Crystal oscillators and vibration/shock

[Phil Hobbs]

Hi, all,

I'm designing a mixed-signal board for a fire safety system intended to
go on large pieces of agricultural equipment. It's doing DSP-type
things, so it needs a decent clock.

Naturally I thought of using an external XO. However, these gizmos are
liable to have a pretty rough life, with lots of vibration and impact
and stuff.

Sooo, what do you use for a +-50 ppm, low jitter oscillator that will
survive being dropped on the ground from a few metres up?

Thanks

Phil Hobbs

--
Dr Philip C D Hobbs
Principal Consultant
ElectroOptical Innovations LLC
Optics, Electro-optics, Photonics, Analog Electronics

160 North State Road #203
Briarcliff Manor NY 10510

hobbs at electrooptical dot net
http://electrooptical.net

[John Larkin]

On Fri, 21 Jul 2017 15:00:19 -0400, Phil Hobbs
<pcdhSpamM...@electrooptical.net> wrote:

>Hi, all,
>
>I'm designing a mixed-signal board for a fire safety system intended to
>go on large pieces of agricultural equipment. It's doing DSP-type
>things, so it needs a decent clock.
>
>Naturally I thought of using an external XO. However, these gizmos are
>liable to have a pretty rough life, with lots of vibration and impact
>and stuff.
>
>Sooo, what do you use for a +-50 ppm, low jitter oscillator that will
>survive being dropped on the ground from a few metres up?
>
>Thanks
>
>Phil Hobbs

Survive damage, or bounce without a big phase hit?

https://www.dropbox.com/s/h9jpjyk6act2h8w/MVC-001X.JPG?raw=1

https://www.dropbox.com/s/bdetgafn7zuc45g/MVC-004X.JPG?raw=1

The thing would drop out of phase-lock when someone unsnapped a
front-panel SMB connector. The springs fixed that.


--

John Larkin Highland Technology, Inc
picosecond timing precision measurement

jlarkin att highlandtechnology dott com
http://www.highlandtechnology.com

[Phil Hobbs]

On 07/21/2017 03:08 PM, John Larkin wrote:
> On Fri, 21 Jul 2017 15:00:19 -0400, Phil Hobbs
> <pcdhSpamM...@electrooptical.net> wrote:
>
>> Hi, all,
>>
>> I'm designing a mixed-signal board for a fire safety system intended to
>> go on large pieces of agricultural equipment. It's doing DSP-type
>> things, so it needs a decent clock.
>>
>> Naturally I thought of using an external XO. However, these gizmos are
>> liable to have a pretty rough life, with lots of vibration and impact
>> and stuff.
>>
>> Sooo, what do you use for a +-50 ppm, low jitter oscillator that will
>> survive being dropped on the ground from a few metres up?
>>
>> Thanks
>>
>> Phil Hobbs
>
> Survive damage, or bounce without a big phase hit?
>
> https://www.dropbox.com/s/h9jpjyk6act2h8w/MVC-001X.JPG?raw=1
>
> https://www.dropbox.com/s/bdetgafn7zuc45g/MVC-004X.JPG?raw=1
>
> The thing would drop out of phase-lock when someone unsnapped a
> front-panel SMB connector. The springs fixed that.
>
>

Fun.
For this one, I'm mostly concerned about the resonator cracking due to
shock.

Cheers

[Klaus Kragelund]

We do 40G hammer tests on products with crystals and resonators. I have never seen a problem with those components, but plenty with leaded caps and PCBs cracking

Cheers

Klaus

[Jon Elson]

Phil Hobbs wrote:

> Hi, all,
>
> I'm designing a mixed-signal board for a fire safety system intended to
> go on large pieces of agricultural equipment. It's doing DSP-type
> things, so it needs a decent clock.
>
> Naturally I thought of using an external XO. However, these gizmos are
> liable to have a pretty rough life, with lots of vibration and impact
> and stuff.
>
> Sooo, what do you use for a +-50 ppm, low jitter oscillator that will
> survive being dropped on the ground from a few metres up?

Rubber shock mounts?

Jon

[Phil Hobbs]

Dropping a crystal on the floor can easily be 5000 g. It'll be a lot
less with the thing on a board in a box, but I haven't used an XO in a
ruggedized unit before.

