Here's another trick I used in one of my rigs, chirp-free keying by means of a counter.
In previous entries, it was indicated that counters could be used to reach a certain frequency, in particular in the LF and MF bands.
Here's an example, well, this example actually reflects the design of my 600m keyed exciter.
To reach 500kHz the signal of a 4MHz ceramic resonator VFO (74HCT04 in Pierce configuration, with some gates used as buffers) is divided by 8. This involves preferably a ripple-counter. My exciter design involves a 7493 which has got an "enable" circuit. And here is where keying is done. The oscillator, since it is using a different IC, is not affected by the keying and therefore does does not create any chirp.
This concept does not provide "click prevention", however, I believe clicking is more acceptable than chirp.
Showing posts with label howto. Show all posts
Showing posts with label howto. Show all posts
Sunday, October 3, 2010
Thursday, September 23, 2010
The Subharmonic (Frequency Doubling) Mixer
This is one of my favorites, the subharmonic mixer. It sound complicated, it may be, but, to build it is not.
How, let's have a look on the options of mixing.
Mixing of a signal with a local oscillator (LO) can be done in two ways, either enable the bypass of a signal in the "rhythm" of the LO or shorten the signal in the "rhythm" of the LO. Usually this is done by a single "non linear element", e.g. diode. In the positive 0-180 degrees of phase, the diode is open, letting through the signal, in the negative 180-360 degrees of phase, the diode is closed, blocking the signal. Alternatively, shorting the signal to ground using the diode would provide the same results, a sum and a difference of the two frequencies.
So far so good...
But where's the doubling?!
Well, here it comes. Assume that two anti-parallel diodes are used as a mixer. Given that the LO is adjusted to the correct level, the first of the two diodes is open for one 1/4 of the period (90 degrees) on e.g. 50% of the positive half-period of the LO and the second of the two diodes is open for another 1/4 of the period (90 degrees) on e.g. 50% of the negative half-period of the LO. This would correspond to the mixer's diodes being open during the phase angles 45-135 degrees and 225-315 degrees. Compared to the 0-180 degrees a single diode would be open, the frequency is effectively doubled.
There are some advantages to this approach. In a direct-conversion receiver, the LO is far of the receiver's front-end, the preamp would therefore be unaffected. Also, a lower frequency oscillator is less critical in design. Any variation of the LO will have double impact on the operating frequency... this has pros and cons, e.g. VXO.
I like using sub-harmonic mixers, try 'em out for yourself!
How, let's have a look on the options of mixing.
Mixing of a signal with a local oscillator (LO) can be done in two ways, either enable the bypass of a signal in the "rhythm" of the LO or shorten the signal in the "rhythm" of the LO. Usually this is done by a single "non linear element", e.g. diode. In the positive 0-180 degrees of phase, the diode is open, letting through the signal, in the negative 180-360 degrees of phase, the diode is closed, blocking the signal. Alternatively, shorting the signal to ground using the diode would provide the same results, a sum and a difference of the two frequencies.
So far so good...
But where's the doubling?!
Well, here it comes. Assume that two anti-parallel diodes are used as a mixer. Given that the LO is adjusted to the correct level, the first of the two diodes is open for one 1/4 of the period (90 degrees) on e.g. 50% of the positive half-period of the LO and the second of the two diodes is open for another 1/4 of the period (90 degrees) on e.g. 50% of the negative half-period of the LO. This would correspond to the mixer's diodes being open during the phase angles 45-135 degrees and 225-315 degrees. Compared to the 0-180 degrees a single diode would be open, the frequency is effectively doubled.
There are some advantages to this approach. In a direct-conversion receiver, the LO is far of the receiver's front-end, the preamp would therefore be unaffected. Also, a lower frequency oscillator is less critical in design. Any variation of the LO will have double impact on the operating frequency... this has pros and cons, e.g. VXO.
I like using sub-harmonic mixers, try 'em out for yourself!
Frequency dividers and their use
Some more off the howto-stuff, I lately started. Sometimes I was mentioning that a frequency should be divided... now, that is strange! So far we had it about multiplication of frequencies by some sort of factors, and now division?!
Yes, rather simple, yes really!
In the digital world, some thing are out there called counters. There are a couple of different counters available. The most primitive of all is the Flip-Flop, which is counting to 2.
