Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss

Inductive digital isolator

126 views
Skip to first unread message

Andrzej Ekiert

unread,
Sep 24, 2012, 6:08:04 PM9/24/12
to
Hi all,
The reason for this post is that I want to safeguard my idea from any
patenting attempts. Releasing the details to the public domain should do
the trick, so here they are:

Imagine a transformer built from two coils routed on a PCB on two
different layers (top and bottom for example). The thickness of the PCB
provides the isolation required. Transmission is performed by shorting one
coil, forming a closed loop - one end is grounded, the other connected to
a microcontroller pin. The micro grounds the pin and that's one state.
When the micro puts the pin in high impedance, the other state is
transmitted. A micro on the other side measures the inductance of the
other (coupled) coil. By observing a variation in inductance it can tell
if the coil on the other side is shorted or not.

Should anyone be interested, I have it implemented. We do this on
Microchip's microcontrollers and measure the inductance with the CTMU
hardware peripheral. Works at 80kbps full-duplex on a PIC24F. I have
written an application note and posted it on my company website. We call
the thing "Freesolator", since it requires almost no components (one
resistor, once you have CTMU on board). We have demo boards and can make
the source code available. See http://protronik.pl/freesolator-en.html

Hope you don't find this post too spammy. I couldn't think of a better
method of undeniable disclosure.

Cheers,
Andrzej Ekiert

Joerg

unread,
Sep 24, 2012, 6:17:39 PM9/24/12
to
Looks like you have re-invented the inductive proximity sensor :-)

--
Regards, Joerg

http://www.analogconsultants.com/

Andrzej Ekiert

unread,
Sep 24, 2012, 6:31:19 PM9/24/12
to
Dnia 25-09-2012 o 00:17:39 Joerg <inv...@invalid.invalid> napisał(a):

> Looks like you have re-invented the inductive proximity sensor :-)
>

Well, there were other inspirations as well. But the application area may
also matter - I don't think this was a known method for data transmission,
so it might have been patentable. Now it surely isn't ;-)

ae

Jim Thompson

unread,
Sep 24, 2012, 7:02:19 PM9/24/12
to
On Tue, 25 Sep 2012 00:31:19 +0200, "Andrzej Ekiert"
<dspi...@tlen.pl> wrote:

>Dnia 25-09-2012 o 00:17:39 Joerg <inv...@invalid.invalid> napisa?(a):
>
>> Looks like you have re-invented the inductive proximity sensor :-)
>>
>
>Well, there were other inspirations as well. But the application area may
>also matter - I don't think this was a known method for data transmission,
>so it might have been patentable. Now it surely isn't ;-)
>
>ae

My method as well... keep it _art_, but _privately_ documented.

I'm expert witness in a case right now where a heathen patented a
block diagram... science fiction, then is suing now that technology
has caught up with fiction :-(

...Jim Thompson
--
| James E.Thompson, CTO | mens |
| Analog Innovations, Inc. | et |
| Analog/Mixed-Signal ASIC's and Discrete Systems | manus |
| Phoenix, Arizona 85048 Skype: Contacts Only | |
| Voice:(480)460-2350 Fax: Available upon request | Brass Rat |
| E-mail Icon at http://www.analog-innovations.com | 1962 |

I love to cook with wine. Sometimes I even put it in the food.

Jamie

unread,
Sep 24, 2012, 8:04:22 PM9/24/12
to
you're too late! way to late.

Jamie

Sjouke Burry

unread,
Sep 24, 2012, 9:39:18 PM9/24/12
to
"Andrzej Ekiert" <dspi...@tlen.pl> wrote in
news:op.wk5uf...@jabba.lan:
Use an ethernet isolation transformer, they come in very small
footprints and are made for the purpose
(datatransmission with isolated in/out).

