Standard SMT crystal size chart?

7 views
Skip to first unread message

Matt Mets

unread,
Apr 10, 2013, 5:45:18 AM4/10/13
to shenzhe...@googlegroups.com
Hey all,

I'm hunting for crystals to use in a couple of board designs, and am
having a bit of trouble figuring out what packages are most common for
different frequencies/tolerances. For instance, I selected a 3215
package for a 32.768KHz crystal, but later found out that there don't
seem to be any 10ppm versions in that package. For the 3225 size,
however, there are some 10ppm versions, but they are of course more
expensive. For another design, I'm looking for a 16MHz crystal with
slightly less stringent tolerance, and there seem to be plenty of 3225
parts, but only a few exotic 3215 ones (with an extra lead, like this:
http://item.taobao.com/item.htm?id=17282685496).

It seems when I go between taobao, digikey, and the 华强北 markets I get
slightly different answers about what is and isn't available :-). Anyone
have advice on 'standard' sizes to use? I'm primarily interested in
small-sized packages, since we are making embedded things.

Cheers,
Matt

Akiba

unread,
Apr 11, 2013, 9:53:52 PM4/11/13
to shenzhe...@googlegroups.com, shenzhe...@googlegroups.com
I mostly use 3225 packages when available. 32.768 kHz packages vary but for smt, I try to use the smallish 2 lead packages. Id justice up a pack of 500 or 1000 at hua Qing bei and use those. Then if you find better pkgs, spin the board.

Sent from my iPad
> --
> You received this message because you are subscribed to the Google Groups "Shenzhen Hackers" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to shenzhenhacke...@googlegroups.com.
> For more options, visit https://groups.google.com/groups/opt_out.
>
>

Matt Mets

unread,
Apr 12, 2013, 4:15:08 AM4/12/13
to shenzhe...@googlegroups.com
Thanks Akiba! It's looking like 3225 for MHz frequencies and 3215 or
other 2-leaded for lower frequencies. I'll stick with those for now, and
maybe actually buy a roll of them at some point.

bunnie

unread,
Apr 12, 2013, 4:28:56 AM4/12/13
to shenzhe...@googlegroups.com
I haven't done the research to justify my hunch, but there should be a
physical limit on crystal size vs cut and frequency, as the physical
principal of crystal operation is a mechanical resonance. Intuitively,
it's a bit like asking how big a bell has to be to ring at a certain
frequency -- lower frequencies just need bigger chunks of quartz.

http://www.ndk.com/en/products/search/crystal/index.html

Has one vendor's table of frequency range vs. package size. Each vendor
will vary a bit based upon their process capability, but my hunch is you
won't see any major deviations in frequency range vs. package size
across multiple vendors, caveated vs. novel crystal resonance modes or
unusual temp/stability requirements.

-b.
--
^`'~*-,._.^`'~*-,._.^`'~*-,._.^`'~*-,._.^`'~*-,._.^`'~*-,._.^`'~*-,._.^`'

Matt Mets

unread,
Apr 14, 2013, 1:17:13 AM4/14/13
to shenzhe...@googlegroups.com
Thanks Bunnie, that looks like an excellent reference chart.

I was also assuming the resonance-vs-size model, but was a bit confused
that the 32KHz ones are actually in smaller packages than the 16MHz
ones. It's possible that they are using different cuts/overtone modes
that have a large effect on the size.

Wikipedia also mentions that MHz crystals are normally packaged with
springs, while lower frequency ones can be mounted at nodal points; that
might also have a big effect on packaging size. You know what? I'll try
popping the tops off of a few and see what they look like.

Cheers,
Matt
Reply all
Reply to author
Forward
0 new messages