- In Germany, the only radio frequency that is available for things like remote door locks is 433MHz. So, as my remote door locking system is EU certified, I strongly assume (couldn't verify, unfortunately) that is operates at that frequency. .

It will be, 98% probability. This is the European standard low power license exempt frequency for telemetry systems

And indeed: as soon as the empeg was running or in standby mode, the mouse behaved erratically! When I disconnected the empeg from AC power without moving the base station, the effect went away.

This does seem to show that there is some interference from the empeg, I have to admit.

Moving the base station more than a few centimeters away from the empeg also cured the problem, so the empeg doesn't radiate much. However, as the remote door locking sender uses very tiny batteries, it sends a *very* low power signal, which means the receiver in the car has to be *very* sensitive.

It is. The receivers for the 433MHz band data transmission system often have quoted sensitivities of around -107dBm, witch is technical speak for absolutely sod all power. Bear in mind this is for receiving a good signal, as well. Much less power on the right frequency will cause interference, if the transmitter is close enough.

- find out with which power a remote usually sends. Can anybody who experiences the problems try to ask their car manufacturers or even try to find that in your manuals?

The output power varies slightly with application and country. However, in the UK this band is regulated under the MPT1340 exemption, covering both 418MHz and 433MHz, and limits the output power to less than 1mW effective radiated power. One TX module I have used had a power output with a 1/4 wave antenna of something like 500uW. This really isn't a lot.

- try to get access to a measuring receiver to actually measure the radiation (frequency and power) of the empeg. Empeg folks: do you have any data from your FCC tests that could be helpful here?

This has been done. Believe me, this has been done. Ad nauseum. It's a pain in the arse, the tests you have to go through to be able to sell any electronics nowadays. And, unless you have access to (1) an electromagnetically clean isolation chamber, and (2) a good RF spectrum analyser and probes (the one the RF test company kindly lent us for a few days cost them in excess of $30000, and that was the cheap spare), you won't even be able to find the emissions of the empeg over the RF hash produced by everything else around you, never mind measure it properly.

The MK1 was a bit of a bugger due to the USB controller. It radiated dozens of harmonics of the 16MHz fundamental oscillator, with peaks at multiples of 48MHz, all the way up to at least 2GHz. I had a hell of a time trying to kill off these parasitic oscillations during the original design phase, and even though we got it under the EEC and FCC limits, I couldn't stop it completely. The manufacturer of the chip finally admitted they had a problem with the thing, but couldn't really suggest any good complete fix.

However, the Mk2 uses a different USB chip, with a completely different oscillator, and doesn't suffer from the same problem.

That said, any complex computing device will produce spurious emissions. All you can do is try to keep the output level down to whatever standard the thing has to meet. Unless we went to something like Tempest classification shielding, a little RF will leak. In the case of most PC equipment, a lot will leak.

I suspect it's only because of the close proximity and extreme sensitivity of the receivers being affected that allow them to in fact BE affected. The frequency they use isn't one that the empeg intentionally produces, but is either a harmonic of one of the oscillators, the bus clock, or the core clock, or a mixing product of two or more of them.

I don't recall any peaks in the emission spectrum around that frequency, but it might be that a multiple of the 200MHz or so CPU core clock is to blame. It in fact probably isn't a signal on the same frequency, but a fairly broadband one near it, at extremely low power. The box is pretty well shielded, and unfortunately there isn't a lot you can easily do to make it any better at that part of it's job. Most of the RF leakage will be through the front panel, and covering that with a stainless steel sheet rather defeats the object, I'm afraid.

Look on the bright side. If your in-car entertainment system used, for instance, a 433MHz intel processor, you'd have no chance.

One thing does occur to me now that I think about it. Does this effect occur when the empeg is completely shut down, ie with the ignition off? Under these circumstances the only thing that should be active is a small part of the PSU, the PIC that controls it, and the RTC, none of which should produce harmonics anywhere near 433MHz at power levels above the nanowatt range.

When the empeg is running or in standby mode with the light flashing the CPU is active, and if it's that producing the putative interference it shouldn't occur when the unit is off.

This only applies to the MK2 of course. The MK1 is completely dead when the ignition is off, as it uses a different PSU circuit.

Patrick

Opinions expressed in this email may contain up to 42% water by weight, and are mine. All mine.