Cheers

[John Larkin]

On Fri, 21 Jul 2017 16:13:10 -0400, Phil Hobbs
<pcdhSpamM...@electrooptical.net> wrote:

>On 07/21/2017 03:59 PM, Jon Elson wrote:
>> Phil Hobbs wrote:
>>
>>> Hi, all,
>>>
>>> I'm designing a mixed-signal board for a fire safety system intended to
>>> go on large pieces of agricultural equipment. It's doing DSP-type
>>> things, so it needs a decent clock.
>>>
>>> Naturally I thought of using an external XO. However, these gizmos are
>>> liable to have a pretty rough life, with lots of vibration and impact
>>> and stuff.
>>>
>>> Sooo, what do you use for a +-50 ppm, low jitter oscillator that will
>>> survive being dropped on the ground from a few metres up?
>> Rubber shock mounts?
>>
>> Jon
>>
>
>Dropping a crystal on the floor can easily be 5000 g. It'll be a lot
>less with the thing on a board in a box, but I haven't used an XO in a
>ruggedized unit before.
>
>Cheers
>
>Phil Hobbs

I'm thinking you won't have a problem, any more than with other parts.
We don't get RMAs with failed Xo's but we get a lot of boxes back that
are bent and bashed.

Drop a few and see.

If it's agricultural, maybe it will usually hit dirt.

[Joe Chisolm]

On Fri, 21 Jul 2017 15:00:19 -0400, Phil Hobbs wrote:

> Hi, all,
>
> I'm designing a mixed-signal board for a fire safety system intended to
> go on large pieces of agricultural equipment. It's doing DSP-type
> things, so it needs a decent clock.
>
> Naturally I thought of using an external XO. However, these gizmos are
> liable to have a pretty rough life, with lots of vibration and impact
> and stuff.
>
> Sooo, what do you use for a +-50 ppm, low jitter oscillator that will
> survive being dropped on the ground from a few metres up?
>
> Thanks
>
> Phil Hobbs

Take a look at this App note from SiTime

https://www.sitime.com/support2/documents/AN10032-Shock-Vibration-Comparison-MEMS-and-Quartz-Oscillators.pdf

I use their mems oscillators in one product. Not because of vibration but cost and power is a little better.

--
Chisolm
Republic of Texas

[rickman]

Phil Hobbs wrote on 7/21/2017 3:00 PM:
> Hi, all,
>
> I'm designing a mixed-signal board for a fire safety system intended to
> go on large pieces of agricultural equipment. It's doing DSP-type
> things, so it needs a decent clock.
>
> Naturally I thought of using an external XO. However, these gizmos are
> liable to have a pretty rough life, with lots of vibration and impact
> and stuff.
>
> Sooo, what do you use for a +-50 ppm, low jitter oscillator that will
> survive being dropped on the ground from a few metres up?

They use crystals in lots of ruggedized equipment. Why do you think this
will be a problem. I think if you have demanding specs for your time base
it may be a concern, but if you are only using a typical crystal you likely
won't have an issue.

Here is something from SiTime showing how "bad" quartz crystals are compared
to their MEMS solutions.

https://www.sitime.com/support2/documents/AN10032-Shock-Vibration-Comparison-MEMS-and-Quartz-Oscillators.pdf

Here is more info from TI.

http://www.tij.co.jp/jp/lit/an/snaa296/snaa296.pdf

They both test shock to 500 gs.

--

Rick C

[TTman]

urvive being dropped on the ground from a few metres up?
>
> They use crystals in lots of ruggedized equipment. Why do you think
> this will be a problem. I think if you have demanding specs for your
> time base it may be a concern, but if you are only using a typical
> crystal you likely won't have an issue.
>
> Here is something from SiTime showing how "bad" quartz crystals are
> compared to their MEMS solutions.
>
> https://www.sitime.com/support2/documents/AN10032-Shock-Vibration-Comparison-MEMS-and-Quartz-Oscillators.pdf
>
>
> Here is more info from TI.
>
> http://www.tij.co.jp/jp/lit/an/snaa296/snaa296.pdf
>
> They both test shock to 500 gs.
>

Not seen anything special in any of the military/aerospace/inflight
refueliling stuff we built,including the latest 18Kw tank(battle)
PSU/distribution system...

---
This email has been checked for viruses by Avast antivirus software.
https://www.avast.com/antivirus

[John Miles, KE5FX]

On Friday, July 21, 2017 at 12:00:31 PM UTC-7, Phil Hobbs wrote:

> Sooo, what do you use for a +-50 ppm, low jitter oscillator that will
> survive being dropped on the ground from a few metres up?
>

Might be a good fit for the LMK61E2. They are resistant to vibration,
although they still jump around if you apply a lot of physical or thermal
stress directly to the case.