Robert Baer

unread,
Sep 25, 2012, 12:29:19 AM9/25/12
to
Jim Thompson wrote:
> On Tue, 25 Sep 2012 00:31:19 +0200, "Andrzej Ekiert"
> <dspi...@tlen.pl> wrote:
>
>> Dnia 25-09-2012 o 00:17:39 Joerg<inv...@invalid.invalid> napisa?(a):
>>
>>> Looks like you have re-invented the inductive proximity sensor :-)
>>>
>>
>> Well, there were other inspirations as well. But the application area may
>> also matter - I don't think this was a known method for data transmission,
>> so it might have been patentable. Now it surely isn't ;-)
>>
>> ae
>
> My method as well... keep it _art_, but _privately_ documented.
>
> I'm expert witness in a case right now where a heathen patented a
> block diagram... science fiction, then is suing now that technology
> has caught up with fiction :-(
>
> ...Jim Thompson
To the best of my knowledge, one cannot patent a block diagram - not
even as a design patent.
And design patents are essentially worthless since one minor aspect
can be changed to generate a new design patent.
Crudely put, the shape of a paper clip cannot be patented in a way to
protect its use; change the wiggle or bend here to make a different
design and thus create competition that cannot be negated by "interference".

Tim Williams

unread,
Sep 24, 2012, 11:32:35 PM9/24/12
to
"Sjouke Burry" <s@b> wrote in message
news:XnsA0D92526C1FBEsj...@213.75.12.10...
> Use an ethernet isolation transformer, they come in very small
> footprints and are made for the purpose
> (datatransmission with isolated in/out).

Not nearly as much isolation as FR-4, though -- certainly not reinforced.
With that kind of isolation, you can drive IGBTs on any supply with
impugnity.

Tim

--
Deep Friar: a very philosophical monk.
Website: http://webpages.charter.net/dawill/tmoranwms


John Larkin

unread,
Sep 24, 2012, 11:47:03 PM9/24/12
to
You could put an LED on one side of the board and a photodiode on the
other. Punch a hole in any power/ground planes to let the light
through the FR4.




--

John Larkin Highland Technology Inc
www.highlandtechnology.com jlarkin at highlandtechnology dot com

Precision electronic instrumentation
Picosecond-resolution Digital Delay and Pulse generators
Custom timing and laser controllers
Photonics and fiberoptic TTL data links
VME analog, thermocouple, LVDT, synchro, tachometer
Multichannel arbitrary waveform generators

miso

unread,
Sep 25, 2012, 12:15:52 AM9/25/12
to
The only thing that matters in the patent is the "claims" section.

Incidentally, if you don't want someone to patent an idea but rather
have it open source, just patent it yourself and don't enforce the
patent. There is no better prior art than a patent with that prior art.




Jim Thompson

unread,
Sep 25, 2012, 12:39:27 AM9/25/12
to
On Mon, 24 Sep 2012 20:29:19 -0800, Robert Baer
<rober...@localnet.com> wrote:

>Jim Thompson wrote:
>> On Tue, 25 Sep 2012 00:31:19 +0200, "Andrzej Ekiert"
>> <dspi...@tlen.pl> wrote:
>>
>>> Dnia 25-09-2012 o 00:17:39 Joerg<inv...@invalid.invalid> napisa?(a):
>>>
>>>> Looks like you have re-invented the inductive proximity sensor :-)
>>>>
>>>
>>> Well, there were other inspirations as well. But the application area may
>>> also matter - I don't think this was a known method for data transmission,
>>> so it might have been patentable. Now it surely isn't ;-)
>>>
>>> ae
>>
>> My method as well... keep it _art_, but _privately_ documented.
>>
>> I'm expert witness in a case right now where a heathen patented a
>> block diagram... science fiction, then is suing now that technology
>> has caught up with fiction :-(
>>
>> ...Jim Thompson
> To the best of my knowledge, one cannot patent a block diagram - not
>even as a design patent.

But they did. Your patent office at its finest.

