The frequency is determined by C13 and R16. The bigger the values (either of them), the lower the frequency of the local oscillator. In principle the circuitry of the radio works in different bands, too. Regarding MW or LW you probably need a different antenna. Not an expert here, but MW and LW have such long wavelengths that you don’t aim for getting an antenna long enough that one wavelength (or an integer fraction of it) fits onto the antenna. The low frequency antennas act more like coils and pick up the magnetic component of the radio waves directly. If you do the research, please let me know what you found out 🙂 All the best, Micha

  13. Hello, can I change the tuning range to medium wave or long wave? Which are the components in your schematic that need to be adjusted, changed in value? I know your project was not aimed at having a tuning coil, but if I aim to insert one, where should I place it.? Many thanks!

  14. I love this project, thank you so much for publishing it online. I am trying to build one these as a learning project. I’ve built the RF mixer and the bandpass filter but I really can’t tell if the bandpass filter is working correctly. When I connect the output of the BP filter to a signal analyzer, I do see the 100kHz signal (signal generator channel 1 at 10 MHz to mixer’s antenna input and channel 2 at 10.1 MHz to the mixer’s LO input). Channel 2 is set to 5 Vpp and square waveform. I can see a weak signal at the BP filter’s output. The signal is pretty weak but it is there. The part that I am unsure about is that I still see very strong 10 MHz and 10.1 MHz signals on the spectrum analyzer. I was expecting those signals to go away or at least be significantly reduced. Does it sound like I have something wrong with my filter?

    • Hi! Thank you for the kind words. Can you test the RF mixer and the bandpass separately? Then it is easier to find out what the problem is. The RF mixer has gain 1 (a bit less). Here you can use let’s say a sine of 1V at the RF input and your 5V square wave at the LO input. In your spectrum analyzer you should see distinct peaks at the sum and the difference of the input frequencies. When you test the bandpass filter, be sure to attenuate the input signal by a LOT, because the bandpass is also a very strong amplifier. As a test input to your bandpass try a sine wave of sth between 1 mV and 5 mV. Probably your signal generator can’t go that low – best use a passive attenuator between generator and the circuit. Hope this helps. All the best – Micha

      • Thank you for the reply. I took your advice and tested each section in isolation. I found that I made several mistakes throughout my various perfboards. I fixed those and got a pretty decent looking filter response using the FFT on my oscilloscope.

        I built the demodulator circuit and was able to hear a weak 1200 hz tone injected via the antenna input. I used another SDR and found a strong signal in the 41m band and tuned the 74xx to that frequency. I could barely hear an audible noise with the volume at max, so something must still be wrong somewhere. I am going to build a second version on a breadboard. If that is successful, I will compare measurements against my original perfboard version to troubleshoot it. Does that sound like a good course of action?

  15. Hi Micha,
    hello from Italy.
    I am building yr project, but I have a problem.
    The resistences of the phase splitter seem with wrong values. May u check ?
    Thx in advance


      • Micha, thx for the reply.
        Well….I’ve tried 2N2222,BC109c,2N3904, but the output on the collector is low amplitude and very dirty, while the emitter output is good.
        I don’t understand…
        Anyway, this evening I’ll try with a transistor dedicated to collector output.
        I am using Manhattan-style build.
        Thx again, I’ll let u know about it!


      • heya, just a wild guess: You did not, by any chance, measure the voltage at the collector of the phase splitter by directly connecting it to a 50R input of your oscilloscope?
        Also: You don’t have to try other transistors. The 3904 is perfectly capable to act as a phase splitter in this frequency region. Maybe check for short circuits, or if the transistor is wired the right way. I sometimes mix up collector and emitter, and still get (poor) signal gain.

      • changed the 470R with 1500R. Now works fine…. Mistery of faith….
        Thanks again for your cooperation Micha!
        I proceed with mixer and filter.
        I’ll get in touch with you soon, if u permit!

        p.s. how post a photo?

