HiFi AM Transmitter using a TBA120 FM demodulator

[Semir_DE]

Hello everyone,

with ever more AM stations closing down it has become necessary to generate signals for older AM only radios. While thinking about building such a device I remembered that I still had some TBA120S ICs in a drawer that I used to use for FM demodulation in various projects. There are other ICs like the µPC1391 that can also be used in the same manner as I will describe below, but I think the TBA120 is probably the most common such chip. Please refer to the attached images Fig.1 through Fig.7.

The TBA120 contains two main parts as outlined in Fig.1:

- A limiting IF amplifier with high gain.
- A multiplier used as a frequency discriminator

In FM applications the IF signal is fed to the IF amplifier at pin 14 where it is amplified and amplitude limited. Then the signal is passed to the multiplier stage. The TBA120 demodulates the FM signal by multiplying two instances of the same FM IF signal with each other. The first one is the direct output of the FM amplifier and the other is a 90° phase-shifted version coming from an LC resonant circuit connected to pins 7 and 9. Since the phase shift of a resonant circuit changes when it is off its centre frequency the signal at pins 7 and 9 will change its phase with respect to the main IF signal depending on the deviation of the FM carrier signal. The multiplier acting as a frequency discriminator will turn this change in phase into a change of its output voltage. This is a brief description of how this circuit usually works when used for its original purpose. When used as an AM modulator the two parts above will be assigned different functions, however. The TBA120 was never intended to be used like that but it works very nicely nevertheless. Please check out Fig.2 in the attachment for the circuit diagram.

In my project I use the multiplier as the actual AM modulator. The audio signal is fed to Pin 7. Pin 9 will be AC grounded by means of a 100nF capacitor and an offset voltage adjustable via RV1 is applied to that pin for setting the 50% carrier level. The modulated signal is available at pin #8. This signal will have a waveform that is not a perfect sine wave, so a low pass filter consisting of C13, L1 and C14+C16 has been added to suppress harmonics. C14 should be tuned for maximum signal amplitude. For the upper MW band an L1 value of 150-220µH should be used while for the lower end of the band a value of 330-470µH is needed.

The IF amplifier is re-purposed in this design as well and is used as an oscillator generating the AM carrier signal. In order to achieve this a crystal or ceramic resonator is placed in a feedback loop between pins 6 (IF output) and 14 (IF amp input). A simple 1MHz crystal can be used but is not the best option as 1000kHz does not fall into the 9kHz frequency spacing in Europe. This could lead to an annoying 1kHz beat tone in the evening hours when broadcasts on 999kHz are being picked up by the same radio. I prefer to use ceramic resonators with frequencies in the AM broadcast band. These resonators can be tuned by around +/- 3kHz off their nominal frequency so with a little bit of luck a resonator with a frequency that fits exactly into the 9kHz scheme can be found.

This store has a wide selection of ceramic resonators:

http://www.oppermann-electronic.de/h...r__filter.html

Please scroll down on the page to the green table and you will find resonators ranging from 190kHz (LW) to 1.58MHz (upper MW). Unfortunately this site is only available in German.

As a second option the circuit can also be used with a traditional LC tuned circuit but this is not as stable and the frequency will exhibit slight drifts over time. The third option is to use an external carrier signal which can be fed into the circuit via connector P2. I used this option in my AM Stereo transmitter.

After passing through the low pass filter the AM signal is fed to an LT1227 or AD844 amplifier IC. These chips are wide band video amplifiers capable of driving a 75Ω or 50Ω load thus the output signal can be connected to a 75Ω or 50Ω coaxial cable for wired distribution to a large number of radios or to feed a simple antenna for short range over-the-air distribution. The output level of the LT1227 is around 10-15mW and can be adjusted with RV2. The output power is more than enough for even a large collection of radios. In Fig.3 and 4 you can see the output signal of the transmitter. The trapezoid shows the almost perfect linearity of the output signal even at 100% modulation which is important for a low distortion level of the modulating audio signal.

I have made a PCB design for this project with the free KiCAD software. In Figs. 5 and 6 you can see the boards and an assembled version. Please PM me for the project data.

The possibility to connect a standard coaxial cable to the transmitters output allows to physically separate the transmitter from the antenna, so it can be situated near a stereo system while the antenna can be in the attic or another room.

In Fig.7 you can find a simple but very effective circuit to tune a wire with a length of 3-5m to any AM band frequency. Please note that the value of the inductor depends on the chosen frequency and the length of the wire used as an Antenna. For Long Wave this should be around 1-2mH for the lower MW band the value is around 470-330µH and in the upper MW band 220µH should be sufficient. The capacitor C1 should be as small as possible in order to maximize the output of the antenna. With the 10mW of the transmitter this circuit gives excellent results inside a house but the signal will fast disappear when leaving the premises.


