Diode Quest

[mstechca]

Your quest isn't over :lol: (or maybe it is)

Anyways, I'm happy to inform some of you that the varactor trick has partially worked with my receiver.

Here's why:

The resistors I used for each bit from the 4-bit counter are as follows:

1.2K
2.7K
4.7K
6.8K

I deliberately went over 1K for each value and tried to exponentially increase each resistor value based on what remaining resistance values I had on hand.

When I used my computer (I wrote a Qbasic program) to calculate all possible resistances based on each counter state, I was happy to see that the voltage increments were small enough (about 6%) if the counter went through each bit combination in order.

So that I believe is not too much of a percent. My supply voltage is switchable between 6 and 9V because I want to make my circuit work with a 9V battery for as long as possible. (please don't tell me to buy duracell or energizer, or some other premium brand battery)

Audioguru made a transmitter circuit showing a diode and capacitor in series. I made it so that it appeared compatible with my receiver.

Here were the results I got when I tried these values for diodes and the coupling capacitor:

test #1
Diode: 1N4007
Capacitor: 150pF
Additional capacitance added: None
Result: No signal

test #2
Diode: 1N914B
Capacitor: 150pF
Additional capacitance added: None
Result: No signal

test #3
Diode: 1N914B
Capacitor: 22pF
Additional capacitance added: None
Result: No signal

test #4
Diode: 1N914B
Capacitor: 22pF
Additional capacitance added: 4.7pF
Result: Signal present only on two values (not bits).


It makes me think that a diode can only produce a small capacitance.

When I tested my circuit with just a standard LC circuit (no varicap add-on), and when C is 4.7pF, I got white noise in, which is good for now.

I also added a 50% voltage divider. (two 680K resistors), because if I had all high bits or all low bits, I can't see how the circuit would work without them. The center of the divider as well as the output voltage from my counter is connected to the center of the varicap circuit. (between the reverse biased diode and capacitor).

I begin to think that changing the diode to one compatible with higher frequencies might do the trick. What do you think?

My objective is to get any signal in when I press the button. When I press it again, I want the signal (or at least the frequency) to change.

 

[audioguru]

The resistors I used for each bit from the 4-bit counter are as follows:

1.2K
2.7K
4.7K
6.8K

I wish you would stay in the same thread or post a schematic of what you have done. I can't remember your receiver so I don't know where you connected the varicap.

Your digital to analog converter resistors at the outputs of the counter have values far too low. As explained in the other thread, the output resistance of the counter requires resistor values 10 times to 100 times higher for accuracy. The resistance of the varicap diode you are controlling is infinite since it is supposed to be reverse biased.

I also added a 50% voltage divider. (two 680K resistors), because if I had all high bits or all low bits, I can't see how the circuit would work without them.

I don't know why you added a voltage divider of two 680k resistors and don't know where you connected it. A 1N4007 diode was shown to have a capacitance of 5.5pF with 9V reverse bias, 10.5pF with 3V reverse bias and 19pF with only 1V reverse bias.

The center of the divider as well as the output voltage from my counter is connected to the center of the varicap circuit. (between the reverse biased diode and capacitor).

That is probably your problem. In the other thread I showed a 100k coupling resistor between the controlling voltage source (the resistors at the output of the counter) and the varicap, for RF isolation. Your resistors were RF shorting the varicap.

I begin to think that changing the diode to one compatible with higher frequencies might do the trick. What do you think?

Did your receiver circuit ever work with a 10pF cap directly across the coil? Then it should also work the same with the 1N4007 varicap in series with the 150pF cap and both in series from the coil to ground. Of course in order for the varicap and 150pF cap to be across the coil then the supply must have a 1000pF ceramic bypass cap nearby. Also in order for the varicap to have 10pF then its reverse voltage fed to it through a 100k isolation resistor must be about 3V for a 1N4007.

The capacitance of a reverse biased 1N914 diode is far too low for your circuit.

 

[mstechca]

Here it is.

The circuit in white is my regen detector that I will consistently use, because IT WORKS.

