How can I improve this cheap AM transmitter?

[MaximRecoil]

It's very low-powered; it only has a range of about a foot. I use it to listen to old radio shows, which are mostly talking, on an AM radio. I feed it an audio signal from my PC's line-level speaker output jack and power it with a 6 VDC, 2-amp "wall wart" type AC adapter power supply.

With the transmitter's volume/gain pot turned up all the way, and with my PC's Windows volume control slider at 50%, it has a good signal-to-noise ratio (I can barely hear any static when it's at a comfortable listening volume), and talking sounds okay, but when music plays there's noticeable distortion. If I turn the PC's volume control down to 25% (and turn up the AM radio's volume control to compensate), there's less distortion, but a worse signal-to-noise ratio. If I also turn the transmitter's volume/gain pot down to about 50%, distortion pretty much disappears, but the signal-to-noise ratio gets really bad.

As far as I can tell, the transmitter only has 3 active components, i.e., the 3 transistors (they are all S9018, at locations BG1, BG2, and BG3). The lower PCB is just for the optional batteries which I'm not currently using; it doesn't do anything else). Are there better transistors that I could use and/or anything else I could do to improve the amount of low-noise volume I can get before it starts to audibly distort?

[Kim Christensen]

You may be able to compress the audio using the application on your PC. That way you could have the volume set to an optimal level for all situations.
I assume you've oriented the radio relative to the transmitter for max signal? (turned the loopsticks for max coupling)
Do you have a schematic for this circuit?

[MaximRecoil]

You may be able to compress the audio using the application on your PC. That way you could have the volume set to an optimal level for all situations.

What do you mean by "compress"? If you mean file size compression, the stuff I was just listening to is already highly compressed (~16 kbps Opus). If you mean compress the dynamic range, since this stuff was recorded from radio broadcasts to begin with, it already has a compressed dynamic range.

I assume you've oriented the radio relative to the transmitter for max signal? (turned the loopsticks for max coupling)

Yeah, reception isn't a problem. One of my radios has a signal strength meter, and it shows pretty much maximum strength. Like I said, I can get it to a good listening volume and barely hear any noise/static at all; the problem is the distortion when music plays (the distortion isn't really noticeable when it's just people talking though). It's the same type of distortion you get when you turn anything up beyond its clean amplification capabilities, which is what makes me wonder if different transistors would help.

Do you have a schematic for this circuit?

Surprisingly, I do (along with a PCB layout diagram). I didn't think I did, else I would have included it in my original post:

[Kim Christensen]

mean compress the dynamic range

Yes, that's what I meant.

It's very low-powered; it only has a range of about a foot.

Yeah, reception isn't a problem. One of my radios has a signal strength meter, and it shows pretty much maximum strength.

These two statements are very contradictory to me... Sounds like it doesn't output much power at all, even for something like this. I did notice that the PCB has a connection for an external antenna. You might want to connect a longish wire to the "monitor" connection. Loudness of AM reception is very dependent upon received signal strength, so you may just be overdriving it to compensate.

Like I said, I can get it to a good listening volume and barely hear any noise/static at all; the problem is the distortion when music plays (the distortion isn't really noticeable when it's just people talking though). It's the same type of distortion you get when you turn anything up beyond its clean amplification capabilities, which is what makes me wonder if different transistors would help.

It's a very simple circuit. The fidelity of the design won't be that great. Was this a kit or did it come pre-assembled?

[MaximRecoil]

These two statements are very contradictory to me... Sounds like it doesn't output much power at all, even for something like this.

When it's within range the reception is great, which is why there's hardly any static at all; much stronger reception than any actual AM radio station here. I have it sitting just a few inches from the radio's internal ferrite rod antenna (or external one in the case of my radio that has the signal strength meter).

I did notice that the PCB has a connection for an external antenna. You might want to connect a longish wire to the "monitor" connection. Loudness of AM reception is very dependent upon received signal strength, so you may just be overdriving it to compensate.

I didn't know it had a connection for an external antenna. I wonder why it's called "monitor." I'll definitely try that though.