Semir

[JimMac53]

Semir,
What a great project! I need to read up on the TBA120, but on paper and from the tests you've conducted, it'll keep those radios singing for a long time. What are the AM power restrictions in Germany?
Regards,

[Andrew2]

Very interesting. I like to see novel uses of mundane chips and this is a useful one. It's ringing a distant bell, though. I think it may have come up in Pat Hawker's TT column many years ago, but I'm not sure. I've used the TBA120 as a synchronous AM demodulator several times (I think I've got one in my latest homebrew Topband transceiver) and it works well in that role.
I'll go through my old datasheets and see what pops up.

EDIT: Ah, TT Nov 85. Using the 120 as an DSB generator. Adjusting R4 causes unbalance and out pops AM.

[Ian - G4JQT]

Splendid work Semir! That modulation looks excellent. What is the audio bandwidth and what determines it in your design? Maybe it's C12, but maybe that's just to keep any stray RF out or the audio source.

Of course modulation beyond about 6 kHz is often removed by the receiver IF bandwidth, but some sets have (or can be adjusted to have) very wide bandwidths and sound great.

I think AM broadcasting in the UK has only about five years left - no inside knowledge, just reading the signs - and interest in our hobby will be just for the die-hard enthusiasts, as 405-line TV is now. So, we need stable, elegant, relatively cheap-to-construct circuits like this to keep our radios 'alive'.

Incidentally, how frequency stable are ceramic resonators, especially when they're being pulled? Once MW is completely clear in Europe (although I think Spain will hang on to it longer than most), of course any quartz crystal operating in the band will be OK as there won't be any co or adjacent channel interference, except perhaps at night.

[Maarten]

Nice work! Does it matter which TBA120 version is used? I see to recall there were at least a handful different versions.

[orbanp1]

Originally Posted by Andrew2

...
I've used the TBA120 as a synchronous AM demodulator several times (I think I've got one in my latest homebrew Topband transceiver) and it works well in that role
...

This also very interesting!

Would you care to share the schematics of the synchronous demodulator with the TBA120?

Thanks, Peter

[Andrew2]

I'll see if I can dig it out, Peter.

[Andrew2]

Here it is, Peter.

I think I used it more or less as shown. Basically the IF is applied to the limiting amplifier which scrubs all the modulation off and leaves a plain carrier to switch the multiplier section. The unlimited IF also goes into the multiplier and is demodulated to audio. You can do a similar trick with the CA3189. Apologies to Semir for diverting his thread.

[Semir_DE]

Originally Posted by JimMac53

What are the AM power restrictions in Germany?

Hi Jim,

glad you like the project. Unfortunately in Germany we are technically not allowed to broadcast in the AM band at all. Only museums and other similar organizations can get a permit for a certain frequency at an annual fee. In real life with the use of energy saving lights, DSL modems and power line networking devices there is so much "hash" on the AM band that a low power TX will not be distinguishable at a distance of a few 100m because all the noise will be far more powerful than the few mW such a device will put on the airwaves.

In my house I run three 10mW AM Stereo transmitters 24/7 to feed my collection of AM Radios and tuners. The content is from internet radio stations and DAB+ radio broadcasts.

I wish the EU would come up with a similar regulation like FCC Part 15 in the US. This would make life a lot easier for us vintage radio collectors.

----------------
Semir

[Semir_DE]

Originally Posted by Ian - G4JQT

what determines it in your design? Maybe it's C12, but maybe that's just to keep any stray RF out or the audio source.

Of course modulation beyond about 6 kHz is often removed by the receiver IF bandwidth, but some sets have (or can be adjusted to have) very wide bandwidths and sound great.

Incidentally, how frequency stable are ceramic resonators, .

Hi Ian,

You are correct in assuming that C12 is the main limiting factor for audio bandwidth Actually it is the time constant formed by the combination of R3 and C12. You are also correct in your assumption that C12 is there to eliminate RF from coming out of the TBA120. Most TBA120 chips except the original TBA120 without any suffix letters have internal capacitors that feed the IF signal into pins 7 and 9. While this is desirable in the original FM demodulator application it poses a problem in the AM modulator solution since it would cause harmonic distortion in the RF signal at twice the carrier frequency. So the main purpose of C12 is to short circuit any carrier RF to ground thus eliminating this issue. Since the input sensitivity of pin7 is quite high an audio signal at line level has to be reduced in magnitude by a factor of about 10:1 this is done by the resistive divider R1/2 and R3. Since R3 is of quite low value at 100Ω the time constant of R3 and C12 has a 3 dB cut off point of about 160kHz which is the theoretical available audio bandwidth. So the usual 20Hz-20kHz should pass through the circuit without any significant attenuation.

