Nice. I love those old large-case crystals, they look so much better than the modern HC-49 ones. But.... Could you please show how to make a transistor-based Hartley/Colpitts oscillator for those small 32,768KHz watch crystals? I've tried a couple of times but never got them oscillating properly, I guess actually I should say "at all" ;-) .
Interesting devices... quartz crystals. There's almost an equivalent mechanical construction. With a few of the measured crystal parameters, ohms law, and some complex number algebra, you can work out the crystal's series resonant frequency, parallel resonant frequency, and impedance at resonance.
Can you manage to get a "simple" Colpitts to oscillate at any frequency between 200 kHz and 24 MHz without having to alter components? I've been trying to make a "simple" crystal tester to prove working or not, any crystal picked up or fund at junk sales. More often than not, old crystals are burnt out due to having too much crystal current passed through them, but it would be nice to have a go/no-go tester. I have managed a tester that has an eBay counter PCB that will count the output of the oscillator, but it's finding an oscillator circuit with wide enough characteristics to work with a crystal between 200 kHz and 24 MHz. Any hints anyone - A circuit is preferable to a description, please. As it is, I am using three oscillators with their outputs parallel with three sets of inputs where crystals can be poked in and one lives in hope that one of the three oscillators will spring into life when the unknown DUT is inserted, if neither of the three oscillators work, the conclusion is that the crystal is dead. Back at home testing crystals, about 1 crystal in 20 tested will not oscillate, but the crystal is good and works, this is usually the low-frequency crystals that are below 1 MHz. Thanks, KW.
@@IMSAIGuy Yes, that's virtually the same as the multi-oscillator input I have on my crystal tester, but no LED indicator, only the outputs of all three oscillators go to the input of a cheap eBay counter PCB, if any of the three oscillators work with the crystal on their inputs, the counter will indicate the crystal op freq. showing a working or not working crystal. The multi oscillators allow for a range of frequencies of crystals to test, usually 200 kHz - 500 kHz, then 500 kHz to 6 MHz, and then up to 24 MHz. Anything higher than 24 MHz is (in my case) an overtone crystal and not fundamental.
Instead of trying to get the crystal to oscillate, sweep it using NanoVNA. Given the price of the latter, you get a whole lot of performance very cheaply. A crystal response sweep will indicate resonant frequencies of various modes and their Q factor - pretty much everything you’d need to characterize a crystal. As for crystal oscillators that can take a wide range of crystals: the trick is to inject wideband noise on the input of the crystal. Just high enough to get the sense amp move enough on the output to make the Schmitt output switch. One or two edges later the filtered crystal’s response overwhelms the noise source. In order not to overdrive the crystals, the drive side needs a current sensor (a shunt), an amplifier for it, and an AGC driven from it to limit the drive power to values that won’t over stress the crystal. All in all getting one device oscillating from tens of kHZ to tens of MHz requires a bit of finesse if you just want to stick a crystal in and have it oscillate. Never mind that the crystal may not oscillate at the frequency you want. Eg 3rd overtone crystals may have a strong fundamental resonance that a wideband oscillator may latch into.
@@absurdengineering Unfortunately carrying a VNA and cables is not really practical when you are at a junk sale looking at crystals in a box that someone is selling. The idea of a small box with three oscillators who's outputs are paralleled and connected to a cheap eBay LED counter is ideal because it is all in one small box, and it'll not only tell you if the crystal works, but also what its frequency is.
What I don't understand is, how these crystals are 'excited' to make them oscillate. I was always under the impression you fed voltage in on one pin, and a pulse train came out of the other pin. That doesn't seem to be the case. I'm missing something. Can you do a video on the theory of how they work? That would be great! Thank you 🙂
The oscillator circuits feeds energy to the crystal. This energy supplies one of the crystal’s mechanical oscillation modes. The piezoelectric action acts as a transducer between the electrical and mechanical domains. The crystal element is a mechanical equivalent of an LC tank with a Q so high that it’s impossible to obtain using bulk passive components in the same volume.
At 2:00, I may have missed something, but i think you demonstrated putting an inductor in parallel with the 680k resistor. Wouldn't that short out the bias to the base?
@@IMSAIGuy -- I think I'm confused. Here's a circuit of a colpitts oscillator with a capacitor in series with the coil to prevent it from shorting out the DC bias... ua-cam.com/video/TSKq5l7uuz4/v-deo.htmlsi=2wMvvdYsijFNO-xK
Very informative video. I have tried it with 27MHZ crystal , it is working but having frequency modulation issue . Can you give any clues but the cause of issue?
this mean as well this quartz would be good as a antenna for 1.001Mhz, would be nice to experiment by removing the top cap (and short the lead) and look at directivity antenna
It would be a good ultrasonic transducer but as an antenna its electrodes are way too tiny. It’d act as an antenna tuning tank if you attached a piece of a very long wire to it. 1MHz wavelength is 300m.
