Go Down

Topic: RFI with standalone atmega328 (Read 913 times) previous topic - next topic

So, weird things are going on. I out together a standalone arduino for my simple home security system. Didn't feel like having my board hooked up to it all the time. Anyways, when I was using the board, I had no problems whatsoever. But now with a standalone unit, there seems to be some very strange interference. It was running, and I went to the kitchen to make dinner. I have a gas stove with one of those automatic igniters. When it was igniting the flame, I notice my arduino alarm kept going off. I checked this over and over, and everytime the igniter sparked, the alarm went off. I know the voltage that runs those sparks is high, so I thought maybe some weird rfi or emp. That's not where the weird stops. Whenever the microwave door is shut, the same thing happens...
Anyone have any idea wth is going on?? Kinda freaked out here...

I should specify...whenever the microwave door is closed, it sets it off. Even weirder, I have an LCD display to show if there is motion outside or if someone opens a door downstairs. Whenever one of these weird occurrences...occur...the display doesn't show anything. The piezo speaker is the only thing that goes off.

oric_dan

Do you have long unshielded wires connecting everything in the house?

Grumpy_Mike

Yes this is not weird it is perfectly normal. Impulse interference is caused by sparks, the door switch sounds like it is arcking.
The cure is, supprise surprise, more decoupling on power supplies, and surge suppression on long leads.

Ahh I see. I do have a pretty lengthy cat5 cable connecting the reed switch and pir to the arduino. Maybe adding a capacitor to that length of wire could suppress any surge? Its bundled up and maybe has some kind of inductance going on which could pick up the impulse. As far as power source....i have the standalone set connected to my arduino boards 5v output. I can still use my boards for projects and power my system until I feel like digging out a 5v regulator. What kinda still bugs me is that the LCD doesn't show what tripped. But the piezo is part of an interrupt so I can have that sound and blink an indicator led, so maybe the impulse is short enough to trigger the interrupt but not long enough for it to register on

oric_dan

You're kind of in the universe now where you need an oscilloscope to look at what
is actually going on. The truth is it's not easy to track down the real culprits via
guesswork alone. There are multiple problems your system can be having,

- long wires acting like antennas.
- high source impedances on signals, which compounds problem #1.
- really nasty spikes in environment meaning all I/O lines will have to have EMI
  filters added.
- may be necessary to try different types of EMI filters.
- noise getting onto sensors, meaning best to add filters out there.
- noise conducted in from the power mains through whatever power supply you're using.
- not enough existing filtering on your bd [eg, bypass and electrolytic caps].
- bd would work better in shielded box.
- on and on.


retrolefty

Both the digital cable TV boxes in my house (one in bedroom, one in living room) are sensitive to the sparking from a hand operated mechanical charcoal starter we use to light candles. One click and the TV screen freezes with random pixelation weirdness and about two seconds later all is well. The one in the living room also does that when the microwave door is closed and started. And these cable boxes are all metal encased devices made by Motorola, not some fly by night company.

RFI can be a strange animal to live with.

Lefty

winner10920

How are your sensors/devices hooked up? Perhaps isolating them with optoisolators as well as the extra  supply decoupling will solve that, if you can't optoisolate perhaps just getting some toiroidal cores and passing your longer wires a few times thru can limit the interference
also not sure how good of an idea but to ground your ckt ground to earth ground may suppress some interference,  just make sure that the supply you are using will not be harmed by that

cjdelphi

RFI is my Arch Nemesis!

Over the years (with exception to my desktop PSU) I've had horrible dealings with RFI, light switches, ovens anything that arcs.

So i've been converting all things electronic that run on DC to run from DC, eg, the Modem 11dc from an AC switching adapter, well that AC was causing a small unwanted amount of noise, So i hooked it up to a 12v small lead acid battery (there's a DC > DC regulator inside the modem rated for 22volts) so now i get a cleaner Data Transfer signal, Tesla was a genius but he never stopped ONCE to think about people who work with microcontrollers and IC's lol..

There's a sound / clap sensor circuit I made years ago, everytime you clapped the mic picked up the sound using a flip flop, triggered a relay, but the relay caused a spark causing the circuit to turn off again! what a badly designed circuit! - I had to run it off a 9v battery to make it work properly.


winner10920

Lol I've never had a problem with rfi, my arch nemesis is static electricity and the atleast 5 ics it has destroyed are maimed costing me alot of time

MarkT

Static is usually due to having nylon carpets or nylon clothing - eliminate both is a good start, using proper anti-static workmat and wrist-strap...
[ I won't respond to messages, use the forum please ]

Grumpy_Mike


Both the digital cable TV boxes in my house (one in bedroom, one in living room) are sensitive to the sparking from a hand operated mechanical charcoal starter we use to light candles. One click and the TV screen freezes with random pixelation weirdness and about two seconds later all is well. The one in the living room also does that when the microwave door is closed and started. And these cable boxes are all metal encased devices made by Motorola, not some fly by night company.

RFI can be a strange animal to live with.

Lefty

As you might know I used to design set top boxes. They are all tested for standard immunity to EMI and pass all the tests. The criteria is that the system has to recover from any disturbance within a certain time. While a spark is a short lived disturbance it is interfering with highly compressed data so it does the maximum damage to data flow. The spark has many time more energy in the band you are picking up than the actual signal, you can't protect against "in band" interference because there is no way to discriminate it from the real signal. On an old TV system it just caused a blip on a tiny part of a single frame but the more compressed the data is the more signal it wipes out.
Swings and roundabouts of data compression I suppose.

Go Up