It's a very simple circuit. The fidelity of the design won't be that great. Was this a kit or did it come pre-assembled?

The sound is pretty good for spoken word, and even with some types of music there's no noticeable distortion. I first noticed the distortion with some music that had prominent bass in it; the bass notes were very noticeably distorted. Then I noticed it again while listening to a recording of an old Art Bell radio broadcast, during his "bumper music" ("Right Back Where We Started From" by Maxine Nightingale), and it wasn't bass that was distorting this time, but rather, the higher pitched instrumentals at the start of the song.

It was a kit that I assembled. I used all the parts that came with it except for the off-brand electrolytic capacitors. I had some new Nichicon capacitors of the right values here already so I used those instead.

Aren't the transistors amplifying the signal? If so, could they be replaced with more capable transistors? They wouldn't need to be all that much more powerful.

[Kim Christensen]

AM is supposed to have a limited bandwidth. I noticed that the circuit doesn't have an audio bandpass filter on the input, so this may be the source of the problem with different music vs voice. Audio is normally limited to 5 kHz for AM broadcast. It should probably have a lower audio frequency cutoff as well.
ie: By restricting the bandwidth, it would reduce the overall audio level of the music while leaving the audio level of voice pretty much the same.

[MaximRecoil]

AM is supposed to have a limited bandwidth. I noticed that the circuit doesn't have an audio bandpass filter on the input, so this may be the source of the problem with different music vs voice. Audio is normally limited to 5 kHz for AM broadcast. It should probably have a lower audio frequency cutoff as well.
ie: By restricting the bandwidth, it would reduce the overall audio level of the music while leaving the audio level of voice pretty much the same.

Like I mentioned in my first post, the distortion goes away, even on the problematic music, if I turn the signal down enough, but that decreases the signal-to-noise ratio to the point that you get a lot of "static" that you don't want to hear along with the audio that you want to hear. It needs a little more clean amplification than it has, maybe 10% more.

Connecting a wire (about 10 feet long) to the "monitor" through-hole didn't do anything that I could notice; didn't increase the reception distance or anything.

I did notice that I get a little less distortion (along with a pretty good signal-to-noise ratio) if I turn up the volume in Windows and decrease the volume pot on the AM transmitter, for example, Windows volume at 100% and transmitter volume at about 40% (which is about the opposite of how I normally have it). It seems to do better with a hotter input signal that it doesn't have to amplify as much, than it does with a weaker input signal that it has to amplify more.

[807]

That circuit modulates the free running oscillator, so will probably introduce a lot of FM. This could make the audio sound distorted.

Do you have a scope that you could connect to the monitor output? If you also have an audio generator, you could look at the modulation depth to see if you are getting anywhere near 100%.

[MaximRecoil]

That circuit modulates the free running oscillator, so will probably introduce a lot of FM. This could make the audio sound distorted.

Do you have a scope that you could connect to the monitor output? If you also have an audio generator, you could look at the modulation depth to see if you are getting anywhere near 100%.

I don't have an oscilloscope, but clipping seems to be the cause of the audible distortion.

I have two ways to control the input signal level: the Windows volume control slider and the pot on the transmitter. I can make the audible distortion disappear if I set the pot to 50% and the Windows slider to 25%, but that results in a lot of static when I turn the radio up loud enough to hear well because the signal is too weak.

If I set the pot to 100% and the Windows slider to 25%, everything sounds good and there's hardly any static, except for some distortion in certain types of music. I can also make everything sound very distorted, including talking, by having both the pot and the Windows slider set to 100%. I don't have enough power to reach a good listening volume while having both low-noise and no audible distortion.

Is there anything in this transmitter that amplifies the input signal? If there's no amplification going on, that means I only have a line-level signal being converted to radio waves, which is pretty weak.

[radiolistener]

For a medium wave transmiter large antenna is mandatory required. Your mistake is that you dont' connect antenna to two terminals marked as "Monitor", this is why you have weak signal on receiver.