-- john, KE5FX

[Phil Hobbs]

Interesting, thanks.

Cheers

[Phil Hobbs]

On 07/21/2017 05:15 PM, rickman wrote:
> Phil Hobbs wrote on 7/21/2017 3:00 PM:
>> Hi, all,
>>
>> I'm designing a mixed-signal board for a fire safety system intended to
>> go on large pieces of agricultural equipment. It's doing DSP-type
>> things, so it needs a decent clock.
>>
>> Naturally I thought of using an external XO. However, these gizmos are
>> liable to have a pretty rough life, with lots of vibration and impact
>> and stuff.
>>
>> Sooo, what do you use for a +-50 ppm, low jitter oscillator that will
>> survive being dropped on the ground from a few metres up?
>
> They use crystals in lots of ruggedized equipment.

Sure. What I'm trying to ascertain is how ruggedized it needs to be.

> Why do you think this will be a problem.

Because I've broken crystals by dropping them on the floor in the past.
Not SMT ones, which is why I asked.

I think if you have demanding specs for your
> time base it may be a concern, but if you are only using a typical
> crystal you likely won't have an issue.
>
> Here is something from SiTime showing how "bad" quartz crystals are
> compared to their MEMS solutions.
>
> https://www.sitime.com/support2/documents/AN10032-Shock-Vibration-Comparison-MEMS-and-Quartz-Oscillators.pdf
>
>
> Here is more info from TI.
>
> http://www.tij.co.jp/jp/lit/an/snaa296/snaa296.pdf
>
> They both test shock to 500 gs.

Thanks. The floor tests to many times that. ;)

Cheers

[Phil Hobbs]

Yikes, $16 in reels. Nice part though.

Cheers

[rickman]

Don't drop your crystals before you put them in equipment. ;)

--

Rick C

[John Miles, KE5FX]

On Friday, July 21, 2017 at 3:29:55 PM UTC-7, Phil Hobbs wrote:
> Yikes, $16 in reels. Nice part though.

They have a pretty decent AT-cut crystal inside, or at least I'm
guessing it's AT-cut since it turns over at about 65C. I've been
tinkering around with a proportional oven design for it.

DigiKey marked them up to $50/1 a couple of weeks ago, while Mouser
was still selling them for about $20/1. I emailed DigiKey and asked
them what they were smoking and if I could order some too. Turned
out to be a listing mistake and they corrected the price, but even
$50 wouldn't be too outrageous if it can "emulate" a $250 custom OCXO.
I probably shouldn't say that in public where TI might see it...

-- john, KE5FX

[Phil Allison]

Phil Hobbs wrote:
>
> Hi, all,
>
> I'm designing a mixed-signal board for a fire safety system intended to
> go on large pieces of agricultural equipment. It's doing DSP-type
> things, so it needs a decent clock.
>
> Naturally I thought of using an external XO. However, these gizmos are
> liable to have a pretty rough life, with lots of vibration and impact
> and stuff.
>

** Vibration may be the bigger issue, most crystals hate it - standard HC49 ( tall)types in particular.

It is standard practice to wrap the radio receivers for RC models that use IC engines in soft foam to prevent crystal failures.

Vibration in the range of a few hundred Hz seems to be worst, the metal support clips holding the crystal inside the can simply sheer off if vibrated hard at their mechanical resonance.

A friend failed to follow this practice with his 1/8 scale gas car and suffered crystal failure in a few minutes one day. The max rpm of the 3.5cc engine was about 30,000 or 500Hz.

He gave me the dead 29MHz crystal and I opened the can with a saw for fun.


.... Phil

[Mikko OH2HVJ]

Phil Hobbs <pcdhSpamM...@electrooptical.net> writes:

> Hi, all,

> Sooo, what do you use for a +-50 ppm, low jitter oscillator that will
> survive being dropped on the ground from a few metres up?

SiTime MEMS-stuff. AT-cut SMD crystals can take quite some shocks, but
the 32.768/32.000 kHz RTC crystals are more prone to die on shocks.

The shock accelerations can usually be limited to reasonable numbers by
thinking a bit the shock direction and PCB bending and installing
methods.

ADXL001 is a nice part for measuring what kind of accelerations you can
see on PCBs, even though they're limited to 500g. Used these in one
wrist device and got bruises on my thigh while testing it by hitting
myself..