Now, how does a Flip-Flop help to divide a frequency? Very simple, have your oscillator's signal on the clock input... .... ah well, it has been written before, check this out.
So, here you got it, cascading Flip-Flops will result in a division by 2^x, as cascading doublers did for multiplication. The cascade of Flip-Flops is known as Ripple-counter.
You may ask yourself, if division by odd numbers would be easily possible, as easily as multiplication was. The answer is NO. One can use hexadecimal or decade counters to divide a frequency, just as it is done with a ripple-counter, however, there is a disadvantage. A ripple counter ensures a 50% duty cycle, i.e. HIGH for half a period and LOW for the other half, which is symmetrical and can be smoothed to a sine-ish waveform by a low-pass filter. Any other counter-divider, however, will result is a duty cycle being off. So, what you really want to do is, ensure that your final division is EVEN and use a Flip-Flop as the last stage. Example: we would like to divide a frequency by 10. We would first use a decade counter, counting to 5. The resulting pulse would be sent to a Flip-Flop, so that the total amount of division would be 10. Always try to have your duty cycle as close as possible to 50%. Assume you need a division by 21. Use a decade counter to divide by 7 and cascade it with a counter to 3. This will give a 33% duty cycle, still relatively symmetrical, compared to a 14% duty cycle in reverse order....
Yes, rather simple, yes really!
In the digital world, some thing are out there called counters. There are a couple of different counters available. The most primitive of all is the Flip-Flop, which is counting to 2.
Now, how does a Flip-Flop help to divide a frequency? Very simple, have your oscillator's signal on the clock input... .... ah well, it has been written before, check this out.
So, here you got it, cascading Flip-Flops will result in a division by 2^x, as cascading doublers did for multiplication. The cascade of Flip-Flops is known as Ripple-counter.
You may ask yourself, if division by odd numbers would be easily possible, as easily as multiplication was. The answer is NO. One can use hexadecimal or decade counters to divide a frequency, just as it is done with a ripple-counter, however, there is a disadvantage. A ripple counter ensures a 50% duty cycle, i.e. HIGH for half a period and LOW for the other half, which is symmetrical and can be smoothed to a sine-ish waveform by a low-pass filter. Any other counter-divider, however, will result is a duty cycle being off. So, what you really want to do is, ensure that your final division is EVEN and use a Flip-Flop as the last stage. Example: we would like to divide a frequency by 10. We would first use a decade counter, counting to 5. The resulting pulse would be sent to a Flip-Flop, so that the total amount of division would be 10. Always try to have your duty cycle as close as possible to 50%. Assume you need a division by 21. Use a decade counter to divide by 7 and cascade it with a counter to 3. This will give a 33% duty cycle, still relatively symmetrical, compared to a 14% duty cycle in reverse order....
The VXO
Oscillators using crystals, aka XO, are supposed to be a reliable source of signal having a stable frequency. OK, compared to VFOs (Variable Frequency Oscillators), using inductors and capacitors, that is safe to say, however, there are some remarkably stable VFO designs out there...
The whole game of combining crystal frequencies, at least for A1A or F1A purposes is, to avoid the potential drift a regular VFO would possibly suffer from. Well, amateur radio, as we all know, is not a channelized game however. So, what are those crystals and combinations of xtal-qrg good for when it comes down to usability?
Well, first of all, there are some frequencies of major interested, such as the QRP callings QRGs. A rig covering a single frequency could still be very useful, depending on the frequency and its use. Just remember the good old times, everyone was using a color burst crystal (3.579MHz) and had a great share of fun.
A crystal oscillator can provide so much more than just a single frequency. And this is why:
a crystal is functioning as some sort of L-C (inductor-capacitor) resonant circuit (check out the internet for more info). Such circuits can be influenced by additional reactance, e.g. a capacitor and/or inductor in series, which will bend the resonant frequency either up or down.
Using the right amount of (variable) inductance and/or (variable) capacitance, a crystal's resonance can be pulled by a certain percentage, which can be quite a bit depending on the frequency.
This frequency pull can further be enhanced by using two or more crystals in parallel. Now we are talking "super VXO" (please check the internet, there is some very good documentation available from Japan).
