embedlets-developer — Outpost Embedlet container developer list

[Embedlets-dev] Embedlets Group's Strategy

[Ted K. <tk...@ya...>]

For those who are interested, here is a document that discuss the main parts of
the Embedlets Group's strategy.  Up until this point, it has been a somewhat
informal strategy but this is in the process of changing fairly quickly and the
following materials will probably be used as a starting point for developing an
official Embedlets road map.



WHY THE EMBEDLETS GROUP WAS FORMED

The reason that the Embedlets Group was initially formed was to discover 'the
most likely markets for Embedded Java' because, even though Embedded Java
systems like the TINI came out in 1999, Embedded Java was not having the amount
of market uptake that it should have. One of the more critical things we
eventually discovered is discussed the following 'Bridging Technology' section.

It is important to point out, however, that the Embedlets Group was designed to
be a place where Embedded Java vendors and Embedded Java open source developers
could work together to promote all facets of Embedded Java.  To this end, we
are definitely following a principle that Sun Microsystems continuously
stresses which is 'cooperate on specifications and compete on implementations'.

We are currently still in the 'specification creation' phase but we will soon
be bringing the 'compete on implementation' phase online and a vendor's area
will be added to the Embedlet's web site to accommodate this in the near
future.



EMBEDDED JAVA IS A BRIDGING TECHNOLOGY

The background information that explains why Embedded Java is currently a
bridging technology can be found on the Embedlets main web page:

http://embedlets.org


This diagram:

http://embedlets.org/images/flounder.jpg


And if that is not enough for you, in this OpenOffice presentation:

http://tkosan.javadevices.org/misc/embeddedjava/embedlets_jacknet_presentation.sxi


What this information states is that (at this point in time) the Embedded Java
community's main strength is that it is a bridging community which sits between
the traditional Embedded Systems community and the Enterprise/CS software
development community.

These two very different communities need to be able to communicate with each
other before the vision of 'billions of devices being brought to the internet'
can become a reality.  They are so different, however, that direct
communication between them is virtually impossible and so a translator
community is needed.

By definition, an Embedded Java developer has an intimate understanding of both
these communities so is in an excellent position to provide this needed
translation function.

As the flounder diagram shows (http://embedlets.org/images/flounder.jpg), we
are a very small but, potentially, a very important community.



THE EMBEDLETS GROUP'S STRATEGY

Once the above insights were realized, forming a multifaceted strategy for
putting Embedded Java 'on the map' was relatively straight-forward.  Here are
the main steps in this strategy:


1) Form a small group of Embedded Java developers that will then create the
open software and hardware foundational technologies and knowledge base needed
to bridge the traditional Embedded System's communities and the Enterprise/CS
software communities.  

Status: The Embedlets project was officially formed in December of 2002, the
Embedlets container code base was mostly functional by late Spring of 2003 and
we are about to figure out how JAPL devices should be integrated with
Embedlets.  The initial group was purposely kept small so that focus could be
maintained but we will soon be enlarging the group's membership (see #3 below).



2) Select an initial target market for this technology, create a class of
product that will be desired by this market and then develop a strategy for
approaching this market with the product class.  

Organizations that need to integrate their processes with their backend
Enterprise systems (like 6 Sigma organizations) was selected as the initial
target market.  The class of product we are in the process of developing is the
'Enterprise Outpost' and the unique strategy that was developed to approach
this market was to create an Enterprise Outpost reference implementation, write
white papers and articles about Enterprise Outposts targeted toward J2EE
developers about it and demonstrate Enterprise Outposts to J2EE developers at
the most promising venues (like the JavaOne conference).

Status: Using the criteria of quality, functionality, price and maturity, TStik
has been selected as the Embedded Java system to be used for the Enterprise
Outpost reference implementation and we are currently in the process of
figuring out a reasonable way to package it.  As soon as the packaging and
production issues are resolved, and step 3 below is put into action, we will
then move forward with creating the white papers, articles and demonstrations
needed to market the Enterprise Outpost concept.

<note>
Even though the Enterprise Outpost is still the most promising technology to
move forward with, it appears that during the next year, the work we are doing
with it can possibly be leveraged to create a Home Automation Outpost.
</note>


3) As soon as steps 1 and 2 have reached a reasonable level of maturity,
increase the size of the Embedlets group by starting to openly advertise our
existence in the most popular Embedded Java lists.  Also, encourage traditional
Embedded Systems developers to join the group by providing them with free
online educational materials which are designed to teach them about Java in
general and how Java can be used to great advantage in Embedded Systems.

Status: It is looking like the month of December will be a good time to start
officially advertising our existence in the main Embedded Java lists.  Since we
appear to be the only vendor-neutral open source Embedded Java community in
existence, there is a reasonable probability that we can recruit the lion's
share of this community for membership in the Embedlets group.

As for the educational materials for Embedded Java beginners, we have a free 7
week online Java beginner's course ready to add to the Embedlets web site along
with a 5 week TStik/TINI beginner's class that is a follow-on module to this. 
A muvium beginner's module is also currently in the works and hopefully people
will use the 7 week beginner's class as a core around which modules can be
built for any Embedded Java systems that people are interested in.

Last, but not least, a version of BASIC (called JBASIC) has been developed
which has been designed to run on most Embedded Java systems and make them
*much* easier to use by Java beginners.  Hopefully this will be a technology
that can be used to help bring the large BASICStamp, PICBasic, etc. communities
over to the Embedded Java platform (finally!).

The educational materials, along with a beta version of the JBasic environment,
will be ready to add to the Embedlets site by sometime in December.



4)  After step 3 has been put into play, start encouraging J2EE (and other
backend and CS style developers) to join the group so that we can help them to
start leveraging the Embedded space.


