[TriEmbed] Electric Fence monitoring

Wed Jul 1 06:42:16 CDT 2020

I like the ideas presented for the Electric Fence monitoring.   I've seen the issue of lightning strikes addressed in other products.   Here's what I suggest:  

1) 10 mm separation without solder mask on your PCB between your lightning signal and arduino ground.   Solder mask puts a smooth surface on the PCB.  Without solder mask the exposed PCB insulator (ie the FR4 dielectric) surface appears rough and distance across it is greater.  Increased separation improves the chance that you can attenuate the lightning signal before it causes damage.  

2) Carbon composition resistor with 2 Watt size manufactured by Allen Bradley.  The 2 watt package size gives 10 mm separation.  Allen Bradley gives a carbon composition resistor with no voids.  No voids give no sparks inside the resistor.  No sparks inside give no damage to the resistor.  No damage gives resistance value that remains constant.    

3) An off-the-shelf lightning strike arrestor with a ceramic insulator separating terminals by maybe an inch or two.   Be sure and install these with exposure to rain.  Water on the surface of the insulator reduces the voltage potential required to jump the gap.  My neighbor attached something like these to his electric fence and the result was that his electric fence charger survived lightning strikes.   https://www.tractorsupply.com/tsc/search/lightning%20strike%20arrestor

4) Neon lamps dissipate energy.  For example, https://www.digikey.com/product-detail/en/cml-innovative-technologies-ltd/6838/2216-6838-ND/12089136.  

Except for the off-the-shelf lightning arrestor,  my suggestions are based on my best recollection of conversations I had circa 1991 with an office-mate that designed a 10Base5 Ethernet transceiver.  A system including these transceivers might have 100 such transceivers attaching to a thick coaxial cable running 500 meters outdoors.  The transceiver was designed to possibly survive and certainly fail safe - nobody wanted to electrocute anyone that happened to be typing on their keyboard when lightning struck.  

-John Edward Moore

From: TriEmbed [mailto:triembed-bounces at triembed.org] On Behalf Of Josh Wyatt via TriEmbed
Sent: Saturday, June 27, 2020 9:27 PM
To: paulmacd at acm.org
Cc: TriEmbed
Subject: Re: [TriEmbed] Electric Fence monitoring

So, a couple of folks emailed mentioning lightning strikes - and that's a real risk, especially with a super-long antenna going almost directly into your arduino. One presented solution was to implement some kind of optocoupler, with the LED powered by harvested energy from the strike (clever). Another idea is to use beefier diodes instead of a zener, for your clamping circuit.

I've had a few lightning strikes on my property, within 100 feet of the house. Each time, the nearby strike took out the shunt regulator circuit in my garage door opener, which powers the safety optical beam and sensor, at the bottom of the door. It was basically a 100 ohm resistor in series with a .5 watt, 5.1v zener, to provide a cheap low-current 5.1v supply to the remote sensors. Each time lightning struck nearby, the zener shorted. The first time, I replaced it with the same zener... until the next strike. The next time, I used a beefier (5 watt) 5.1v zener... until the next strike. Finally, I replaced that zener with 8 series-connected 1n4007 diodes, forward-biased. All strikes since then, no damage! 

There are also various spark gaps that can shunt lightning strikes.

Your mileage may vary...

-j

On Sat, Jun 27, 2020 at 7:23 AM Josh Wyatt <dragojdw at gmail.com> wrote:

Your solution looks pretty good to me; I might consider adding a couple of capacitors to clean the waveform up a bit if you're just looking to detect the pulse. Something like a .1uF between R2 and D2, and maybe a .01uF or .001uF from A0 to ground.

Thanks,
Josh

On Fri, Jun 26, 2020 at 11:19 PM The MacDougals via TriEmbed <triembed at triembed.org> wrote:

My electric fence monitoring project is making progress.

Attached is a scope shot of the signal I am trying to monitor.

There are two issues I want to address before putting this signal on the A0 pin of my ESP8266 development board (D1 mini).

The spike to ~5 volts is concerning.  I think I can use a 3.3v Zener diode to suppress this.  Installed between my signal and ground with the banded end away from ground.

The negative lobe of the signal is also concerning.  Can I just use a diode in series here?  Banded end toward the A0 pin.  I can adjust the voltage up a bit to compensate

for voltage drop across the diode.

I have drawn up a schematic of what I think is the solution.

Thanks for any comments.

---> Paul

_______________________________________________
Triangle, NC Embedded Computing mailing list

To post message: TriEmbed at triembed.org
List info: http://mail.triembed.org/mailman/listinfo/triembed_triembed.org
TriEmbed web site: http://TriEmbed.org
To unsubscribe, click link and send a blank message: mailto:unsubscribe-TriEmbed at bitser.net?subject=unsubscribe

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.triembed.org/pipermail/triembed_triembed.org/attachments/20200701/da117b0c/attachment.htm>