(Newbie AM Project) Square v. Sine: Which is best for transmitting?

[Gip-Gip]

I've been brought the task of making an AM transmitter that'll essentially be used to test/demo radios inside my house (due to the lack of AM stations where I live). This is my first radio-transmitter project and building it semi-custom has brought up a few questions;

• Is a sine wave better at transmitting than a square wave?
• If one's more-or-less superior, how superior is it?

and most importantly

• Where in the circuit do you hook the antenna up to?

(I apologize if the last question has been answered elsewhere on the forum, I though I might as well ask it here bc it's convenient)

It wouldn't be that hard to add a "sine stage", but if the result is barely (if at all) better there's no reason to add it.

BTW, by sine vs square, I mean a sine carrier "generator" vs a square carrier "generator"(a 555 for example)

The rather newbie schematic is in the attachments, feel free to point out anything that seems odd

EDIT: am.pdf is the current schematic

[grumpydoc]

I think that, at present you are out of your depth.

Carriers for radio transmitters are always sine waves.

Why do you wish to demo AM radios anyway?

[Rerouter]

Sine is going to be better because it will be a clear, ideally single frequency tone,

Square contains many frequencies, which may be beneficial if you where viewing the demodulated output on a spectrum analyser, but for a go / no go test, sine will work better.

[Gip-Gip]

I think that, at present you are out of your depth.

Carriers for radio transmitters are always sine waves.

Why do you wish to demo AM radios anyway?

1. I was thinking like a 555 powered crude radio when I thought of a square wave. I know fundamentally radio waves are sine, but I was looking for the difference between a square and sine  "generator", per-se

2. I have a bunch of old cheap AM/Multiband radios I wish to restore, and once I'm done restoring them I don't want them to rot in my closet.

Sine is going to be better because it will be a clear, ideally single frequency tone,

Square contains many frequencies, which may be beneficial if you where viewing the demodulated output on a spectrum analyser, but for a go / no go test, sine will work better.

That's the answer I needed! Thank you sir!

[grumpydoc]

I think that, at present you are out of your depth.

Carriers for radio transmitters are always sine waves.

Why do you wish to demo AM radios anyway?

1. I was thinking like a 555 powered crude radio when I thought of a square wave. I know fundamentally radio waves are sine, but I was looking for the difference between a square and sine  "generator", per-se

2. I have a bunch of old cheap AM/Multiband radios I wish to restore, and once I'm done restoring them I don't want them to rot in my closet.

Sine is going to be better because it will be a clear, ideally single frequency tone,

Square contains many frequencies, which may be beneficial if you where viewing the demodulated output on a spectrum analyser, but for a go / no go test, sine will work better.

That's the answer I needed! Thank you sir!

I think that's the answer that you wanted but not the one you needed.

Get yourself a cheap signal generator which covers the AM band - 2MHz will do it which is pretty low for that sort of thing.

Square waves have a very high harmonic content - you will not only be transmitting in the AM band but will splat all over other bands as well, not to mention that any such transmitter will be illegal.

Thankfully with a 555 the edges won't be all that sharp so you shouldn't splat too far up the spectrum - maybe into the low 100's of MHz.

E.g
http://www.ebay.com/itm/FY3224S-24MHz-USB-Dual-channel-Arbitrary-Waveform-Function-DDS-Signal-Generator-/322359328969

If you hook it to any antenna it will still technically be illegal but any old bit of wire on the output should be OK if you have the AM set nearby.

[Gip-Gip]

I think that's the answer that you wanted but not the one you needed.

Get yourself a cheap signal generator which covers the AM band - 2MHz will do it which is pretty low for that sort of thing.

Square waves have a very high harmonic content - you will not only be transmitting in the AM band but will splat all over other bands as well, not to mention that any such transmitter will be illegal.

Thankfully with a 555 the edges won't be all that sharp so you shouldn't splat too far up the spectrum - maybe into the low 100's of MHz.

E.g
http://www.ebay.com/itm/FY3224S-24MHz-USB-Dual-channel-Arbitrary-Waveform-Function-DDS-Signal-Generator-/322359328969

If you hook it to any antenna it will still technically be illegal but any old bit of wire on the output should be OK if you have the AM set nearby.

If I put a LC filter on the output of the crystal oscillator I plan to use, would that work?

[kg4arn]

I don't think your schematic will generate AM primarily.  Looks to me like the audio input will just shift the DC level of the carrier (the final is a difference amplifier), not change its amplitude (until it limits).  I suppose AM may be generated as a by product of the non-linearity as the final op amp limits.