The circuit in yellow is my add-on. I took audioguru's version of his varactor tuned transmitter and modified it slightly to suit my circuit, because my (LC tank) inductor is grounded. Like he said, I didn't use a limiting resistor between the voltage source and the junction between the capacitor and the diode.

The bit #'s are the bits on the counter output. Bit 1 is the first bit (LSB?), the one that turns on first when I activate the clock on the counter.

After pressing the button about 14 times (with a 4.7pF capacitor tied to the LC inductor which isn't shown), I got a signal.

If I didn't use a divider network with the 680K resistors, then I would have a problem when all bits are on or off.

Assume there were no 680K resistors:

If they are all on, all resistors will be internally (according to the 74HC393) tied to VCC, and there would be no resistor tied to ground.

If they are all off, all resistors will be internally (according to the 74HC393) tied to ground, and there would be no resistor tied to VCC.

The goal is to provide a voltage, and my way of doing that is to always use a 2-resistor voltage divider.

and the R-2R ladder is something I will pass up on, because I see no benefit with it.

 

[bloki]

For those experimenting with ordinary diodes to be used as varicaps this site may be interesting:
**broken link removed**

 

[audioguru]

Hi Stubborn,
1) The varactor diode is supposed to be reverse biased.
2) Without a 100k isolating resistor, the output of the resistors of the counter were a 680 ohm RF short across the tuned circuit.
3) Where's your supply bypass cap?
4) Use a diode that has enough capacitance to do the job.
5) There isn't a requirement to waste up to 9mA and lose accuracy in very low value resistors from the counter.

I won't comment on the rest of the circuit since you say it works.
Try it like this:

 

[bloki]

Althought diode 1N4007 has large enough capacitance ratio compared with 1N914 in this circuit it will be impossible to tune receiver through all FM band. I think that left pin of 150pF capacitor has to be connected to collector instead to emitter.

 

[Nigel Goodwin]

Althought diode 1N4007 has large enough capacitance ratio compared with 1N914 in this circuit it will be impossible to tune receiver through all FM band. I think that left pin of 150pF capacitor has to be connected to collector instead to emitter.

I agree, but Audioguru specifically said that he was leaving the original circuit alone, as it's 'supposed' to work? - doesn't look much like a worker to me?.

 

[audioguru]

Hi Bloki,
If the 150pF cap was connected to the collector then the circuit would be a transmitter. MStechca is using his circuit as a receiver. Its 1mH tuning coil seems awfully large to tune the FM band.

 

[bloki]

Collector coil is only 100nH and I do not agree the receiver will become a transmitter. To tune collector tank to 100MHz you need a total parallel capacitance of 25pF. Trick of those regenerative receivers is to set a working point that it does not oscillate continiously. Quenching oscillation at his peaks is driving it to oscillate.

 

[audioguru]

Collector coil is only 100nH and I do not agree the receiver will become a transmitter. To tune collector tank to 100MHz you need a total parallel capacitance of 25pF.

Most of these simple FM transmitters tune the FM band with a 0.1uH inductor and 5pF to 10pF of stray capacitance in parallel with a trimmer capacitor of 5pF to 30pF.

Trick of those regenerative receivers is to set a working point that it does not oscillate continiously. Quenching oscillation at his peaks is driving it to oscillate.

MStechca used to have super-regen receivers with self-quenching at an ultrasonic frequency but now he made this one that probably doesn't even oscillate.

It is odd that it is a receiver without an antenna.

 

[mstechca]

I found a mistake in my upload. both inductors should be 0.1uH.

and yes, I will up my resistance values at the counter.

and I have a comment for Audioguru:

Isn't your "corrected" varactor equivalent to LC in parallel?

Mine should be equivalent to a capacitor connected between emitter and VCC.

I am still deciding on the bypass capacitor. Experimentation so far is a good answer to me.

and why isn't my diode reverse-biased when the cathode of my diode is connected to VCC?

 

[audioguru]

I found a mistake in my upload. both inductors should be 0.1uH.

