4-20 ma signal

[robh 12]

At what distance would most people feel unconfortable using a 4-20 mA analog control signal or feedback signal?

[BobLfoot 301]

More than distance I am concerned with environmental degradation. Using a 4-20 alongside a welder I'd use a much shorter distance than in a VFD Drive control application were the line ran along open belt.

[ECSI 17]

About three years ago I had a customer using a transmitter/receiver to send a 4-20 signal over a dedicated telephone line. The distance was approximately 1.5 kilometres. A power surge took out the transmitter one day so I hooked up the 4-20 directly to the phone line, boosted the 24 VDC supply up to 28V and it has been working great ever since.

[DanW 41]

Miles. Go for it. Any top brand name transmitter has a chart that shows the loop resistance that the transmitter output can drive versus the power voltage available on the loop. Ignore the statement about the 250 ohms for communication, which HART related. I wouldn't have any reservations about running as far as the loop resistance calculation conforms to the chart values. Wire resistance tables are avialable on the web. The resistance of the receiver device should be a known. The longest run I've seen was 3,950 feet from the panel to the well head (they had a surveyer determine it for them) and then 600' down a deep well. I suggested using an instrument transformer off the well power line, they declined and trenched almost 4,000' for the transmitter signal. This was 10 years ago before wireless was common. It's still working today. The other issue with long runs is bad cnnections in junction boxes, which are likely over long runs, can create problems. Dan

[paulengr 44]

Once in a while you see "20mA digital current loops" mentioned. This is a telegraph system. There is no "standard" but generally <5mA is "0" and >=10mA is "1". The encoding format is similarly not really standardized but these days when I see it, it is usually just regular old ASCII. It's essentially an alternative to RS-485. Ranges are same as telegraph...up to several miles if you've got enough voltage to handle the line resistance.

[panic mode 246]

it's a simple calculation: 1) understand the loop or how the circuit is closed. we need this to calculate resistance, something like: > from (+) to transmitter (this is one length of wire) > from transmitter to (input+) (this is another length of wire) > from (input-) to (-) wguch is shor piece of wire and so not really important 2) you must know the voltage (24V for example). use it to determine max loop resistance for biggest signal (that would be 20mA): Rmax=U/I=24/0.02=1200 Ohm 3) determine resistance of analog input (read datasheet). should be 250 Ohm for most cards 4) calculate wire resistance (remove input resistance from the calculated maximum) Rwire=Rmax-Rin=1200-250=950 Ohm 5) what is resistance of your cable (depends on wire size). let's say you use AWG20 and resistance is 34.1 Ohm for every 1000meters (you need to verify this value). this means that max wire length is Lwire=Rwire(ohm)*1000(m)/34.1(ohm)=27.8km or about 17.3 miles. in real world this will be less because we didn't take into account min resistance of the transmitter (which has it's own "analog input" for closed loop monitoring). this should be less than 250 Ohm (read the datasheet). but let's make it 350 just for sake of argument. Rwire = 950-350=600 Ohm ---> wire length = 17.5km or 10.9miles. 6) split that in half (we have two wires in a cable - to and from transducer) so your cable length is in last case some 5.45 miles IF supply voltage is 24V and used cable is only AWG20. I was always using AWG18 or AWG16 (need some mechanical strength for pulling) and the longest cable I had was just under 2 miles.

[BobLfoot 301]

Panic I seem to recall a Belden cable that was 8 conductor by 24 awg and had a #10 bare steel wire for "ground" / mechanical strength in the center. This stuff was awesome for long signal runs. Can't recall the part number thoguh. Let's say I am having the first of many "senior" moments.

[paulengr 44]

Almost right. Don't forget to include the reactance of the cable and the desired sampling rate. In a DC-only world, range is simple a non-issue. Frequency effects do become an issue when you start attempting to go several miles. That is one part of the reason Ethernet is spec'd at 100 meters (the other reason is once they established 100 meters as a limit, the timing parameters were similarly adjusted...so even if you tried to use CAT 6, you still couldn't exceed 100 meters). In long (2-5 mile) cables, it is not unusual to find that you can't sample faster than once per second on a straight run. On a particular 3 mile cable I'm familiar with it had several receivers and my predecessor stuck a lightning arrestor on the incoming AND outgoing sides at each station. These added a lot of reactivity to the system and made it take about 1 minute to stabilize when the transmitter readings changed. Of the two readings, the tank level wasn't really an issue. But current on the pump was...it fooled us a couple times because what was actually a very fast increase (something stuck in the line) looked like a gradual increase (more flow).