> And design patents are essentially worthless since one minor aspect
>can be changed to generate a new design patent.
> Crudely put, the shape of a paper clip cannot be patented in a way to
>protect its use; change the wiggle or bend here to make a different
>design and thus create competition that cannot be negated by "interference".

We go to court Nov 5

John Devereux

unread,
Sep 25, 2012, 1:53:43 AM9/25/12
to
It might well be the best way to stop someone else patenting it.

But it would not be a good way to let other people use it who might want
to. The mere fact that it is patented would put off most people I
think. What about explicitly releasing it into the public domain?
Although this seems to be discouraged for some reason. A better way
might be to use one of the "open hardware" sites that are springing up.


--

John Devereux

Andrzej Ekiert

unread,
Sep 25, 2012, 2:30:17 AM9/25/12
to
Dnia 25-09-2012 o 03:39:18 Sjouke Burry <s@b> napisał(a):

> Use an ethernet isolation transformer, they come in very small
> footprints and are made for the purpose
> (datatransmission with isolated in/out).

The whole point of this method is not to use any components one needs to
buy. You only pay for circa 3cm2 of PCB space, so building an isolated
data link between two microcontrollers is almost free.

Regards,
Andrzej

BubbleSorter

unread,
Sep 25, 2012, 2:35:41 AM9/25/12
to
Or just use the built in magnetometer on your iPad.

(lots of Android devices too)

Andrzej Ekiert

unread,
Sep 25, 2012, 2:43:02 AM9/25/12
to
Dnia 25-09-2012 o 06:15:52 miso <mi...@sushi.com> napisał(a):

> Incidentally, if you don't want someone to patent an idea but rather
> have it open source, just patent it yourself and don't enforce the
> patent. There is no better prior art than a patent with that prior art.

I did talk to a patent lawyer to estimate costs involved in having
something patented internationally. A costly hobby I would say. And if it
comes to court, then a small company cannot afford to defend.

My opinion on patents is that the whole purpose of this system has been
circumvented in the recent decades. Patents now do not protect individual
inventors. They are only useful to corporations that already have money
and lawyers. Also, the patent databases are polluted with useless or
obvious "inventions". Worse - anything you do, you may fall into a patent
trap. I would like to see this whole system thrown to trash.

Regards,
Andrzej

Jan Panteltje

unread,
Sep 25, 2012, 4:18:46 AM9/25/12
to
On a sunny day (Tue, 25 Sep 2012 08:30:17 +0200) it happened "Andrzej Ekiert"
<dspi...@tlen.pl> wrote in <op.wk6ho...@jabba.lan>:

>Dnia 25-09-2012 o 03:39:18 Sjouke Burry <s@b> napisa=C5=82(a):
>
>> Use an ethernet isolation transformer, they come in very small
>> footprints and are made for the purpose
>> (datatransmission with isolated in/out).
>
>The whole point of this method is not to use any components one needs to=
> =
>
>buy. You only pay for circa 3cm2 of PCB space, so building an isolated =
>
>data link between two microcontrollers is almost free.
>
>Regards,
>Andrzej

Not new I think, you could do it capacitively too, 2 close PCB layers,
1 pull down resistor NO CURRENT

PIC out--- | |------ PIC in
|
R
|
///

Tauno Voipio

unread,
Sep 25, 2012, 1:37:52 PM9/25/12
to
Before you burn all your assets, have a look at Analog Devices
ADUM -series devices.

--

Tauno Voipio

k...@att.bizzzzzzzzzzzz

unread,
Sep 25, 2012, 1:43:23 PM9/25/12
to
On Mon, 24 Sep 2012 20:29:19 -0800, Robert Baer <rober...@localnet.com>
wrote:

>Jim Thompson wrote:
>> On Tue, 25 Sep 2012 00:31:19 +0200, "Andrzej Ekiert"
>> <dspi...@tlen.pl> wrote:
>>
>>> Dnia 25-09-2012 o 00:17:39 Joerg<inv...@invalid.invalid> napisa?(a):
>>>
>>>> Looks like you have re-invented the inductive proximity sensor :-)
>>>>
>>>
>>> Well, there were other inspirations as well. But the application area may
>>> also matter - I don't think this was a known method for data transmission,
>>> so it might have been patentable. Now it surely isn't ;-)
>>>
>>> ae
>>
>> My method as well... keep it _art_, but _privately_ documented.
>>
>> I'm expert witness in a case right now where a heathen patented a
>> block diagram... science fiction, then is suing now that technology
>> has caught up with fiction :-(
>>
>> ...Jim Thompson
> To the best of my knowledge, one cannot patent a block diagram - not
>even as a design patent.

You can't patent a block diagram, per se, and it would be silly to do so. You
certainly can use a block diagram in the teachings. The claims section is
text, so no, a block diagram won't do there, either. Why would you want to?

> And design patents are essentially worthless since one minor aspect
>can be changed to generate a new design patent.

Depends on what it is. Design patents certainly are worthwhile in some
circumstances, like the shape of the iPhone.

> Crudely put, the shape of a paper clip cannot be patented in a way to
>protect its use; change the wiggle or bend here to make a different
>design and thus create competition that cannot be negated by "interference".

Of course, it would be stupid to file a design patent when the function needs
protection.

k...@att.bizzzzzzzzzzzz

unread,
Sep 25, 2012, 1:44:39 PM9/25/12
to
There are places to publish such things that are far cheaper than a patent.

>But it would not be a good way to let other people use it who might want
>to. The mere fact that it is patented would put off most people I
>think. What about explicitly releasing it into the public domain?

See above.

k...@att.bizzzzzzzzzzzz

unread,
Sep 25, 2012, 2:31:19 PM9/25/12
to
They're certainly nice but they cost like they have part numbers starting with
"LTC".

Tauno Voipio

unread,
Sep 25, 2012, 3:13:56 PM9/25/12
to
You gets what you pays for ...

--

TV


k...@att.bizzzzzzzzzzzz

unread,
Sep 25, 2012, 3:53:04 PM9/25/12
to
On Tue, 25 Sep 2012 22:13:56 +0300, Tauno Voipio
Not true. ...on either end of the scale.

Phil Hobbs

unread,
Sep 25, 2012, 7:09:00 PM9/25/12
to
Such as Circuit Cellar.

>
> >But it would not be a good way to let other people use it who might want
> >to. The mere fact that it is patented would put off most people I
> >think. What about explicitly releasing it into the public domain?
>
> See above.
>
> >Although this seems to be discouraged for some reason. A better way
> >might be to use one of the "open hardware" sites that are springing up.

Cheers

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
845-480-2058

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

k...@att.bizzzzzzzzzzzz

unread,
Sep 25, 2012, 8:55:47 PM9/25/12
to
Technical Disclosure Bulletin. ;-)

http://en.wikipedia.org/wiki/Defensive_publication

<...>

rickman

unread,
Sep 26, 2012, 3:22:19 PM9/26/12
to
The Washington area Consultants Network group of the IEEE recently
sponsored a "Patent" town hall discussion with a patent consultant, a
patent lawyer and two representatives of the patent office. One of the
things I learned was that a great way to establish prior art is to file
a preliminary patent application. I don't recall the exact term used
for this filing, so I may not have it correct, but it only costs $300
and clearly establishes prior art so that no one else can ever patent
the idea. I think the "preliminary" filing expires in a year so you
also can't patent it if you don't follow up within the year.

Rick

legg

unread,
Sep 27, 2012, 11:03:20 AM9/27/12
to
http://www.nve.com/Downloads/il71x.pdf
http://www.analog.com/static/imported-files/data_sheets/ADuM5200_5201_5202.pdf
http://www.analog.com/static/imported-files/data_sheets/ADuM4160.pdf
and maybe
http://www.silabs.com/Support%20Documents/TechnicalDocs/si8410.pdf

There are also a number of commercially available chipsets designed to
facilitate magnetic data coupling. I don't know how they compare in
price to a pair of PICs.