Good.

and yes, I will up my resistance values at the counter.

Get rid of your 680k resistors and add the 100k RF isolation resistor.

Isn't your "corrected" varactor equivalent to LC in parallel?

Of course. Then it tunes a station and rejects others.

Mine should be equivalent to a capacitor connected between emitter and VCC.

I am still deciding on the bypass capacitor. Experimentation so far is a good answer to me.

You still don't understand the function of a supply bypass capacitor.
With the tuning capacitance connected to VCC, then the circuit would work only if the battery is brand new so has a low internal resistance and the wires to the battery are very short so they have low inductance.
With a supply bypass capacitor to make the RF potential of VCC and ground the same, then the tuning capacitance can be connected to ground or to VCC and even with a weak battery it will work the same.

why isn't my diode reverse-biased when the cathode of my diode is connected to VCC?

Sorry, I thought the anode was at VCC. The diode should be connected to ground like in every other circuit.

With the diode connected to VCC and with the same supply voltage for the counter, the tuning is backwards. When the counter is reset then the resistors will have zero voltage output and the capacitance of the diode is at its minimum and the RF frequency is at its highest.

With the diode connected to ground like I show it, when the counter is reset then the radio will tune to its lowest frequency. Then clocking will increase the frequency. Isn't that what you want?

You are using an ordinary diode to replace a varactor which has a much wider tuning range. The tuning range with an ordinary diode isn't enough, unless you increase the supply voltage of the counter for the diode to have a wider tuning range. With the diode connected to ground like I show it, you can increase the supply voltage of the counter if you use a CD4024 counter with its much higher VCC rating. :lol:

 

[mstechca]

With a supply bypass capacitor to make the RF potential of VCC and ground the same, then the tuning capacitance can be connected to ground or to VCC and even with a weak battery it will work the same.

Then I shall use a few uF's for the capacitor, just so that the AC-VCC and AC-GND are the same.
Notice how I called it AC-VCC and AC-GND?
I did that because the VCC coming from a battery is NOT the same as battery's ground. but in AC, things change.

Get rid of your 680k resistors and add the 100k RF isolation resistor.
...

With the diode connected to VCC and with the same supply voltage for the counter, the tuning is backwards. When the counter is reset then the resistors will have zero voltage output and the capacitance of the diode is at its minimum and the RF frequency is at its highest.

With the diode connected to ground like I show it, when the counter is reset then the radio will tune to its lowest frequency. Then clocking will increase the frequency. Isn't that what you want?

I don't care which way the tuning is. All I care about now is to have each bit combination produce a frequency. (or at least most of them).

btw, my circuit I am modifying is on PCB.

You are using an ordinary diode to replace a varactor which has a much wider tuning range. The tuning range with an ordinary diode isn't enough, unless you increase the supply voltage of the counter for the diode to have a wider tuning range. With the diode connected to ground like I show it, you can increase the supply voltage of the counter if you use a CD4024 counter with its much higher VCC rating.

Diodes are cheaper. In fact, I am willing to switch to zener if it is necessary. I have used 1N5221's before (tiny red diodes) in my very first superregen, but they tend to break down easy (even on the smallest circuit mistake).

If I were to replace the diode with a diode with a higher tuning range, should I go for the 1N5221 or some other diode?

I don't want to blow my money on a varicap and find out later that it is useless.

Also, my supply voltage must be switchable between 6 and 9V, because I want my receiver to work on an almost dead battery. and so far, my receiver works from 9V to 6V. The only thing I might have to do is alter the jumper setting in my circuit.

so your idea of higher VCC isn't going to go well with me or my circuit.

 

[audioguru]

With a supply bypass capacitor to make the RF potential of VCC and ground the same, then the tuning capacitance can be connected to ground or to VCC and even with a weak battery it will work the same.

Then I shall use a few uF's for the capacitor, just so that the AC-VCC and AC-GND are the same.

NO, NO, NO! A few uF's cap is useless at these high RF frequencies due to its inductance. You need a 1000pF ceramic disk cap with very short wires.
Use an additional 100uF to 1000uF to bypass audio frequencies.