According to this schematic:

your transmitter is designed for low impedance output, just because output coil of transformer has low impedance. This means that your transmitter is designed for low impedance antenna, such as magnetic loop. So, in order to get better efficiency you're needs to connect magnetic loop antenna instead of dipole. The simple way is just to connect large wire loop to "Monitor" terminals of your transmitter. Note, that wire loop should have enough cross-section of the loop, this is very important parameter for magnetic loop antenna.

As alternative you can wound 1-2 loop around your receiver with a wire and connect that coil to "Monitor" terminals of transmitter. In such way you can maximize reception signal level for your receiver and minimize for other devices which is not covered with such wire loop.

PS: I also notice, that there is connection between output coil and GND terminal of the power supply connector. This connection is intended to use power supply cable as a part of antenna. But this is not a good idea, I would suggest to disconnect it and use separate antenna antenna wire connected to both "Monitor" terminal. But don't expect much from such modification, it just allows you to avoid "strange behavior" of your transmitter which depends on power supply cable length in current implementation.

Another thing which needs to be taken into account is power supply. For AM transmitter it should be linear power supply with a good filtering. If you use switching power supply such as a phone charger or other SMPS, it can flood your music with a weird noise.

[MaximRecoil]

The simple way is just to connect large wire loop to "Monitor" terminals of your transmitter. Note, that wire loop should have enough cross-section of the loop, this is very important parameter for magnetic loop antenna.

As alternative you can wound 1-2 loop around your receiver with a wire and connect that coil to "Monitor" terminals of transmitter. In such way you can maximize reception signal level for your receiver and minimize for other devices which is not covered with such wire loop.

You say "monitor terminals" but I only see one terminal that's marked "monitor," so I'm not sure how I'm supposed to connect a loop of wire. Do I connect both ends of the wire loop to the monitor terminal, or do I connect one end to the monitor terminal and the other end to the ground (GND) terminal that's beside it? I tried it both ways. With both ends of the wire loop connected to the monitor terminal, there was no difference at all. With one end connected to the monitor terminal and the other to the GND terminal, it didn't work at all (zero reception). When I removed it reception was instantly restored.

I don't think it's a reception problem to begin with. The included ferrite rod transmitting antenna is only a couple inches from the radio's ferrite rod receiving antenna, and there's hardly any static. The audio comes through loud and clear, that is, until certain notes of certain songs play and you hear distortion.

The problem seems to be lack of volume in the line-level signal from my PC. Line-level signals aren't meant to be used as-is, they are meant to be amplified, and if this transmitter doesn't amplify the signal (I can't seem to get an answer to the question of whether it does or not, so I'll just assume it doesn't), then that's a problem. In order to get sufficient volume it has to be over-driven, resulting in clipping/distortion.

Another thing which needs to be taken into account is power supply. For AM transmitter it should be linear power supply with a good filtering. If you use switching power supply such as a phone charger or other SMPS, it can flood your music with a weird noise.

I'm using a switching power supply that gives the same results as powering it with batteries. I have another switching power supply that I tried too, but it introduced a constant buzzing sound so I ditched that one. I was going to order a linear power supply, but before I did I tried that switching one that I already had on hand, and since it sounded the same as the batteries I didn't bother ordering a linear power supply.

[radiolistener]

You say "monitor terminals" but I only see one terminal that's marked "monitor," so I'm not sure how I'm supposed to connect a loop of wire.

There are two terminals for antenna on your PCB, they marked as GND and Monitor. I suggest you to break wire between power supply GND and GND terminal of antenna. It minimize effect of power supply cable on your transmitter efficiency.

[radiolistener]

Do I connect both ends of the wire loop to the monitor terminal

No

or do I connect one end to the monitor terminal and the other end to the ground (GND) terminal that's beside it?

Yes. These both terminals are intended for antenna connection. Note that antenna always needs two terminals.

I tried it both ways. With both ends of the wire loop connected to the monitor terminal, there was no difference at all. With one end connected to the monitor terminal and the other to the GND terminal, it didn't work at all (zero reception).

