Will this Circuit fry my Crystals?

[Fionaa]

Will this circuit fry an max 500 uW Drive Level XTAL?

 

[Diver300]

Short answer, no it won't as crystals are really difficult to overdrive.

Long answer, no it won't and here is why. That circuit will run that crystal at around 50 pF load capacitance so it's equivalent circuit is an inductor that tunes at 8 MHz to the 50 pF, so the impedance is around 400 Ohms.

I would expect an amplitude of around 3 V pk - pk on the crystal and so that is about 1 V rms, so the current is about 2.5 mA rms.

The ESR of the crystal will be something like 50 Ohms, so with 2.5 mA flowing, there will be around 300 μW dissipated in the crystal, which is quite a bit less than the 1 mW which is the typical maximum for an HC49 crystal.

https://portal.iqdfrequencyproducts.com/products/details/hc49.pdf

Small crystals will have lower drive levels but they don't get damaged quickly by overdriving. They can age faster if overdriven.

As for what the circuit is doing, it looks like an unmodulated transmitter. There is no control of the amplitude and the direct connection of the antenna to the oscillator circuit will make the frequency unstable. Two of those in the same room will phase-lock to each other. There will be a lot of harmonics generated.

If that is being used as a transmitter, there is unlikely to be any legal way to use it outside of a screened room.

 

[Fionaa]

Short answer, no it won't as crystals are really difficult to overdrive.

Long answer, no it won't and here is why. That circuit will run that crystal at around 50 pF load capacitance so it's equivalent circuit is an inductor that tunes at 8 MHz to the 50 pF, so the impedance is around 400 Ohms.

I would expect an amplitude of around 3 V pk - pk on the crystal and so that is about 1 V rms, so the current is about 2.5 mA rms.

The ESR of the crystal will be something like 50 Ohms, so with 2.5 mA flowing, there will be around 300 μW dissipated in the crystal, which is quite a bit less than the 1 mW which is the typical maximum for an HC49 crystal.

https://portal.iqdfrequencyproducts.com/products/details/hc49.pdf

Small crystals will have lower drive levels but they don't get damaged quickly by overdriving. They can age faster if overdriven.

As for what the circuit is doing, it looks like an unmodulated transmitter. There is no control of the amplitude and the direct connection of the antenna to the oscillator circuit will make the frequency unstable. Two of those in the same room will phase-lock to each other. There will be a lot of harmonics generated.

If that is being used as a transmitter, there is unlikely to be any legal way to use it outside of a screened room.

Its an CW Transmitter thats Missing the Harmonics Filter, and the crystals ESR is 80 Ohms. Thank You for your Help

 

[Nigel Goodwin]

I'd like to see a high value resistor from gate to ground, I don't like to see the gate floating.

 

[Fionaa]

I'd like to see a high value resistor from gate to ground, I don't like to see the gate floating.

Will 1 Megaohm work?

 

[Nigel Goodwin]

Will 1 Megaohm work?

Yes, it's not critical - 1Meg would be fine, or 2.2Meg, 10Meg - just a grid 'leak' to stop the gate floating.

 

[Tony Stewart]

Short answer, no it won't as crystals are really difficult to overdrive.

Long answer, no it won't and here is why. That circuit will run that crystal at around 50 pF load capacitance so it's equivalent circuit is an inductor that tunes at 8 MHz to the 50 pF, so the impedance is around 400 Ohms.

I would expect an amplitude of around 3 V pk - pk on the crystal and so that is about 1 V rms, so the current is about 2.5 mA rms.

I have never burnt out a crystal and understand that microwatt power levels would never be felt as heat.

So I ran a simulation now to see if the average motional resistance, Rm would dissipate more than the 500 uW rating in this design, but the instantaneous complex peak power is much higher , although decaying slowly.

A previous simulation of a discrete model of a crystal with a Q of 10k proved to me that this amplifies the internal motional capacitance voltage that could stress the breakdown voltage at the lattice structure and produce a pixel arc and slowly degrade the frequency and ESR.

