Needs the accurate 12MHz Crystal on Audio FX Sound Board

General project help for Adafruit customers

Moderators: adafruit_support_bill, adafruit

Please be positive and constructive with your questions and comments.
Locked
User avatar
SLB_Frank
 
Posts: 2
Joined: Mon Nov 22, 2021 2:12 pm

Needs the accurate 12MHz Crystal on Audio FX Sound Board

Post by SLB_Frank »

Dear Sir/Madam,

We are using the Adafriut Audio FX Sound Board 16MB. we load BPSK or FSK waveforms into it and make it as a component of communication system. but we fund that some of boards are not working. We find that the root cause results from the non-accurate 12MHz crystal on the board. What are the crystal parameters ( like working temperature, frequency stability, frequency tolerance and etc. ) used on this board?

I tried to look for an accurate 12MHz crystal to replace it. but I cannot find anyone with the same footprint.
Do you have an accurate 12MHz crystal ( working temperature: -40 to 85C, frequency stability: <= 30ppm, frequency tolerance: <=30ppm) which we can buy it from you?

thanks and regards,
Frank Fan

User avatar
adafruit_support_mike
 
Posts: 67454
Joined: Thu Feb 11, 2010 2:51 pm

Re: Needs the accurate 12MHz Crystal on Audio FX Sound Board

Post by adafruit_support_mike »

They’re standard AT-cut quartz crystals (the most common kind).

Oscillators have to treated as complete systems though, so changing the crystal won’t guarantee stability or accurate frequency. The timing characteristics of the amplifier have a strong effect on accuracy and stability. Voltage regulation of the supply power is another strong influence, as is temperature. The main difference between ‘drifts a few seconds per day’ oscillators like the DS1307 and ‘drifts a few seconds per year’ ones like the DS3231 is thermal control and temperature compensation.

Most ICs with connections for an external crystal use a minimal Pierce oscillator topology, and the effect of temperature on the silicon inside the package has a larger effect on frequency accuracy and stability than the thermal sensitivity of the crystal. Frequency accuracy also depends on the values of the capacitors between the ends of the crystal and GND.

Before swapping the crystals, make sure the power is well regulated and the temperature is reasonably well controlled. The variation from those will overwhelm any effects you’d see from a new crystal.

User avatar
SLB_Frank
 
Posts: 2
Joined: Mon Nov 22, 2021 2:12 pm

Re: Needs the accurate 12MHz Crystal on Audio FX Sound Board

Post by SLB_Frank »

Thanks for your reply.

I really appreciate it if you could directly answer my following questions.

1. what is the Frequency Stability (?ppm), Frequency Tolerance (?ppm) and Operating Temperature of the 12MHz crystal on the Adafruit FX Audio Sound Board with 16MB?

2. could you please recommend me an accurate one with the same footprint?

Thanks and regards,
Frank

User avatar
adafruit_support_mike
 
Posts: 67454
Joined: Thu Feb 11, 2010 2:51 pm

Re: Needs the accurate 12MHz Crystal on Audio FX Sound Board

Post by adafruit_support_mike »

The typical spec for crystals is 30ppm absolute accuracy, and AT-cut crystals have a parabolic thermal variation of 50ppm between -20C and 70C (It’s actually a cubic curve centered near 20C, but the magnitude of the frequency error is roughly parabolic). The standard drfit equation is -0.034 +/- 0.006ppm / (25 - T)^2

You can find crystals with slightly better absolute tolerances.. down to about 10ppm.. but that’s only marginally relevant. Quartz crystals have two resonant frequencies: one where its impedance is lowest (series resonance), and one where its impedance is highest (parallel resonance). Above or below those frequencies, the crystal’s impedance is capacitive. Between those frequencies its impedance is inductive. At both resonant frequencies, the impedance is purely resistive.

Phase shift through the crystal is close to -90 degrees at series resonance and close to +90 degrees at parallel resonance. An oscillator circuit’s feedback loop provides about 270 degrees of phase shift, and the whole circuit operates at whatever frequency makes the crystal supply enough phase shift to bring the total to 360 degrees.

Only a few oscillator topologies are capable of running at the crystal’s series or parallel resonant frequencies. All the others operate somewhere between the two. The Pierce topology is series-resonant, and the version used by most ICs will operate 10 degrees to 40 degrees away from the crystal’s series resonant frequency because that reduces sensitivity to things like noisy supply power.

So having the crystal’s series resonant frequency within 10ppm of the nominal value doesn’t really have much effect on the oscillator circuit’s resonant frequency. At best, it improves the chance that several identical (within the limits of the components) oscillators will have nearly the same resonant frequency. It has very little effect on the actual frequency though. That mostly depends on the phase shift contributed by all the other components in the feedback loop.

Tuned oscillators take advantage of that by putting a variable capacitor at the crystal’s input side. By varying that capacitance, they can change the phase shift provided by the rest of the loop and ‘pull’ the crystal to a higher or lower frequency in its inductive region.

That means there’s no such thing as an ‘accurate crystal’ in the way you’re using the term: something that will make the oscillator circuit’s operating frequency land closer to the ideal value. None of the properties used to specify a crystal control the exact frequency of an oscillator circuit independently from the rest of the circuit.

Locked
Please be positive and constructive with your questions and comments.

Return to “General Project help”