Antenna wires should be connected to both - Monitor and GND terminals. One side of wire to Monitor terminal, second side of wire to GND terminal. You can experiment with length of wire and winding count for better reception. Magnetic loop needs to use at least 10-20 meters length of wire and loop needs to be at least 1x1 meter or more.

As I remember, resonant frequency of 10 meter piece of wire used as magnetic loop with 12-400 pF variable capacitor is about 3.5-7 MHz. For MW band it needs at least twice longer wire. So, don't expect that it will work good with 1-2 meters of wire like it works for VHF band.

If it stops to work, it means that your loop impedance is too small for that circuit. Use more long loop with more windings. If it stops to work, it may be due to too small impedance of your loop - try more longest wire and more winding in a loop. You can also try to use two long wires as a dipole, but it needs much more size than magnetic loop antenna (about 150-200 meters of straight wire).

Keep your receiver within transmitter antenna loop, it maximize energy transfer through near field.

Also note that magnetic loop should be large enough, for example it can be wire around your room. At medium wave good efficiency antenna needs to have dimension about 150-200 meters.

[NiHaoMike]

Try shorting all 3 terminals of BG3 to effectively bypass that amplification stage, you'll then have to turn up the volume to compensate.

[MaximRecoil]

There are two terminals for antenna on your PCB, they marked as GND and Monitor. I suggest you to break wire between power supply GND and GND terminal of antenna. It minimize effect of power supply cable on your transmitter efficiency.

Like I mentioned in my previous post, I get the exact same results powering the transmitter with the AC adapter power supply as I do with four AA batteries (which takes the power supply and its cable completely out of the equation).

Antenna wires should be connected to both - Monitor and GND terminals. One side of wire to Monitor terminal, second side of wire to GND terminal. You can experiment with length of wire and winding count for better reception. Magnetic loop needs to use at least 10-20 meters length of wire and loop needs to be at least 1x1 meter or more.[/quote]

Thanks. I don't have enough wire here to experiment with large loop antennas, and I definitely can't make a magnetic loop antenna, for more reasons than just a lack of wire.

Try shorting all 3 terminals of BG3 to effectively bypass that amplification stage, you'll then have to turn up the volume to compensate.

So the BG3 transistor is functioning as an amplifier? If so, is there a type of transistor with more amplification that I could replace it with, that would work as-is in the circuit? I don't need the amplification to be all that much stronger, since the way it is right now sounds good for everything I listen to through it aside from certain notes in certain music.

By the way, what do you mean by shorting all three terminals? Do you mean to short them to each other, like would happen if you wrapped a single piece of wire around all three of the transistor's legs with it still in circuit?

[radiolistener]

Like I mentioned in my previous post, I get the exact same results powering the transmitter with the AC adapter power supply as I do with four AA batteries (which takes the power supply and its cable completely out of the equation).

any cable or wire affects RF fields, and you cannot predict how it can affect. It can be better or worse, just unpredictable, this is why I recommend to break connection between GND terminal of antenna and GND terminal of power supply. In such way you can eliminate random effect of the power supply wire and concentrate on antenna experiments.

[jwet]

The circuit is kind of a minimalist AM transmitter- and not terrible.  I don't know why it only transmits a foot- it should cover a few hundred feet even with a poor antenna.  The two upper transistors are an oscillator operating in a somewhat balanced way- this isn't bad.  The third transistor pulls on this pair to increase current, this is AM modulation- it shouldn't affect frequency too much at least first order.  The 3.3K resistor in the emitter circuit of Q3 sets the modulation "depth"- it sounds like you're getting decent modulation.  Decreasing the value of the 3.3K would increase modulation- you could play with this.  The way the circuit is designed, it would also increase output power.  This kind of puts you in a box- it would be nice to set RF power output separate from modulator depth.  You could experiment with putting something like a 1K resistor to ground from the emitter of the two upper transistors- this would get them making a carrier at some higher power.  You would then use the BG3 circuit as is with the 3.3k to further pull on two upper emitters.  So I guess the change would be to bridge the lower Q and R6 with a 1K and see if that let's you have a strong carrier with decent modulation.  I hope this makes sense.  The pot setting stuff is not important, open this wide open and use the PC volume control.  You said you don't have a scope, if you have voltmeter, it might be fruitful to probe around the terminals of the transistor and make sure that there all operating in their linear range.   If you get the bias values for the nodes, post them and we can have a look.  I played with little am transmitters as a kid- I made room bugs and other gadgets in the 60's.  Good luck.