What's the benefit? Assume we are going back to the example for the 40m band, used in the xtal-combi-post. Assume we pull the 8.867MHz oscillator by +/- 5kHz (which is no problem on that frequency at all), the resulting mix with a 1.843MHz frequency would allow for a range of 7.019 to 7.029Mhz, representing a very useful portion of the 40m CW range. With a super-VXO, this range could be from approx. 7.000 to 7.040Mhz.
This is what crystal combinations are all about. A stable VXO converting into a "cheap" I.F., and we got us a
cheap receiver having a crystal CW-filter.
BTW, this is what this entry was all about.
The whole game of combining crystal frequencies, at least for A1A or F1A purposes is, to avoid the potential drift a regular VFO would possibly suffer from. Well, amateur radio, as we all know, is not a channelized game however. So, what are those crystals and combinations of xtal-qrg good for when it comes down to usability?
Well, first of all, there are some frequencies of major interested, such as the QRP callings QRGs. A rig covering a single frequency could still be very useful, depending on the frequency and its use. Just remember the good old times, everyone was using a color burst crystal (3.579MHz) and had a great share of fun.
A crystal oscillator can provide so much more than just a single frequency. And this is why:
a crystal is functioning as some sort of L-C (inductor-capacitor) resonant circuit (check out the internet for more info). Such circuits can be influenced by additional reactance, e.g. a capacitor and/or inductor in series, which will bend the resonant frequency either up or down.
Using the right amount of (variable) inductance and/or (variable) capacitance, a crystal's resonance can be pulled by a certain percentage, which can be quite a bit depending on the frequency.
This frequency pull can further be enhanced by using two or more crystals in parallel. Now we are talking "super VXO" (please check the internet, there is some very good documentation available from Japan).
What's the benefit? Assume we are going back to the example for the 40m band, used in the xtal-combi-post. Assume we pull the 8.867MHz oscillator by +/- 5kHz (which is no problem on that frequency at all), the resulting mix with a 1.843MHz frequency would allow for a range of 7.019 to 7.029Mhz, representing a very useful portion of the 40m CW range. With a super-VXO, this range could be from approx. 7.000 to 7.040Mhz.
This is what crystal combinations are all about. A stable VXO converting into a "cheap" I.F., and we got us a
cheap receiver having a crystal CW-filter.
BTW, this is what this entry was all about.
Wednesday, September 22, 2010
Frequency multipliers and their use
Sometimes, one may want to multiply a frequency by factors other than two (doubling, see earlier post).
Here is an example why you actually may want to do this. What about a crystal controlled 30m band (10.100...10.150MHz) signal generator to generate a frequency of 10.118MHz using cheaply available xtals?
The spectrum of a square wave signal consists of odd harmonics. Hence odd multiplications are available by simply "clipping" the original signal, in other words, convert it into a square wave, and filter out the harmonic of interest.
In the digital age, one may think of using a digital gate, e.g. XOR or NOT (inverter), to generate a nice frequency fence in the first place.
Back to our example, in order to create our signal, we would build a square-wave generator using a 7.3729MHz crystal, filter out the 22.1187Mhz contribution and mix it with 12.000MHz.
Closing with an academic one: One may want to multiply a signal by 6. This could be done by filtering the 3rd harmonic and double it by means of a diode doubler...
Here is an example why you actually may want to do this. What about a crystal controlled 30m band (10.100...10.150MHz) signal generator to generate a frequency of 10.118MHz using cheaply available xtals?
- 3 * 7.3729 - 12.000 = 22.1187 - 12.000 = 10.1187
The spectrum of a square wave signal consists of odd harmonics. Hence odd multiplications are available by simply "clipping" the original signal, in other words, convert it into a square wave, and filter out the harmonic of interest.
In the digital age, one may think of using a digital gate, e.g. XOR or NOT (inverter), to generate a nice frequency fence in the first place.
Back to our example, in order to create our signal, we would build a square-wave generator using a 7.3729MHz crystal, filter out the 22.1187Mhz contribution and mix it with 12.000MHz.
Closing with an academic one: One may want to multiply a signal by 6. This could be done by filtering the 3rd harmonic and double it by means of a diode doubler...
Frequency doublers and their use
Occasionally, one might have seen that I mentioned a frequency had to be doubled. Why is that?!
Assume we want to generate a signal in the 12m band (24.890...24.990MHz). We could aim for a frequency of 24.915Mhz. This frequency could easily be synthesized by means of cheap computer crystals as follows:
use a 20.000MHz crystal in an oscillator and mix this signal with the 4.915MHz signal
or
use a 10.000MHz crystal in an oscillator, frequency double the generated 10.000MHz, resulting in 20.000MHz and mix that with the 4.915MHz.