Actually this common collector Colpitts design is about the worst for starting with oscillators. It's not easy to understand even to many professionals, difficult to design properly, often just doesn't work, bias of the transistor affects the frequency, has large distortion, suffers heavily from pulling etc. Even its commin base variant is much easier to understand and design.
Am I the only one who is disappointed by the fact that the crystal is 0,03 percent off (0.03 percent in the US). I expect the crystal to be exact, with nine zero's behind the local decimal sign.
The oscilloscope is also not a perfect device to measure frequency and might be off by a bit. They usually get better if you have more waves on screen but then you won't see the details of the waveform as nicely as in the video. If you want to be sure about the frequency a frequency counter is much more accurate as it counts over a much larger timespan
@@IMSAIGuy I have been thinking, I love low prices and I know that higher quality certified calibrated stuff adds cents to every Dollar spent. So, I was thinking, if not one but two crystals are used and both of them have a known error in the frequency, but since you are using two different crystals the second can be used to correct the first. Or is this to far off?
Nice. I love those old large-case crystals, they look so much better than the modern HC-49 ones. But.... Could you please show how to make a transistor-based Hartley/Colpitts oscillator for those small 32,768KHz watch crystals? I've tried a couple of times but never got them oscillating properly, I guess actually I should say "at all" ;-) .
yes this would be nice :)
use a 74C14 like this: ua-cam.com/video/FExN8ZaufCI/v-deo.html
@@IMSAIGuy That's cheating ;). Can't have an IC as the oscillator if I ever get a round to make a fully transistorised wall-clock....
@@MatsEngstrom electronics.stackexchange.com/questions/299392/low-frequency-pierce-oscillator-with-transistor
ua-cam.com/video/jyZQdZF6KIE/v-deo.html
I got a couple old tv boards yesterday. Complete with delay lines and huge Crystals. Also has a few of those stacked capacitors on it.
Could u make a video explaining how to design it? The component values and so on
got a circuit from a book, tried larger capacitors until it worked
Darn! I was going to do a video on a Colpitts oscillator for testing crystals and you beat me to it. Now people will think I copied you!
also: ua-cam.com/video/s7rcZAto1BU/v-deo.html
Interesting devices... quartz crystals. There's almost an equivalent mechanical construction. With a few of the measured crystal parameters, ohms law, and some complex number algebra, you can work out the crystal's series resonant frequency, parallel resonant frequency, and impedance at resonance.
Can you manage to get a "simple" Colpitts to oscillate at any frequency between 200 kHz and 24 MHz without having to alter components?
I've been trying to make a "simple" crystal tester to prove working or not, any crystal picked up or fund at junk sales.
More often than not, old crystals are burnt out due to having too much crystal current passed through them, but it would be nice to have a go/no-go tester.
I have managed a tester that has an eBay counter PCB that will count the output of the oscillator, but it's finding an oscillator circuit with wide enough characteristics to work with a crystal between 200 kHz and 24 MHz.
Any hints anyone - A circuit is preferable to a description, please.
As it is, I am using three oscillators with their outputs parallel with three sets of inputs where crystals can be poked in and one lives in hope that one of the three oscillators will spring into life when the unknown DUT is inserted, if neither of the three oscillators work, the conclusion is that the crystal is dead. Back at home testing crystals, about 1 crystal in 20 tested will not oscillate, but the crystal is good and works, this is usually the low-frequency crystals that are below 1 MHz.
Thanks,
KW.
ua-cam.com/video/s7rcZAto1BU/v-deo.html
it won't cover your entire range without some component changes
@@IMSAIGuy Yes, that's virtually the same as the multi-oscillator input I have on my crystal tester, but no LED indicator, only the outputs of all three oscillators go to the input of a cheap eBay counter PCB, if any of the three oscillators work with the crystal on their inputs, the counter will indicate the crystal op freq. showing a working or not working crystal.
The multi oscillators allow for a range of frequencies of crystals to test, usually 200 kHz - 500 kHz, then 500 kHz to 6 MHz, and then up to 24 MHz. Anything higher than 24 MHz is (in my case) an overtone crystal and not fundamental.
Instead of trying to get the crystal to oscillate, sweep it using NanoVNA. Given the price of the latter, you get a whole lot of performance very cheaply. A crystal response sweep will indicate resonant frequencies of various modes and their Q factor - pretty much everything you’d need to characterize a crystal.
As for crystal oscillators that can take a wide range of crystals: the trick is to inject wideband noise on the input of the crystal. Just high enough to get the sense amp move enough on the output to make the Schmitt output switch. One or two edges later the filtered crystal’s response overwhelms the noise source. In order not to overdrive the crystals, the drive side needs a current sensor (a shunt), an amplifier for it, and an AGC driven from it to limit the drive power to values that won’t over stress the crystal. All in all getting one device oscillating from tens of kHZ to tens of MHz requires a bit of finesse if you just want to stick a crystal in and have it oscillate. Never mind that the crystal may not oscillate at the frequency you want. Eg 3rd overtone crystals may have a strong fundamental resonance that a wideband oscillator may latch into.
