[RFI] Help on speed drive motor controller RFI

Ian White GM3SEK gm3sek at ifwtech.co.uk
Mon Jan 29 03:54:01 EST 2007

Tom Rauch wrote:
>> Ask to see the documents for the declaration of "CE"  conformance, 
>> apply to every country in the European Union.
>> On the face of it, it seems impossible that such a system  could meet 
>> regulations for conducted emissions of 15kHz and its  harmonics 
>> using RFI filtering.
>Having had first hand experience with this exact problem, put me on the 
>record as disagreeing. A controller can easily pass conducted emissions 
>in CE or FCC testing by a wide margin and be horrible near an amateur 
>or commercial HF receiving system.

Please read what was actually written, Tom.

You have taken the discussion in a different direction, making 
additional points that do not disagree with anything I said.

The original writer, EA5RS, wrote (emphasis added): "The drive has NO 
RFI FILTERS fitted." Taking him at his word, I replied that it seems 
impossible that a system containing literally NO RFI FILTERS could meet 
CE emission standards. If this system has been incorrectly or falsely CE 
certified, that would give Juan some powerful leverage.

But if he actually meant maybe it might contain some filtering, but it 
isn't *effective*, then that moves us on to the next range of issues 
that you discuss.

On those issues, please put me on record as AGREEING with all of the 

>I've cleaned up speed controllers (three phase from a German company) 
>and filter and suppression help from the manufacturer is useless. They 
>don't understand radiated noise, and the filters they sell have 
>excellent suppression below the HF bands but don't do much at all at 
>HF. The fact it is FCC approved or CE approved doesn't mean a thing 
>when the controller is near radio gear, especially if it is used in or 
>near an HF system.

That's certainly true. Such product standards are grossly inadequate to 
protect weak-signal reception of all kinds. As we regularly witness, all 
the political leverage is in the opposite direction, towards relaxing 
standards and decreasing protection.

It's also very true that many engineers don't understand EMC or RF. 
There's a whole new generation of electrical engineers who can do clever 
digital things involving load currents of tens up to thousands of amps, 
but have no idea what havoc they can cause to the RF spectrum. They have 
even less clue about how to prevent it.

It then falls to a generation of more experienced engineers to fix the 

>I've made systems I've worked on so noise-free you can hold a receiver 
>next to the case and not hear the controller, but I had to design 
>custom filters in RF tight boxes for the motor leads, control, and 
>power leads. The entire controller had to be enclosed in a properly 
>designed box with copper screening over the vents and everything 
>entering and leaving through multi-section shielded filters. The 
>filters also had to withstand the high transient voltages and high 
>starting currents of the inductive load while not causing the peak 
>current to be out of spec for the controller. It was designed to look 
>into an inductive load and there were cautions about using it into a 
>capacitive load beyond a certain capacitance.
>The power line is not the issue, the line to the load is.

Jim already put his finger on that problem of long lines to the load, 
and also the introduction of ground loops.

This highlights a further weakness of product standards: they are 
supposed to apply to the complete installation, but it's obviously 
impossible to put Juan's whole apartment block into a test chamber.

Thus the mandated tests focus on the individual system components, and 
fail to confront the installation issues that can be much more 
important. Some standards require only a small token length of attached 
cables,  while others completely ignore the problem. This can easily 
create situations where a subsystem like a motor controller can meet the 
strict letter of the standards in a test chamber, but will cause serious 
RFI in any practical installation.


73 from Ian GM3SEK

More information about the RFI mailing list