      • weird … but okay … should also work with 1.5k, because the next stage is relatively high impedance. You can send me a picture to the email adress in the /contact section 🙂 Cheers, Micha

  16. Pingback: OTA: LimeRFE Shipping Begins, LibreCellular Launches, OpenWifi, and More – MyriadRF·

    • Interesting question. Well … is it relevant at all? All residual RF/LO signal components don’t survive the IF-Filter anyways … if you build the receiver as a whole. I am not quite sure if I understand your question correctly, but I had an idea, maybe this goes in your direction: You could use a 74hc4052 which is a dual-multiplexer. You could use it as in the following picture: https://acidbourbon.files.wordpress.com/2021/10/2021-10-29_14-07.png Then you would in principle get a differential output. The LO feedthrough should appear as common mode signal on both, the pos and neg IF output and thus cancel out at the differential IF receiver. I hope this helps 🙂 All the best, Micha

      • Sorry I guess I should have put the reason for me asking. Your design has become a little obsession of mine. I struggled to build a working version but finally worked through all my issues and got it working. I learned so much by building this and working through my problems. One thing that I had read about was image response. I finally came to understand this when I realized that was the reason I was hearing the signal 100 kHz above and below the RF signal. I did some more reading and thought I could improve upon this by adding a second IF. My goal was to have the first IF at 35MHz and the second at the original 100 kHz. The image response for the first IF would be out of the tuning range I’m trying to cover (0 to 30 MHz). This should eliminate the problem that I was experiencing with hearing the signal in two places. I’m the process of building the 35 MHz IF filter but I couldn’t figure out how to design a MFB filter the way you did for 100kHz. Everything I tried when I modeled it in ltspice seemed to top out at 10MHz. I switched directions and made a 3rd order LC filter instead. When looking at the output from the mixer through the filter, I see that it worked pretty well. However I was seeing significant feedthrough of the LO in the IF output. It was lower power than the 35 MHz IF but still pretty significant. Some other references said that it could contribute to desensitisation and noise. I was going to try and make a better LC filter, possibly 5th order. I ran across a description of a single balanced mixer and saw that you could use two of them to make a double balanced mixer. This is supposed to significantly reduce the RF and LO feedthrough in the IF. It just got me thinking how to adapt your design to make a single and then double balanced mixer.

      • Oh wow! Someone really went “down the rabbit hole”! 😀 😀 😀 Yes, building a narrowband active filter is way harder for 35MHz than for 100 kHz. You’d need very expensive OpAmps for that. LC is the way to go here (I guess). If you want to use a “professional” mixer, you could go for the excellent SA612. It can mix fully differential RF with fully differential LO and output fully differential IF.

  17. Pingback: a simple ringmodulator-ish FX pedal | a c i d b o u r b o n·

  18. Hello!

    My name Mario García from Uruguay:

    About the “RF-AMP”, I´m thinking of using S9018 transistors instead of BFR92… Do you think it is a good idea?


    • I don’t know this transistor, but judging from the datasheet it is quite a bit slower than the BFR92. You might not get the desired amplification at the high frequencies.

      • Hi, thanks for aswering me…

        Yes, It´s true, the S9018 is not a substitute for BFR92.

        BFR92 works about 5,5Ghz and S9018 FQ is 1,1Ghz, and the project works at 30MHZ, max. 50Mhz,

        I´m having troubles at finding BFR92 or BFR96 here in Uruguay.

        As soon as I build the “74xx defined”, I promise you a youtube video.

        I´ll be back! 🙂

      • Wow, that’s great! Love to see your video. How about you first build the radio without the extra HF amplifier? From my experience you can receive the strong stations just fine without it.

  19. Yes, you´re right, that´s what i will do.

    But with only 4 transistors extras, a few resistors and capacitors, I can improve the performance… and almost without disheveled!!!

    But you´re right, let´s do it step by step…

    As soon as I receive the components I buy on the web, I get my hands on the matter.