Ted

__________________________________
Do you Yahoo!?
Protect your identity with Yahoo! Mail AddressGuard
http://antispam.yahoo.com/whatsnewfree

Re: [Embedlets-dev] Enclosure for TStik

[Ted K. <tk...@ya...>]

Marc wrote:

> I designed a 100mm*100mm Euroboard enclosure. It¹s extruded aluminum (it¹s a
> customized version of a Hammond case). The end plates have punched-out holes
> to fit a Tilt-400 or Tilt-400 Pro. Oh, and it¹s clear anodized.

Sheesh Marc!  If I would have known it was my turn to have a wish granted on
the list, I would have asked for free bars of gold!  This is excellent news... 


Do you happen to have a picture of a prototype unit you can post?



> I¹ve been on the fence regarding actually doing the first production run...I
> certainly don¹t currently need as many units as I¹d need to produce to make
> it viable. If there are others that could use a case like this, contact me
> off-list and we¹ll talk.
> 
> Cost would be in the range of US$20/unit.

How many units would be needed in order to make it worth your while to do a
production run?

As it turns out, since our goal is to make available an open Outpost reference
implementation, we will need to discuss as much of the business issues as
possible in the open list in order for our strategy to work.


And believe it or not, the Embedlets group developed a strategy last December
and January for finally putting Embedded Java 'on the map' and almost
everything we have done since then has aligned with this strategy.


But, since a number of the people on this list joined the group after this
period, it would probably be a good idea to re-state the strategy (especially
since we are nearing the point when we are going to start kicking it into 4th
gear).


I will do this in a separate post.


Ted 


__________________________________
Do you Yahoo!?
Protect your identity with Yahoo! Mail AddressGuard
http://antispam.yahoo.com/whatsnewfree

Re: [Embedlets-dev] Enclosure for TStik

[Marc N. <ma...@ge...>]

I designed a 100mm*100mm Euroboard enclosure. It=B9s extruded aluminum (it=B9s =
a
customized version of a Hammond case). The end plates have punched-out hole=
s
to fit a Tilt-400 or Tilt-400 Pro. Oh, and it=B9s clear anodized.

I=B9ve been on the fence regarding actually doing the first production run...=
I
certainly don=B9t currently need as many units as I=B9d need to produce to make
it viable. If there are others that could use a case like this, contact me
off-list and we=B9ll talk.

Cost would be in the range of US$20/unit.

-marc



On 14/11/03 02:08, "Bruce Boyes" <bb...@sy...> wrote:

>> Are you saying here that for industrial use, TStik systems do not need t=
o be
>> tested at all?
>=20
> Not clear. I discussed this with Scott and apparently "computing equipmen=
t"
> needs to be tested. We should maybe chat with the local lab and ask exact=
ly
> what is included in this.
>=20
>> The two most likely markets for this to happen in appear to be the Indus=
trial
>> Automation market and the Home Automation market and we must simply be
>> prepared
>> to do whatever it takes to crack one or both of these markets.
>=20
> I'd bet that anything sold for use in the home is subject to much more
> regulation.
>=20
>> ENCLOSURE FOR TSTIK
>>=20
>> A number of us here in Southern Ohio have been racking our brains trying=
 to
>> figure out the best kind of enclosure to wrap around TStik in order to t=
urn
>> it
>> into an Enterprise Outpost.
>>=20
>> The two main sources for enclosures we have been looking at are DigiKey =
and
>> Mouser and the top 3 enclosure manufactures that they carry are Hammond,=
 Bud
>> and Serpac.
>>=20
>> The TStik socket boards are 3.9"L x 3.9"W x roughly 2.75"H and so far we=
 have
>> not found a box that looks like it will match this dimension perfectly. =
 Our
>> strategy with a pre-made box is to cut the connector openings in the box
>> using
>> a CNC mill.
>=20
> All our socket boards are Euroboards, 100 mm in one dimension and 100 or =
160
> mm in the other. They conform to the international Euroboard specificatio=
ns
> for size and corner holes. As a result there are hundreds of off the shel=
f
> enclosures designed for such boards. BICC/VERO, Schroff, Rose/Bopla, BUD,=
 and
> others. You can also purchase a variety of extrusions with 100-mm board
> guides.
>=20
>> If we can not find a pre-made box, then another alternative is to design=
 our
>> own box along with the injection mold that goes with it and if worse com=
es to
>> worse, I will machine the mold myself by hand if needed.
>=20
> Packaging is not our specialty. With that disclaimer, here's my $0.02:
>=20
> Any enclosure will have a big impact on FCC testing. Plastic doesn't work=
 at
> all without a shielding spray, usually a nickel compound. Connectors whic=
h
> pierce the enclosure would need to be considered initially so that they d=
idn't
> hurt the integrity of the enclosure more than necessary.
>=20
> For a variety of reasons we are looking at punched and formed steel sheet=
,
> with pressed in threaded bosses. The top and bottom would mate together i=
n
> sort of a clamshell fashion. In low volume it's relatively affordable and=
 the
> NRE is very reasonable. Such a design can also be customized at low cost =
(say
> you want 100-1000 units with different connectors, colors, and silkscreen=
).
> Epoxy paint with silkscreened legends looks good and is also affordable. =
Plus
> steel is both an electromagnetic and electrostatic shield. Our own metal
> enclosure is on my "todo" list with a number of other things ahead of it.
>=20
> For much higher volume, custom molded plastic with shielding spray, molde=
d
> bosses with threaded mounting inserts, etc, would be the way to go. In ou=
r
> limited experience, machining plastic is surprisingly costly. Most plasti=
cs
> are not easy materials to machine. The enclosures we sell now are machine=
d
> plastic, I think it's polystyrene. We lose about 5% of the raw stock to
> machining accidents when end mills grab the plastic (which flexes easily)=
 and
> gouge it. Some of this is due to the type of machining we do - pockets fo=
r the
> LCD and keypad:
> http://www.systronix.com/access/encl_photos.htm
> The end panels you see in the photos are actually un-clad PC board materi=
al.
>=20
> So far, solving the enclosure problem hasn't been a high enough priority,=
 and
> we haven't had enough demand to make it higher. I do have some prototypes=
,
> some sketches in lab books, and notes from talking with a couple of local
> sheet metal shops.
>=20
> I googled 'Euroboard enclosures" and found this:
> http://www.schroff.us/support/faq/techcenter.asp?file=3D6&cat=3D6
> <http://www.schroff.us/support/faq/techcenter.asp?file=3D6&amp;cat=3D6>
>=20
> The euroboard form factor was created to standardize the size of printed
> circuit =20
> boards, which in turn would allow for the use of standardized enclosures.=
 By
> working to a standard an OEM can by components from multiple suppliers wi=
th
> the confidence that they would all be mechanically compatible with one
> another. =20
> Open architecture products that use the Euroboard form factor include, VM=
E,
> VME64 extensions, Multibus, MultibusII, CompactPCI, VXI and PXI along wit=
h an
> untold number proprietary. Designs of OEMs who have taken advantage of th=
e
> mechanical =20
> integrity of the form factor.
> =20
> The dimension of a euroboard is defined by IEC 60297-3 (1984-01).
>=20
> 1 rack unit or 1U =3D 44.45mm or 1.75
>=20
> Standard euroboard heights
> Reference U height      Printed board height
> 3U                      100mm (3.937")
> 6U                      233.35mm (9.187")
> 9U                      366.7mm (14.437")
>=20
> Standard euroboard depths
> 100mm   160mm           220mm
> (3.937")        (6.299")                (8.661")
> and other larger ones...
>=20
>  From the values above you can combine heights and depths but the most co=
mmon
> euroboard sizes are:
> 3U (100mm) x 160mm
> 6U (100mm) x 160mm
> 6U (233.35mm) x 220mm
>=20
> Note that TILT is 100x100 which is a standard (but not most common) Eurob=
oard
> size. STEP is 100x160, and so are SaJe, uCAN2, DPB2, HSM320/550, etc. The
> JSimm backplane is a 100x100 mm Euroboard.
>=20
> All our boards can mount in SnapTrack which then clips on to a DIN rail. =
Many
> of our customers use this to add our systems to an existing NEMA box whic=
h
> already has the DIN rail in it, and already houses other equipment. Here'=
s
> SnapTrack:
> http://www.systronix.com/access/snaptrack.htm
> We sell more units of SnapTrack than we do our machined enclosures.
>=20
> You can also search for IEC 60297-3 at http://webstore.ansi.org
> <http://webstore.ansi.org/> to purchase the spec, for example:
> http://webstore.ansi.org/ansidocstore/product.asp?sku=3DIEC+60297%2D3+Ed%2E=
+1%2E
> 0+b%3A1984
>=20
> Here are some more enclosure links:
> http://rosecat.phoenix-mecano.ltd.uk/Index.html
>=20
> http://www.schroff.de/internet/html_e/index.html
>=20
> Regards
>=20
> Bruce


