Searching \ for '[EE]: Potting compound to protect IP?' in subject line. ()
Make payments with PayPal - it's fast, free and secure! Help us get a faster server
FAQ page: piclist.com/techref/index.htm?key=potting+compound
Search entire site for: 'Potting compound to protect IP?'.

Exact match. Not showing close matches.
PICList Thread
'[EE]: Potting compound to protect IP?'
2005\09\21@214342 by Jesse Lackey

flavicon
face
Hello all,

I've just finished a product that uses a specialized dc/dc controller
(designed for a wholly unrelated field) with some further tweaks to
solve a particular thorny problem.

It took three major redesigns to get here, 100s of hours, etc.

But... it could be easily copied.  And can't be patented.

So... at the very least I'm going to (carefully) sand off the markings
on the IC.

But it would be great to seal everything up as well.  Only the dc/dc
section need be protected, there is no need to encapsulate everything.
Ideally just a tablespoon of viscous goo poured over it.

Now I realize there are limits to what can be protected, the goal is to
make it hard enough that casual poking around won't yield any clues.
The product is a specialized enough thing that there aren't many
companies in the business field that would be interested in it, so it
doesn't need to withstand some concerted, well-equipped "attack".

Has anyone been in this situation before?

Any advice appreciated!
Thanks
Jesse


2005\09\21@222341 by Mauricio Jancic

flavicon
face
I have recently tried "de-poting" a burned motorola RFID reader. It was a
succesfull job. Its very easy. Just a hot air gun and a screwdriver. Heat it
up and then remove piece by piece of the resin.

Then, I tried to do it in a working device and I have a non-poted working
RFID reader, just for the fun of it... :)


Mauricio Jancic
Janso Desarrollos - Microchip Consultants Program Member
spam_OUTinfoTakeThisOuTspamjanso.com.ar
http://www.janso.com.ar
(54) 11 - 4542 - 3519

{Original Message removed}

2005\09\21@222357 by Chen Xiao Fan

face
flavicon
face

Hot glue is often used in power supply design to secure the
discrete components again vibration. Still it is not good for
your application. So maybe some resin/hardener can be used.

For example, in one fieldbus terminator I developed some years
ago, I use Wevo PU403FL resin and the corresponding hardener.
There are some other types like the Epoxylite EIP 4236 Resin
and hardener which may be better for your application. CY221
is the most commonly used potting compound here and maybe it
is already good for you. It is cheaper than Epoxylite.

Regards,
Xiaofan

{Original Message removed}

2005\09\21@232500 by Chen Xiao Fan

face
flavicon
face
Just take note all the resin/hardeners we used are for functional
use intellectual property protection. IP is a bit confusing to use
here since it does help the Ingress Protection (IP65/67 etc).
I think dedicated attacker will always find a way to circumvent your
protection methods for such power supply circuits. There are plenty
of hackers out there specializing in read out the code protected PICs
or other MCUs as well. I think Microchip has a standard
disclaimers in front of the datasheet named "Note the following
details of the code protection feature on Microchip devices".

The last clause is as following:
************Quote from Microchip datasheets *****************
Neither Microchip nor any other semiconductor manufacturers can
guarantee the security of their code. Code protection does
not mean that we are guaranteeing the product as "Unbreakable".
************End of quote*************************************

Regards,
Xiaofan

{Original Message removed}

2005\09\21@232900 by Chen Xiao Fan

face
flavicon
face
Sorry there are some mistakes in the previous post.
The following is better.
"Just take note all the resin/hardeners we use are for functional
use and not meant for intellectual property protection. IP is a bit
confusing to use here since they do help the Ingress Protection
(IP65/67 etc)."

-----Original Message-----
From: Chen Xiao Fan
Sent: Thursday, September 22, 2005 11:25 AM

Just take note all the resin/hardeners we used are for functional
use intellectual property protection. IP is a bit confusing to use
here since it does help the Ingress Protection (IP65/67 etc).

