FM transmitter (mod4)

[audioguru]

Here is how the bridge rectifier should be connected:

 

[Hero999]

I've changed Q3, it is still getting hot (without any resistor at the emitter)

I should've noticed that problem before.

This circuit has a transistor biasd on with a short circuit (at DC) to both rails, it's almost certain to be unstable and cook itself.

I haven't tried to add the resistor yet.

You might be lucky, this time your transistor might have a slightly lower gain, there again if you turn the heating on or use it on a hot summer's day it will probably blow again.

What's happing is a classic thermal runaway scenario. Q3 is biasd on slightly, it gets warmer so its gain increases slighly with increasing temperature so more current flows from the emitter to collector until it blows.

Adding the resistor breaks this cycle by lowering the gain slightly and more importantly limiting the current.

 

[audioguru]

Q3 is supposed to be driven at the resonant frequency of the tuned circuit at its collector. Then it won't have the dead short of the coil at DC at its collector, it will have the high impedance load of the tuned circuit in parallel with the 75 ohms of the antenna.
Q3 also operates a little in class-C because the input coupling capacitor becomes charged by the input signal. Then it operates cooler.
When it has a signal, a lot of the power from Q3 will go to the antenna without adding to the heating of Q3.
Then Q3 won't be hot, just warm.

 

[Hero999]

I don't think so audioguru.

Q3 is being operated in class A, it's always conducting some current because pulses from the oscillator stage simply aren't strong enough to turn it fully off at any part of the cycle.

Even if it was operating in class C while the ocillator is running, this is still very dodgy because if the oscillator decides to stop the amplifier won't fail safe, it will just blow up.

 

[audioguru]

I don't think so audioguru.

Did you calculate the heating when the oscillator is stopped?
The 47k base resistor will have about 8.2V across it. At room temperature the 2N3904 will have a typical gain of 200 so its collector current is 34.9ma and its power dissipation is only 307mW.
If it is at its max temperature then its gain is 280 and its dissipation is only 430mW. Its max rated dissipation is a lot more at 625mW so it is fine.

Q3 is being operated in class A, it's always conducting some current because pulses from the oscillator stage simply aren't strong enough to turn it fully off at any part of the cycle.

I disagree. The pulses from the oscillator are from a low impedance emitter follower so have plenty of current to charge the coupling capacitor so that Q3 is cutoff over some of its cycle.

Even if it was operating in class C while the ocillator is running, this is still very dodgy because if the oscillator decides to stop the amplifier won't fail safe, it will just blow up.

Lots of circuits blow up if their oscillator fails, but I show above that this one just gets hot (not too hot). What about an inverter that makes mains AC from a car battery? If its oscillator stops then hundreds or thousands of DC amps flow.

 

[Hero999]

Please don't base your assumptions purely on the typical values listed on the datasheet.

Did you calculate the heating when the oscillator is stopped?
The 47k base resistor will have about 8.2V across it.
At room temperature the 2N3904 will have a typical gain of 200 so its collector current is 34.9ma and its power dissipation is only 307mW.

Yes 200 is typical but what if it's the maximum which is 300?

That's 52.3mA giving a power dissipation of 471mW, causing the transistor to rise to 119.2°C (assuming 25°C abient and a rise of 200°C/W given on the datasheet).

If it is at its max temperature then its gain is 280 and its dissipation is only 430mW. Its max rated dissipation is a lot more at 625mW so it is fine.

So the at the higher temperature gain has increased by 40%, the gain will go up to 420. Now it's passing 73.3mA, giving a power dissipation of 659.5mW causing the transistor to rise to 131.9°C which is above it's maximum rating of 125°C and given you're supposed to derate by 5mW/°C the transistor is toast!

I disagree. The pulses from the oscillator are from a low impedance emitter follower so have plenty of current to charge the coupling capacitor so that Q3 is cutoff over some of its cycle.

I doubt it because when I built a simple oscillator like this the output was less than 2Vpp or so and the emitter voltage was even lower, far too low to completely turn the transistor off.

Lots of circuits blow up if their oscillator fails, but I show above that this one just gets hot (not too hot). What about an inverter that makes mains AC from a car battery? If its oscillator stops then hundreds or thousands of DC amps flow.

Firstly the MOSFETs will turn off, unless the oscillator jams on in which case that's what fuses are for!

 

[audioguru]

Yes 200 is typical but what if it's the maximum which is 300?

That is at a collector current of only 10mA. The gain drops at higher current.

So the at the higher temperature gain has increased by 40%, the gain will go up to 420. Now it's passing 73.3mA, giving a power dissipation of 659.5mW causing the transistor to rise to 131.9°C which is above it's maximum rating of 125°C and given you're supposed to derate by 5mW/°C the transistor is toast!

It is rated to have a dissipation of 625mW in an ambient temp of 25 degrees C. So its current can be 625mW/8.8V= 71mA then its current gain would need to be 413. I don't think any will have a current gain so high at the high current, even when hot.

I doubt it because when I built a simple oscillator like this the output was less than 2Vpp or so and the emitter voltage was even lower, far too low to completely turn the transistor off.

My 'scope can't show 100MHz but the base of the output transistor needs a voltage swing of only -0.8V to turn it off. The current through the coupling capacitor needs to be only 191uA (in the 47k base resistor)to turn it off.
The oscillator transistor operates at an average current of 9mA. So the emitter resistor has an average of 1.98V across it and the emitter's voltage swing must be pretty high.

 

[Hero999]

I don't think any will have a current gain so high at the high current, even when hot.