[bd139]

Some visualisations for anyone coming across this. Without delving too deep into mathematics, a square wave can be represented as a number of sine waves added together:



The same is true in reverse obviously.  Ergo if your square wave is hitting an antenna, it is actually broadcasting at the fundamental frequency (the square wave frequency) and a number of harmonics above that. For example if you are sending a signal at 1MHz with a square wave, there are signals being sent as well at 3x, 5x and 7x that frequency:



Now there's really nothing wrong with the oscillator or transmitter using square waves. In fact some of them start with that as they are very efficient when it comes to amplification, but you need to remove the harmonics from the output for a number of reasons. Mainly because it will annoy anything running at the harmonic frequencies (interference) but also because not all of them make it to the antenna and will be reflected back and blow up your amplifier.

This is done with a multiple stage low pass LC filter usually.

Even if your amplifier generates a pretty good sine wave, it's good practice to add an LC filter to get rid of any distortion as the harmonics are still transmitted potentially.

If you want to build a simple AM transmitter, the simplest approach is:

1. Use a 1MHz crystal can oscillator to generate a square wave. A 555 is being pushed pretty hard in the AM band. They're really only designed for 100KHz max. They're also terribly unstable.
2. Apply that to a common emitter amplifier.
3. Use an op-amp to control the amplifier's supply current driven from audio (this is crude but effective). 
4. Stick a low pass filter on the output of that.
5. Attach a very long bit of wire as an antenna.

There is a nice example here of the sort of thing working: http://ludens.cl/Electron/amtx/amtx.htm

This is purely by coincidence very close to something I built.

[HB9EVI]

Better keep the wire serving as antenna as short as possible; even as hamradio operator one is not allowed to send on broadcast frequencies. Even with some LC-Filter on the output you cannot be sure, that no harmonics are reaching the antenna - to reach -60dB attenuation of unwanted harmonics, you need at least a 5th order LC, which should actually be swept through an analyzer before usage.
If you place your transmitter e.g. 20cm next to the radio you're working on, a short wire of a few cm will give you enough signal to work.
As an AM-modulator you can use a NE602/SA612 like SM5ZBS is showing it: http://elektronikbasteln.pl7.de/am-modulation-mit-einem-sa612-oder-sa602-ne612-ne602.html.
Pot P1 allows to adjust the amount of carrier suppression; of course for an AM-radio you need the carrier to receive a clear signal, so match the the setting of P1 on you scope screen.

[bd139]

You're not going to get much range on AM broadcast with a long wire. I think off my DDS I get about 20-30 feet before the signal trails off. The wavelength is way too long.

I suppose it depends if you give a crap about the rules or not

[Howardlong]

If you want to build a simple AM transmitter, the simplest approach is:

1. Use a 1MHz crystal can oscillator to generate a square wave. A 555 is being pushed pretty hard in the AM band. They're really only designed for 100KHz max. They're also terribly unstable.
2. Apply that to a common emitter amplifier.
3. Use an op-amp to control the amplifier's supply current driven from audio (this is crude but effective). 
4. Stick a low pass filter on the output of that.
5. Attach a very long bit of wire as an antenna.

^ Precisely this, also known as Plate Modulation to us old farts, even though I've almost never gone near tubes. The op amp in this case would also act as a summing amplifier to provide DC bias, and its gain adjusted for the right modulation depth. Also, add a mini step 3a, AC couple with a cap to the LPF, or the LPF will effectively short out your op amp's DC biased output.

[bd139]

+1 for the coupling cap and better op amp explanation. Good idea. When I built it I used a pretty naff Pi filter so there is no inductor between collector and ground so the caps all charge up to the collector potential. Your antenna is at the bias voltage then so not ideal.

I had no idea that's what plate modulation is! Not a tube guy either although I just bought a VTVM to play with and kill myself with probably too.

[Gip-Gip]

After redesigning the circuit I'm more confident in the design, but I'm not 100% sure where to put the LC filter (if I'm supposed to put in an LC filter in the first place) The schematic's in the attachments as always

[Gip-Gip]

I think this circuit is the way it's supposed to be. In this schematic the chopped signal is ran through a LC oscillator, which spits out an AM'd sine wave in simulations.

(Schematic in attachments)

[Richard Crowley]

Is there some compelling reason to try to design an AM transmitter with no experience?
When I go to Google image search and enter: part 15 am transmitter circuit
I see hundreds of examples of low-power AM transmitter circuits that have likely been proven.

If you are in the US (which we assume from your flag) then there is an FCC rule ("Part 15") which permits you to "transmit" on the AM broadcast band using no more than 100mW and an antenna no longer than 10 ft. and as long as you don't interfere with licensed broadcast signals. At least those were the rules last time I checked

Ref: https://www.fcc.gov/media/radio/low-power-radio-general-information#UNLICENSED
Ref: https://www.google.com/search?q=part+15+am+transmitter+circuit&tbm=isch

There are also kits of working circuits, some for likely less than you could buy all the parts.
For example:  http://www.ebay.com/itm/152702773666

[Gip-Gip]

Is there some compelling reason to try to design an AM transmitter with no experience?
When I go to Google image search and enter: part 15 am transmitter circuit
I see hundreds of examples of low-power AM transmitter circuits that have likely been proven.