You are using an ordinary diode to replace a varactor which has a much wider tuning range. The tuning range with an ordinary diode isn't enough, unless you increase the supply voltage of the counter for the diode to have a wider tuning range. With the diode connected to ground like I show it, you can increase the supply voltage of the counter if you use a CD4024 counter with its much higher VCC rating.

Diodes are cheaper. In fact, I am willing to switch to zener if it is necessary. I have used 1N5221's before (tiny red diodes) in my very first superregen, but they tend to break down easy (even on the smallest circuit mistake).

If I were to replace the diode with a diode with a higher tuning range, should I go for the 1N5221 or some other diode?

You didn't look at the datasheet for a 1N5221 3.3V/20mA zener diode.
Low voltage zener diodes are lousy for anything. With a reverse voltage of only 2.6V, it typically leaks 5mA. Its capacitance is far too high and its range isn't enough for your application.

You should use a proper varactor diode for tuning. If you can't afford mail-order then go to the dollar store and buy a "scanning FM radio" for $1.00. It is a lousy radio but it comes with stereo earphones (the radio is mono) and two cheap AAA battery cells. Its varactor diode tunes the entire FM band with less than 3V. I didn't look but its parts are probably tiny surface-mount.

Also, my supply voltage must be switchable between 6 and 9V, because I want my receiver to work on an almost dead battery. and so far, my receiver works from 9V to 6V. The only thing I might have to do is alter the jumper setting in my circuit.

Your 74HC393 counter has an absolute max supply voltage of only 7V. It will blow-up if powered from a 9V battery. It would need a low-dropout 5V regulator then the tuning range would be very small with an ordinary diode.

 

[Nigel Goodwin]

If you can't afford mail-order then go to the dollar store and buy a "scanning FM radio" for $1.00.

Rather pointless taking a good working radio to pieces for it's parts, and then use them in an inferior radio?.

Reminds me of the UK TV show "Blue Peter", where they crush biscuits up to use to make more biscuits? - strange idea?.

 

[Styx]

Reminds me of the UK TV show "Blue Peter", where they crush biscuits up to use to make more biscuits? - strange idea?.

what the whole idea of Blue Peter strange or just the making something out of the same thing

 

[audioguru]

Rather pointless taking a good working radio to pieces for it's parts, and then use them in an inferior radio? - strange idea?.

It's a lousy radio with a 70kHz IF done with an opamp active filter, so it doesn't have much image rejection.

You should see the "spying amplified-ear" product my son got at the dollar store for only a buck. It has lotsa good stuff inside including stereo earphones and many spare button alkaline battery cells.
It's a lousy amplifier without enough gain but its sound is clear and noise-free. :lol:

 

[Nigel Goodwin]

It's a lousy radio with a 70kHz IF done with an opamp active filter, so it doesn't have much image rejection.

But it's still going to wipe the floor with a super-regen, particularly one that's been altered to stop it working correctly!.

 

[mstechca]

NO, NO, NO! A few uF's cap is useless at these high RF frequencies due to its inductance. You need a 1000pF ceramic disk cap with very short wires.
Use an additional 100uF to 1000uF to bypass audio frequencies.

How do I calculate the capacitor's inductance?
I can calculate the AC resistance of a capacitor based on a frequency.

wait a sec... you say that a few uF's make too much inductance. Why do you suggest adding 100uF to 1000uF?

Your 74HC393 counter has an absolute max supply voltage of only 7V. It will blow-up if powered from a 9V battery. It would need a low-dropout 5V regulator then the tuning range would be very small with an ordinary diode.

I originally worked with the CD4024 before I even introduced a varactor circuit into my super regen. Isn't the CD4024 slower than the 74HC393?
HC does mean high-speed cmos.

I hope that my 74HC393 allows me to go to 9V and violates datasheet standards.

 

[DigiTan]

wait a sec... you say that a few uF's make too much inductance. Why do you suggest adding 100uF to 1000uF?

That's obviously a typo.