[Flexradio] Odd RFI Problem

Sat Dec 5 14:13:45 EST 2009

Wow.  There is a Ph.D dissertation in there somewhere!

All kidding aside, this is a great post about how to methodically track down a problem and open yourself up to new ways of thinking about a problem.  The scientific method is alive and well.

Scott, I have a question.  Do you shunt the coax shield on the run into your shack from the antenna before the choke?  I have done this with some great success.  When you choke those common mode currents with a choke, giving that potential a low impedance path to ground will help with the efficiency of the choke.  On my coax egress it looks something like this

Antenna--->ICE lightning arrestor (grounded)-->coax shield grounded at bulkhead-->|(enters house)-->1:1 RF choke-->switches & radios

-Tim

From: Scott Myers [mailto:ac8de at ameritech.net]
Sent: Saturday, December 05, 2009 1:55 PM
To: FlexRadio at flex-radio.biz
Cc: 'Dudley Hurry'; Tim Ellison
Subject: RE: [Flexradio] Odd RFI Problem

Thanks for the responses Tim, Dudley and Brian.  I apologize as this is a longwinded e-mail as I went through some troubleshooting.  The eureka moment is near the end, if you would like to skip all the rambling and details.  Bottom line: RFI causes the printer driver to disconnect the printer from the network.  When the printer software re-establishes the connection to the printer, it causes PowerSDR to freeze.  RFI source has not been determined.  Read on (painfully) for details.

Well, I did some more detective work.  I moved my choke outside my shack.  It was right in the shack before the feedline went outside.  The choke I am using as a Balun Designs 1115ud.  The details of which are at:
http://www.balundesigns.com/servlet/the-63/balun-designs-1-cln-1-Isolation/Detail
The choke is now about 15 feet away from the shack.  Probably about 20 feet from the amp as the coax runs.  This made no difference.  So I added a second 1115u (not D version) choke back inside the shack.  This also had no effect, but I am choked to the max to stop any common mode, that is certain!

For the record, the antenna in use is a Butternut HF9V-X that is about 100 feet from the shack/house.  The shack is in the basement below grade.  The feedline is pulled through conduit underground below the very extensive radial field; 64 radials 40 feet long.  I use 9913 foam (RG-8U) coax throughout my system.

I do use the DPC latency checker.  My latency never gets above 100 uS under the most busy clicking sessions.  The machine is built for quality and speed.  Even under this dropping of communications with the Flex post TX, the latency doesn't rise, so that has been ruled out.

I have added another ferrite choke to the firewire cable that was large enough to loop the cable through twice.  This was no help.  I put this at the computer end, as the two on the stock Flex provided firewire cable are at the rig end currently.

Another observation of note; The TX must be in excess of 20-30 seconds for the problem to occur.  On a short TX, the problem doesn't occur, even at PEP 750 watts out on RTTY.  The networked printer I mentioned doesn't lose communications for until at least that 20 second mark is reached also.  It is as if the RFI has to build up.  Seems odd.

Certainly, going to a better grade firewire cable is a good idea and I may do this.  But I will work on getting the RF out of the shack first, as I should.  I want to try nipping it at the source.  I need to get myself a RF sniffer in here or build something capable of doing it.

I changed a 2 foot coax jumper between the transmatch and the SO-239 on the wall to a 60 foot cable I had lying around.  No change.  So feedline length had no effect.  I changed back to the 2 footer.

As far as grounding the computer to the station, I read the article and was hesitant at doing this.  WAS is the operative word there.  I have taken some care to insure that the house AC line grounds are not common with the RF grounds and my metal sealtight conduit that runs up the wall and out of the shack.  My thoughts would have been before this that by grounding the stations RF ground to the computer, I am making the computer the common point between the house ground and my station's RF ground.  I would have thought it would be necessary to use a converter plug to lift the ground of the computer from the house's AC ground circuit.  But upon further consideration, I got out what schematics I have of equipment.  I noted right off in the schematic for my AL-80B that the AC ground is common with the RF ground of the SO connectors ,at least by its schematic.  Hmm.  I do not have the schematic for the Flex-3000, as it is not in the manual.  But looking under the hood, (and correct me if I am wrong) it would appear that the AC ground is common with the BNC connector's ground, making RF and AC grounds common in this too.  So while we all take great care to try to eliminate ground loops, it would appear that the standard practice (probably due to UL and NEC requirements) in ham gear is for the AC ground to be common with the RF ground point on the back.  So while I have tried to isolate my house ground system from my RF ground system, those efforts would appear to be for naught, as it is all made common through the rig and amp anyway.  What does it mean?  Probably nothing.  This was just a revelation that I didn't realize and from an engineering standpoint and something I hadn't considered before now.

The big weakness in my shack is the length of the my main ground strap with runs to my RF ground system.  It too is 15 feet long and runs out that conduit along with my feedlines.  All components attach to the strap it in a star pattern with equal length braided 1" wide conductors.  The entire ground strap is a 1" braided strap.  Not ideal.  One theory would be that I have gotten to a situation that a resonance/high impedance is set up on it and now that more power is added, it is rearing its ugly head.  Looks like I could be drilling into the basement floor.  Still a few things to try, but it's not looking good for the home team in avoiding that intensive bit of labor and expense.  If I do so, 3-4 copper pipes and 3" wide copper strap will be involved.

So now to drill a hole in the back of the computer and add a strap and grounding screw.  Such fun.  But maybe I'll get lucky and this will solve the problem and I won't have to reinvent my RF ground system.

I added a big snap-on MFJ ferrite to the printer's network cable at the printer end with the cable through it 4 times.  No change on helping the printer's connection.  At this point, I almost overlooked that fact that when I plug the printer back into the network that the PowerSDR will freeze!  PowerSDR isn't doing anything but sitting there idling in RX mode.  There is no rise in DPC during this.  So I watched closely after a TX and it is true that PowerSDR only freezes when the printer re-establishing its connection with the computer's driver.  (It's an HP Officejet 7310 running their Hp Solution center driver software)  If I leave the computer unconnected from the network where it doesn't lose communications because it isn't connected in the first place, then PowerSDR never freezes!!!!  Just turning it off works too.  Eureka!  So it seems I need to get rid of the RFI to keep some parts of the network from having problems that then stops the printer driver and then freezes PowerSDR upon reconnect!  But in the short term, I can at least unplug the printer/turn off and keep using RTTY on 20M and not have any PowerSDR freezes.

So I can now conclude that something about the printer software re-establishing communications with the printer after being interrupted by RFI causes PowerSDR to freeze, even though there is no DPC latency issue.

There are some of the network wires that run through the wall within 4 feet of the coax conduit run.  I estimate that the conduit runs within 3 feet of the data closet where the DSL modem, router and switch reside.  I have now tried ferrite chokes on every part of the printers network cable, tried it on different wall connections and put ferrite on the printer's power supply cable, both on the high voltage and low voltage side.  No changes.  I must now look for the source of the RFI.

Ok, I'm now officially mentally exhausted for the day with this one but I can now operate RTTY with the amp on 20M without PowerSDR freezes.  Now to sniff out where the RFI comes from exactly.

73,
Scott AC8DE