This gives me reason to doubt if it average power level from thermal effects or over-voltage inside the Xtal lattice on peak reactive power levels.
Anyone?

 

[Nigel Goodwin]

I have never burnt out a crystal and understand that microwatt power levels would never be felt as heat.

A friend of mine did, years ago, and he killed multiple ones.

He built a simple one valve AM medium wave radio transmitter, and cathode modulated it from a transistor audio amplifier via a transformer, he used a large pentode for the transmitter, I think it might have been an 807? - and ran a pirate radio station using it.

He only transmitted certain times a week, for a few hours at a time, and even had phone ins, using a phone box elsewhere on the estate, with runners bringing him the details

So as the oscillator was the PA the crystal was rather abused, and he killed a number of them. Obviously adding a separate oscillator valve would have cured the issue, but that would have meant rebuilding it

Interestingly the quality was quite good (for AM ) and he had a niche local audience.

 

[Diver300]

I have never burnt out a crystal and understand that microwatt power levels would never be felt as heat.

So I ran a simulation now to see if the average motional resistance, Rm would dissipate more than the 500 uW rating in this design, but the instantaneous complex peak power is much higher , although decaying slowly.

A previous simulation of a discrete model of a crystal with a Q of 10k proved to me that this amplifies the internal motional capacitance voltage that could stress the breakdown voltage at the lattice structure and produce a pixel arc and slowly degrade the frequency and ESR.

This gives me reason to doubt if it average power level from thermal effects or over-voltage inside the Xtal lattice on peak reactive power levels.
Anyone?

I have wondered if the heat dissipation in the ESR is a good theoretical basis for modelling the stress in the crystal, and I don't think it is, but it maybe the best that is available.

The component model of a crystal as an inductor in series with a capacitor (plus the ESR and the static capacitance) only models the characteristics near the main resonance. Spurious responses and overtones are not modelled at all.

Similarly the physical movement of the crystal isn't modelled by just having an inductor and a capacitor, and I really don't know how you could put limits on the physical movement.

 

[Fionaa]

I made this now using this Crystal: https://jlcpcb.com/partdetail/YXC_CrystalOscillators-X49SD5MSD2SC/C2198

 

[Diver300]

The frequency has changed a lot. Normally that's something you have to decide early on. One of the main reasons to have a crystal is that it defines the frequency with good accuracy.

The 5 MHz HC49/4 crystal may be more difficult to get to oscillate than the 8 MHz HC49 crystal. The actual quartz in an 8 MHz crystal will be around 0.2 mm thick, while the quartz in a 5 MHz crystal will be around 0.32 mm thick. They oscillate in thickness-shear mode (https://en.wikipedia.org/wiki/Crystal_oscillator) and the oscillation is strongest in the middle of the disk of quartz, and reduces towards zero at the edges.

The 5 MHz crystal will have to be thinner at the edges to make that happen, so that is an additional restriction to the oscillation. The 8 MHz crystal in an HC49 crystal will most likely be a flat disk, possibly with a slight thinning at the edges.

The HC49/4 mm crystal can't be round like it is in the HC49 housing, so the sides of the disk have to be cut off. The crystal is oriented so that the motion is along the crystal, not across, but the motion is still affected.

The maximum ESR of the crystal specified by JLCPCB goes up sharply at low frequencies, showing that the design is compromised at low frequencies.

 

[Fionaa]

The frequency has changed a lot. Normally that's something you have to decide early on. One of the main reasons to have a crystal is that it defines the frequency with good accuracy.

The 5 MHz HC49/4 crystal may be more difficult to get to oscillate than the 8 MHz HC49 crystal. The actual quartz in an 8 MHz crystal will be around 0.2 mm thick, while the quartz in a 5 MHz crystal will be around 0.32 mm thick. They oscillate in thickness-shear mode (https://en.wikipedia.org/wiki/Crystal_oscillator) and the oscillation is strongest in the middle of the disk of quartz, and reduces towards zero at the edges.