There are a couple of ways to double the frequency of a radio frequency signal.
One of the most simple ways is using just simple (fast switching) diodes in a sort of rectifier circuit. A rectifier folds the negative valley of an AC signal to positive. We obtained two positive humps per cycle, meaning that the resulting AC signal has twice the frequency of the original signal. Yes, it as easy as that.
There are other methods, e.g. using XOR digital gates and a phase shifter, please check the internet for other solutions.
As a closing remark, frequency doubler can be cascaded. With the occasional amplification, one can easily build multipliers for factors 2^x, e.g. 4=2*2, 8=2*2*2, 16=2*2*2*2....
Assume we want to generate a signal in the 12m band (24.890...24.990MHz). We could aim for a frequency of 24.915Mhz. This frequency could easily be synthesized by means of cheap computer crystals as follows:
- 24.915 = 20.000 + 4.915 = 2 * 10.000 + 4.915
use a 20.000MHz crystal in an oscillator and mix this signal with the 4.915MHz signal
or
use a 10.000MHz crystal in an oscillator, frequency double the generated 10.000MHz, resulting in 20.000MHz and mix that with the 4.915MHz.
There are a couple of ways to double the frequency of a radio frequency signal.
One of the most simple ways is using just simple (fast switching) diodes in a sort of rectifier circuit. A rectifier folds the negative valley of an AC signal to positive. We obtained two positive humps per cycle, meaning that the resulting AC signal has twice the frequency of the original signal. Yes, it as easy as that.
There are other methods, e.g. using XOR digital gates and a phase shifter, please check the internet for other solutions.
As a closing remark, frequency doubler can be cascaded. With the occasional amplification, one can easily build multipliers for factors 2^x, e.g. 4=2*2, 8=2*2*2, 16=2*2*2*2....
Numbers and Combinations thereof (experts, please ignore this entry!)
OK, slowly but surely, I got behind the mystery that is perceived in the numbers I occasionally publish.
The mystery goes as follows:
Use the cheapest available material (i.e. electronic components) in order to be active in the ranges in which radio amateurs are allowed to operate. That's it! No secret message in here!
Example:
We want to operate (transmit or receive) in the 40m band CW (Morse code or telegraphy) section, we can easily use a superhet (supersonic heterodyne) design mixing two different frequencies such as 8.867Mhz and 1.843MHz resulting in two mixing products, the sum which results in 10.71MHz and the difference resulting in 7.024MHz. Using a low pass filter, one will remove the sum of the two frequencies and end up with the difference, which is at a very convenient place on the telegraphy portion of the 40m amateur radio band.
The trick, or secret if you insist, is to use two standard crystals, which are cheap to obtain anywhere, and mix those up.... that's all.... In a superhet receiver design, one of those frequencies can be used as an intermediate frequency filter, using simple ladder filters, made from those cheaply available standard crystals.
Please, go through all the numbers I published so far and tell me if there is anything wrong with those... Some combinations are to be found on this blog, some here and even more here.
As I said, experts, please ignore this....
The mystery goes as follows:
Use the cheapest available material (i.e. electronic components) in order to be active in the ranges in which radio amateurs are allowed to operate. That's it! No secret message in here!
Example:
We want to operate (transmit or receive) in the 40m band CW (Morse code or telegraphy) section, we can easily use a superhet (supersonic heterodyne) design mixing two different frequencies such as 8.867Mhz and 1.843MHz resulting in two mixing products, the sum which results in 10.71MHz and the difference resulting in 7.024MHz. Using a low pass filter, one will remove the sum of the two frequencies and end up with the difference, which is at a very convenient place on the telegraphy portion of the 40m amateur radio band.
The trick, or secret if you insist, is to use two standard crystals, which are cheap to obtain anywhere, and mix those up.... that's all.... In a superhet receiver design, one of those frequencies can be used as an intermediate frequency filter, using simple ladder filters, made from those cheaply available standard crystals.
Please, go through all the numbers I published so far and tell me if there is anything wrong with those... Some combinations are to be found on this blog, some here and even more here.
As I said, experts, please ignore this....
Subscribe to:
Posts (Atom)