=20
------------------------------------------------
Marc Nicholas - Geekythings, Inc.
Vox: 416.543.4896 SMS: 416...@pc...

Re: [Embedlets-dev] Home Automation space - X10 update

[Ted K. <tk...@ya...>]

Gregg Wrote:

> I have done several MIDP apps for my Nextel phone.  Let me know what you are 
> thinking about accomplishing and I'll try and help out.

Well, how do you feel about playing around a bit with JXME?

http://jxme.jxta.org


Like I told Jac, I think I should be able to integrate your X10 API into the
JackNet Appliance Controller application without too much trouble. 

If we can then get some Java cell phones to remotely control X10 appliances
using this setup, then we will have achieved secure remote access into the home
using a general-purpose and freely available framework.


I know that proprietary solutions exist that have these kind of capabilities
(like https://www.gotomypc.com , which, ironically, is backed by Sun... ) but
none of these proprietary solutions provide an open framework that developers
can freely build upon.

If this actually works, I think that the Embedlets group is in an excellent
position to assist the J2ME community to expand their horizons by enabling them
to start writing MIDP applications targeted at the Home Automation space.


JavaOne session and BOF proposals are due November 21st and it was my initial
thought that this year we would focus on an Enterprise Outpost session and an
Embedlets BOF.  If this experiment looks like it is going to work, however, we
might also want to think about working some J2ME/Home Automation oriented
content into the proposals too.


My plan is to post a JavaOne proposals email to the list sometime on Sunday and
we can talk more about it then.


Ted

__________________________________
Do you Yahoo!?
Protect your identity with Yahoo! Mail AddressGuard
http://antispam.yahoo.com/whatsnewfree

Re: [Embedlets-dev] Home Automation space - X10 update

[Ted K. <tk...@ya...>]

Jac wrote:

> Yep. Got five midp 1.0 phones and one iappli phone laying around for 
> testing purposes. (MIDP implementations tend to have the most anoying 
> implementation differences [some people would call them bugs] so you need 
> to test on each device you want to deploy on.

Wow Jac, I did not know you were into Java Cell phones so deeply!  

Ok, next question then. I know you are a member of the JXTA community, have you
had a chance to play around with the JXME code at all?

If you were able to get JXME running on one of your phones, I think I could
modify the JackNet Appliance Controller to support X10 modules without too much
trouble (using the X10 Java APIs that were recently donated).

This would make a great proof-of-concept demo to show that we really do have
the capability to remotely and securely access the home using a Cell Phone.


Ted

__________________________________
Do you Yahoo!?
Protect your identity with Yahoo! Mail AddressGuard
http://antispam.yahoo.com/whatsnewfree

Re: [Embedlets-dev] Home Automation space - X10 update

[Gregg G. W. <gr...@sk...>]

>Does anyone on the list have any Java Cell phone programming experience?  I
>currently have 2 students that are bringing themselves 'up to speed' on
>CLDC/MIDP but we could probably use all of the help we can get in this area.

I have done several MIDP apps for my Nextel phone.  Let me know what you are 
thinking about accomplishing and I'll try and help out.

