AM modulator with a 4.9152MHz TTL osc

[RoGeorge]

Tried once to make an "AM modulator", one of those low power MW TX for indoors.  The result was awful:  unstable frequency, distorted sound, too little power even when the receiver was sitting on the same bench, though still a great tinkering fun. 

Conclusion was that next time to use a stable and well isolated oscillator.

Well, found recently a 4 pins IC from the TTL era, a 4.9152MHz square wave quartz oscillator (fabrication date 1983, 40 years ago - and if that frequency seems arbitrary, it was to derive all the standard baud rates for serial communications over RS232 using nothing but TTL binary counters ICs).  Anyway, clipped the alligators from a 5V supply directly to the pins of the osc IC, then probed its output with the oscilloscope.  It was working just fine. 

Now, 4.9152MHz is not a broadcast frequency for AM radio.  In EU the MW broadcast band is allocated between 531-1602kHz, and has 9kHz spaced channels (or 526.5-1606.5kHz with 10kHz space between channels in USA).  However, when dividing 4.9152MHz with an integer (easy to do with TTL counters), the divided F might coincide with one of the MW broadcast channels.

Code: [Select]

Fosc   Divider   Result    Close to   F error    Error
(MHz)    (N)      (kHz)       (kHz)     (kHz)    (% of the 9kHz channels spacing)
4.9152    3    1638.400    1638.000     0.400     4.4
          4    1228.800    1233.000    -4.200   -46.7
          5     983.040     981.000     2.040    22.7
          6     819.200     819.000     0.200     2.2
          7     702.171     702.000     0.171     1.9
          8     614.400     612.000     2.400    26.7
          9     546.133     549.000    -2.867   -31.9

Fosc divided by 3, or 6, or 7 will be very close to an ideal MW channel.  I care about matching standard broadcasting frequency because more recent AM receivers no longer have continuous tuning, they can only tune in jumps of 9kHz.

Most tempting was to divide by 6 to get 819.2kHz as the AM carrier.  There are ready made TTL counters with 6 (SN7492).  Though, a counter with 6 will produce an asymmetric waveform.  To get a symmetric 819kHz square wave, could have used a frequency doubler, then divide with 6 first, then again with 2 (using the same 7492).  That would mean a few more TTL gates to make a frequency doubler, so yet another IC.

Then I've found this unusual dividers idea from a PDF by Charles Wenzel, where a tuned LC is inserted in a typical D-type FF divider by 2, so to make it divide with ratios other than 2. 



I've used same idea to make a divider by 3, together with 100pF, 100uH and a 54LS74 (in fact an M555TM2 i happen to already have), then used the remaining FF as a divider by 2.  Tested the divider on a breadboard, and it was very stable.  Tried it with LS, H, N and HCT types of 7474, and at Vcc between 2.5-5.3Vcc.

For the next stages, I've seen some schematics are using a single transistor as a class C common emitter RF amplifier, where the AM modulation is obtained by varying the collector voltage, while others are using a differential pair with a transformer output as RF amplifier, while the AM modulation is obtained by varying the tail current of the differential pair.  Not sure which one to choose, and if 5V will be enough.


TL;DR got a quartz-stable carrier at 819.2kHz, square wave and with 50% duty factor, as Q and NQ from a TTL flip-flop.

What kind of modulator+RF amplifier schematic will fit well with this TTL carrier?
Is 5V enough for the RF amplifier (most I've seen are using 12V or more for the RF amplifier)?

[duak]

A 1970s monolithic solution is to use the common MC1496 balanced modulator/demodulator.  Here's a design for an AM modulator with a single +5V rail:  https://www.ee-diary.com/2022/09/single-supply-mc1496-am-modulator.html

Some caveats:
- I haven't tried the circuit so I don't know if the values are reasonable and within spec.  Consider it food for thought.
- you'll want to filter the output of the oscillator as the purity of the carrier signal determines what frequencies are sent down the transmitting chain.
- the output of the MC1496 is balanced differential so you are getting twice the voltage swing of a single ended modulator.  A balun transformer could probably drive some sort of antenna for local low-power use.

Cheers,

[RoGeorge]

That should work, just that I don't have any MC1496.  From the internal schematic of the MC1496 (datasheet, Onsemi AN531) looks like a typical Gilbert cell.  As a last resort, that could be done with discrete transistors.

