High voltage/ surge resistant galvanic isolator?

[Amper]

Hi,

i have a 4-20mA industrial waterlevel sensor on a 1000m cable running through a forrest. By experience its common for lightning to hit close and cause damage, usually no direct hits but i want to be sure.
The sensor gives waterlevel information to a small powerplant, there is no issues regarding safety but reliability is one as there is no personel to repair anything.
First priority is stopping a stronger surge to damage the PLC or other controls.

Just using usual antenna sure protectors with fuses and gas surge supressors, then filtering and protecing again by TVS diodes works for most cases but a few meters distance would be nice to keep it outside of the Building. Signal wise its easily solveable by two microcontrollers and a bit of fiber but how can i supply the sensor?

Stuff that came to my mind so faar is an external solar panel with supercaps, a motor-generator using a long shaft or laser/LED via lightpipe.
Those things may work for the 500mW but i dont really like any of them.

Any ideas?
Solid state is pretty much a must and life to be expected >100kh

[nfmax]

4-20mA is a high power load these days. Could it operate at a lower duty cycle, say turning on for 1 second every minute? That would make the solar panel + battery option more feasible

[tszaboo]

You want an easy and trusted solution. So you go to IBM Phoenix contact's website and select a 4-20mA isolator with the required protection level. Nobody ever got fired for choosing IBM Phoenix contact.
https://www.phoenixcontact.com/online/portal/pi?1dmy&urile=wcm:path:/pien/web/main/products/list_pages/Loop-powered_isolators_P-13-15-05-06/e5e3e532-9d42-4cdc-aeb7-0913bb7ac5e1

[Amper]

My issue is that i wont get fired but i will have a big loss in income in winter operation a day of outage equals roughly my monthly sailary in my dayjob and a 1600km trip with overnight stay to fix the thing 

Reducing the duty is an option actually, maybe even for power via light. Ill have to check what sort of time the sensor needs to settle.

[T3sl4co1l]

If it doesn't have a remote ground connection, and its insulation hasn't been compromised, then there's only the equivalent capacitance of cable to ground, loading it in the common mode.  That's what, 30nF total? (~30pF/m being ballpark typical.)  And it's a transmission line so for brief impulses (if we take a typical induced lightning waveform of the 8/20 or 1.5/50us type, compared to the >3.3us electrical length of the line, we should expect some TL behavior), the source impedance corresponds to the transmission line characteristic, so, some hundred ohms give or take.  Which, if this is buried, then the impedance may be lower actually, but also very lossy because earth forms the opposite electrode (so we expect lots of losses at high frequency, further rounding-off the impulse); or if suspended in air on poles or whatever, then a higher impedance (~200 ohms?) as wire over ground plane (ground again being earth, but with some air inbetween, the frequency response may be fair).  And for intermediate cases (wire laid on ground? where rats/chipmunks can gnaw it??), something inbetween.  Also, an elevated line may be subject to additional electric fields, because, line of sight and all that.

So, I guess we could potentially expect quite a high voltage at the source (multiple kV?), but the result should be that it has quite a high impedance (low 100s ohms) at much of a distance, so the peak current isn't devastating, pretty ordinary fare for TVSs and such.  Having a grounded receiver is no problem, and all the transients can be filtered out.  Presumably the sensor's bandwidth only needs to be fractional Hz, right?  And it's not sending digital code or anything?

If that assumption (no grounding) is violated, then the signal will both become imbalanced (also violating the current loop assumption), and the source impedance may drop (having a direct connection, maybe induced lightning is able to spark into it instead -- speaking of, sparks act to sharpen a waveform so there may be more high frequency content in this case).  Isolation may be required to deal with this.

Also, if the cable is 24 AWG, it'll have a good 160 ohms total loop resistance.  Pretty big, but well within the compliance range of typical 4-20 receivers.  If we expect the average lightning strike near the middle of the run, that's half or 40 ohms per line as the source resistance, plus whatever impedance it has due to environment.  This sets the minimum, and again isn't much of a problem -- most protective devices are suitable for say 20 ohm surges, a typical test level for industrial and ITE circuits.  (Or, if we take the equivalent of both lines in parallel, that is indeed 20 ohms common mode.)

Tim

[voltsandjolts]

Water level sensor at 1km sounds achievable with battery powered radio transmitter, assuming the duty can be very low, like once per hour or so.

[NiHaoMike]

Have you considered an alternative way to sense the level such as a hose to the bottom that is pressurized with compressed air and then reading what pressure it takes to equalize the pressure of the water? Could easily have multiple for redundancy.

Other than that, a telecom surge protector would probably be the best fit for those signal levels.