2005\09\22@045157 by Mike Harrison

flavicon
face
On Wed, 21 Sep 2005 21:43:41 -0400, you wrote:

{Quote hidden}

Sanding off numbers and/or Potting is a complete waste of time for stopping people copying, and adds
to your build costs on every unit.
It may add a few hours or days to the time it takes someone to do it but if they have decided to do
it they will do it anyway.
 

2005\09\22@082742 by Mauricio Jancic

flavicon
face
Yep. I agree... Braking the code of a pic cost between 1K and 3K U$ dollars,
depending on the chip. The process takes 1 week..... That's a lot less that
what you would need to pay for a full development, if you want to do the
thinks the wrong way.

Regards,


Mauricio Jancic
Janso Desarrollos - Microchip Consultants Program Member
.....infoKILLspamspam@spam@janso.com.ar
http://www.janso.com.ar
(54) 11 - 4542 - 3519

{Original Message removed}

2005\09\22@085409 by Electron

flavicon
face

At 09.27 2005.09.22 -0300, you wrote:
>Yep. I agree... Braking the code of a pic cost between 1K and 3K U$ dollars,
>depending on the chip. The process takes 1 week..... That's a lot less that
>what you would need to pay for a full development, if you want to do the
>thinks the wrong way.

any PIC? even 18F and 30F?

2005\09\22@090641 by Mark Rages

face picon face
On 9/22/05, Electron <electron2k4spamKILLspaminfinito.it> wrote:
>
> At 09.27 2005.09.22 -0300, you wrote:
> >Yep. I agree... Braking the code of a pic cost between 1K and 3K U$ dollars,
> >depending on the chip. The process takes 1 week..... That's a lot less that
> >what you would need to pay for a full development, if you want to do the
> >thinks the wrong way.
>
> any PIC? even 18F and 30F?
>

fifty bucks and a UV eraser: http://www.bunniestudios.com/?page_id=13

--
You think that it is a secret, but it never has been one.
 - fortune cookie

2005\09\22@094209 by Mauricio Jancic

flavicon
face
Don't know particularly which one. I requested a quote just to know if a
service was for real and they told me that about pics. They to quoted
precisely for other MCU that I asked, but on pics I was not acurate.

Regards,


Mauricio Jancic
Janso Desarrollos - Microchip Consultants Program Member
.....infoKILLspamspam.....janso.com.ar
http://www.janso.com.ar
(54) 11 - 4542 - 3519

{Original Message removed}

2005\09\22@115242 by Jesse Lackey

flavicon
face
Sorry I should have been clearer... the PIC in the project does
virtually nothing.  The IP is the switcher design.  The point is to make
it hard to identify what chip is being used.  I could care less if the
PIC code is extracted.

Mike, do you have firsthand experience with a product that was de-potted
and cloned, or are your comments anecdotal?  I personally don't see how
this design is going to be quickly reverse-engineered without IC
markings on the key part.

J


Mauricio Jancic wrote:

{Quote hidden}

> {Original Message removed}

2005\09\22@121558 by Alan B. Pearce

face picon face
>I personally don't see how this design is going
>to be quickly reverse-engineered without IC
>markings on the key part.

The key to this is making assumptions about such parts once you have reverse
engineered the circuit diagram. The connections to the chip tell you a lot
about the part, and them some "back of cigarette packet" scribblings about
estimated requirements gleaned from watching waveforms soon narrows down the
likely item.

One way people attempt to get around this is to use a standard chip rotated
inside a package so it is no longer a standard pin out, but often this can
be recognised from the pin out sequence, especially if the production
company is known to have a relationship with a certain chip manufacturer.


For this reason you see why (in other types of circuits) people have a
tendency to try and hide IP inside FPGA devices.