Another thing not yet decided is the antenna.  Has to be something reasonably small to keep indoors.  The receiver will be in the same room, so I imagine a loop antenna will not be good, because it has to be re-oriented if the receiver is moved in another corner of the room.  I guess an omnidirectional antenna would be preferable, but I know nothing about antennas.  All I know is that a simple wire randomly hanged on the wall was not good enough.

[iMo]

I saw AM TXes with an audio power amplifier IC, the output of the amplifier modulating the drain voltage of an IRF5xx RF output stage (the PA powered off the audio amplifier's IC output).
The TTL output level should be ok for modulating the IRF5xx. You may get 4-10W at 3-7MHz out of it easily, imho (provided you are the licensed ham)..

[Wallace Gasiewicz]

Take a look at some AM CB radio schematics.  Many of them use a big "modulator" to supply power to the collector of the output transistor to achieve good AM modulation. This can also be done thru a transformer.  This audio signal is applied to the final collector thru a choke that keeps the RF out of the modulator . This is similar to the old "plate modulation" of AM tube transmitters.
The modulator is just two big PNP transistors, they are just audio freq BJTs . One driving the other bigger one to supply collector current to the RF final.Audio signal is applied to the modulator by a commonly available audio amp.The RF is supplied to the final thru the base of the final transistor.Either BJT or more commonly these days a 520 type mosfet is used as a final (usually much cheaper).
Remember almost all of the transmit power goes thru the "modulator" so it needs to be a big transistor that is heat sinked well..   Depends on how much output you want, obviously.
The output of the final transistor must be matched to your antenna, usually 50 ohms, The output circuit can be quite complicated, but since you are using only one frequency a tuned Pi network would suffice.
Most of these things are 12 V.
Edit:  If the audio amp is big enough,it can supply all the collector power needed for the final without using a "modulator". This is done on lower power radios.

[A.Z.]

what about using a modulator like the one shown here

http://www.qrp.gr/minimaltx/

[retiredfeline]

Well, found recently a 4 pins IC from the TTL era, a 4.9152MHz square wave quartz oscillator (fabrication date 1983, 40 years ago - and if that frequency seems arbitrary, it was to derive all the standard baud rates for serial communications over RS232 using nothing but TTL binary counters ICs).  Anyway, clipped the alligators from a 5V supply directly to the pins of the osc IC, then probed its output with the oscilloscope.  It was working just fine. 

Active crystal oscillators are still being made and available. Usually square or rectangular metal can.

[joeqsmith]

Tried once to make an "AM modulator", one of those low power MW TX for indoors.  The result was awful:  unstable frequency, distorted sound, too little power even when the receiver was sitting on the same bench, though still a great tinkering fun. 
...

Just curious why AM rather than FM?   If you don't care about stereo, simple enough to make a reasonable transmitter.   This is one I had made several years ago for reasons I no longer remember.  Looks like I used my favorite jellybean NPN, 2N4401.

[RoGeorge]

AM because the modulator was meant to play voice only (audiobooks), so no need for stereo or FM sound quality.  Additionally, the 88-108MHz FM is almost fully occupied, while the MW band is almost empty now.  Some countries even discontinued the AM broadcasting entirely.

Truth is I don't really need an AM modulator, it's rather a pretext to experiment with analog radio. 


I've just found in the box of scrapped mains sources a wall-adapter with a transformer in good shape.  Small in size, measured turns ratio is 10.5:1 (the wall adapter was from a 120Vac to 9Vdc/200mA), and the DC resistance is 570 and 4.78 ohms.  Want to use it for injecting the audio signal into the collector of a final RF transistor (common emitter, class C, like in the draft schematic).

Since the intended RF power is very small, it might work to connect the secondary of the transformer directly, as if the modulation transformer would be a headset, without an additional audio amplifier, but I didn't try that in practice yet.

[RoGeorge]

Depending on the collector load and surrounding values, in simulation V(y) can go as high as 100-200V, much higher than the transistor's V_CEO from the datasheet.

Can this voltage fry a "normal" 45-60V max V_CEO transistor?
Which one from these medium power transistors would be more suitable?
   - SF128 (metal can TO-39, NPN, Ic=500mA, Pd=600mW, Ft>=60MHz, Vceo=60V, HF or Video)
   - BD139 (TO-126, NPN, 80V, 1.5A, 1.25W/12.5W, B=50/100)
   - BF459 (TO-126, NPN, Vce0=300V, Ic=100mA, Ptot=6W, fT=90MHz, high voltage, video amp)