-----
gr...@cy...  (Cyte Technologies Inc)

RE: [Embedlets-dev] Enclosure for TStik

[Christopher S. <cs...@oo...>]

Sorry - typo, I was referring to the Saje card from Systronix. It is built
to Euro card standards.
>
> Topic tags:[ARCH][JAPL][WIRING][DOCS][MGMT][STRATEGY][NEWBIE]
> _______________________________________________
>
> Chris wrote:
>
> > Schroff has this kind of Euro-card packaging as well. I have
> used them for
> > some time with success.
> >
> > http://www.schroff.us/home.asp
> >
> > The Sage product line fits this standard nicely. Unfortunatley
> JStik does
> > not seem to conform from what I can see (Bruce?).
>
> I could not find the Sage product page on the schroff site and
> their realtime
> locater gave the following message:
>
> <message>
> Realtime Product Locator
>
> This product is currently not available in our distributor channel. Please
> contact the Schroff sales office for product availability at
> (401) 732-3770
>
> </message>
>
> Do you happen to have a link to the Sage product page?
>
>
> Thanks,
>
> Ted
>
> __________________________________
> Do you Yahoo!?
> Protect your identity with Yahoo! Mail AddressGuard
> http://antispam.yahoo.com/whatsnewfree
>
>
> -------------------------------------------------------
> This SF. Net email is sponsored by: GoToMyPC
> GoToMyPC is the fast, easy and secure way to access your computer from
> any Web browser or wireless device. Click here to Try it Free!
> https://www.gotomypc.com/tr/OSDN/AW/Q4_2003/t/g22lp?Target=mm/g22lp.tmpl
> _______________________________________________
> Embedlets-developer mailing list
> Emb...@li...
> https://lists.sourceforge.net/lists/listinfo/embedlets-developer
>

Re: [Embedlets-dev] Home Automation space - X10 update

[Jac K. <j.k...@th...>]

On Fri, 14 Nov 2003, Ted Kosan wrote:

> Does anyone on the list have any Java Cell phone programming experience?  
> I currently have 2 students that are bringing themselves 'up to speed'
> on CLDC/MIDP but we could probably use all of the help we can get in
> this area.

Yep. Got five midp 1.0 phones and one iappli phone laying around for 
testing purposes. (MIDP implementations tend to have the most anoying 
implementation differences [some people would call them bugs] so you need 
to test on each device you want to deploy on.

Regards,

Jac

-- 
 Jac Kersing            Technical Consultant   The-Box Development
 j.k...@th...                         http://www.the-box.com 

[Embedlets-dev] Home Automation space - X10 update

[Ted K. <tk...@ya...>]

After reading all of the X10 input from people on the list, I ended up
purchasing one of X10's FireCracker kits to get started with:

http://www.x10.com/automation/firecracker.htm


I received a note from my local post office today that they received my
FireCracker kit on Friday and they are waiting for me to pick it up (Murphy's
law seems to raise the probability to 90%+ that a package that I really want to
receive will arrive at the post office on a Friday so that I can not pick it up
until Monday!)


Gregg has kindly donated his FireCracker-compatible X10 Java API to the
Embedlets project and as soon as I pick my FireCracker kit up I am going to
start playing with it and coming up with a plan for placing it into the CVS.

Kelly has also donated all of his X10 materials to the project too and we are
currently working out a way for him to get the zip archive of it to me.


If the FireCracker seems solid enough, I am also going to start integrating X10
modules into the JackNet project.  The JackNet Appliance Controller is a great
proof-of-concept device, but there are SO many inexpensive UL/FCC-approved X10
devices available out there that the opportunity they offer appears too good to
pass up.

I appreciate Kelly's observations about the difficulties that people can
encounter with X10 modules, but it appears that the Home Automation industry
has solved this problem by developing a nation wide network of X10 installers
that are trained in best-practice X10 installation techniques.

Since the Home Automation industry seems to have a handle on this already, I do
not think that we need to worry about it too much.


At this point I think we are safe in assuming that X10 solves the in-home
appliance controlling problem and JXTA solves the secure remote access to the
home problem.

The last thing that appears to be needed in order to have a shot at creating a
marketable technology for this space is secure access to an in-home JXTA edge
peer from a Java enabled Cell Phone.

Does anyone on the list have any Java Cell phone programming experience?  I
currently have 2 students that are bringing themselves 'up to speed' on
CLDC/MIDP but we could probably use all of the help we can get in this area.


Ted

__________________________________
Do you Yahoo!?
Protect your identity with Yahoo! Mail AddressGuard
http://antispam.yahoo.com/whatsnewfree

RE: [Embedlets-dev] Enclosure for TStik

[Ted K. <tk...@ya...>]

Chris wrote:

> Schroff has this kind of Euro-card packaging as well. I have used them for
> some time with success.
> 
> http://www.schroff.us/home.asp
> 
> The Sage product line fits this standard nicely. Unfortunatley JStik does
> not seem to conform from what I can see (Bruce?).

I could not find the Sage product page on the schroff site and their realtime
locater gave the following message:

<message>
Realtime Product Locator

This product is currently not available in our distributor channel. Please
contact the Schroff sales office for product availability at (401) 732-3770

</message>

Do you happen to have a link to the Sage product page?


Thanks,

Ted

__________________________________
Do you Yahoo!?
Protect your identity with Yahoo! Mail AddressGuard
http://antispam.yahoo.com/whatsnewfree

[Embedlets-dev] intro

[James T. <Jam...@Su...>]

<lurk mode="off">

Just want to say hi.

I recently joined this list based on some recent and fun work with Ted.

</lurk>

--
Java == platform independence
XML  == application independence
JXTA == network independence
 
Secure End-to-End Computing

RE: [Embedlets-dev] Enclosure for TStik

[Christopher S. <cs...@oo...>]

Schroff has this kind of Euro-card packaging as well. I have used them for
some time with success.

http://www.schroff.us/home.asp

The Sage product line fits this standard nicely. Unfortunatley JStik does
not seem to conform from what I can see (Bruce?).

