Re: About outta spec signals...

Howard Fidel

I don't know what the FCC does, but I have been through this on the product validation side.  I have worked on several EMI certifications. You need at least 3 dB margin on the spec to get approved without testing multiple devices, usually 3 to show they comply. Otherwise a sample of 1 is enough. It is all statistics. I know of devices with tens of thousands of installed units that are out of compliance. Nothing has happened. No reported EMI issues. I think the FCC will only investigate if there is a complaint. Keep in mind the prior discussion, where the standards actually allow higher emissions for older requirement. I found a DC to AC inverter that was significantly out of compliance in production. I used it in a product and it failed EMI due to the inverter, which was CE marked and FCC compliant. I notified the manufacturer, but they claimed I used it wrong and was my problem! Actually they tested it incorrectly. They used a resistive load, but if you powered a PC or other device with a switching power supply, its input impedance caused high radiation. I believe that the FCC is underfunded like many agencies, and they can only react to problems, not search for them.


On 8/9/2018 9:14 AM, Jack Purdum via Groups.Io wrote:
I guess my point is that, if a commercial producer got a "surprise" inspection of their production run, should that inspection be on just one unit? If the inspection of 100 units shows 1 that is just 1dB out of spec while the other 99 are in spec, should the FCC insist on a design change? That was my point: I don't know the procedures the FCC uses for "passing" a rig. Clearly, any commercial producer is going to pre-test any sample unit going to the FCC for certification and make damn sure it passes inspection. Indeed, I think the FCC would be stunned if a unit sent to them tested out of spec. However, that's an entirely different process than random sampling of units coming off a production line. Is -43.0000000000000000000001dB out of spec? You say it is, I say I don't know what the FCC would do in that case.

As to my example, statistics shows that very few things are absolute. If the manufacturer tests every unit coming off the line and one fails, by all means they should not release the unit into the market. It's like the astronaut who was asked what goes through your mind at the moment of liftoff and he says: You think about the 15 million parts that hold this bomb together, all of which were let to the lowest bidder. So it is in producing anything: The limiting factor is the weakest link in the chain.

If FCC certification is a submitted sample of 1 versus a random sample of 100 conducted by the FCC, statistics says those are two different procedures. I simply don't know which the FCC uses.

Jack, W8TEE

On Wednesday, August 8, 2018, 11:52:58 PM EDT, Jerry Gaffke via Groups.Io <jgaffke@...> wrote:

The regs make it pretty clear, any harmonics or spurs that have more power than 43dB down from the carrier are out of spec.
If you measure exactly -43dB then get a more accurate instrument, I doubt it is exactly -43.0000000000000000000000dB.
I doubt *they* care about exactly, and exactly where *they* would care on a 10W rig I don't know.

If a manufacturer sells 10,000 units and 5 are out of spec, then those 5 are out of spec.
But some of the caps dried out in transit?  Those rigs are still out of spec.
And your graduate level statistics class is not going to do us any good here.   ;-)

Bottom line:  Future uBitx's should meet spec.
On present units it would be best to clean them up, though our power levels are low enough
that they may cause less bother than a 100W rig that does meet spec.
I for one will clean mine up.

Jerry, KE7ER

On Wed, Aug 8, 2018 at 08:06 PM, Jack Purdum wrote:
When I hear a target of -43dB, what does that really mean? That number or lower? That number plus or minus 3 sigma? I don't know and probably don't want to find out. Still, within -2dB of the target? I wonder.

Join to automatically receive all group messages.