This is mostly about theory so I don't have a circuit or program to share. But I've noticed when doing something simple like put a switch on the end of a long wire (100ft) and using it to ground a pin, noise can be picked up and seen with a DSO. When using pullup I can understand a noise spike above 5v on the line. But I see dips as well. How can a long wire cause a dip below 5v (without closing the switch) from noise? I realize a cap fixes this problem but I'm just confused how the voltage dips or sags?
What's a DSO? Digital Storage Oscilloscope?
Noise is typically random so it's usually equally-distributed positive & negative. A capacitor can help, a lower-value pull-up can help, or shielded wiring can help.
DVDdoug:
What's a DSO? Digital Storage Oscilloscope?
Yeah sorry.
DVDdoug:
Noise is typically random so it's usually equally-distributed positive & negative. A capacitor can help, a lower-value pull-up can help, or shielded wiring can help.
Well I don't understand how noise can drain current.
A long wire is an antenna, which picks up signals from house wiring, radio/TV/towers, nearby motors, etc.
You need two wires for a switch (signal and ground), so use twisted pair or shielded cable to reduce noise.
jremington:
A long wire is an antenna, which picks up signals from house wiring, radio/TV/towers, nearby motors, etc.You need two wires for a switch (signal and ground), so use twisted pair or shielded cable to reduce noise.
I understand a long wire picks up noise. But how does it drain current from the arduino (negative noise)?
The noise signals are AC (with no DC path), so alternating positive and negative voltages and currents are induced between antenna and ground.
When I work with long wires, such as reed switch, I use optoisolators to decouple most of the spurious thingies on the line.
freakdaddy:
This is mostly about theory so I don't have a circuit or program to share. But I've noticed when doing something simple like put a switch on the end of a long wire (100ft) and using it to ground a pin, noise can be picked up and seen with a DSO. When using pullup I can understand a noise spike above 5v on the line. But I see dips as well. How can a long wire cause a dip below 5v (without closing the switch) from noise? I realize a cap fixes this problem but I'm just confused how the voltage dips or sags?
Simple arithmetic! A noise pulse can be positive or can be negative, or can be a series of both. Either case it arithmetically adds/subtracts from the assumed value.
Paul
freakdaddy:
This is mostly about theory so I don't have a circuit or program to share. But I've noticed when doing something simple like put a switch on the end of a long wire (100ft) and using it to ground a pin, noise can be picked up and seen with a DSO. When using pullup I can understand a noise spike above 5v on the line. But I see dips as well. How can a long wire cause a dip below 5v (without closing the switch) from noise? I realize a cap fixes this problem but I'm just confused how the voltage dips or sags?
There are two ways interference is picked up by long wires - capacitive and inductive.
Capacitive pickup will increase with the length of the wire and is due to man-made and natural electric
field changes (lightning, mains wiring, spark plugs in car engines, radio transmissions etc).
Inductive pickup ought to be minimal if the wire is floating at the far end, and the loop area is minimal (as
with twisted pair). If the wiring is a big loop with significant area, lots of inductive pickup is possible, from
changing magnetic fields due to radio transmissions, heavy electrical machinery, lightning, even solar flares.
If the wire is grounded at both ends it forms a loop with the ground and will pick up much more inductive
pickup. Again the longer the wiring and the larger the loop area the worse the pickup.
The usual cure for pickup is to use differential signalling (most pickup is common-mode only), and opto-isolation
(to prevent destructive voltages passing through - lightning is the main culprit here). Of course optic fibre and
wireless signalling are both effective.
And actually grounding both ends of a wire may be counter productive because there will always be a potential difference between the two grounds as well as AC ground currents. All depends on location.
Paul
freakdaddy:
This is mostly about theory so I don't have a circuit or program to share. But I've noticed when doing something simple like put a switch on the end of a long wire (100ft) and using it to ground a pin, noise can be picked up and seen with a DSO. When using pullup I can understand a noise spike above 5v on the line. But I see dips as well. How can a long wire cause a dip below 5v (without closing the switch) from noise? I realize a cap fixes this problem but I'm just confused how the voltage dips or sags?
100 feet of wire has substantial inductance. A noise spike can "ring" the inductor and create all kinds of strange signals.
Solution: twisted pair. Twist the signal and ground wires together (chuck the two wires into a drill motor, have someone else hold the two wires at the far end and twist it up).
Note that this will make the wire pair shorter, so start with enough extra to end up with the length you want.
Pardon the crude pencil drawing.. When I did theater production videos, we would routinely run hundreds of feet of shielded, twisted-pair microphone cables. Since these typically operate at 1V P-P, they would be sensitive to any generated noise.
The theory is that twisted pairs will cancel common-mode interference. In other words, a magnetic induction or other induced noise will be seen as + on one line and - on the other, cancelling out the interfering noise. (The ground at the switch end is optional).
This is how I would wire a switch 100 ft from an Arduino. If I had to. (If I had to control something that far away, I would use a 433MHz transmitter/receiver pair).
The remote ground connection injects noise directly onto the wire, lose it. Remote ground is not local ground.
Agreed. My drawing may introduce a ground loop, so lose the ground at the switch end.
SteveMann:
Pardon the crude pencil drawing.. When I did theater production videos, we would routinely run hundreds of feet of shielded, twisted-pair microphone cables.
Those typically use balanced signals, so most noise is cancelled out.
wvmarle:
Those typically use balanced signals, so most noise is cancelled out.
With the very low signals from a microphone you need both a shield and balanced pair to take out enough
noise to be usable. You also need a special layer inside the shield to reduce cable microphonic effects caused by friction inside the cable as it flexes and moves.
Digital signals are much more robust, typically just twisted pair and differential signalling will do, as in ethernet,
RS485, LVDS etc etc.
I do seem to remember those microphones have a battery in them.... Must be for some kind of amplifier, to get to the typical +/- 1V signals.
Trying to transmit the signal directly from the microphone is quite hopeless indeed.
Condenser microphones need a power source either from "phantom power" on the audio wires or a battery. Dynamic microphones make their own power.
Grounding shields on long cables at both ends - if you don't, hope yo don't get an HV earth fault at the far grounded end, and you're holding the shield at tother. In many countries there are regulations/legal requirements, about running cables between buildings, for example.
raymw:
In many countries there are regulations/legal requirements, about running cables between buildings, for example.
Regulations are for mains power cables, not for signal cables.