At the end of the day the limiting factor for frequency response will most likely be the IF section in the radio itself as you have correctly pointed out. Normal AM valve radios were designed to support an audio bandwidth of 4.5 kHz, this means their IF bandwidth would be about 9-10 kHz. Some radios have, however, a facility to switch IF bandwidth to a wider value, these would be predestined to work with a full bandwidth signal like the one from my transmitter. Please bear in mind, however, that regular broadcasts are limited to an RF bandwidth of 9kHz i.e. 4.5kHz audio bandwidth. A transmitter that supports 10kHz or more will have an RF spectrum that will cover more than one European MW 9kHz RF channel slot i.e. interfere with up to 3 regular AM stations. While this an academic discussion these days with diminishing number of MW broadcasts, it shows why radios were designed to have limited RF/IF bandwidth. An exception being the US AM Stereo receivers that have much larger AM IF bandwidths and are often switchable between "Narrow" and "Wide" these receivers will have an audio bandwidth in the wide position of around 10kHz and sound much like regular FM stereo receivers but in AM!

On the subject of ceramic resonators, well they are quite stable. Not as good as crystals obviously but much better than pure LC circuits. I have been running an AM Stereo transmitter at 1260kHz here for about two years now and have not noticed it to drift significantly over this time period. The hardware is situated in my basement and temperature is relatively constant there at about 20°C. The main problem is finding a suitable resonator with the "right" frequency.

Just in case you're interested I also have designed a frequency generator that can be programmed to any frequency ranging from 100kHz to 2MHz. This uses a DDS chip AD9833 and an Arduino to control it. This circuit features the same stability as a crystal oscillator but can be set up to deliver any desired frequency such as LW, AM IF, MW etc. I will post this project here in the coming weeks including the software code for the Arduino and the PCB data.

[Semir_DE]

Originally Posted by Maarten

Nice work! Does it matter which TBA120 version is used? I see to recall there were at least a handful different versions.

Hi Maarten,

greetings to Haarlem, I have been there many times. Very nice place and I especially like the Wilma an Albert steak house...

You are correct there are several versions of the TBA120. My circuit works with the TBA120, 120A, 120AS and 120S. I have not tested TBA 120U. The TBA120T does not work because it has a 500Ω resistor between pins 7 and 9.


..Semir

[orbanp1]

Originally Posted by Andrew2

Here it is, Peter.
...

Hi Andy,

Thanks for the schematics!

Way back I tried an NE561 PLL chip as an synchronous detector, as per the suggested application, but was not too impressed.
I might take another go at it (now I have more test instruments), though there are barely any SW station these days you would want to listen!

Thanks, Peter

[Andrew2]

It's a fascinating subject. Back in the early 90's I built a board which took the IF output of a transistor radio and applied it to three different detectors - a simple diode type, a TBA120 type as in the diagram, and a PLL locked type.

Using the diode detector as a base reference, on strong sigs there was nothing to choose between them, but on weak signals the diode detector suffered from that familiar 'noise modulating the audio' graininess. On the same signal, the TBA120 one (which is basically a product detector driven by a hard-limited version of the off-air IF) was better, and the PLL one (prod detector with a re-generated IF locked to the incoming one) was the smoothest. I think I used a 567 chip for the phase detector/osc.
It's still a subject that interests me and I sometimes go on internet 'trawls' to find novel ideas.

[David G4EBT]

Originally Posted by Semir_DE

Hi Jim,

Unfortunately in Germany we are technically not allowed to broadcast in the AM band at all.

I wish the EU would come up with a similar regulation like FCC Part 15 in the US. This would make life a lot easier for us vintage radio collectors.

----------------
Semir

It doesn't matter how low the power level, or whether the equipment is commercially built or homebrew, there is no 'micro-transmitter' equipment that is legal anywhere in the EU, or Switzerland. As Semir says, it's rather a shame that something along the lines of FCC Part 15 doesn't apply, but it doesn't and it isn't going to. The same harmonised restrictions apply throughout the EU, EFTA and Switzerland. No one country's regulatory body, (OFCOM in the case of the UK) is able to deviate from that. Originally it was called the 1999 Radio and Telecommunications Equipment Directive, updated in 2014, which covers:

Austria, Belgium, Denmark, Finland, France, Germany, Greece, Iceland, Ireland, Italy, Liechtenstein, Luxembourg, Netherlands, Norway, Portugal, Spain, Sweden, United Kingdom plus, since 1st May 2004, Czech Republic, Estonia, Cyprus, Latvia, Lithuania, Hungary, Malta, Poland, Slovenia and Slovakia plus, since January 2007, Romania and Bulgaria. There is also an agreement with Switzerland which gives it access to the European Community.

