why does this FM transmitter perform so poorley?

[yada]

So I built this circuit:


Using a SS9018FBU NPN RF transistor, exactly like the schematic but with a few modifications: Instead of the 4.7pf there are three 10pf in parallel. Instead of a microphone I have a 1k resistor and the inductor is 7 turns of 22ga magnet wire 5mm dia X 7mm long (have no idea what the value is but this is critical to the circuit. Putting metal pin in it can double the signal and completely stop it just by moving it 1/2 in to 34 way in.

As you can see from the water fall it makes five signals and they move back and fourth. The bandwidth is huge and the power supply says it pulls 14ma. Trying different transistors that I listed in the other thread makes no difference. I built it using as few wires as possible. I know a ground plane helps but all I have is perf board, so I was going to try building on that and see if this thing "calms down". Is the microphone substitute resistor the wrong size? I have no idea why this thing acts like this.

[Earendil]

Could you describe the power supply too? Do you use batteries? Have you tried with batteries?

[hamster_nz]

As a first guess, the transistor's base will be at about 5V, so the emitter will be at about 4.5V, so current through R4 will be about 10mA. So 14 mA total power seems a little higher than expected at first glance. Double-check R2 and R3?

Also, without a MIC, you can just remove R1&C1 - they are only there to bias the microphone, and couple the signal from the microphone to the transistor's base.

[hamster_nz]

Also, the 10pF caps are in parallel. Doesn't that give 30pF rather than 3.3pF?  Try two in series. Should give about 5.0pF.

[hamster_nz]

Is the leftmost resistor on the picture R3? It looks to be 1200 Ohm, 5600 Ohm.

[kony]

Breadboarding this is not exactly the brightest idea. Get some scrap piece of singlesided copper clad and try again wire-nest style.

[DimitriP]

Breadboarding this is not exactly the brightest idea. Get some scrap piece of singlesided copper clad and try again wire-nest style.

It's about time someone said this !!!

[yada]

I'm perf boarding it now. First thing I just fixed was that I had the resistor (in place of microphone) connected to wrong side of the 4.7nf cap. Now it pulls 30ma and the frequency moves all over the place. Going to try and fix the parallel caps. I always get that mixed up, series lowers value.

This is my first RF circuit so I'm learning the hard way (bread board to see power consumption first)! 

[nev23]

7 turns is way too much; try 3.
this circuit performs poorly because it is very simple. Moving or touching the antenna will severely de-tune or possibly stop the oscillator. even moving the PSU wires will affect it.
You need copper clad board and christmas tree construction. Stick tiny scraps of board on top of the main board to make circuit nodes. keep all wires short.
It will still be unstable; there is no buffering from the effects of the antenna, and no temperature compensation.
The only way to remove the effect of the antenna is to set the oscillator at half frequency and use a doubler. Or use a PLL. In either case, a buffer amplifier will also be needed.
Output power from your present circuit will be at most about a milliwatt, so the range will only be a few metres.

[bobaruni]

When biased correctly and constructed well, I have seen this type of circuit go a few hundred metres with decent FM quieting.
Try changing R4 to a higher value  (47R is way too low) and add a cap in series with the antenna so as not to load the LC tank too much or tap the coil about half way for connection to the antenna.
See below a very successful tried and tested version of the circuit you presented, just ignore the mic pre-amp circuit as it's not required unless you need extreme mic sensitivity.
http://www.talkingelectronics.com/projects/FM-Bug/FM-Bug.html

[mmagin]

I built something similar to that from a kit when I was a kid.  Something I discovered later on when I had access to a sufficiently fast scope was that it exhibits AM modulation in addition to FM modulation, which is not too surprising for such a simple circuit.

Now that I think of it, how is it frequency modulating, anyway?  Is it relying on varactor-like effects in the transistor?

[voltz]

As stated, the construction is not good enough for a VHF transmitter. Breadboard is out. It needs to be on a better base, like copper clad.

Its such a simple circuit with no buffering, any movement close to the antenna will effect the frequency of the transmitter. Any Power supply variance will also effect the frequency and any mechanical movement / temperature changes of the components will also effect the frequency. Hence the better construction required to help stability and for other reasons.

Your waterfall SDR display is receiving way too much signal from the transmitter. Did you connect it directly? If so, dont. just let the SDR pick it up with its own antenna or use an attenuator. Reduce the signal to the SDR. You should end up with a single line (not a broad bright white band as in the pic).

[bobaruni]

Is it relying on varactor-like effects in the transistor?

Yes, I think so, and definitely there will be some AM but this will be limited out by a good FM tuner, it's not ideal but all you should expect from a single transistor.

[vk3yedotcom]

One thing I find is that although beginners may be attracted to super simple circuits, performance and even ease of construction actually gets a bit easier with a few more parts.

This is because you're not asking too much from each transistor or you're only asking it to do one thing (and you can optimise component values for that duty).

So for a simple FM bug I'd recommend something with 2 or 3 transistors.