@@absurdengineering Unfortunately carrying a VNA and cables is not really practical when you are at a junk sale looking at crystals in a box that someone is selling.
The idea of a small box with three oscillators who's outputs are paralleled and connected to a cheap eBay LED counter is ideal because it is all in one small box, and it'll not only tell you if the crystal works, but also what its frequency is.
english book mentioned about this oscillator in the past
this is a very old circuit originally using vacuum tubes
Thank you for the info. 👍
What I don't understand is, how these crystals are 'excited' to make them oscillate. I was always under the impression you fed voltage in on one pin, and a pulse train came out of the other pin. That doesn't seem to be the case. I'm missing something. Can you do a video on the theory of how they work? That would be great! Thank you 🙂
you don't excite them. they are a simple LC resonant circuit.
it acts like an LC circuit. most crystals have multiple resonant modes, so you will need an external LC circuit to select the correct overtone.
The oscillator circuits feeds energy to the crystal. This energy supplies one of the crystal’s mechanical oscillation modes. The piezoelectric action acts as a transducer between the electrical and mechanical domains. The crystal element is a mechanical equivalent of an LC tank with a Q so high that it’s impossible to obtain using bulk passive components in the same volume.
At 2:00, I may have missed something, but i think you demonstrated putting an inductor in parallel with the 680k resistor. Wouldn't that short out the bias to the base?
read up on inductor reactance. at that frequency it acts as a resistor.
Is there some capacitor in series with the inductor so as not to short out the DC bias thru the inductor?
@@IMSAIGuy -- I think I'm confused. Here's a circuit of a colpitts oscillator with a capacitor in series with the coil to prevent it from shorting out the DC bias... ua-cam.com/video/TSKq5l7uuz4/v-deo.htmlsi=2wMvvdYsijFNO-xK
@@Music-zh7fm I see it both ways, there is no one way
I had a Radio Shack crystal calibrator kit with a 100khz crystal that was bigger than those 😉
Fascinating tutorial...cheers !
Very informative video. I have tried it with 27MHZ crystal , it is working but having frequency modulation issue . Can you give any clues but the cause of issue?
that might be a third harmonic crystal you are using
If you wanted to shove it through a Schmitt trigger at 50% duty... Would you just have to adjust the transistor bias?
this mean as well this quartz would be good as a antenna for 1.001Mhz, would be nice to experiment by removing the top cap (and short the lead) and look at directivity antenna
It would be a good ultrasonic transducer but as an antenna its electrodes are way too tiny. It’d act as an antenna tuning tank if you attached a piece of a very long wire to it. 1MHz wavelength is 300m.
Whats the name of the book if you dont mind sharing thanks
ARRL Handbook
Nice video
Name of the book please you are referring
this is the ARRL Handbook ua-cam.com/video/OBNl-6nFyL0/v-deo.html
Thanks 🤩
Does a crystal take ac or dc current
it is a passive component, similar to an inductor and capacitor in parallel
To bad you don't have matched capacitors.
: )
I'm using a sine wave osci and it uses 2 same value capacitors.
What book is this ?
ua-cam.com/video/u4DOKTi3Wdo/v-deo.html
Actually this common collector Colpitts design is about the worst for starting with oscillators.
It's not easy to understand even to many professionals, difficult to design properly, often just doesn't work, bias of the transistor affects the frequency, has large distortion, suffers heavily from pulling etc.
Even its commin base variant is much easier to understand and design.
Am I the only one who is disappointed by the fact that the crystal is 0,03 percent off (0.03 percent in the US). I expect the crystal to be exact, with nine zero's behind the local decimal sign.
you pay for exactness. you can pay for xtals that are better. this was used as a simple microprocessor clock so no need to be very exact
to be really good you have to control the temperature: ua-cam.com/video/h-KXTXe9U8s/v-deo.html
ua-cam.com/video/EiAsVbYYdZ0/v-deo.html
The oscilloscope is also not a perfect device to measure frequency and might be off by a bit. They usually get better if you have more waves on screen but then you won't see the details of the waveform as nicely as in the video. If you want to be sure about the frequency a frequency counter is much more accurate as it counts over a much larger timespan
@@IMSAIGuy I have been thinking, I love low prices and I know that higher quality certified calibrated stuff adds cents to every Dollar spent. So, I was thinking, if not one but two crystals are used and both of them have a known error in the frequency, but since you are using two different crystals the second can be used to correct the first.
Or is this to far off?
@@vanhetgoor no, the best way that it used, is to add a variable capacitor into the circuit, you can tune the xtal to the correct freq.
Perhaps you should start your video's, by explaining the what and why of your video