The 5 MHz crystal will have to be thinner at the edges to make that happen, so that is an additional restriction to the oscillation. The 8 MHz crystal in an HC49 crystal will most likely be a flat disk, possibly with a slight thinning at the edges.

The HC49/4 mm crystal can't be round like it is in the HC49 housing, so the sides of the disk have to be cut off. The crystal is oriented so that the motion is along the crystal, not across, but the motion is still affected.

The maximum ESR of the crystal specified by JLCPCB goes up sharply at low frequencies, showing that the design is compromised at low frequencies.

I couldnt find those bigger Crystals at JLCPCB

 

[Diver300]

I couldnt find those bigger Crystals at JLCPCB

It doesn't surprise me. The taller crystals are a bit old fashioned. You should bear in mind that the lower frequency crystals in small packages may not work as well as the same frequencies in larger packages.

 

[Fionaa]

i could try Ceramic Resonators too for normal AM at 1 MHz https://jlcpcb.com/partdetail/HCI-HCTB1_1_000CRBHRL/C19181338 do they have any similar limits?

 

[Diver300]

i could try Ceramic Resonators too for normal AM at 1 MHz https://jlcpcb.com/partdetail/HCI-HCTB1_1_000CRBHRL/C19181338 do they have any similar limits?

I'm not familiar with ceramic resonators, so I can't help.

 

[Fionaa]

i could try Ceramic Resonators too for normal AM at 1 MHz https://jlcpcb.com/partdetail/HCI-HCTB1_1_000CRBHRL/C19181338 do they have any similar limits?

Nevermind it works, i built something on an breadboard with those smaller crystals

 

[Tony Stewart]

I have wondered if the heat dissipation in the ESR is a good theoretical basis for modelling the stress in the crystal, and I don't think it is, but it maybe the best that is available.

The component model of a crystal as an inductor in series with a capacitor (plus the ESR and the static capacitance) only models the characteristics near the main resonance. Spurious responses and overtones are not modelled at all.

Similarly the physical movement of the crystal isn't modelled by just having an inductor and a capacitor, and I really don't know how you could put limits on the physical movement.

This is what lead me to hypothesize that it is the high voltage breakdown of the weakest lattice in a crystal that burns which is like a partial discharge on a dust particle in a high e-field.

Here in this series mode crystal XO sim you can see the peak HV across the smallest capacitance, Cm = 100 fF , meanwhile even if the thermal resistance was 1000'C/W it would only rise 0.6 'C with ESR Pavg = 0.6mW yet 5kVpp in voltage stress. after only a few milliseconds. https://tinyurl.com/2dg4sq7u Now scaling up in voltage and power drive seems to me the thermal model is harmless until it arcs.

 

[Diver300]

I think that the physical stress is what matters, and that isn't modelled by a tiny capacitor and a huge inductor.

A crystal is a mechanical resonator, and the frequency is determined by the mass and the mechanical stiffness. The Q of the crystal is very high, which makes external influences like load capacitance have very little effect on the frequency.

In the inductor + capacitor model the values are chosen to give the correct effect of changing the load capacitance. The motional capacitance doesn't represent a physical capacitor, it's just part of a model that gives the correct frequency and adjustability of the crystal.

The motional capacitance can vary a lot within a batch of crystals, maybe by a factor of 2. If the capacitance and inductance in the model are changed by that factor it will change the voltage across the capacitor by a similar factor. That change in voltage in the model doesn't represent any real-life change, so I don't think that the peak voltage on the capacitor in the model needs to be considered, as it's not a real voltage.

 

[Nigel Goodwin]

I think that the physical stress is what matters,

I agree, the main failure mode of crystals (as with my old friend) is over driving them, and the excessive mechanical movement cracks or damages the crystal wafer.

 

[Lightium]

If you get into constructing a crystal model that models the forces a good link I think would be: Electrical to mechanical forces