Chris

>
> Topic tags:[ARCH][JAPL][WIRING][DOCS][MGMT][STRATEGY][NEWBIE]
> _______________________________________________
>
> Bruce wrote:
>
> > Here are some more enclosure links:
> > http://rosecat.phoenix-mecano.ltd.uk/Index.html
>
>
> Ok, we were looking in the wrong place for enclosures!  What the
> implications
> of going with the Euroboard standard were for TStik, etc. did not
> fully sink in
> until I started reading the links you submitted.
>
>
> And wow! take a look at these extruded aluminum housings:
>
> http://www.rose-bopla.com/PMX/RSBPCatg.nsf/65680fe57e02219d85256bb
c00506958/e5046498ec7af42885256bdc00516cdc?OpenDocument

Rose sells this stuff by the meter and it can be cut to length which is
perfect.  I have not been able to determine prices yet but I sent an email
to
them asking who their USA distributer is so we will find out soon.


Ted

__________________________________
Do you Yahoo!?
Protect your identity with Yahoo! Mail AddressGuard
http://antispam.yahoo.com/whatsnewfree


-------------------------------------------------------
This SF. Net email is sponsored by: GoToMyPC
GoToMyPC is the fast, easy and secure way to access your computer from
any Web browser or wireless device. Click here to Try it Free!
https://www.gotomypc.com/tr/OSDN/AW/Q4_2003/t/g22lp?Target=mm/g22lp.tmpl
_______________________________________________
Embedlets-developer mailing list
Emb...@li...
https://lists.sourceforge.net/lists/listinfo/embedlets-developer

Re: [Embedlets-dev] Enclosure for TStik

[Ted K. <tk...@ya...>]

Bruce wrote:

> Here are some more enclosure links:
> http://rosecat.phoenix-mecano.ltd.uk/Index.html


Ok, we were looking in the wrong place for enclosures!  What the implications
of going with the Euroboard standard were for TStik, etc. did not fully sink in
until I started reading the links you submitted.


And wow! take a look at these extruded aluminum housings:

http://www.rose-bopla.com/PMX/RSBPCatg.nsf/65680fe57e02219d85256bbc00506958/e5046498ec7af42885256bdc00516cdc?OpenDocument

Rose sells this stuff by the meter and it can be cut to length which is
perfect.  I have not been able to determine prices yet but I sent an email to
them asking who their USA distributer is so we will find out soon.


Ted

__________________________________
Do you Yahoo!?
Protect your identity with Yahoo! Mail AddressGuard
http://antispam.yahoo.com/whatsnewfree

[Embedlets-dev] Enclosure for TStik

[Bruce B. <bb...@sy...>]

>Are you saying here that for industrial use, TStik systems do not need to be
>tested at all?

Not clear. I discussed this with Scott and apparently "computing equipment" 
needs to be tested. We should maybe chat with the local lab and ask exactly 
what is included in this.

>The two most likely markets for this to happen in appear to be the Industrial
>Automation market and the Home Automation market and we must simply be 
>prepared
>to do whatever it takes to crack one or both of these markets.

I'd bet that anything sold for use in the home is subject to much more 
regulation.

>ENCLOSURE FOR TSTIK
>
>A number of us here in Southern Ohio have been racking our brains trying to
>figure out the best kind of enclosure to wrap around TStik in order to turn it
>into an Enterprise Outpost.
>
>The two main sources for enclosures we have been looking at are DigiKey and
>Mouser and the top 3 enclosure manufactures that they carry are Hammond, Bud
>and Serpac.
>
>The TStik socket boards are 3.9"L x 3.9"W x roughly 2.75"H and so far we have
>not found a box that looks like it will match this dimension perfectly.  Our
>strategy with a pre-made box is to cut the connector openings in the box using
>a CNC mill.

All our socket boards are Euroboards, 100 mm in one dimension and 100 or 
160 mm in the other. They conform to the international Euroboard 
specifications for size and corner holes. As a result there are hundreds of 
off the shelf enclosures designed for such boards. BICC/VERO, Schroff, 
Rose/Bopla, BUD, and others. You can also purchase a variety of extrusions 
with 100-mm board guides.

>If we can not find a pre-made box, then another alternative is to design our
>own box along with the injection mold that goes with it and if worse comes to
>worse, I will machine the mold myself by hand if needed.

Packaging is not our specialty. With that disclaimer, here's my $0.02:

Any enclosure will have a big impact on FCC testing. Plastic doesn't work 
at all without a shielding spray, usually a nickel compound. Connectors 
which pierce the enclosure would need to be considered initially so that 
they didn't hurt the integrity of the enclosure more than necessary.

For a variety of reasons we are looking at punched and formed steel sheet, 
with pressed in threaded bosses. The top and bottom would mate together in 
sort of a clamshell fashion. In low volume it's relatively affordable and 
the NRE is very reasonable. Such a design can also be customized at low 
cost (say you want 100-1000 units with different connectors, colors, and 
silkscreen). Epoxy paint with silkscreened legends looks good and is also 
affordable. Plus steel is both an electromagnetic and electrostatic shield. 
Our own metal enclosure is on my "todo" list with a number of other things 
ahead of it.

For much higher volume, custom molded plastic with shielding spray, molded 
bosses with threaded mounting inserts, etc, would be the way to go. In our 
limited experience, machining plastic is surprisingly costly. Most plastics 
are not easy materials to machine. The enclosures we sell now are machined 
plastic, I think it's polystyrene. We lose about 5% of the raw stock to 
machining accidents when end mills grab the plastic (which flexes easily) 
and gouge it. Some of this is due to the type of machining we do - pockets 
for the LCD and keypad:
http://www.systronix.com/access/encl_photos.htm
The end panels you see in the photos are actually un-clad PC board material.

So far, solving the enclosure problem hasn't been a high enough priority, 
and we haven't had enough demand to make it higher. I do have some 
prototypes, some sketches in lab books, and notes from talking with a 
couple of local sheet metal shops.

