[Elecraft] 14.03 MHz Continuous Tone/Carrier?@#$

[Alan Biocca]

My experience with network gear is that different designs use 
different frequencies for the oscillators and such, so it is likely 
that both of you are correct! Newer gear tends to have higher 
frequencies and lower power in the general trend, so changing or 
upgrading gear may help the level of interference. Interestingly, 
lower priced gear generally uses higher levels of in-chip integration 
and has less effective antenna area. Of course really low cost gear 
may skip some filtering, so testing is really the only way to tell. 
Perhaps we should collect info on which switches seem to be better or 
worse and see if there is a correlation between different people's 
experience and maybe provide some useful info.

If the birdies were net traffic related they would vary over time, 
and probably would not be referred to as 'birdies' which implies a CW 
note. The thing we are most likely to hear is the oscillators feeding 
the chips - elements with higher power and more radiative capability 
over multiple traces on the pcb, etc. The chips that do the 
encode/decode generally develop the different frequencies for 
different network speeds internally, so those signals would be lower 
power and have negligible antennas and not tend to be radiated. 10 
and 100 megabit twisted pair network signals are transformer coupled, 
and with the FCC certification requirements of this gear those 
transformers are likely designed to help reduce EMI at HF, I don't 
know the precise characteristics, but common mode EMI is something 
they are designing to avoid. I have not looked at the details of Gig 
hardware interfacing yet, but the requirements of network ground 
independence generally demand isolated connections, so they are 
likely transformer coupled as well, and are also designed to reduce 
EMI to meet specs.

None of the noises I hear seem to be net traffic related. Those would 
tend to be very brief most of the time.

The design of these network system components varies with their 
bitrate, manufacturer, model and age. Higher bitrate capable devices 
tend to have different basic oscillators. In general the higher speed 
devices have higher frequency oscillators which are often above HF.

Each link has multiple speed capabilities in modern switches, 
independent of the other links.

As Leigh hints, Many stations have inadequately balanced feedlines 
(whether coax or balanced lines), so noise and carriers in the shack 
conduct out to the antenna and come back into the receiver rather 
easily. An effective balun can help here, and ferrites on the lines 
coming out of the switch.

Distance to the antenna and radio all matter to some degree, as both 
radiated and conducted RFI are attenuated by distance. Distance to 
the radio should not matter, but many stations use inadequate baluns 
and so are quite susceptible to common mode noise and EMI at the 
radio. (Note that no balun at all is also an inadequate balun). Jim's 
paper, link below, has a lot of good information in it, and agrees 
that most baluns are inadequate (less than 5,000 ohms choking Z).
.
In my shack these birdies are all extremely weak and readily covered 
by minor band noise. Two switches and the cable modem and cable entry 
and several computers plus the wifi are all within a six foot sphere 
including my HF station, and the antenna is 30 feet away fed by 6 
feet of coax to a substantial balun to 30 feet of balanced line to 
the lower leg of the inverted L fed against radials. Just a data 
point, each station will be different. Network gear here is Linksys 
and Netgear both of 100 megabit fairly recent vintage. Cable modem is 
probably the only 10 megabit device, though perhaps the networked 
printer is also 10, also within this six foot sphere.

As Leigh mentions, the best way to force things to a certain speed is 
to use a 10 megabit hub. However this does force to half duplex which 
is a pretty low level of performance. Adequate for your internet 
connection but it will take a long time to do your backups or 
transfer large files over it. Also note that hubs that do 100 
megabits often have a funny mode where different ports run at 
different speeds and the hub speed shifts, so you will have a real 
mix, and of course switches negotiate each port independently. The 
net speed is 10 megabits when these hubs have at least one 10 megabit 
device connected, and the 100 megabit devices are forced to wait 90% 
of the time so the hub can translate their data to the lower speed.

Another technique is to put things on wifi and reduce or eliminate 
the wired network. 2.4 GHZ doesn't generally bother HF.

73,

-- Alan, wb6zqz


At 11:52 PM 8/14/2007, Leigh L. Klotz, Jr. wrote:
>Jim,
>I know you and I have a different understanding, that he birdies 
>come from 100Mb (my belief) and that they come from 10Mb (yours), 
>but I would like to straight out a couple of issues, perhaps minor ones.
>
>1. I didn't write the second quoted paragraph below about 6' 
>distance from the radio; someone else did, even though your message 
>has my name above it.  You are of course correct on this point, but 
>I didn't make the point.  (I suspect your RFI-Ham PDF file does say, 
>though, that if you have not taken care, your rig's coax and other 
>surrounding objects may be part of the antenna and so noise from 
>something in proximity to the rig may be an indicator of other problems.)
>
>2. Regardless of whether the birdies come from 100BaseT or 10BaseT, 
>if you force your equipment all to the same, your switch will be 
>forced to use the same.  And better yet, if you use a hub, forcing 
>one jacked piece of equipment to 10MB (or 100MB in your view, 
>doesn't matter for this case) is enough to force all to the same.
>
>I've asked a number of ethernet luminaries to explain to me the 
>source of the birdies, and gotten blanks.
>
>Leigh/WA5ZNU
>
>
>Jim Brown wrote:
>>On Tue, 14 Aug 2007 15:25:06 -0700, Leigh L Klotz, Jr. wrote:
>>
>>
>>>Also 21.060.  It is from 100 megabit ethernwt devices.  Switch 
>>>your shack to 10 megabit ethernet and it will be only your 
>>>neighbors QRP transmitters with end-fed long wire antennas (CAT5 
>>>wiring) that you hear.
>>>
>>
>>Not that easy. If your 100MBit system is connected to a 10 MBit 
>>device (like most Internet modems), it will carry that traffic as 
>>10 MBit traffic, and you'll hear the birdies.
>>See my RFI tutorial for more details and fixes.
>>http://audiosystemsgroup.com/RFI-Ham.pdf
>>
>>>I have two 100 megabit switches within six feet of the radio, and can
>>>barely hear the spur on 14.030, so experience varies.
>>
>>No, experience doesn't vary, it has NOTHING to do with proximity to 
>>your radio. What matters is proxmity to your ANTENNA, the degree to 
>>which that RF trash is suppressed by the router, and the ANTENNAS 
>>connected to the router (the Ethernet cables and the power cable)!
>>
>>73,
>>
>>Jim Brown K9YC