Cumulus Support

I've found some 4 core shielded cable but it's not twisted pair. I've got some telephone cable which I'm pretty sure is twisted pair. That's what I've been looking for. I cabled the house for telephones many years ago but now we used DECT phones. As I recall I had 10 or 15 metres spare. Actually, I think shielding may be more important than twisted pair.

Hi Gina
When I suggested earthing the pole I think my idea of earthing and yours were quite different. I am not suggesting that the lack of a better earth than you suggested is the reason for your problem, but only saying that an earth may not be an earth as far as RF interfernece is concerned. I noted a comment by another poster saying your 9" spike was not an effective earth. I remember many years ago as a Electronics Technician with the Royal Australian Signal corp construction squadron I had to put down a signal earth. This earth consisted of about 10metres of copper about 50mm" x 6mm long with laterals of approx 1m about every 400mm buried in a mixture of carbon and earth and it was still not sufficient to reduce the radiation from our secure comms room. I tend to agree with most posters that you have a faulty unit as your fault finding is almost got to the dog chasing its tail stage. I know the feeling about not wanting to admit defeat, but without a workshop of test equipment the only chance you have of finding the problem is plain blind luck. Before you say anything, I had my 67th birthday about 2 months ago, so I have experienced the frustration of the intermittent fault many times in my career. I hope you find the problem, but I wonder how much time you can waste before you bite the bullet and replace the unit.
gordonwh40

As I read it, the idea of earthing was to prevent the build-up of static electricity. With nothing to leak the charge away it could build up until it overcame the gap between mast and ground, then a spark would occur producing interference.

19:45 20 Apr 2010 New cable on wind vane unit, mast back up and cable plugged into tx unit. NOTE. Cable shield not yet earthed and no suppression components fitted as yet. I've ordered chokes and also clip-on ferrite beads and hoping they'll come tomorrow.

I put some new batteries in my hand-held oscilloscope and been look looking at the voltage on the cable shield w.r.t. ground. There is a 50Hz field resulting in around 20mV RMS. On top of that is a small continuous interference of higher frequencies but just occasionally I have been seeing pulses of interference going way off the scale. Difficult to tell but I would estimate the burst as 1ms long. This is not the WS transmitter - that's indoors without its batteries. I'm about to solder a wire onto the negative battery terminal ready to connect to the cable shield.

Later... Earth wire soldered to tx unit and brought out through small hole in case. Unit reassembled, batteries replaced and unit installed outside with earth wire connected to cable shield. Indoors, the console was reset to re-recognise the signal and date/time/pressure set up again.

Log:-
10:50 21 Apr 2010 TX unit disconnected and brought indoors. PCB further cleaned and photo taken. Batteries removed and earth wire soldered onto battery negative.
15:45 21 Apr 2010 TX unit reassembled including batteries and re-installed outdoors with earth wire connected to cable shield. Wind and rain sensors connected. Console reset and recalibrated.

Seventeen hours and no spikes!!!





Give it another few hours and we can probably say you have cured it. I don't remember it going this long without a spike before now.

Will you still install the chokes and capacitors to act as filters on the input lines from the wind gear? Might as well do. Then you could disconnect your shield and test them out!!



EDITED OUT LIVE UPDATED GRAPH OF GINA'S SPIKES SINCE IT BECAME OUT OF DATE AND CONTRADICTED THE COMMENTS AS OTHER PARAMETERS HAD BEEN CHANGED AND A SUBSEQUENT SPIKE AROSE BECAUSE OF THAT. This made my comments look even 'madder' than usual.

Yes, looking good so far

The chokes have just arrived in the post so I can make up the rf filters now. I'll connect the capacitor common to the battery negative. Seems a good idea to check out the rf fiters even if the cable shielding has cured the problem. It would give an alternative solution to other people getting spikes. I also have clip-on ferrite beads so could try that first - it would be an easier option for others if it works.

LATER... Removed earth connection to cable shield and put ferrite bead on cable, looped so cable goes twice through bead. Here's my test log giving times :-
10:45 22 Apr 2010 Wind sensors cable unplugged from TX unit. One direction wire had broken - now repaired. Ferrite bead on cable (looped), no earth.
11:10 22 Apr 2010 Wind sensors cable plugged back in to TX unit.

