How to use a Si diode as AM detector

[BkraM]

Hello,

I'm trying to build an AM radio. All circuits i have found use an Ge diode as detector as it has a low voltage barrier. I'don't have access to a Ge diode only to a Si. I'm wondering if it is possible to adjust the circuit so a Si diode wil do the trick. Would an amplification stage of the RF signal with a BJT after the tank circuit, prior to the demodulation stage work? Is there another trick i could use to make it work?

I'm fairly new to electronics, so simple solutions are the best.

Thanks,

 

[ronsimpson]

With a long antenna and a strong radio station the Si diode will work.
You will need a strong station with a Ge diode.
These radios are not good for weak signals.

 

[crutschow]

A Schottky diode may work better since it has a lower forward drop. Here is a source for Ge diodes.

 

[BkraM]

Thanks for the replies. Is a amplification of the rf signal no option so a Si diode can be used?

 

[crutschow]

Amplification is certainly an option. That's basically what standard Superhet AM radios do. They have constant frequency IF amplifier/filter stages to provide high amplification and tight filtering of the incoming signal prior to detection by the diode.

Amplifying an AM signal in the 1MHz frequency range is quite simple and can even be done with a low noise, high frequency op amp.

 

[ronsimpson]

Thanks for the replies. Is a amplification of the rf signal no option so a Si diode can be used?

Yes, all (most) radios have a amplifier or many amplifiers before the detector.

 

[BkraM]

would a simple class a-amplifier work?

 

[crutschow]

would a simple class a-amplifier work?

Yes, if it's properly designed.

 

[Willen]

Hi BkraM,
Seeing your name I guessed you are around india, isn't it? And I felt you have components accessibility problem like I have. Then I can suggest you few amazing tricks for you and it was experimented by me.

I got double audio then Ge diode when I replaced it with 1N5819. Yes it is Schottky Power diode of 1A. It had 170mV forward voltage drop only (very lower than its datasheet). If you cannot find this diode in shop then don't worry! You can find this diode in local made (Chinese) cell phone charger (cheap wall wart) power supply. Open the charger and see DC output series diode, it looks like 1N4007 (same packaging). I tried 25 pcs of Ge diode but sound was almost half than this 1N5819 diode.

I tried another few amazing tricks- use any condenser Mic instead of Ge diode (or use its FET by shorting Gate-Drain for little high audio). For detail, here is my tiny page- www.techlib.com/area_50/Readers/Rajkumar/index.html

In't it amazing? And another amazing thing is I am using normal earpiece of cell phone which is little more sensitive than few others.

Another amazing thing is I am receiving thousands of miles far stations (overloaded) from China and India... So I used fixed capacitor.

I tried few Si diodes and sound was empty for my ears without external amplifier. Later I added an tiny amplifier and got again almost nothing.

I am using my entire metal roof as an antenna which is able to glow little a VERY bright red LED.

Do not use transistor based amplifier because even if you used two transistors, sound is still low to drive 8 ohms speaker. Here is an amplifier designed by Charles Wenzel (admin of www.techlib.com/electronics) which is an another amazing device. It has TL431 shunt regulator chip which looks like a small transistor. It can drive 8 ohms normal speakers of radio and sound if room filling. I am also using it.

(I am talking about Crystal Radio AM detector)

 

[BkraM]

Willen, Thanks for your reply.
I'm not from india, a local european
As my name was allready in use i reversed it.

Using your entire roof should give you a extremely good reception. I have a steel rim aound mine, could try the same!

I could have acces to a Ge diode, it's just that i suspect i could do without, and learn something on the way.
I think i go for an amplification stage first, and if not succesfull try one of your sugestions.

During the day i was wondering if it would be possible to bias the Si diode with say, 0.5 V so it is allready close to the voltage barrier?
In that case a pre-amplification would be unnescesarry.

Thanks,

Mark

 

[crutschow]

..................................

During the day i was wondering if it would be possible to bias the Si diode with say, 0.5 V so it is allready close to the voltage barrier?
In that case a pre-amplification would be unnescesarry.

Yes, biasing the diode with a small DC voltage through a large resistor should improve the sensitivity.

 

[Willen]

...then this could be a helpful page-http://www.techlib.com/electronics/crystal.html (He tested almost all types of diodes and made a chart here of result.)