[RoGeorge]

That's a Peltz oscillator.  If R6 is too big, it might stop.

I suspect the oscillator might stop intermittently when modulated with music, and such the audio distortions.

- First, connect an antenna, a big loop of wire between the pads 'Monitor' and 'GND', as told before.
- Then, if you still have distortions, I guess it's because the oscillator doesn't oscillate continuously.  Try reducing R6, for example put another 2...3k in parallel to reduce R6 value, so to increase the current and make the Peltz oscillator more stable.

[MaximRecoil]

The circuit is kind of a minimalist AM transmitter- and not terrible.  I don't know why it only transmits a foot- it should cover a few hundred feet even with a poor antenna.

Reception falls off rapidly and drastically if you move the transmitter more than about a foot from the radio. Before I ordered this thing I watched a YouTube review of one, and his had the same ~1 foot range, so I didn't expect anything more than that to begin with.

The two upper transistors are an oscillator operating in a somewhat balanced way- this isn't bad.  The third transistor pulls on this pair to increase current, this is AM modulation- it shouldn't affect frequency too much at least first order.  The 3.3K resistor in the emitter circuit of Q3 sets the modulation "depth"- it sounds like you're getting decent modulation.  Decreasing the value of the 3.3K would increase modulation- you could play with this.  The way the circuit is designed, it would also increase output power.  This kind of puts you in a box- it would be nice to set RF power output separate from modulator depth.  You could experiment with putting something like a 1K resistor to ground from the emitter of the two upper transistors- this would get them making a carrier at some higher power.  You would then use the BG3 circuit as is with the 3.3k to further pull on two upper emitters.  So I guess the change would be to bridge the lower Q and R6 with a 1K and see if that let's you have a strong carrier with decent modulation.  I hope this makes sense.  The pot setting stuff is not important, open this wide open and use the PC volume control.

Thanks for the detailed information. I might try swapping out the 3.3K resistor for a lower value one first, since that would be the easiest thing to try, and I might already have some resistors on-hand to try.

You said you don't have a scope, if you have voltmeter, it might be fruitful to probe around the terminals of the transistor and make sure that there all operating in their linear range.   If you get the bias values for the nodes, post them and we can have a look.  I played with little am transmitters as a kid- I made room bugs and other gadgets in the 60's.  Good luck.

I do have a voltmeter, but I don't know what I'd be looking for when probing around the terminals of the transistor with it. I don't know what "bias values for the nodes" means either.

I made some direct-feed recordings from my radio's headphone jack into my PC's soundcard's line-in jack while it was playing broadcasts from my transmitter (which I fed from a different PC):

https://app.box.com/s/uzhdxkmmeea60seaa0twxp3ct26ro0q1

File #1 is a 20-second clip from the song Tom's Diner; Windows volume is at 100% and the transmitter's pot is at 100%. I included this to show that there isn't a reception problem, since there's hardly any static and it sounds almost as good as the source audio which is from a CD (the source audio sounds "brighter" for lack of a better term, due to containing higher frequencies than the recording from the AM transmitter), which I included for comparison (file #5). I chose it because it's just talking/singing, no musical instruments, and the source is high quality.

File #2 is an extreme example of the type of distortion I've been talking about (100% Windows volume, 100% pot). The source audio (which I also included; file #6) is someone's recording off the radio, so it's not very clean audio to begin with, but it's typical of what I usually listen to through this transmitter.

File #3 is the same audio clip, but with the input signal levels set to where I usually have them (100% Windows volume, 50% pot), because it gives me enough volume with low noise, talking sounds fine, but there is still some distortion in the music.

File #4 is the same audio clip, but with the input signal levels turned down enough to eliminate audible distortion in the music (25% Windows, 50% pot). This results in a poor signal-to-noise ratio, i.e., prominent static.