    I Write you soon!

    Thanks Again!

  20. Pingback: A 74xx-Defined Radio (2021) - newsonetop.com·

  21. Pingback: A 74xx-Defined Radio (2021) - PROJIN NEWS·

  22. Pingback: A 74xx-defined radio | Majorpricedrop.com·

  23. Pingback: A 74xx-Defined Radio (2021) - Tech Learn To Earn·

  24. Pingback: TTL & Microprocessors | Consort3's Blog·

  25. I am going to lay out a PCB for this for fab at Osh park. If I can, I will make it an arduino shield and post the files on Github. Is there any interest in me doing this?

    • Hi! Please make a PCB layout. If you get it working with your PCB and are happy with the results, then I will link to your Github repo from my blog. 🙂 Have fun, all the best, Micha

      • Ok, I will start the work. My plan is straight away to change out the PLL device for the frequency synthesizer from Adafruit. I will also take a look at what can be done with some logic level MOSFETs for the mixer rather than the 74HC4051 device. I will simulate that in NGSPICE to be sure I am getting what I want.

      • Good luck with the MOSFETs. In case you don’t like the 74HC4051 because it is too “big” and has too many unused pins, you can try out this IC: 74LVC1G3157 . It is just a tiny 6 pin SMD device which features exactly the one needed analog switch, switching between two inputs. I guess you can’t get it any simpler than that. Cheers, Micha

      • I am not so worried about how big the IC is, but rather availability. I will have a look at the MOSFETs tonight. I can’t solder the SOT23-6 you mention, so I won’t use it. To that end, do you think 1206 SMT parts are ok for most people?

        BTW, This is a nice piece of work you have done. You should be proud. BTW, I am a HAM… been licensed since 1974.

      • Oh wow 🙂 Thanks for the kind words from a guru. (I was still waiting 15 years to be born in 1974, haha!).
        I consider 0603 parts hand-solderable. 1206 is in a weird in-between place. You don’t really save a lot of space in comparison to through hole parts. And those people who are actually afraid of SMD soldering will not attempt to solder 1206 either. (but that is just my guess). Just do what you deem right. It is your board. If it works and you can assemble it without too much hassle – then it is a good board.

      • That shield really does look pretty neat and DIY friendly! How did you do it, 45 dB is a lot. Did you build a 3904 cascode?

      • I have Cadence OrCAD for schematics and Allegro for PCB design. I use these tools in my work. For the 45 dB of gain, I used a single 2N3904. The trick is to find the best operating point where the DC gain is the highest and operate there. I did this in the simulator. When I first did this, I thought the simulator was lying to me, so I went and built it (it was a BC337). It was true!

        I have simulated your bandpass filter and the peak is around 30 kHz. Is that what you intended? I am also curious as to why you did not choose 455 kHz for the IF? That is more common. My first radio design used that frequency and a ceramic resonator like this:


        I don’t know what the bandwidth of this resonator is, but obviously we want around 5 kHz of bandwidth.


        I gave it some more thought and I think I am sticking with the 4051 mux for now.

      • Ah yes, I know Allegro, we have that at work. Interesting that your simulation yields 30 kHz instead of 100 kHz. I did mine in LTSpice. Maybe you have a better/different SPICE model for the NE5532. Why did I not choose the more common 455 kHz? – Well I knew that you could buy resonators for that IF (i actually have some in my cupboard) but my goal was to build a receiver without classical radio parts. An IF of 100 kHz seemed okay for my purpose – and this being my first and so far only radio project, and having no previous contact with the HAM world, I left it like that.

      • There is no reason at all why the IF has to be at 455 kHz, so 100 kHz is just fine. I have a radio receiver which has an IF of 456 kHz and tunes 100k kHz to 30 MHz, but there is big hole in coverage around 500 kHz as such. It cannot receive there. They made a different and very similar model with an IF of 175 kHz which allowed reception at 500 kHz which is a maritime distress frequency IIRC.

