Analog sensor long range

[CAI_Support]

good point.  however, if the lightening is happening every week, the cost of replacing $50 item can add up quickly.
I thought the analog signal is converted to digital then sending through the long cable. Long cable can introduce a lot of analog noise even without lightening.
 
In military, when transfer signal in noisy situation, it sometimes use +15V to -15V as phase change indicator, since most noise are going either + side or - side, rarely has noise from +15V to -15V. Using that phase change to indicate a bit change, combined with chksum to eliminate noise interruption totally.

 

[BTCAD]

Where I live, we are plagued by summer thunderstorms - and I've observed cloud-to-cloud lightening that's induced a voltage and current high enough to cause a visible arc on a length of wire only several hundred feet long.  For that reason alone I'd favor an RF solution with batteries (and solar if budget is not a problem).

RF link are great but when 50 or 100 sensors are running the same time without any line of sight, we may use MESH network. But mesh network configuration is not an easy task. We can use something like Xbee but i don't know if it's a good idea.

 

[BTCAD]

However,
                 I think i can use something better to make communication between my wc32 and sensors. I can use Coax-lan converter over 6000ft without any problem 30 mbps BW ! 
 
 
So the idea is to connect a pair of lan-coax convert then link a wc8 at the far end. The wc8 can be powered by a little 12v battery and a little solar pannel. I can use a Picaxe 08m2 as a timer, it can power on the the wc8 to send Webset value of sensors, turn on the lan converter then power off both devices. After that, entering in sleep mode to save energy.
 
What do you think about that?
 
This is really easy to do, we use standard IP communication and coax is cheap and easy to use. I already used a lot of coax-lan converter even up to 5800' and it works VERY well. 

 

[CAI_Support]

coax cable is great, you can bury it underground, so that no lightening damage at all.

 

[BTCAD]

Or using a coaxial surge protector 1000j or more?
 
 
Where will the surge protector perform the best? At the one end or at the middle of the span?

 

[rossw]

The wc8 can be powered by a little 12v battery and a little solar pannel. I can use a Picaxe 08m2 as a timer, it can power on the the wc8 to send Webset value of sensors, turn on the lan converter then power off both devices. After that, entering in sleep mode to save energy.

I do this now, and it works well. (Picaxe monitors battery voltage, charge state etc, and turns on WC8, microwave link radio, IP camera and other devices, as required)

 

[rossw]

coax cable is great, you can bury it underground, so that no lightening damage at all.

 
This is a fallacy. Just because it's buried does NOT protect from lightning damage.

 

[BTCAD]

Sounds good. I will try a little demo at home when picaxe will be arrived.

 

[LarrylLix]

Also, unless you pay about $1 per foot it's (coax cable) not likely to be rated for underground usage.
 
Long cable shields should never be grounded  (earthed) at both ends. Differences in ground potential will induce currents into the signal conductors.
 
Power equipment faults and ground leakage (including lightning hits like RossW indicated) will be shared by the cable shield and can burn the cable up. The buried part may be protected from lightning noise, somewhat, but other problems can be created.

 

[LarrylLix]

Use MOVs across the conductors and ground the shield at the major electronics end only (CPU board and PS) . Use Movs across conductors and not to ground at far end of cable. Use  10,000 or 20,000 joule units.
 
Do not create an opportunity for lightning surges to find a remote path. Try to get them to ground as near to the source as possible.
 
Put up a lightning rod to discharge lightning build-up elsewhere, not through your equipment.
 
 
EDIT: I see Ross had already covered this stuff previously.
 
  Consider a pair of modems back to back, with isolation protection and  conversion circuitry. You can't run microwatt signals from a WC8 board for 2km. Induction will destroy the board almost every time there is a lightning strike or power grid fault in the neighbouring area.

 

[rossw]

Power equipment faults and ground leakage (including lightning hits like RossW indicated) will be shared by the cable shield and can burn the cable up. The buried part may be protected from lightning noise, somewhat, but other problems can be created.

 
Close-by lightning has two effects on cables.
1. Directly induced voltage. This is a function of the MAGNETIC field. The orientation of the cable to the lightning (and therefore the magnetic field), and the proximity are the primary influence on the induced voltage. Being buried in the ground makes almost NO DIFFERENCE to the magnetic effect. The ONLY way to avoid it is the use of proper magnetic shielding. (Eg, steel conduit)
 
2. Potential difference to ground. This is a function of the conductivity of the medium around the cable - (in this case, the ground itself). Current from the lightning bolt will be conducted away from the strike site in a typically uniform, radial pattern. Lets say the ground has a resistance of 1 ohm per metre. A lightning strike discharging 50KA may be passing easily 1000 amps though a given piece of soil, thus developing 1000 volts difference across a mere 1 metre of distance. (This is how cows have been electricuted by lightning without even being struck directly). If you have a cable that runs 1000m and it just happens to be the wrong orientation, you can easily have MANY THOUSANDS of volts potential difference from one end to the other, just because of the ground current. This can frequently be more than the cable insulation can withstand, and you get punchthrough. (This is all significantly simplified, the current dissipates exponentially with distance).

 

[LarrylLix]

Exactly! We have also seen his from loose neutrals on a high voltage system. The cows don't like it!
 
Years back we had a linesman detect a bad street light cable with his balls! They dug it up and fixed it exactly where he indicated. Eeeeewh!

 

[BTCAD]

sensors must be used only during winter. So I think i will only disconnect both ends cables.  But it was interresting, I thought that both end grounded gived a better protection.
 
Does a pair of twisted cable will be abetter choice?

 

[rossw]

Does a pair of twisted cable will be abetter choice?

 
Yes, No and Maybe.
 
It depends on the type of signal you're trying to send, the construction of the cable, how it's shielded and protected etc.
Twisted-pair helps reduce the common-mode noise.
If either (or both) ends are isolated (eg, like ethernet which is transformer-isolated), that will help further.

 

[BTCAD]

ok.
 
 
Xbee looks like to be easy to use with picaxe. I'm confused now :s There are a lot of possibilities to do this project.