05-01-2017, 01:07 PM
You mean like a MW aerial coil and capacitor, both in parallel to ground???
Mike
Mike
Testing crystals
|
05-01-2017, 01:07 PM
You mean like a MW aerial coil and capacitor, both in parallel to ground???
Mike
Sort of yes and no, I suspect you'll need to buffer the oscillator then take a tunable output from the buffer.
In my limited experience with Xtals they a manufactured to resonate at the frequency stamped on the can, eg: if it says 1MHz on the can then that's it, that's the frequency it's designed to operate at, no other, even though in practice it might, oscillator circuits produce all sorts of harmonics/mixed frequencies and general crud, where ever you want it's designed frequency to end up in a particular application then I would say make sure it ends up there via a tuned tank/filter etc unless you intend to use the oscillator as an harmonic marker. On another note, here's an old favourites link of mine, if you open it, it should open in the middle of a chapter on the TPTG oscillator....keep scrolling down...should end up with a chapter on quartz xtals and oscillators etc, I've just checked the link and there were no page interruptions when I scrolled through it: https://books.google.co.uk/books?id=f9QXAAAAYAAJ&pg=PA54&dq=tptg+oscillator&hl=en&sa=X&ved=0ahUKEwjyloPD9qrRAhWHHsAKHZZQBwYQ6AEINDAC#v=onepage&q=tptg%20oscillator&f=false Lawrence.
05-01-2017, 03:08 PM
(05-01-2017, 12:54 PM)Crackle Wrote: Hi Terry My fault, Mike. There is nothing wrong with the formula but I omitted to tell you to format the cells to display the result as a percentage. (If you look carefully, 0.000197316 is the same figure as +0.0197316% - it's just the position of the decimal point that is different.) Select one block of cells, then look for a button on the ribbon or toolbar with % on it. Click that and then adjust the number of decimal places to suit. I used 3 but you might prefer 4. I wouldn't go any further because the additional jumble of digits isn't going to be of any use to you: it just makes the results harder to take in. An alternative method is to select the cells as before, then press CTRL 1 to call up the format menu. Select 'Number', then 'Percentage' from the list provided. Select the number of decimal places in the box which appears, then click 'OK'. It will then be possible to quickly sort out the wheat from the chaff. Any that indicate an error of more than about, say, ±0.05%, should be taken as indicating that either the crystal or the test method is suspect. "What I want to understand is how to get the crystals to give out lower harmonics, so that they can be used as the manufacturer intended." I'm puzzled by this as 'what the manufacturer intended' was for the crystal to be used at the frequency marked on it! If it is a third overtone crystal it can be made to oscillate at its fundamental frequency, as you have proved in your tests. If it has been made so that its marked frequency is also its fundamental frequency then it won't produce any usable sub harmonics. From what Trevor had to say on the subject it looks as if would be a world first if you could come up with a circuit that did use sub-harmonics in a useful and repeatable manner! But ask yourself why anybody would bother. It costs no more to have a crystal manufactured for the frequency you want as it does for a crystal for a frequency you don't want! I don't know how much it costs for a one-off sample these days but in the days that I used such things they were remarkably cheap, I thought. For a MW crystal, a parallel tuned circuit, as you suggested, should do very well. Either in the anode/collector circuit of the oscillator or AC coupled from the output. A ferrite rod aerial for the inductor would be ideal as you can then pick up the oscillator with a portable radio!
05-01-2017, 03:10 PM
One thing to watch out for in xtal circuits is xtal current, it can be destructive to the xtal.
Lawrence. (05-01-2017, 03:08 PM)Terry Wrote: My fault, Mike. There is nothing wrong with the formula but I omitted to tell you to format the cells to display the result as a percentage. (If you look carefully, 0.000197316 is the same figure as +0.0197316% - it's just the position of the decimal point that is different.)Hi Terry OK I understand what you are doing now regarding the %. The manufacturer of the second crystal in the last table I posted, has marked the crystal as 4463.619 kHz, yet when tested in the circuit I made, it was oscillating at 10071.5 kHz. Why, or how, would a crystals properties change so that it would start oscillating at about 3 times its original frequency. My assumption was that it may be possible to utilise the crystal to produce lower harmonics, which could fall within the MW band. My understanding of this was based on the conversations I had some while ago with a guy who sold me the crystals to make Trevor's circuit he posted on UKVRR in 2011. The crystals were marked at quite a high frequency (21.8Mcs). But being as I cant get either to oscillate reliably in my circuit I wonder if he was making it all up. I guess they need to be tried in a circuit like Trevor's pantry TX to be sure. Mike
05-01-2017, 05:17 PM
I did do some experiments with tuned circuits a few years ago.
For a random selection of crystals you need a circuit that can work with both parallel or series crystals. You need a circuit with two places to plug the crystal in. A series tuned crystal can simply be added in parallel with an emitter resistor and the gain stage can then be adjusted to get a clean output. A parallel tuned crystal will need to be in series with the resistive feedback loop and again the gain stage adjusted for a clean output. Manually tuned circuits can then be added in the two crystal positions according to the type of crystal found in the first test to see what harmonics can be tuned. A bit of bespoke test equipment needs to be made for this task.
Hi Ref
Any chance you can please post a very simple circuit, preferably transistor. Thanks Mike p.s. thanks for the link Lawrence.
05-01-2017, 05:57 PM
The original circuit I did was run at 50Khz and acted as a proximity sensor for a boundary fence using a basic OP amp and a parallel tuned circuit shunting out local radio stations and phone signals. The previous design picked up a radio station from overseas and did not work.
All I used could be honed down to two capacitor coupled one transistor amplifier stages. Once you link the output to the input you get a square wave. Add a volume control and turn it down until it stops oscillating. You can vary the gain by using a pot and capacitor in place of the emitter decoupling as well giving a low impedance control of gain. Adding a chain of resistors and points to connect crystals will them make the feedback through the "volume" control more complex as will adding tuned circuits to the emitter resistors. You get one emitter resistor that is inverting and another that is non-inverting. For the really cunning constructor there is JFETS and detector diodes that can be used to make an AGC loop with the FET added across one of the emitter resistors. A test set can be made as complicated as the constructor likes starting with a simple two transistor square wave generator and growing as far as is needed for the task in hand. OP amp chips are expensive for high frequencies but a couple of transistors will be fine. I have never made a full circuit and it remains a block diagram and no more. Just remember that parallel tuned crystals are high impedance at resonance and series tuned crystals are low impedance at resonance.
05-01-2017, 06:02 PM
05-01-2017, 06:11 PM
The video shows how to do it with an external signal generator and a fast scope.
A dip in the scope waveform will be seen for a series tuned crystal and a dip in the waveform will show a parallel tuned crystal. |
Users browsing this thread: |
1 Guest(s) |