Testing crystals

[pwdrive]

I've know idea about harmonics below a fundamental frequency, I always thought an xtal resonated at it's fundamental or overtone, the oscillators output frequency can be multiplied by adding a tank circuit tuned to one of the harmonics. The Tri-Tet was popular at one stage:

https://books.google.co.uk/books?id=e_oZ69GAuxAC&pg=PA155&lpg=PA155&dq=tri+tet+oscillator&source=bl&ots=vZpG-pmX3t&sig=cNCqSz9vi8U_-gaglf-wbDpejfs&hl=en&sa=X&ved=0ahUKEwiEpbGPrKjRAhWHDsAKHbksDDYQ6AEIUjAJ#v=onepage&q=tri%20tet%20oscillator&f=false

Lawrence.

[Crackle]

I was just getting into that book at the Harmonic oscillators bit, when I came to my limit of what I was allowed to see.

Mike

[Terry]

Mike, I think you are getting a little confused here. You seem to be under the impression that all crystals can be used at sub harmonic frequencies. I am not a crystal oscillator expert but I cannot see any logical reason why they should.

I think the confusion stems from crystals made for higher frequencies that are cut to oscillate strongly on an overtone (harmonic) of their fundamental frequency: this is an entirely different situation, though, and does not apply to all crystals! A 30MHz 3rd overtone crystal might oscillate at 10MHz but there is no way it should be expected to generate 3.333MHz!

Looking at your results for all the crystals which are close to their indicated frequencies highlights an interesting inconsistency. Those that are slightly high show an error of, typically, +0.02% whereas many of those that are low are close to two orders of magnitude worse - around -0.65%. Some are very bad.

Could this be a difference between series and parallel resonant crystals or is your oscillator over-driving the under performing examples, I wonder?

The last column in your table doesn't make it easy to compare accuracy, so I've converted them all to percentages, to make it easier to see the wood from the trees. It is now easy to see that most of the third overtone crystals are very accurate indeed at 0.001% or better.

If you want to incorporate the same function in your own spread sheet, the formula I used is (this is in cell I3):

=IF(H3>2,H3-3,IF(H3>0.5,H3-1,H3-0.333333333))

(I had great fun doing an OCR on your graphic - virtually all of the numbers were correct but most of the decimal points vanished! A .csv file of the original will be useful if you post any more - change the .csv to .txt if it upsets the upload process.)

The crystals which are slightly above their marked frequency can probably be got spot on by adding a small value trimmer in parallel.

As for the bad and non performers, try a different oscillator design and see what happens!

   

[Murphyv310]

Hi.
Personally I wouldn't build a Micro TX without a tank coil circuit. It's advantages are that it will filter out harmonics and due to that will produce a sine wave at it's output.

One for Terry, I can assure you that you can get sub harmonics from "fundamental" crystal frequencies, they are very low amplitude and usually disregarded, I know of no application that they have been utilised in. I do agree that perhaps Mike is getting confused with overtone crystals which in the main are used for VHF applications where the third harmonic is used, it's easier to make a 3rd overtone crystal for 144 MHz say than a "straight" 144 MHz Crystal.

[Crackle]

Hi Terry
You sure are right, I am confused and with good reason. I admit to knowing next to nothing about the properties of crystals prior to starting this thread..
One of the points I was attempting to show was; if you look at the last crystal on the left half of the table, it would seem clear that the manufacturers expected the crystal to produce a lower third harmonic. The fundamental frequency is 2.92 times (almost 3times) the marked frequency. Yet it does not produce any lower harmonics that I can detect on my radio. Maybe it would in a different oscillator circuit.

Mike

[Crackle]

Hi Terry
Is your formula correct? Expressing the test frequency as a percentage of the marked frequency for the top crystal on the right half gives a result of 100.0197316%. This can be tested by working it backwards and multiplying the marked frequency (2534) by 100.0197316 and dividing by 100 the answer equals the test frequency.
   
xls file attached

  Crystals.xls (Size: 19.5 KB / Downloads: 3)

Mike

[Terry]

Mike, re: "The fundamental frequency is 2.92 times (almost 3times) the marked frequency", this is an error of -8% on the third overtone of the marked frequency which is a ludicrous figure for the accuracy of a crystal oscillator! The crystal seems to be active but something is badly wrong - thus my comment that the crystal is being over driven.

A different oscillator circuit may well highlight the problem.

"Expressing the test frequency as a percentage of the marked frequency for the top crystal on the right half gives a result of 100.0197316%": so it does, Mike, but what I am doing is calculating the error - think carefully about what an error of +100% really means!

The actual error is +0.0197316% but, for clarity, I'm only showing the result to 3 decimal places which rounds this figure up to +0.020% as shown.

In an ideal world, all the figures in the Error column would be 0% ...

Note also that I have manipulated the figures for the 3rd overtone crystals to show the error relative to the fundamental, rather than the marked, frequency. I don't think expressing the actual frequency as a -66.6% error on the marked frequency would be of much use to you or anybody else!

By the way, how accurate is your frequency counter and how are you measuring it? Any error here will have an effect on these results.

Finally, I note that all the crystals with a fundamental frequency in the Medium Waveband seem to show some attempt at activity but that's as far as it goes. Again, a different oscillator - or even different component values - might tell a different story. Why not try something with a MW tuned circuit in it, which is how I would expect it to be used in any micro Tx? I think it unlikely that all of these crystals are duff, so persevere: I'm sure you'll get the right answer in the end!

[pwdrive]

I reckon most of those xtals will be ok, not a lot to go wrong with them normally, typical defects I know of are xtal fracture and poor electrode contact.

I also wonder if tin whiskers have ever been encountered in them.

Lawrence.

[Crackle]

Hi Terry
When I put your formula into cell I3 the result was 0.000197316. not quite the same as your answer of +0.0197316%.

  Crystals3.xls (Size: 16.5 KB / Downloads: 1)

Am I doing something wrong?

I find it difficult to get any meaning out of these figures and they dont seem to be expressing the measured frequency as a function of the marked frequency, which is what the tests is all about. The degree of inaccuracy is fairly immaterial, because as you say it can probably be adjusted.

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.


Mike

[pwdrive]

The oscillator will probably be chucking out all sorts, try a tuned tank at the output tuned to the xtals marked frequency.

Lawrence.