Looks like the ferrite bead approach is a no-go, as we already have a spike. It's not that surprising. I'll make up the LC rf filters next and see how that works.

Having broadband problems here. Can only get sporadic Internet access today. Modem is invisible to the Virgin Media techies. Here I have large scale packet loss and mostly can't download pages, especially if the content has graphics. Still now and again trying to follow your 'travails' with your Watson gismo. I think you had it sorted with the earthed signal cable.

EvilV wrote:Having broadband problems here. Can only get sporadic Internet access today. Modem is invisible to the Virgin Media techies. Here I have large scale packet loss and mostly can't download pages, especially if the content has graphics. Still now and again trying to follow your 'travails' with your Watson gismo. I think you had it sorted with the earthed signal cable.

Oh dear, sorry to hear that

Had to go out this afternoon and only just got back. I've re-connected the shield to earth (battery -ve) and removed ferrite bead.
17:45 22 Apr 2010 Re-connected earth to shield and removed ferrite bead.

I may make up the rf filters tomorrow to try.

EvilV wrote:EDITED OUT LIVE UPDATED GRAPH OF GINA'S SPIKES SINCE IT BECAME OUT OF DATE AND CONTRADICTED THE COMMENTS AS OTHER PARAMETERS HAD BEEN CHANGED AND A SUBSEQUENT SPIKE AROSE BECAUSE OF THAT.

Good point. That's the problem with the live graph. Sorry to mess up your post Back as it was now with the earthed shield so not expecting to see any more spikes for now.

With 17 hours free of spikes before and now another 4, I do seem to have found the answer to the problem - using shielded (screened) cable and connecting the shield to battery -ve (the ground of the tx unit).

I have decided to leave things as they are for the next 24 hours to confirm that the problem is indeed solved. I may try the LC filter sometime but having thought about it for a while, I see little point in pursuing other solutions. I do have other things I would like to spend my time on.

It does seem that the temperature spikes have been caused by EM interference picked up by the long cable from the wind sensors to the transmitter unit and that by using screened/shielded cable this has been overcome.

A ferrite bead on the cable is NOT a solution to the problem.

Hi Gina
You finding that the problem was interference induced into a long cable will be usefull for any of us who are contemplating major extensions to the signal cables. Will save a lot of work and frustration by using earthed shielded cable for everyone else. Congratulations.
gordonwh40

gordonwh40 wrote:Hi Gina
You finding that the problem was interference induced into a long cable will be usefull for any of us who are contemplating major extensions to the signal cables. Will save a lot of work and frustration by using earthed shielded cable for everyone else. Congratulations.
gordonwh40

Yes indeed - and as I am about to extend my wind cable it's particularly relevant for me. So thanks, Gina, for sharing your long saga!

My environment is very isolated, with no other dwelling within 400m, and there are no power lines or any obvious sources of EM interference, so I will try first the unshielded cable that I have to hand - but now I know what to do if spikes appear . . . .

I'll be pleased if my results are a help to others

Here's the temperature plot over the last 42 hours, ie. from when I first connected to cable shield to battery negative.

Conrgatulations Gina. You have certainly identified the cause and solved the problem. I see a tiny blip on your temperature graph at around 0830 this morning. Without the screened cable, I have little doubt that would have been another big spike.

The next question is where is the interference coming from?

Can you record it any way? I'm wondering if your digital oscilloscope can preserve the spikes. If so, maybe you could make a directional antenna and gradually track it down by rotating the antenna over time and preserving the spike traces. If all else fails in the detection and recording business, you have a pretty effective ready made unit in your Fine Offset broadband interference detector.

Some kind of coupling transformer and a directional antenna fed into the socket of the F.O. wind input, and you'd be there, I think. The only hard part would be to work out how to make an effective directional antenna without having any idea what frequency you are dealing with. It is probably low frequency sparking or something, so research into the kind of antennas used for lightening detection might be useful. I once made a Direction Finding set up which worked very well, and it was just a high Q loop antenna with a variable capacitor in parallel with it. On the other hand, you can get a pretty good null with a ferrite rod antenna, so, you could wind a coil onto a ferrite rod, couple that into your detector / recorder (F.O.) and rotate the ferrite over time to identify the line that goes through the signal interference source. Unfortunately, it could be in either direction of the bearing you finally get, but nonetheless, it would give you an idea.

http://www.homingin.com/joemoell/80intro.html