[Amper]

@T3sl4co1l

Thanks, i have actually never thought of it as a transmission line but yes, the capacitance will do a lot i guess. The cable is just laid down on the forest floor and not in the air like the antenna i always assumed. Maybe i can help it a lot just putting a large (and low inductance) Capacitor between the conductors at both ends and ground one, essentially like X capacitors.

Also i may have to specify about the cable a bit, since putting one down is not easy, not passable for vehicles and 150m difference in altitude, we decided on a 4x4mm² heavy mining industry cable. Its close to indestructible by anything below a few hundred kg and with sharp edges. It also means at least the DC resistance is very low. Its actually enough to hook up a pretty beefy welder on the other end.
This also explains why radio and pressure is not really an option, radio wont work around obstacles well and usually is not that reliable, also the water source has no power outlet, solar doesnt work well enough in woods and mountains with snow. We actually wanted to install a small water turbine up there and try RF but its pretty much abandoned since freezing will be an issue...
Pressure is a close one but we are not measuring the level over the entire distance, its a 1000m long pipeline thats always filled to the top, you can imagine the top end like a funnel, the sensor helps us adjust the power plant in such a way to keep the funnel full by varying the amount taken no matter how much is coming from the source.
Using pressure for a backup works but the pipes pressure loss depends on many factors and a difference of 10m (1/15 of the total we have) will result in 1/15 loss in power, thats way to expensive so sacrifice

[antenna]

Mechanical solution....

Reservoir "funnel" has a float. The float is connected to monofilament line that is fed through the pipe. The pipe is terminated in a horizontal section equal in length to the reservoir head prior to being connected to the power plant. The end of the monofilament line is attached to small drag cone to catch water in pipe and has a magnet embedded in it. The horizontal pipe is covered in a magnetic field viewing film making the location of the drag cone (thus reservoir head) visible. When the water goes down in the reservoir, the magnet moves further down the pipe along a scale marked on the horizontal section.  Add a screen at the power plant in case the mono breaks releasing the drag cone. If this happens, the magnet will not be visible on the film simplifying the troubleshooting and repair.

[Circlotron]

but how can i supply the sensor?

Masthead RF amplifiers are often powered by the 2 wires that carry the signal. Maybe you could convert the sensor signal to a variable frequency at the sensor end that could be isolated by an inductor in the feedline?

[Amper]

Mh, in this case i can use optics for the signal and magnetics for the power.
Maybe thats actually the easiest way, just make a transformer that has properly separated windings like a high voltage transformer, not stacking the windings ontop of each other. This way i can even separate them by a larger shield (like a metal can) and protect the outside by the usual sparkgaps. In case things go wrong the sensor side will fry but nothing relevant can get through the can.

Mechanical solutions are intersting but honestly not really realistic sadly. There is abrasion in the water (slight but existing), the pipe is .8m in diamteter so there is quite some room for moovement and if it fails, this length of strong cable will close the nozzles of the turbine or even worse, keep them from closing.

[David Hess]

I would first try using isolation transformers and AC excitation with a bridge rectifier at the load.  A transformer can be used at each end and multiple transformer can be used in series for greater isolation.  This leaves designing and building the circuit to generate the AC excitation and return the 4 to 20 milliamp signal.

A sure thing would be delivering the power and returning the signal optically over fiber but the cost would be greater.

[Buk]

Maybe too expensive given your length, but we used stuff like this

to protect overground cables from both lightening strike (with regular stakes and grounding) and rodent attack.

Worth a thought.

[mrpelikan]

These are intended for exactly this problem and for best protection you should have a device at either of cable.
(Maybe something similar is available a bit closer to your location though...    )

https://www.intech.co.nz/product/lpn-ovp-2/

[Terry Bites]

Go wirelss. It will be a lot cheaper than cable. No loops, no surges, full isolation. EG Monit Alta wifi, https://www.monnit.com/products/sensors/interface-meters/0-20-ma-current-meter/

[antenna]

I have a Dakota Alert MURS driveway alarm a mile away. vehicles and deer trigger it multiple times a day. With 6 lithium AA batteries, it runs like that for almost an entire year.  With the small current a wireless system would use, solar is the way to go. I know, you said snow, but with that low of consumption, mount two panels dead vertical, one facing the morning sun and the other facing the evening sun. Snow will not stay on a vertical panel and they, although not in optimum position, will provide plenty of power!

[trobbins]

Perhaps elaborate on what interface you have at the PLC end, and the local earthing scheme there that you have to manage EPR.  As tszaboo indicates, start with identifying the best specified interface and the manner in which incoming cable is bypassed to earth and isolated by series impedance.