Your technique, however, may be original. If it works and saves you
money, use it.

Publication is always an alternative to patenting, but it doesn't
absolve you of the responsibility for patent searching, prior to
commercial use. You didn't mention that this was the case and in doing
so may be placing other implimenters at hazard.

RL

Andrzej Ekiert

unread,
Sep 27, 2012, 2:54:30 PM9/27/12
to
Dnia 27-09-2012 o 17:03:20 legg <le...@nospam.magma.ca> napisał(a):

> There are also a number of commercially available chipsets designed to
> facilitate magnetic data coupling. I don't know how they compare in
> price to a pair of PICs.

I know the market. I have used iCouplers in my designs. I have even been
digging out the information how they do their work internally, to make
sure my method was different. A pair of PICs can be "free", if you need
them anyways to do the other tasks on both sides of the link - in that
case using an additional IC for isolation is always more expensive.

> Publication is always an alternative to patenting, but it doesn't
> absolve you of the responsibility for patent searching, prior to
> commercial use. You didn't mention that this was the case and in doing
> so may be placing other implimenters at hazard.

I also did some patent searching, but you need to remember, that no search
can be complete. There may exist patents missed because:
1) they didn't show up in patent search engines, for whatever reasons;
2) they are filed, but not published yet;
3) there are so many patents to check, that the analysis is practically
non-feasible;
4) the patent in question may have been filed in a country you don't even
thought to check.

Ensuring non-infringement is hardly possible and in my case the search is
more expensive than the expected revenue. Again, I would love to see the
whole patent system thrown to trash, to free myself from that kind of
problems. As soon as I start to be afraid to implement my own independent
invention, the purpose of that system is subverted.

Regards,
ae

josephkk

unread,
Sep 28, 2012, 10:50:51 PM9/28/12
to
I suggest you use Creative Commons as well. It is reasonably funded and
is capable of being proper "witness" should some nutter (patent troll) try
to steal it.

?-)

josephkk

unread,
Sep 28, 2012, 10:53:28 PM9/28/12
to
True, but expensive.

?-)

josephkk

unread,
Sep 28, 2012, 10:56:47 PM9/28/12
to
On Tue, 25 Sep 2012 08:43:02 +0200, "Andrzej Ekiert" <dspi...@tlen.pl>
wrote:

>Dnia 25-09-2012 o 06:15:52 miso <mi...@sushi.com> napisa?(a):
You are a registered member of the disorganized group now.

?-)

rickman

unread,
Oct 3, 2012, 10:18:52 AM10/3/12
to
But for the purpose of the OP, you don't need to follow through on the
patent, just filing establishes prior art. A patent can be expensive to
defend, but you don't need to defend it if you are just trying to make
it available to all, you only need prior art to prevent someone else
from patenting it.

Rick

legg

unread,
Oct 3, 2012, 7:46:41 PM10/3/12
to
On Tue, 25 Sep 2012 00:31:19 +0200, "Andrzej Ekiert"
<dspi...@tlen.pl> wrote:

>Dnia 25-09-2012 o 00:17:39 Joerg <inv...@invalid.invalid> napisa?(a):
>
>> Looks like you have re-invented the inductive proximity sensor :-)
>>
>
>Well, there were other inspirations as well. But the application area may
>also matter - I don't think this was a known method for data transmission,
>so it might have been patentable. Now it surely isn't ;-)
>
>ae

There are also a number of commercially available chipsets designed to
facilitate magnetic data coupling. I don't know how they compare in
price to a pair of PICs.

Your technique, however, may be original. If it works and saves you
money, use it.

Publication is always an alternative to patenting, but it doesn't
absolve you of the responsibility for patent searching, prior to
commercial use. You didn't mention that this was the case and in doing
so may be placing other implimenters at hazard.

RL
0 new messages