--
mikko

[Steve Wilson]

Phil Hobbs <pcdhSpamM...@electrooptical.net> wrote:

> On 07/21/2017 05:15 PM, rickman wrote:
>> Phil Hobbs wrote on 7/21/2017 3:00 PM:
>>> Hi, all,
>>>
>>> I'm designing a mixed-signal board for a fire safety system intended
>>> to go on large pieces of agricultural equipment. It's doing DSP-type
>>> things, so it needs a decent clock.
>>>
>>> Naturally I thought of using an external XO. However, these gizmos
>>> are liable to have a pretty rough life, with lots of vibration and
>>> impact and stuff.
>>>
>>> Sooo, what do you use for a +-50 ppm, low jitter oscillator that will
>>> survive being dropped on the ground from a few metres up?

> > Why do you think this will be a problem.

> Because I've broken crystals by dropping them on the floor in the past.
> Not SMT ones, which is why I asked.

> Cheers

> Phil Hobbs

SMT may be your answer. Some excerpts:

SMT military oscillator withstands 100,000-g shock

IQD's HGXO Series of surface mount oscillators is designed primarily
for military applications such as smart munitions and projectile
electronics although the devices can also be used in demanding
industrial applications. Designed around a hermetically sealed
high-shock crystal and a CMOS compatible integrated circuit, the new
device is able to survive shock levels of up to 100,000 g and
vibration levels of 20g at 10-2000 Hz swept sine in accordance with
MIL-STD-202G. Housed in an industry standard 7-mm x 5-mm four-pad
ceramic package with metal lid, five supply voltage versions are
available at 5.0 V, 3.3 V, 3.0 V, 2.5 V and 1.8 V, thus providing
compatibility with the latest generation of low power chipsets.

Output frequencies can be specified between 460 kHz and 5 MHz with a
15-pF HCMOS drive capability. Frequency tolerance is available down
to +/-10 ppm whilst frequency stability can be specified at +/-40
ppm over the full military temperature range of -55 C to 125C down
to +/-10 ppm over commercial temperature ranges coupled with low
frequency ageing characteristics.

https://www.ecnmag.com/product-release/2013/07/smt-military-oscillator-
withstands-100000-g-shock

High-shock quartz crystal oscillators

The classical quartz crystal oscillator is historically one of the
most fragile components in an electronic system. This is not
surprising since the quartz crystal resonator within the oscillator
was composed of a large crystal such as a large round-blank AT-cut
crystal mounted by metal clips inside of a metal housing. This
construction could not survive shocks much beyond 50 to 100 g. While
these crystal oscillators are superb for large benchtop instruments
and similar devices, they are not well suited for applications where
the device can expect high shocks such as handheld devices and
munitions. In these cases, the accelerations can be on the order of
thousands or even tens of thousands of g's. Clearly, the classical
construction is not adequate for these applications.

The impetus to change the construction of quartz crystals and
oscillators came from the continuing drive to miniaturize
electronics. A key step in this miniaturization took place in 1970
with the development of the photolithographic and chemical milling
processes for manufacturing quartz crystals. These processes,
adopted from those used in the silicon industry, allow the precise
milling of quartz crystals with dimensions under 1 mm and features
as precise as a few microns. Another important step in this
miniaturization was the development of the ceramic package for
firmly mounting the crystal in a rugged housing. Together, this
manufacturing and construction technique has become the de facto
standard for miniature quartz crystals.

Miniaturization vs. benefits

Fortunately, the miniaturization of the quartz crystal has had the
added benefit of greatly improving their shock and vibration
survivability. Because of its small size, the resonator has low
mass, and so the force on the resonator is low. Using strong
mounting materials, the resonator is held firmly in place - the
force due to acceleration is not sufficient to cause the crystal to
dismount. Further, because of its small size (short blank size or
short tuning-fork tines), the shear forces within the resonator are
low and hence they can survive high shocks without breaking.

Another added benefit of the small size is that the frequency of the
lowest flexure mode of the resonator can be on the order a few
kilohertz or higher. This has at least two benefits.

First, for shocks that have a characteristic time of about 1 ms or
longer, the shock can be treated as a quasi-static impulse - at any
given time the shock can be approximated as a static acceleration.
Because of this, the build-up in acceleration is sufficiently slow
that it does not excite the flexure modes of the crystal.

Second, since these flexure modes are high in frequency, they will
not be excited under vibration (which normally does not extend
beyond 2 kHz in typical applications). This is important in both
high-vibration applications and when manufacturing boards that are
cut out using a router.

http://www.electronicproducts.com/Passive_Components/Oscillators_Crystals_S
aw_Filters/High-shock_quartz_crystal_oscillators.aspx