I googled 'Euroboard enclosures" and found this:
http://www.schroff.us/support/faq/techcenter.asp?file=6&cat=6

The euroboard form factor was created to standardize the size of printed 
circuit
boards, which in turn would allow for the use of standardized enclosures. By
working to a standard an OEM can by components from multiple suppliers with
the confidence that they would all be mechanically compatible with one 
another.
Open architecture products that use the Euroboard form factor include, VME,
VME64 extensions, Multibus, MultibusII, CompactPCI, VXI and PXI along with an
untold number proprietary. Designs of OEMs who have taken advantage of the 
mechanical
integrity of the form factor.

The dimension of a euroboard is defined by IEC 60297-3 (1984-01).

1 rack unit or 1U = 44.45mm or 1.75

Standard euroboard heights
Reference U height      Printed board height
3U                      100mm (3.937")
6U                      233.35mm (9.187")
9U                      366.7mm (14.437")

Standard euroboard depths
100mm   160mm           220mm
(3.937")        (6.299")                (8.661")
and other larger ones...

 From the values above you can combine heights and depths but the most 
common euroboard sizes are:
3U (100mm) x 160mm
6U (100mm) x 160mm
6U (233.35mm) x 220mm

Note that TILT is 100x100 which is a standard (but not most common) 
Euroboard size. STEP is 100x160, and so are SaJe, uCAN2, DPB2, HSM320/550, 
etc. The JSimm backplane is a 100x100 mm Euroboard.

All our boards can mount in SnapTrack which then clips on to a DIN rail. 
Many of our customers use this to add our systems to an existing NEMA box 
which already has the DIN rail in it, and already houses other equipment. 
Here's SnapTrack:
http://www.systronix.com/access/snaptrack.htm
We sell more units of SnapTrack than we do our machined enclosures.

You can also search for IEC 60297-3 at http://webstore.ansi.org to purchase 
the spec, for example:
http://webstore.ansi.org/ansidocstore/product.asp?sku=IEC+60297%2D3+Ed%2E+1%2E0+b%3A1984

Here are some more enclosure links:
http://rosecat.phoenix-mecano.ltd.uk/Index.html

http://www.schroff.de/internet/html_e/index.html

Regards

Bruce 

Re: [Embedlets-dev] FCC Testing and Enclosure for TStik

[Ted K. <tk...@ya...>]

Bruce Boyes,

> And some equipment does not 
> need to be tested - everything we sell now, for example. 

Are you saying here that for industrial use, TStik systems do not need to be
tested at all?

Specifically, the Enterprise Outpost we are in the process of developing will
not need to go through any testing?




> What's the end goal here? This isn't the sort of thing you do unless you 
> really need to, and in most industrial settings it isn't required.

The end goal is to find a large lucrative market for Embedded Java ASAP and it
looks like TStik is in a good position to finally have a shot at this.

The two most likely markets for this to happen in appear to be the Industrial
Automation market and the Home Automation market and we must simply be prepared
to do whatever it takes to crack one or both of these markets.



On a related topic...

ENCLOSURE FOR TSTIK

A numebr of us here in Southern Ohio have been racking our brains trying to
figure out the best kind of enclosure to wrap around TStik in order to turn it
into an Enterprise Outpost.

The two main sources for enclosures we have been looking at are DigiKey and
Mouser and the top 3 enclosure manufactures that they carry are Hammond, Bud
and Serpac.

The TStik socket boards are 3.9"L x 3.9"W x roughly 2.75"H and so far we have
not found a box that looks like it will match this dimension perfectly.  Our
strategy with a pre-made box is to cut the connector openings in the box using
a CNC mill.


If we can not find a pre-made box, then another alternative is to design our
own box along with the injection mold that goes with it and if worse comes to
worse, I will machine the mold myself by hand if needed.


But before we resort to this, does anyone have a link to a pre-made plastic
enclosure that might fit TStik?


Thanks,

Ted



__________________________________
Do you Yahoo!?
Protect your identity with Yahoo! Mail AddressGuard
http://antispam.yahoo.com/whatsnewfree

[Embedlets-dev] FCC Testing of TStik systems

[Bruce B. <bb...@sy...>]

At 08:03 PM 11/12/2003 -0800, you wrote:
>From: Ted Kosan <tk...@ya...>
>To: emb...@li...
>Subject: [Embedlets-dev] FCC compliance for TStik
>Reply-To: emb...@li...
>
>Has anyone on the list gone through the process of FCC approving an Embedded
>System?
>
>I am going to start to research what it will take to get FCC approval for 
>TStik
>and any pointers anyone has on this process would be welcomed.

Scott here has been through this process on a product which we both worked 
on at a previous company.

We've almost done this on a number of prior products, so we've researched 
it a bit.

There is a testing lab here locally, so it's more affordable than shipping 
things far away. We can rent the facility (with a test engineer)  for a day 
for $X and do as much testing/changing/retesting of as many products as we 
wish, within that day. Then they charge $Y to file each set of paperwork. 
For an average small product this could all total up to $5K to $10K 
assuming you didn't have to completely redesign the system.

There are different types of FCC testing. Industrial and test equipment is 
not as strict as consumer goods for example. And some equipment does not 
need to be tested - everything we sell now, for example. Power tools, for 
another (motor brushes are great arc generators). So I'd guess that most 
machine shop equipment is exempt. (Can you imagine trying to prevent an EDM 
machine or an arc welder from radiating RF?)

You can only test a complete system, so it's not possible to just test 
TStik. It has to include the socket board, power supply, any enclosure, and 
- here's the catch - any cables which will normally be used. Cables are 
usually the worst radiators. This is why you often see ferrite chokes on 
monitor and other peripheral cables. So I'm not sure how you'd test an 
ethernet hub, for example, since it can't possibly include all the cables 
which users might employ. Maybe there's an exception in such cases.