Do not forget to update the thread after you experiment with schematics. I want to learn too from use of Si diode in AM Rx. Some Si diodes are very slow, may be 1N4148 are little faster than power diode. I don't know exactly..

 

[ronsimpson]

Some Si diodes are very slow, may be 1N4148 are little faster than power diode. I don't know exactly..

Power diodes probably have too much capacitance. They likely are too slow.
Use small diodes.

i was wondering if it would be possible to bias the Si diode with say, 0.5 V so it is allready close to the voltage barrier?
In that case a pre-amplification would be unnescesarry.

I found biasing a diode helps bring the week signals but they are hard to hear. It effectively may give you a 0.1V diode.

 

[BkraM]

Got myself a Ge (1N60), Si (1N4148) and Schottky diode (1N5819).

All thee do work, and allow demodulation of local stations.
The Si diode can clearly deal with fewer stations (I found only 1). I would not recommend itwhen building a AM radio for the first time as it will take a lot of tuning to find out whether it works or not.
The Schottky en Ge diode deliver approx the same results, although with the Schottky diode the tank circuit needs to tune to a higher frequency to receive the same station (3000 kHz instead of 1000 kHz).

The selectivity of the circuit at this stage is not very good, stations appear to overlap.
Should i increase selectivity by adding a resistor in the tank circuit (RLC), at the cost of some signal. Or is there a better approach for this?

Thanks,

Mark

 

[JimB]

The selectivity of the circuit at this stage is not very good, stations appear to overlap.

That is to be expected with a crystal set with a single tuned circuit.

Should i increase selectivity by adding a resistor in the tank circuit

Why do you think that adding a resistor will increase the selectivity?
The quick answer is that it will NOT increase selectivity, it can only make it worse.

You could try connecting the diode to a tapping point half way or a quarter way along the coil.
That way the effective resitance of the diode and associated components load the tuned circuit will be reduced, so increasing the Q of the circuit and hence increase the selectivity.
Do not expect miracles from a circuit like this, it is not going to happen.

Or is there a better approach for this?

Look at this video of a one valve receiver...

And a follow-up where he runs through the circuit diagram...

JimB

 

[BkraM]

JimB,

Thanks for your reply. I'll have a look at your suggested videos.

I figured adding a resistor would reduce the amplitude of the reception, but would narrow down the bandwidth.
Requiring more amplification, but giving a better selectivity.

https://www.mathworks.nl/help/control/ug/analyzing-the-response-of-an-rlc-circuit.html

What am i missing here?

Thanks,

Mark

 

[JimB]

OK I see where this is coming from.

Try it if you like, but I suggest that you will be somewhat disappointed.

The detector circuit will still load the tuned circuit, leading to poor selectivity.
The mathcad analysis takes no account of a load on the output.

JimB

 

[MrAl]

JimB,

Thanks for your reply. I'll have a look at your suggested videos.

I figured adding a resistor would reduce the amplitude of the reception, but would narrow down the bandwidth.
Requiring more amplification, but giving a better selectivity.

https://www.mathworks.nl/help/control/ug/analyzing-the-response-of-an-rlc-circuit.html

What am i missing here?

Thanks,

Mark

Hi,

To get better selectivity you need a bandpass filter, one or more. Using one isnt too bad, but using two would work even better. The filters are made out of inductors and capacitors, the inductors can be air core that you wind yourself. One section for example might be a pi section.

The bandpass filter allows some frequencies to pass while rejecting others, and the sharpness can be changed depending on the values of the components. The tradeoff is higher selectivity is hard to tune to the right station and it could drift too, while lower selectivity allows easier tuning but might not be selective enough to block out other channels that are close to the one you are tuning to.

Another method is to use frequency addition/subtraction. That's where you build a local oscillator and mix the output with the signal input, that creates a lower frequency but also allows some tuning. The lower frequency is then used to get the audio. This is a little more complicated because you need a stable local oscillator and a mixer as well as filters.

 

[audioguru]

When crystal radios were used a long time ago there was only one or two AM radio stations in town. Now there are many radio stations in my town and many more AM radio stations in towns all around me. No wonder the stations overlap. You need a REAL radio circuit today to have good selectivity.

 

[BkraM]

So a regenerative receiver should be the way forward for more selectivity.
Thanks for the tips, i'll give it a go.

(I tried adding a resistor before the LC circuit, didn't do much but kill the signal)