Talking Electronics has a good write up and circuits http://www.talkingelectronics.com/projects/FM-BugsIntro/FM-BugsIntro.html

[yada]

When biased correctly and constructed well, I have seen this type of circuit go a few hundred metres with decent FM quieting.
Try changing R4 to a higher value  (47R is way too low) and add a cap in series with the antenna so as not to load the LC tank too much or tap the coil about half way for connection to the antenna.
See below a very successful tried and tested version of the circuit you presented, just ignore the mic pre-amp circuit as it's not required unless you need extreme mic sensitivity.
http://www.talkingelectronics.com/projects/FM-Bug/FM-Bug.html

I thought that used the electrick mic as a 1st stage like a FET so not easily removed.

[bobaruni]

It can be very easily removed, it's just a gain stage for the mic, just couple to mic directly to the base of the oscillator transistor via a small cap (22n-100n should do) but also remember to include the mic bias resistor.

[yada]

It can be very easily removed, it's just a gain stage for the mic, just couple to mic directly to the base of the oscillator transistor via a small cap (22n-100n should do) but also remember to include the mic bias resistor.

This says otherwise:

The electret microphone contains a FET transistor and can be counted as a stage, if you wish. The microphone detects audio in the form of air vibrations that enter the hole (at the end of the mic) and move the diaphragm. This diaphragm is a thin piece of metallised plastic such as mylar and is charged with electrical charges during manufacture.
Next to this is a metal plate containing a number of holes so that the air readily passes through. The relative distance of the mylar diaphragm to the metal plate makes the charges move on the diaphragm (remember static electricity theory: like-charges repel and unlike-charges attract). Some of the charges pass down a lead that touches the metal plate and into a FET amplifier - it looks like a three legged transistor. The FET amplifies the charges and gives a reading on the output lead.
The output must be connected to a supply via a resistor called the load resistor. The FET draws a varying current during its operation and this creates a varying voltage on the output (across the load resistor).
The reason why a FET has been used is due to it having a very high input impedance and does not have any loading effect on the charges.
The output waveform from the microphone will be typically 3 - 30mV in our case, depending on how close it is to the source of the sound. The circuit is capable of detecting a whisper at 10ft (3M) and only very sensitive microphones have been included in the kits.
You can also get medium and low sensitivity devices from suppliers so you have to be careful as they are not labelled.
A 22n capacitor on the output of the microphone couples the signal to the input of the first audio amplifier stage. This capacitor is designed to separate the DC voltage on the microphone from the base voltage on the transistor.

I don't have a mic so I want the circuit to be think it does, like just make a carrier wave.

[yada]

I'm perf boarding it now. First thing I just fixed was that I had the resistor (in place of microphone) connected to wrong side of the 4.7nf cap. Now it pulls 30ma and the frequency moves all over the place. Going to try and fix the parallel caps. I always get that mixed up, series lowers value.

This is my first RF circuit so I'm learning the hard way (bread board to see power consumption first)!

The circuit only pulls 5-6ma bread boarded, so I don't think its oscillating and as far a I can tell its built right. I think this only works if make a proper PCB for it with lots of effort put into the length of the traces.

[stj]

L1 is not an inductor, it's an air-spaced coil.
the antenna should be tapped from the center most winding for best results.

[BigBoss]

I'm perf boarding it now. First thing I just fixed was that I had the resistor (in place of microphone) connected to wrong side of the 4.7nf cap. Now it pulls 30ma and the frequency moves all over the place. Going to try and fix the parallel caps. I always get that mixed up, series lowers value.

This is my first RF circuit so I'm learning the hard way (bread board to see power consumption first)!

The circuit only pulls 5-6ma bread boarded, so I don't think its oscillating and as far a I can tell its built right. I think this only works if make a proper PCB for it with lots of effort put into the length of the traces.

If you want to build up a brand-new PCB based transmitter, I suggest you to use SMD component.They don't have leads and there fore they don't create any trouble.

[4CX35000]

Such simple circuits tend to be only for demonstration of LC resonance and the LC resonance formula were students can measure or hear the output on a radio and even modulate the output to some degree and then using LC formula can calculate the setting of the variable capacitor. But the limitations are vast, such as changing antenna position and the output frequency changes due to the changes in antenna impedance, limited FM deviation and so on. The transistor is simply forming part of the LC oscillator, so changing it will achieve very little.

If you wanting to build a simple transmitter then look around for a design based on the BA1404 IC
http://electronics-diy.com/BA1404_Stereo_FM_Transmitter.php

[john_c]

I think this only works if make a proper PCB for it with lots of effort put into the length of the traces.

The style of prototyping known as "dead bug" makes high frequency circuit construction a lot easier (the name refers to how ICs are placedupside down on top of your copper-clad board). You can get a good idea of what that looks like from Appendix F, here: http://cds.linear.com/docs/en/application-note/an47fa.pdf

[Teledog]

For experimenting, perhaps try a VCO?
Cheap and dirty  (literally)
Lower VHF VCOs are getting harder to find (eBay)
Bias the Vf with a pot & couple a capacitive input to the same.. and  a small pF cap on the output - to an appropriate antenna..
Cheap & easy FM transmitter (will drift like hell, but fun for an experiment!)
G'Luck!