Generating ~500Khz pulses?

[Things]

Hi all, I'm mucking around with the cool little AS3935 lightning sensor IC, however being in the middle of winter, storms are a bit hard to come by up here. 

AMS sell a dev kit for this chip, which includes a little board that lets you emulate various noise and lightning strikes, however it's $240, and I've already got the IC wired up and running.



From what I can see, it just has a basic uC and a DAC, and some sort of antenna. The chip says it's frequency detection is around 500Khz, so I'm trying to figure out a way of basically copying the operation of that board. Could anyone help me out/point me in the right direction? I've never done any RF stuff before, so I haven't got a clue where to start.

Dan

[richard.cs]

I suspect the 500 kHz is a bit misleading - to detect lightning it will be looking for impulse events which have a broad frequency spectrum. I would suspect you can trigger it on many kinds of electrical transients - do you just want to verify it works?

Stick it next to an inductive electrical appliance like a motor and switch it on/off a few times, it'll probably think there's a lightning strike when you switch off. If it doesn't (and still picks up real lightning strikes) then it must do some interesting filtering in software. Other things to try triggering it with are camera flashes / strobe lights, car ignition circuits, etc. Anything that produces a fast-rising pulse, naturally since these are all weaker than lightning it'll need to be much closer.

[madshaman]

HV supply and a jerry-rigged spark gap?

The spark should give nice fast rise-time pulse generating lots of broad spectrum RF energy.

[Things]

Thanks for the replies, the difficulty here is that the chip does indeed do some internal processing, and it's very good at detecting false positives. I suspect the micro on their emulator board is creating pulses similar to what the chip is looking for, I just don't really understand the circuitry running the antenna.

Almost looks kinda like a h bridge setup.

Dan

[Stonent]

555 timer set to 500Khz through a flyback transfomer?

[gnif]

Not sure if this might help, but an interesting read none the less...

http://www.electronicproducts.com/Sensors_and_Transducers/Sensors_and_Transducers/Designing_truly_portable_lightning_detectors.aspx

Clearly states that a 500KHz broadband receiver can be used to detect strikes, but that there is additional filtering required to filter out false positives. This goes way beyond what I know/understand so could be way off mark here, but looks like a good start if you were to build an emulator.

[Things]

Hmm, interesting article, thanks. Once I've got some sort of RF frontend going I can control with a uC, those graphs will definitely be useful.

Dan

[madshaman]

555 timer set to 500Khz through a flyback transfomer?

I'm wondering if the processing would detect this as a false positive, the signal that would generate would be purely 500Khz and its harmonics, whereas an impulse like a lighting strike will fill the spectrum more (although I'm pretty sure by the time the signal gets to the chip that looks at it, it's been effectively [just by virtue of rhe frequency response of the environment and the circuit itself] comb filtered and low pass filtered etc.).

I'm curious to find out what the OP discovers.

[PA0PBZ]

I would expect the datasheet to give some more information about the signal verification, but unfortunately all they say is:

8.9.1 Signal Validation
The watchdog enables the lightning algorithm block in the event any activities are detected at the antenna. As this happens the output signal of the AFE is evaluated by the Signal Validation block, which checks the pattern of the received signal. The signal validation checks the shape of the received signal. In particular, the AS3935 can reject the impulse signals, like spikes, picked up by the antenna.

So I guess you have to find out what the shape of the received signal is supposed to be, maybe you could ask the manufacturer?

[edavid]

If they have an Australian distributor, see if you can borrow the board from them.

[Things]

I'll look into it, but from my understanding of lightning strikes, it'll be looking for a weaker "crackly" type signal for the leaders, followed by a large pulse, and then some more successive pulses after that, as shown in "gnif's" link.

My take on that board is they're using the DAC to allow variable output power, to emulate strikes at different distances. Generating pulses shouldn't be out of the realm of say an ATMega I wouldn't think. I'll have to do some playing around and see what I can get