2005\09\22@122336 by Mark Rages

face picon face
On 9/22/05, Jesse Lackey <jsl-mlspamspam_OUTcelestialaudio.com> wrote:
> Sorry I should have been clearer... the PIC in the project does
> virtually nothing.  The IP is the switcher design.  The point is to make
> it hard to identify what chip is being used.  I could care less if the
> PIC code is extracted.
>
> Mike, do you have firsthand experience with a product that was de-potted
> and cloned, or are your comments anecdotal?  I personally don't see how
> this design is going to be quickly reverse-engineered without IC
> markings on the key part.
>

http://www.digikey.com/scripts/DkSearch/dksus.dll?KeywordSearch&Keywords=BK570-ND

Regards,
Mark
markrages@gmail

--
You think that it is a secret, but it never has been one.
 - fortune cookie

2005\09\22@124328 by Mike Harrison

flavicon
face
On Thu, 22 Sep 2005 11:52:42 -0400, you wrote:

>Sorry I should have been clearer... the PIC in the project does
>virtually nothing.  The IP is the switcher design.  The point is to make
>it hard to identify what chip is being used.  I could care less if the
>PIC code is extracted.
>
>Mike, do you have firsthand experience with a product that was de-potted
>and cloned, or are your comments anecdotal?  I personally don't see how
>this design is going to be quickly reverse-engineered without IC
>markings on the key part.

Unless the 'key part' is particularly obscure, once the basic function and pinout has been
determined, it is usually not hard to figure out what it is with a bit of searching.
If someone is really determined, they may also try de-capping it to establish the manufacturer -
chips quite often have makers marks on the die.

I have de-potted a few things myself, mostly for curiosity - never took more than a day or so.
Potting can be a real nightmare for production, and is best avoided unless necessary for some other
reason than an ineffective attempt at improving security.



2005\09\23@040811 by Alan B. Pearce

face picon face
>Unless the 'key part' is particularly obscure, once
>the basic function and pinout has been determined,
>it is usually not hard to figure out what it is
>with a bit of searching. If someone is really
>determined, they may also try de-capping it to
>establish the manufacturer - chips quite often
>have makers marks on the die.

Or the underside of the chip.

I have an HP 7475 plotter that was going to be dumped as it was faulty, and
HP wanted something like the cost of a brand new plotter to send it from new
Zealand the USA for repair. We had a second identical working one in the
workshop which one of the techs used to swap chips between them to identify
the faulty chip. Tried to get just the chip from HP but they would not play
ball. When I got it I removed the known fualty chip which had the HP chip
number on the top, and looked at the bottom. It was marked 6805, so I
grabbed a Motorola 6805 processor from some other equipment we had, plugged
it in, voila, working plotter in about 5 minutes.

2005\09\24@110123 by Electron

flavicon
face
At 17.15 2005.09.22 +0100, you wrote:
>>I personally don't see how this design is going
>>to be quickly reverse-engineered without IC
>>markings on the key part.
>
>The key to this is making assumptions about such parts once you have reverse
>engineered the circuit diagram. The connections to the chip tell you a lot
>about the part, and them some "back of cigarette packet" scribblings about
>estimated requirements gleaned from watching waveforms soon narrows down the
>likely item.
>
>One way people attempt to get around this is to use a standard chip rotated
>inside a package so it is no longer a standard pin out, but often this can
>be recognised from the pin out sequence, especially if the production
>company is known to have a relationship with a certain chip manufacturer.
>
>
>For this reason you see why (in other types of circuits) people have a
>tendency to try and hide IP inside FPGA devices.

Which incidentally are also the easiest to clone..

2005\09\26@062941 by Alan B. Pearce

face picon face
>>For this reason you see why (in other types of circuits) people have a
>>tendency to try and hide IP inside FPGA devices.
>
>Which incidentally are also the easiest to clone..

maybe - depends on the type of FPGA. I know there are types around that
claim you cannot read them back, and that attempting to retrieve the pattern
by grinding down the metalisation is claimed to be fruitless as the fusing
does not show a pattern when doing this.

More... (looser matching)
- Last day of these posts
- In 2005 , 2006 only
- Today
- New search...