        I did not use the 5532. Back in the day, it was a good low noise op amp for audio. It is really too good for this application because by the time the signal gets there, it has gotten past the point in the circuit where noise is important the most, and that is the first stage which is the phase splitter. I will likely look at a higher gain circuit there and still do the phase split. So, I used a lowly LM358.

        See new stuff posted in my folder. Maybe I made an error in my simulation.

        Another thing cool about the Adafruit oscillator is that it has 3 outputs. One could be used as a beat frequency oscillator which would be very stable and allow demodulation of single sideband.

      • Hi, up on the google drive where the other pictures are is now a quick and ugly area study for the design. It does not show the Arduino connectors nor the Adafruit module, but they will fit. I will be able to put components under the Adafruit oscillator module.

  26. Pingback: A 74xx-Defined Radio (2021) – Cyber Geeks Global·

  27. Hi, I spotted your micro switch from 3-10 MHz range to the 10-30 MHz in the local oscillator section pad, in your video, and this isn’t shown in your full 74xx receiver schematic. Can you give me the actual values of the C13/R16 pair in the local oscillator, as it seems that the 47 pF/10k
    doesn’t cover the whole KW band ? Thank you so much!

    • sorry, I don’t remember, but it might be that the micro switch (SPDT) selects between 10 pF and 47 pF. The switch only affects the capacitor. Hope this helps. Feel free to try it out and verify with an oscilloscope 🙂

      • I have not published my schematic, but will do so, soon. I don’t know what you mean by “micro switch”. I just dropped out the PLL chip for the LO and changed to the Adafruit 2045 as was suggested in the blog. I intend to be able to at least 1 MHz to 30 MHz, if not lower in frequency.

      • I looked at your schematic again today. I see you have found your missing feedback path in the IF filter. Still I don’t believe you’re going to have a good time with that 3904 in the phase splitter being upside down 🙂 all the best, micha

      • The 2N3904 works fine “upside down”. I did that to improve the routing and simulated it. The transistor operates at unity gain in the reverse active region.

        I still cannot get the IF filter to work right.

      • Hah! What a trick. I now remember building audio preamplifiers a long time ago. I accidentally mixed up C and E. Still worked (with reduced gain). I checked the datasheets. Your LF358 has only a tenth of the unity gain bandwidth of the NE5532. Might that limit your performance? I don’t see how the 5532 is “too good” for this circuit – it does not even cost 1€ 🙂

      • All of the design equations yield an amplifier in the wrong center space. I tried a TL062 and it was no better. I am trying a Sallen Key right now.

      • Actually, with some heuristic tweaking, I can get the center of the band to be at 100 kHz, but the Q is low. Probably good enough. This is with either the LM358 or the TL062.

      • if Q ist your problem: Why is your R9 so much bigger than mine? I chose R9 to be 47R exactly for the purpose of making the filter narrow. Furthermore: Are C5 and C6 really 100n (not pico)?

      • I tried your circuit and although the Q is higher, the center frequency is off. I also could not get the model for the 5532 to work.

        I did a design using the standard design equations and still had to tweak a value to get the center on 100 kHz. I can tinker with that resistor to tighten up the Q.

        I would like higher bandwidth ~3 kHz to do better by AM, rather than CW.

        Please see updated schematics and my filter respose.

  28. Hello.
    I’ve build your design but I’m not able to tune into anything. I’m assuming my antenna is the issue here.
    What exactly are you using as the antenna? what is that small red thing that the wire is attached to?

    • I used a circa 10 m long thin copper wire, that I spanned across my room. Nothings special. I don’t know what small red thing you mean … maybe it was a connector or a short piece of thicker wire that I soldered my antenna to. Another question: What ICs did you use? Depending on the logic series they have different speeds. For example a CD4046 is way slower than a 74HCT4046. Cheers, Micha

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.