EMI and ESD - The Ghost in our Machines

10/12/2009 1:15 PM EDT

There are many ways for an embedded system to go SNAFU! Anyone out there that can make claims to creating a perfect system on the 1st pass is ... well ... being disingenuous. However, this does not mean that it is OK to let stuff like this go.

As we live in an age where lower cost is always better, and with the inexorable race to the bottom; more stuff like this will happen.

It is a miracle that a lot of stuff out there works. The poor engineering that I have seen go into products is truly astonishing ... and it still appears to work!

The big question I have is:

Why does a low-tech device like a stove have a microprocessor in it???

10/12/2009 2:54 PM EDT

Hello Jack,
It sounds to me as if there might be some "Muntzing" going on. For you young folks, Muntz was a consumer electronics manufacturer in the 1940's & 50's. The company gained some notoriety in the engineering community because the owner would prowl through the engineering department & snip components out of the circuits to see what would happen, with the goal of making the circuit as cheaply as possible. Unfortunately, this did not always result in stable operation, but it made the TV cheaper, so it was accepted.
I think that we have something similar here.
In my own line of work, aftermarket performance ignitions, one of the major headaches is getting them to lie with some pretty horrendous spark gap transmitters (spark plugs & engine distributors) working near the wiring. Sometimes, it seems as if the RF affects the uC directly at the die level!
At any rate - I'm sometimes amazed at how cell phone transmitters seem to be able to get into everything nowadays!
And as far as the question about why a stove needs a microprocessor in it - it makes it easier to implement programmed cooking time and temperatures, eliminate a lot of mechanical contacts, and adds nice display features. These may not be necessary for everyone, but you could say the same for a variety of products, including all the mostly never used stuff on a lots of cell phones!
Dave Telling

10/16/2009 7:47 AM EDT

In my opinion, the main problem is the lack tests by a third-party notified body. There no laws requiring such any longer. Here in Europe for example, such laws were replaced by the huge fiasco known as the CE mark, which means that the manufacturer guarantees that the product follows the appropriate standards. What it does not mean is that the product actually follows said standards.

Ask any salesman if the product he sells will work or not... of course every single company will happily claim such and put some ugly CE letters on it. Even though they may be absolutely clueless of what technical standards their product is supposed to follow. Or even though they are too cheap to pay for a conformance test.

The authorities typically only call for third-party test when serious accidents have happened, and not when the faulty products are released on the market. This is true even for safety-critical equipment: as long as you vouch for your product and write down on a paper that it won't kill people, it is accepted. Even though your product could turn out to be a doomsday device in reality. No matter the nature of the application, it is really too late to start dealing with the problems once the product is in full production.

The solution is simple: enforce a third-party review of all technical standards applying to the product. The governments should not only list a number of standards as they do today, they should also make sure that all products released live up to the standards.

10/19/2009 5:00 AM EDT

Ask any of the engineers who do embedded what the impedance is?
Why would you want your input port see a low impedance rather than high?
If you use a bipolar transistor as a switch that has series resistor between its base and MCU port, where would you place the resistor on PCB? Close to the transistor or MCU? WHY?
How would you measure crosstalk on your ribbon cable?
Why using interrupts on input lines may be dangerous? What will happen if the input picks up noise and generates much more interrupts than it should? Will the code have enough time to shut down power switches and avoid machine to roll over your fingers?

Interesting questions... Ask them to engineers and see...

I can see from my experience that we are living in “copy and paste” world. We copy from other products in hardware as in software without understanding why it was done like that.

Good advice: Try to unplug all electrical appliances before you leave home.

Or am I getting paranoid??

Maciej

10/26/2009 4:31 PM EDT

I wear an insulin pump that has a small transciever built in to accept readings from my blood glucose meter and to send out a verification that it read the meter correctly. Sometimes when going through the security devices at store entrances and exits, the security device gets triggered. I've never seen any changes to settings in the meter though.

I've also seen products exposed to the effects of personal defense products i.e. tasers that are built to look like cell phones. These generate a high voltage that makes the standard ESD 4KV or 8KV test look like a love tap.

11/14/2011 1:20 PM EST

I was performing ESD on a product a few years back when I noticed that I could change the angle of the ESD gun and change the results while injecting into the exact same screw.

If the ESD gun leaned away from the uP, it passed. If it leaned over parallel to the uP it failed. The issue clearly wasn't the energy going through the screw and into the circuitry, it was the EM fields. The ESD protection devices in the silicon are not going to address this.

I never saw this back in the days of the larger IC geometries. Of course, most of those devices would not switch at ESD speeds. The challenges are changing quickly.