No, most transistors won't but some might, it's the same old story, typical values on a datasheet aren't reliable especially when a transistor's gain is concerned.

My 'scope can't show 100MHz but the base of the output transistor needs a voltage swing of only -0.8V to turn it off. The current through the coupling capacitor needs to be only 191uA (in the 47k base resistor)to turn it off.
The oscillator transistor operates at an average current of 9mA. So the emitter resistor has an average of 1.98V across it and the emitter's voltage swing must be pretty high.

The voltage swing on the emitter is lower than the collector, so that 2V will easilly become 1Vpp which is only -0.5V - far too low to turn off the transistor.

Anyway, can you think of a reason why Q3 blew then?

Apart from thermal runaway there is no other possible mechanism I can think of and given the worst case scenario for gain it makes perfect sense.

 

[audioguru]

I think his Q3 is fine. It got hot (not too hot) when the oscillator stopped. Q3 will work again when the oscillator is fixed.

 

[Hero999]

Well, let's just wait and see then.

 

[bananasiong]

How to know whether the transistor is blow or not? Measure the base and the emitter with digital multimeter?
Is there anyway to check whether the oscillator is working or not? I think I should remove the oscillator and construct at another board to check the error. Is the oscillator formed by:
R6, R7, C5, C6, C7, L1 and Q2?

Add a resistor at the emitter of Q3 can prevent it from being blew? Will this affect the transmission distance?

Thanks

 

[audioguru]

With Q3 in the circuit, it can be checked by disconnecting C12 and shorting the base to the emitter of Q3. If it gets cold then it is good.
You can also remove Q3 from the circuit and measure the base-emitter and base-collector diodes for passing current when forward-biased and blocking current when reverse-biased. Also measure the collector to emitter blocking of current with the base open.

Your list of parts for the oscillator is correct but you should also include C3, C9 and C12.

The collector load for Q3 is the C13 and L2 tuned circuit which is a high impedance in parallel with the antenna which is about 75 ohms. The current gain of the 2N3904 at 100MHz is only about 5. Adding a 100 ohm emitter resistance will decrease its signal output to about 1/8. The 100 ohm resistor could be bypassed with 1000pF for much less signal loss.

Without a signal and without an emitter resistor, a typical 2N3904 for Q3 will have a collector current of 44mA and a dissipation of 384mW. Its rated max is 625mW.
Adding a 100 ohm emitter resistor will reduce Q3's current to 28.5mA, reduce its dissipation to only 175mW and the 100 ohm resistor will dissipate 81mW.

EDIT: Q3 might breakdown if the antenna is too short.

 

[bananasiong]

The antenna from the schematic is 30 inch long (Maybe it should be only 3 inches?), I used only 8cm wire.
How if I add a 56 Ohm resistor by passed by a 1000pF to reduce the power dissipated by the resistor. Can the range be increased as compared to a 100 Ohm resistor?

If I disconnect C12 from the base of Q3 and probe at the end of C12, I will see oscillation right?

Thanks

 

[bananasiong]

Hi,
i have added a 56 Ohm resistor at the emitter of Q3 and a 1nF capacitor by passing the resistor. There is still no output. I think the oscillator is not working.
Is it because of spoit components? Any other reason that cause the oscillator not to oscillate?

Thanks.

 

[Hero999]

Did you replace Q3?

Are you getting a signal from Q2?

The 56hm: resistor won't reduce the range much if it's bypassed by 1nF but it will keep Q3 cool and stable under a wider temperature range.

audioguru's concern about your antenna bieng to short is valid because the voltage across the LC circuit will be well above the supply voltage due to resonant voltage magnification if the Q is too high. The emitter resistor should reduce this problem a little but never the less try increasing your antenna's length to about 750mm.

 

[bananasiong]

Did you replace Q3?

Are you getting a signal from Q2?

The 56hm: resistor won't reduce the range much if it's bypassed by 1nF but it will keep Q3 cool and stable under a wider temperature range.

audioguru's concern about your antenna bieng to short is valid because the voltage across the LC circuit will be well above the supply voltage due to resonant voltage magnification if the Q is too high. The emitter resistor should reduce this problem a little but never the less try increasing your antenna's length to about 750mm.

Yes, i've replace a new transistor. After adding the resistor and capacitor, Q3 is not as hot as previously, still warm. But I still couldn't get any signal at the antenna side.
Increase the antenna to 750mm? That's long.. Can I fold it or make it like a spiral?

Thanks

 

[audioguru]

A quarter-wavelength at 100MHz is 750mm. The antenna on my cell phone is only 15mm long but most of its total length of about 195mm is inside, and it operates at 824MHz to 1990MHz.

The range will be much less with a shorter antenna and the output's tuned circuit will develop a high voltage that might damage the circuit.

EDIT: wave-length.

 

[bananasiong]

A half-wavelength at 100MHz is 750mm. The antenna on my cell phone is only 15mm long but most of its total length of about 195mm is inside, and it operates at 824MHz to 1990MHz.

The range will be much less with a shorter antenna and the output's tuned circuit will develop a high voltage that might damage the circuit.

750mm is a quater right?
You've said that it is not oscillating, si there anything to do with the antenna? And, it worked in the beginning.
Any spoilt component at the oscillator will cause the oscillator mulfunction right? Any other reason?
And for the antenna again, can I fold it or make spiral?

Thanks

 

[audioguru]

You are correct, 750mm is a quarter-wavelength, not a half-wavelength.
Car radio antennas and the rabbit ears on my stereo are 750mm.
It won't be 750mm if it is folded or in a spiral so will have less range.

 

[Hero999]

Have you checked to see if there's any oscillation from Q2?