If you are in the US (which we assume from your flag) then there is an FCC rule ("Part 15") which permits you to "transmit" on the AM broadcast band using no more than 100mW and an antenna no longer than 10 ft. and as long as you don't interfere with licensed broadcast signals. At least those were the rules last time I checked

Ref: https://www.fcc.gov/media/radio/low-power-radio-general-information#UNLICENSED
Ref: https://www.google.com/search?q=part+15+am+transmitter+circuit&tbm=isch

There are also kits of working circuits, some for likely less than you could buy all the parts.
For example:  http://www.ebay.com/itm/152702773666

Since this isn't a time-critical application I feel I can learn some new things by making and debugging my own transmitter (that and I enjoy the design process just as much as the assembly)

Thank you for the legal docs though!

[Rerouter]

For building transmitters i have found it best to imagine the antenna is a terminated transmission line of some resistance. That way you can imagine it like driving a dummy load of that resistance

Your power calculations and what your output looks like on a scope get easier when your testing into a dummy load.

[Gip-Gip]

For building transmitters i have found it best to imagine the antenna is a terminated transmission line of some resistance. That way you can imagine it like driving a dummy load of that resistance

Your power calculations and what your output looks like on a scope get easier when your testing into a dummy load.

do you know any good way to calculate said resistance?

[G0HZU]

The simple/cheap mini AM transmitter circuit in the link below looks better to me.

http://www.techlib.com/electronics/amxmit.htm

This uses three NPN transistors arranged as a diff pair as the heart of the modulator.

The diff pair will naturally output a square wave at RF assuming there is sufficient LO drive from the crystal oscillator but the tuned loop antenna should filter out the worst of the RF harmonics. So this isn't really an issue. The modulation should be quite good on this transmitter with low distortion. I think it will perform better in this respect than the circuit you currently have? Easy and cheap to build as well!

[bd139]

For building transmitters i have found it best to imagine the antenna is a terminated transmission line of some resistance. That way you can imagine it like driving a dummy load of that resistance

Your power calculations and what your output looks like on a scope get easier when your testing into a dummy load.

do you know any good way to calculate said resistance?

Usually you assume the antenna is 50 ohms. It's usual to use a couple of 100 ohm resistors in parallel as a load. However the antenna rarely is 50 ohms precisely so you stick a tuning unit between the antenna and the transmitter which "matches" the impedance of the antenna to the transmitter. Don't worry about that for now though.

[Richard Crowley]

Since this isn't a time-critical application I feel I can learn some new things by making and debugging my own transmitter (that and I enjoy the design process just as much as the assembly)

Just remember that you can inadvertently cause significant interference to other users of the RF spectrum by indiscriminate experimentation or ignorance of the principles.  DIY transmitter design is not a typical beginner project.  Without some experience and specialized test gear, you could get yourself into trouble.

Particularly troubling is the question about transmitting with square waves. That is exactly the kind of problem you can get yourself into by transmitting a significant series of harmonics and interfering with many other communication bands and users.  Be very careful. Not as simple as it looks.

[Gip-Gip]

Just remember that you can inadvertently cause significant interference to other users of the RF spectrum by indiscriminate experimentation or ignorance of the principles.  DIY transmitter design is not a typical beginner project.  Without some experience and specialized test gear, you could get yourself into trouble.

Particularly troubling is the question about transmitting with square waves. That is exactly the kind of problem you can get yourself into by transmitting a significant series of harmonics and interfering with many other communication bands and users.  Be very careful. Not as simple as it looks.

The main reason I asked the question is because there are many videos/tutorials/schematics that use square waves

You're right in that this is kinda a risky project, but the thing's so low powered in the first place I sincerely doubt I'm going to run into any issues

It would be nice to know what special equipment you speak of...

[bd139]

I'm not sure it requires that much specialised test gear. It's amazing what you can get away without if you know how it works and actually come up with well behaved transmitters. If you have an RF power meter and frequency counter (both trivially DIY) then you build something pretty good. If you're worried about harmonics you can get a long way with a crappy absorption wave meter.

The first TX I built was knocked up with a VOM and GDO.

Really it's quite difficult producing something that has enough range to cause RFI problems.

[Gip-Gip]

Couldn't I do a good portion of the tests with the 8 Band radio I have?

[bd139]

Yes and no. The main problem is harmonic output. So you'd need to be able to receive at 3x, 5x, 7x the frequency that you are transmitting.

This is rarely done for simple transmitters. A low pass filter is usually added which reduces these harmonics significantly. You can build one of these easily without having to measure anything really.