In message 20170820181537.a7a834e32b848ea3f63d3943@kinali.ch, Attila Kinali writes:

I see that rule as a way to carve out telco-class rubidiums, and
that's why I think "or" would make much more sense than "and".

--
Poul-Henning Kamp      | UNIX since Zilog Zeus 3.20
phk@FreeBSD.ORG        | TCP/IP since RFC 956
FreeBSD committer      | BSD since 4.3-tahoe
Never attribute to malice what can adequately be explained by incompetence.

On 8/20/17 9:15 AM, Attila Kinali wrote:

I have some experience in this area - in the US it's the USML (United
States Muntions List) that determines what is subject to controls under
ITAR - that's run by the Department of State.  Then there's the
Controlled Commodities List which features in the EAR run by the
Department of Commerce.  The two groups have different objectives.

ITAR tends to focus a lot on "knowledge" as well as "things" - EAR is
more about "things"  - A component might be export controlled, but the
data sheet isn't.

Also, there's a huge difference between "speculating" that something is
so and "knowing" that something is so, in terms of design information or
performance.  If you're interested in building, say, a guidance system
for a ICBM - performance in a UAV might be a good indication that it
would work, but there's no substitute for test in a real missile.

A few years ago, there was a big rewriting of the USML - to make it more
specific in terms of capabilities, etc. rather than fuzzy - folks
seeking export licenses were frustrated by the previous more generic
language (often including the phrase "designed for military purposes" or
similar).  For spacecraft, it got a lot more liberal - before "if it
goes into space, it's ITAR" was the basic rule - afterwards, it's more
about "does it tell someone how to make it, design it, etc." and a lot
more things fell into more of a dual use (EAR is more about dual use)
bucket - just because you're using 6-32 machine screws or 100MHz OCXOs
on your spacecraft doesn't make ALL 6-32 screws or the same OCXO in
other uses export controlled.

This was a godsend for us cobbling together breadboards for things that
might someday go into space in a different form - before, if the work
was funded by a "space technology development" sort of bucket, the
evaluators would tend to say: yeah, that's export controlled, because
the "design intent" is for an eventual space application. That made it
hard to publish papers and reports openly as well as other tedious
administrative aspects - you can't put export controlled information
just anywhere or transmit it any old way, etc.

That process, of course, gets input from both industry and government,
and is not perfect.  But if you are a maker of a specific widget, you
could weigh in on the rule making process and explain why YOUR
particular widget's technology should or should not be controlled.
"should not" if you want to sell it overseas; "should" if you want to
avoid competition.

And then, there's a sort of continuing revision process - as each new
license application (or commodity jurisdiction (CJ) request) makes its
way through the system, that folds back to the actual rules or, more
importantly, their interpretation.

ALso as "work-arounds" for technology become well known, the rules
gradually change.  A good example is things like selective availability
and code-less 2 frequency GPS.

There's also well known hacks - a mfr might claims that their parts are
rad tolerant up to a particular level, even though they're made on a
process which is well known to be much harder.  To claim that they're
hard to a higher level would make them subject to ITAR instead of EAR,
for instance.

Like all regulatory matters, the wheels grind slow and exceedingly fine.
MUCH slower than the advance of technology (nothing has really changed
since you couldn't export pinball machines which had 68000
microprocessors) and that's something you live with.

And then, there's also just plain old typographical screwups - I'm sure
there are rules where it should be AND instead of OR, or someone wrote a
description designed to cover a particular class of "dangerous" stuff,
that inadvertently covers a much wider range. Small handheld software
defined spectrum analyzers have multiple uses - most people use them as
lab equipment, but SOME people use them as a tactical sensor.

So seeing a weird exception isn't all that unusual.

As the export lawyer explained to a group of us engineers a while back -
do NOT read the rules and try to rationally analyze them to figure out
if your thing falls in or outside - the rules are not internally
consistent.

• Especially, do not try to "design around" the rules to avoid controls.
  When your license app goes in, somebody at State (or Commerce) will be
  looking at the totality and might make the decision based on other
  information.

Hi Attila,

the paragraph applies also to low phase noise crystal oscillators.
The limiting equation for phase noise stated in that section 3. A. 1. b. 10 reads as follows:
L(F) > -(126+20log10F-20log10f) for 10 Hz ≤ F ≤ 10 kHz
With f being the carrier frequency in MHz and F the offset in Hz

In the attachment I have added a chart showing the limits for 10 MHz and 100 MHz.
This equation is rather arbitrary and not conclusive, because it assumes a slope of the PN curve close to carrier of -20 dB/decade, while in reality it is -30 dB/decade. This means that even if you do not touch the limits very close to carrier, the real phase noise may intersect at larger offsets. If you look at the numbers, you can easily find that not only a few best in class OCXO will touch or crosss the limits, but many normal "industrial state of the art" OCXO nowadays will...
I have no idea, where I could place my arguments to advocate a change of this unrealistic regulation.

Indeed it may be good that not all customs officers are able (or have the time) to solve that equation.

Best regards
Bernd
DK1Ag

-----Ursprüngliche Nachricht-----
Von: time-nuts [mailto:time-nuts-bounces@febo.com] Im Auftrag von Attila Kinali
Gesendet: Sonntag, 20. August 2017 12:08
An: Discussion of precise time and frequency measurement time-nuts@febo.com
Betreff: [time-nuts] Atomic clocks and Wassenaar agreement

On Sun, 20 Aug 2017 08:50:59 +0000
"Poul-Henning Kamp" phk@phk.freebsd.dk wrote:

As far as I can tell, there is no explicit mention of atomic clocks. But the list of dual use electronics is long and broad.
E.g. Section 3. A. 1. b. 10. covers basically all low noise frequency sources. Including just a simple low-noise XO.

Does anyone have more specific knowledge?

		Attila Kinali

--
You know, the very powerful and the very stupid have one thing in common.
They don't alters their views to fit the facts, they alter the facts to fit the views, which can be uncomfortable if you happen to be one of the facts that needs altering.  -- The Doctor _______________________________________________
time-nuts mailing list -- time-nuts@febo.com To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.