But the main point is there would have to be a complete system defined with 
an enclosure and at least some of the required cables. And the 
certification would only apply to that specific collection of parts. Change 
any part of it and you'd have to retest. So in the case of using power 
cubes, where we use three different models (for different countries), we'd 
have to test with all three power cubes. I'm not sure if that means we 
would have three different certifications or what... I'll ask Scott 
tomorrow and he can tell me how much of the above I've gotten wrong.

What's the end goal here? This isn't the sort of thing you do unless you 
really need to, and in most industrial settings it isn't required.

We have designed equipment to survive in heavy RF environments, which is 
the other side of this coin, and could also be a consideration around 
machine tools such as EDM machines, plasma cutters, etc.

Bruce

RE: [Embedlets-dev] FCC compliance for TStik

[Christopher S. <cs...@oo...>]

Yep,

It is a tough process that usually runs $10,000 + for the independent lab
that has to certify the device. Basically the box is taken out a remote
sight where field strength measurements are made. Then a report comes back
telling you where in the spectrum you failed... 90% of the time a uP box
will fail unless careful rf design has been applied from the start.

Add shielding, power line filters, io filters, improved layout etc.

Repeat 1 to n times until you pass.

Good luck!

Chris

>
> Topic tags:[ARCH][JAPL][WIRING][DOCS][MGMT][STRATEGY][NEWBIE]
> _______________________________________________
>
> Has anyone on the list gone through the process of FCC approving
> an Embedded
> System?
>
> I am going to start to research what it will take to get FCC
> approval for TStik
> and any pointers anyone has on this process would be welcomed.
>
>
> Thanks,
>
> Ted
>
> __________________________________
> Do you Yahoo!?
> Protect your identity with Yahoo! Mail AddressGuard
> http://antispam.yahoo.com/whatsnewfree
>
>
> -------------------------------------------------------
> This SF.Net email sponsored by: ApacheCon 2003,
> 16-19 November in Las Vegas. Learn firsthand the latest
> developments in Apache, PHP, Perl, XML, Java, MySQL,
> WebDAV, and more! http://www.apachecon.com/
> _______________________________________________
> Embedlets-developer mailing list
> Emb...@li...
> https://lists.sourceforge.net/lists/listinfo/embedlets-developer
>

[Embedlets-dev] FCC compliance for TStik

[Ted K. <tk...@ya...>]

Has anyone on the list gone through the process of FCC approving an Embedded
System?

I am going to start to research what it will take to get FCC approval for TStik
and any pointers anyone has on this process would be welcomed.


Thanks,

Ted

__________________________________
Do you Yahoo!?
Protect your identity with Yahoo! Mail AddressGuard
http://antispam.yahoo.com/whatsnewfree

Re: [Embedlets-dev] X10 and 802.11

[Gregg G. W. <gr...@sk...>]

>> or any other Embedded System for that matter, was a very useful thing to
>> do and that next-generation technologies like JINI and JXTA would be the
>> way to go.
>
>Well, those are nice technologies, but I haven't seen any free JINI
>implementations for TINI lately and it's hard to get started with JXTA for
>most people...

Sun's contributed Jini implementation is free.  The Surrogate project has 
finally been approved as a standard, so it can provide the Jini visiblity of 
anything that can be connected to a surrogate host.  The interconnect 
implementation is the issue.  An IP interconnect was the first thing done.  
There was a bluetooth interconnect being discussed a while back.  You can 
implement an interconnect yourself for some other 'wire' transport.

I am going to try and make some time to play with the surrogate stuff RSN...
-----
gr...@cy...  (Cyte Technologies Inc)

RE: [Embedlets-dev] X10 and 802.11

[Jac K. <j.k...@th...>]

On Tue, 11 Nov 2003, Ted Kosan wrote:

> This is good news!  So it appears to be the web server that is killing the
> system and not the X10 communications code.  

Wellllll, ah, I'm probably just lucky with a run time of well over a
year...  The uptime before I had to pull the plug early last year (moving
acros town was hard with power connected) was well over a year also.

> To be honest, I have never thought that running a web server on a TINI,

Well, it sure beats having to roll your own server every time you need to 
communicate with a device. Nice standard way of communication with plenty 
of clients around (even build into cell phones)

> or any other Embedded System for that matter, was a very useful thing to
> do and that next-generation technologies like JINI and JXTA would be the
> way to go.

Well, those are nice technologies, but I haven't seen any free JINI
implementations for TINI lately and it's hard to get started with JXTA for
most people...

Regards,

Jac

-- 
 Jac Kersing            Technical Consultant   The-Box Development
 j.k...@th...                         http://www.the-box.com 

RE: [Embedlets-dev] Outpost rollout initiative, TStik & JINI

[Jac K. <j.k...@th...>]

On Mon, 10 Nov 2003, Kelly Smith wrote:

> I was able to keep track of about for power modules with the CM11A, but
> operation with the CM11A was erratic as Hell, and with or without the
> CM11A plugged in to the TINI.

Hmmm, just wondering, could the difference be the 220V 50Hz power we're 
using over here?

> Best regards, Kelly (been there, done that with X10s, anf gagged) Smith

Regards,

Jac (still there and enjoying the convenience) Kersing

-- 
 Jac Kersing            Technical Consultant   The-Box Development
 j.k...@th...                         http://www.the-box.com 

RE: [Embedlets-dev] X10 and 802.11

[Jac K. <j.k...@th...>]

On Mon, 10 Nov 2003, Kelly Smith wrote:

> I also have a complete disporation on how to "Bullet Proof" the wiring
> in you home that can somewhat minimize the erratic behavior of the AC
> wiring... Keep in mind that every home will be unique.

Have been using X10 for 6 years now without any problems. The advantage is 
no messing around with the wiring is needed (which is very nice when 
you're renting.)

> I had lights going off and on "Just for the Hello of it" at sporadic and
> un-corralatble times eith the TINI unplugged from the CM11A.

Never seen these problems (yet?)

> In short, people will be disappointed with the results, and the TINI (at
> best) can only handle about four X10 modules (on a good day, with no sun
> spots. 

at least 8 modules over here...

