Prevent reverse engineering with light sensor?

[jm_sanz]

Hello everybody,

I have seen someof posts about potting for preventing reverse engineering. However, I would like to know your opinion about other ways of doing this. Considering high-end encryptations levels on the ICs. My concern is not about reversing either the PCB layout or the memory inside the chip but about power converter control stragegies that can be observed by oscilloscope probe connections.

Basically my power converter product will be enclosured inside a potting for thermal dissipation so I am thinking of placing a light sensor (either LDR, or IC) that detects if the potting has been removed. If so, the algorith will not run and memory will be overwritten once some supply voltage is detected.

Have any of you have any expertise on something like this? It might not be the best way of doing this so I am open to other ideas and advices. Maybe a preassure sensor? The solutions cannot be very expensive as the product is close to customer electronics.

E.g. placing multiple sensors for redundancy

Thank you in advance,

Miguel. 

[inse]

Once the light sensor has been found, it will be bypassed before applying power…
Have a look at this thread, what efforts people are willing to spend on reverse engineering and bypassing:
https://www.eevblog.com/forum/reviews/dymo-550-thermal-printer-drm-hacking/

[jm_sanz]

Thanks for the reply!

Yes, it looks like people can go very deep into this things. However, I feel like a LDR that is removed would have a different impedance value (Open circuit) and this can also be taken into account. In the end, they could get into it but for sure this would be a pain in the neck for anyone attempting it.

Any suggestions on how to improve the method?

BR,

Miguel.

[PGPG]

Any suggestions on how to improve the method?

Do power converter need such protection?
If someone will find problems to copy it than he can just make the other converter giving the same voltages.

[inse]

Compared to the RFID topic, a LDR is less than a piece of cake or a walk in the park.
Just an idea, not considered practical feasability.
Embed capacity measurement in the microcontroller (as Microchips CapSense) on traces that end in not populated components.
The capacity depends on the dielectric properties of the potting compound.
Calibrated at initial operation to compensate voids or other irregularities.
But whether that would work stable over lifetime…?

[Haenk]

A component approach is - IMHO - nonsense.
You'd better go with a a hardened MCU, with known-to-work read protection. If you are creating such a great piece of hardware that everyone is wanting to copy it, there are protected MCUs with layers above the flash memory that are almost impossible to remove without destroying the contents.
However, that has to be some genius level software control implemented. Otherwise, probing out the black box in a not too complex system is possible.
After all, you can hardly protect your hardware design, there always is a way to recreate it.
Removing IC markings and potting with something nasty, like 2-component hard epoxy with embedded quartz sand might stop most people.

[Zero999]

Don't bother.

Consider filing a patent, if you have anything worth protecting.

Otherwise you're just making the produce more expensive than necessary.

[jm_sanz]

Hello,

That actually is a brillitan idea. Our uC has a capacitive touch controller that allows measuring capactiances of the capacitive sensors connected to its pins.

As you mention, some feseability tests will have to be carried out but it might work.

Thank you very much.