[RoGeorge]

Seems like the distorts happen only when the audio signal swings too much.  Try making R6 twice as small, or twice as big, and see if there is any difference.  Maybe, try powering the transmitter from another wall-adapter, or from batteries.  Anything between 5-12V should be close enough voltage to power the transmitter.

Were your posted records made with an antenna attached to the transmitter, or without?

[MaximRecoil]

Seems like the distorts happen only when the audio signal swings too much.  Try making R6 twice as small, or twice as big, and see if there is any difference.

I just tried it with a 1K resistor in place of the 3.3K resistor at R6, and that didn't make a difference, so I removed the 1K and put the 3.3K back in there. I also tested the 3.3K resistor with a meter (out of circuit) and it's close enough to its stated value (3.27K). I don't have a resistor that's twice as big; the biggest I have is 5.1K. I guess I could try that.

Maybe, try powering the transmitter from another wall-adapter, or from batteries.  Anything between 5-12V should be close enough voltage to power the transmitter.

I've already tried it with batteries and a different AC adapter; no difference (aside from the different AC adapter introducing a constant buzzing sound in the speakers, so I didn't continue using it).

Were your posted records made with an antenna attached to the transmitter, or without?

Just with the onboard ferrite rod antenna which is positioned a couple inches from the radio's ferrite rod receiving antenna, and perpendicular to it. I don't have an antenna to attach to the monitor/GND terminals. As I mentioned in a previous post, I tried a loop of 18 gauge insulated wire that was about 10 feet long but it made it so there was no reception at all when it was attached. Removing it instantly restored reception. I don't have enough wire here to experiment with big loops.

[PA0PBZ]

What frequency is this transmitter supposed to transmit at? I'm beginning to wonder if you are listening to a harmonic frequency of the transmitter and not the fundamental frequency (harmonic meaning 2 or 3 times the frequency of the transmitter) or maybe even at a mirror of the fundamental frequency. Do you have any means of measuring the output frequency of the transmitter? Did you tune the radio over the full band to see if there is any other spot where you can hear the transmitter?

[MaximRecoil]

What frequency is this transmitter supposed to transmit at? I'm beginning to wonder if you are listening to a harmonic frequency of the transmitter and not the fundamental frequency (harmonic meaning 2 or 3 times the frequency of the transmitter) or maybe even at a mirror of the fundamental frequency. Do you have any means of measuring the output frequency of the transmitter? Did you tune the radio over the full band to see if there is any other spot where you can hear the transmitter?

Its variable capacitor corresponds with the variable capacitor in a typical AM radio, i.e., if you turn its tuning dial all the way to the left, it corresponds with all the way to the left on the AM radio (about 530 kHz). I currently have the transmitter set to about 570 kHz, and the radio can also just barely pick it up when tuned all the way to the right (about 1710 kHz). It's mostly static/noise when the radio is tuned to 1710 kHz, but you can just barely hear the audio signal.

[jonovid]

looking at your the circuit. it seems like an odd way to modulate AM
better have one transistor as your dedicated oscillator circuit and the other transistor as amp/ or buffer  output modulation tank circuit.
from a audio transformer secondary on its power input.

my experience with AM MW Medium-wave bugs is the harmonics are horrendous popping up all over the Medium-wave band at just in the same room.
as the radio receiver.

[radiolistener]

Forgot to say, this transmitter doesn't have frequency stabilization with crystal oscillator and buffering, so it's frequency depends on impedance of your transmitting antenna.

What this means - if you connect some transmitting antenna and signal disappears from your receiver, it doesn't means that transmitter stopped to work due to overload, it may just change frequency. So, you're needs to check if transmitter works on different frequency before conclusion that it stops to work. Just scan entire AM band on your receiver if signal disappears after some change within circuit or antenna.

The good practice is to check signal with a small wire loop (RF sensor) connected to oscilloscope. But if you don't have oscilloscope, you can use your receiver, just scan entire AM band to find the carrier of your transmitter.