AS3935 Lightning Sensor

[lgbeno]

Curious if anyone on the forum has seen this part from AMS:
http://www.digikey.com/product-highlights/us/en/ams-as3935-franklin-lightning-sensor/1966

Looks pretty cool, I can think of a lot of neat projects with it.  Theres no way I'm going to pony up $300 for the simulator which is pretty much necessary if you ever want to debug your code!  They have schematics in the user guide, hardware looks simple enough to make but the big question is what this signature waveform looks like.  I bet a half an hour with a scope and I could figure out how to make my own but getting my hands on one is another story.

Under some stroke of luck is there a EEVBlog member who might have access to this kit and could take some measurements?

[FLL-Freak]

I do not, but I have been using this part for many months. It has some warts, but once you get over them, it works fairly well.

I would not put much confidence in the simulator. Even if you did reverse engineer the signature, it may or may not match real life! I have had strikes that have rattled my teeth be ignored by the unit and some that I could not hear pop up just fine.

My website has some details: www.fll-freak.com/blog

If you do get a few minutes with the simulator, please do share. But I will guess it is a series of pulses that match the normal reverse leader and then the down stroke of most lighting. I have found the timing of these various places on the web.

[vtkkorhanjoh]

Ahh, a user! I have wanted to use this IC for a few months now, i am planning to put it into a project. (It just sounds too cool to miss!) How good is it at accurately detecting lightning strikes? And is the 1-40 km range just arketing bullshit or serious?

[FLL-Freak]

There are several people actively using this chip. Have not seen any commercial use or it.

I think AMS rushed this device to production a bit quick. There are at least two flaws in it that I found. One is that if you configure the device for I2C address 0, the chip does not ACK properly. Best to use a different address or to use the SPI interface.

Second flaw is just incomprehesibly stupid. When you get an interrupt, you must wait 2 ms before reading the register that tells you why there was an interrupt.

As far as false positives and negatives, I would say the chip does fairly well. I have yet to get a false positive. But I have had strikes that rattled my teeth that were considered a "Noise Source" and not a strike.

I am begining to think the antenna is directional and this may be related to the missing strikes.

[kayvee]

I have a couple of them sitting in my drawer, awaiting my weather station project, just as soon as I can find some time. 

Sadly it is also competing with other things on the same list, and then there is the pressure to projects that actually generate income 

[FLL-Freak]

pressure to projects that actually generate income

This is a foreign concept. All my projects consume income.

[vtkkorhanjoh]

So it is accurate enough, the deveice just has some quirky and strange ways of communicating? I want this to be a 100% detecting lightning sensor, where i live we have quite a lot of bad thunderstorms, and I am often out in the wild.
So if i get a fairly omnidirectional antenna with good gain, will that be enough to get accurate and persistent readings?

[FLL-Freak]

First off, I would never count on any electronic device to keep me safe. Use common sense.

The chip does have some quirky behaviour using I2C at address 0.

The chip does have some design flaws that can be worked around.

The chip is NOT a clear 100% never fail sensor. It has false positives and negatives.

The chip may have sensitivity issues indoors, around metals, and electronic devices.

I am not an RF guy (still working on DC!), but this is not a normal antenna. It works at very low frequencies (500kHz) and uses an inductor as the antenna. Not sure you can really make that omnidirection. Not even sure that it is not, but experimentation seems to indicate that it is.