I'm not suggesting anyone reads or tries to understand it, but here are the links:

https://www.tuv-sud.co.uk/uploads/im...20RTTE%202.pdf

DIRECTIVE 2014/53/EU OF THE EUROPEAN PARLIAMENT AND OF THE COUNCIL of 16 April 2014 on the harmonisation of the laws of the Member States relating to the making available on the market of radio equipment and repealing Directive 1999/5/EC

https://eur-lex.europa.eu/legal-cont...4L0053&from=EN

But to get things into perspective, using a well designed and constrructed 'micro transmitter' which has a radiation area of little more than a few metres is quite different from the anarchic anti-social wholesale use of illegal CB radios back in the late '70s/early '80s, which forced lawful model control enthusiasts off 27MHz.

We know that MW broadcasting is in terminal decline, and that's why many restorers, myself included, are keen to use devices which enable them to keep their radios in use. We also know that the worst that could happen is that we might radiate a signal into a neighbouring property, but what's the likelihood of someone having a radio in use on MW on the otherwise empty frequency that we're using?

So being pragmatic, when we engage in such activities, it's with the best of intentions, it isn't anti-social and it doesn't interfere with any essential services. The forum rules recognise this, with Rule F (2) stating: "Small low powered transmitters with an output of less than 100mW e.i.r.p. for home use may be discussed in this forum. The moderators will use their discretion as to what is acceptable".

It's evident from several designs that have appeared on the forum - including Semir's - that a lot of thought and expertise goes into their R&D, and all credit due to them. We're just not on the enforcement authorities radar and nor should we be, given that enforcement bodies must ensure that any enforcement action is proportionate and is targeted only at cases where action is needed.

If the Mini-Mod rumbles on much longer, it will have had almost as long a run as 'The Mouse Trap' - let's hope that Semir's design does likewise. Hopefully it won't be too long before several forum members will have successfully built their own and will update the thread to tell us.

[Mike. Watterson]

Originally Posted by Andrew2

Very interesting. I like to see novel uses of mundane chips and this is a useful one. It's ringing a distant bell, though. I think it may have come up in Pat Hawker's TT column many years ago, but I'm not sure. I've used the TBA120 as a synchronous AM demodulator several times (I think I've got one in my latest homebrew Topband transceiver) and it works well in that role.
I'll go through my old datasheets and see what pops up.

EDIT: Ah, TT Nov 85. Using the 120 as an DSB generator. Adjusting R4 causes unbalance and out pops AM.

An SA612 is cheap, still made, in SMD & DIL. It has onboard voltage regulator for the built in oscillator.
Pair of resistors unbalances the mixer input to give super linear AM.
drive for a push pull output stage allows cancelling of second harmonic.
Very stable & cheap.

However ALL these sorts of circuit are either fixed frequency using a resonator or crystal, or need double tuning. The output stage needs tuned or else there are harmonics up to Short Wave, not friendly.

I've built a few of these. Except people want reliable tuning, so I'm looking at PLL synth VCO on the SA612 oscillator using the cheap easy Motorola synth that was in CBs and Ham Radio. Unlike modern ones that are ALL I2C, it can use an 8 way DIP switch to set frequency, or you can use a two pole 6 way switch with preset caps to tune output stage power amp and diodes to set the PLL synth channel. Obviously channel spacing depends on crystal for the PLL reference.

An FM demodulator IC does work, but the SA612 is designed to suit AM, Audio or FM signals at the balanced input. A resistor to 0V on one input and to an On indicator LED on the other input (so 'regulated') gives the requisite carrier level without injecting hum or noise.

The MW aerial is so inefficient that about 500mW drive is a fraction of a milliwatt RF. So a 12V PSU needed for output stage, a pair of NPN transistors in class B. A zener regulator for the SA612 or the less good FM demodulator IC.

[Mike. Watterson]

Note you probably want a multipole AF filter in front of ANY AM modulator as otherwise the higher frequencies of the audio create problems in the AM receiver.
NO AM radio is HiFi. You must decently limit the audio bandwidth, a single RC stage isn't enough!

There WILL be harmonics and the short aerial will radiate higher ones better! A tuned loop aerial about 40cm diameter hardly radiates harmonics and needs much less power.

[Andrew2]

Mike - you're a man after my own heart. I always pick up a few 612's at radio rallies as they are such a useful workhorse. My topband transceiver uses one for generating both SSB and AM modes. It is used in balanced mode for ssb (well DSB later filtered to ssb) and with a small bias added to get AM. It has another one in the product detector.
With a bit of fiddling it's possible to employ a 612 as a voltage-controlled mic compressor. I must get some more...