> Real world applications with X10, that might endanger the public, are
> out of the question.

X10 is nice, but not when it might endanger the public. Switching a few 
lights on and off should not be a problem. I wouldn't use it to control a 
coffee maker [only drink tea anyway ;-) ], water heater or anything like 
it.

> In short, X10 is _NOT_ a predictable and "rock solid" solution, and
> there are too many variables to contend with on a case-by-case
> installation in a home (much less a large installation like an
> industrial control site).

Don't even think industrial control and X10. Is has not been designed for 
an application like that. Protocol is far too slow for anything that 
demands near realtime control.

Regards,

Jac

-- 
 Jac Kersing            Technical Consultant   The-Box Development
 j.k...@th...                         http://www.the-box.com 

RE: [Embedlets-dev] X10 and 802.11

[Jac K. <j.k...@th...>]

On Tue, 11 Nov 2003, Kelly Smith wrote:

> Mostly exceptions for "Out of memory" and "Unable to connect to remote host"
> from a browser.  Sporadic at best, and even though I sprinkled som GCs in
> critical sections when the Java code (running on TINI) was idle (i.e.,
> waiting for somtehing to do in a thread). Communication data rates to/from
> the CM11A from TINI's serial port were never a problem.

Hmm, as I stated before, I've got it up and running for over 400 days. No 
out of memory errors yet...

> P.S. You will need a cross-over cable from a TINI/TStik to the CM11A.

Or just use a handmade cable with GND, TX and RX (cross-over for those 
indeed) as all other signals are not important anyway (if they are 
connected at all)

Regards,

Jac

-- 
 Jac Kersing            Technical Consultant   The-Box Development
 j.k...@th...                         http://www.the-box.com 

RE: [Embedlets-dev] X10 and 802.11

[Ted K. <tk...@ya...>]

Kelly wrote:

> (2) Re: "By grunt, do you mean that TINI couldn't pump out serial data to
> the CM11A
> module fast enough?  Could you give a short description of where TINI fell
> short?"
> 
> Mostly exceptions for "Out of memory" and "Unable to connect to remote host"
> from a browser.  Sporadic at best, and even though I sprinkled som GCs in
> critical sections when the Java code (running on TINI) was idle (i.e.,
> waiting for somtehing to do in a thread). Communication data rates to/from
> the CM11A from TINI's serial port were never a problem.

This is good news!  So it appears to be the web server that is killing the
system and not the X10 communications code.  

To be honest, I have never thought that running a web server on a TINI, or any
other Embedded System for that matter, was a very useful thing to do and that
next-generation technologies like JINI and JXTA would be the way to go.


I think it is the discovery of these kinds of insights that gives the Embedlets
community most of its current value.


Ted 

__________________________________
Do you Yahoo!?
Protect your identity with Yahoo! Mail AddressGuard
http://antispam.yahoo.com/whatsnewfree

RE: [Embedlets-dev] X10 and 802.11

[Kelly S. <be...@ea...>]

Dear Ted -

(1) Re: "Go ahead and create the .zip file"

 am doing that now, and trying to minimize the file size as much as I can
(th first attempt as 67MB!).  I now have X10.zip down to 24 MB.

There is downloadable executables for X10 on PCs (Active Home, and
Firecracker) at:

http://www.x10.com/support/support_soft1.htm

...so, this won't be included in "X10.zip" that I am organizing for you.

(2) Re: "By grunt, do you mean that TINI couldn't pump out serial data to
the CM11A
module fast enough?  Could you give a short description of where TINI fell
short?"

Mostly exceptions for "Out of memory" and "Unable to connect to remote host"
from a browser.  Sporadic at best, and even though I sprinkled som GCs in
critical sections when the Java code (running on TINI) was idle (i.e.,
waiting for somtehing to do in a thread). Communication data rates to/from
the CM11A from TINI's serial port were never a problem.

Let me know where I can put the X10.zip file, OK? - Best regards, Kelly
Smith

P.S. You will need a cross-over cable from a TINI/TStik to the CM11A.

-----Original Message-----
From: emb...@li...
[mailto:emb...@li...]On Behalf Of Ted
Kosan
Sent: Monday, November 10, 2003 8:33 PM
To: emb...@li...
Subject: RE: [Embedlets-dev] X10 and 802.11


Topic tags:[ARCH][JAPL][WIRING][DOCS][MGMT][STRATEGY][NEWBIE]
_______________________________________________

Kelly wrote:

> Dear Ted - I can zip my entire "X10 Home Controller Escapades" into a
rather
> large zip file.  This includeds the DalSemi/Motorola demo of using
> Motorola's iDEN phone to communicate to an X10 power module via a TINI
> with a CM11A and a servlet from a PC.
>
> I also have a complete disoration on how to "Bullet Proof" the wiring in
> you home that can somewhat minimize the erratic behavior of the AC
wiring...

If you are willing to donate this code and documentation to the project that
would be great!  Go ahead and create the .zip file, send me an email telling
me
the file's size and then we can figure out a way for me to get the file.

Within the next few weeks I will begin adding some X10 information to the
Embedlets website and making a place for it in the CVS.



>Finally, TINI or TStik (at 1 MB) doesn't have the " grunt" to deal with
>many X10's to support it.

By grunt, do you mean that TINI couldn't pump out serial data to the CM11A
module fast enough?  Could you give a short description of where TINI fell
short?


Ted

__________________________________
Do you Yahoo!?
Protect your identity with Yahoo! Mail AddressGuard
http://antispam.yahoo.com/whatsnewfree


-------------------------------------------------------
This SF.Net email sponsored by: ApacheCon 2003,
16-19 November in Las Vegas. Learn firsthand the latest
developments in Apache, PHP, Perl, XML, Java, MySQL,
WebDAV, and more! http://www.apachecon.com/
_______________________________________________
Embedlets-developer mailing list
Emb...@li...
https://lists.sourceforge.net/lists/listinfo/embedlets-developer