4-20mA, 0-5V transition reading?

Guys

This is not a strictly PLC question, but I figure we deal with a lot of sensors input signal so I think this would be a good place to ask.

I have a wind direction sensor with 4-20mA output. So 0deg = 4 mA, 360deg = 20mA.

When the wind sensor turns anticlockwise from 5deg to 355deg without going through 10deg, 90deg, 250deg, 300 deg... (happens only when the wind sensor turns clockwise):

I expect quantums reading from 4.6mA to 18mA like this:

4.6mA
18mA

But then i see something like this:

4.6mA
12mA (transition reading a flicker of about 0.2 second)
18mA

The PLC picks up the 12mA reading and gives a false impression of the wind sensor turning clockwise going through 170deg.

So, I tested all 4-20mA output wind sensor types and they behave the same way, so i am thinking that all 4-20mA signal module has a variable potmeter inside (or some old classical analog transistor inside) and the signal cannot quantum jump from 4 to 20mA, but need a very brief transition say 12mA like this 4mA,12mA, 20mA.

Am I correct?!

If yes, does 0-5V behave the same way?? It must behave the same way, right?

I don't have personal experience with wind sensors, so this answer is somewhat speculative. However, I believe you have correctly identified the problem. I don't see any reason that a 0-5 Volt output wouldn't have the same problem.

One answer to this is to use an absolute encoder on the weathervane instead of an analog sensor. I would think this could be readily mounted to the device.

hi
yes i think you are correct as it is analoge siganl,if you use incremental encoder that give you the exact position with accuracy
depend on no of bits but this need input register if you used it in plc circut,if you use 4-20 ma in plc you can build circut for alarm that did not give the value util we reached the steady state condition ,i did not use it in wind direction but i use it in weighers that i didnot take the weight if there is any vibration
or change in weight value .
good luck

I think that you are up against a limitation in the ability of the Analog Input to physically transition from the max to min value instaniously. This ability is sometimes referred to as the "Slew Rate" when something like an op-amp is involved, but in this case, I would expect that there exists some filtering devices that, however small, are inhibiting the ability to change instaneously. I would want to investigate the input filtering aspect and make certain that you are not using more than whatever is the minimum available. My guess is that this would occur with a voltage input as well.

Tom (above) mentioned the possibility of a vane with an encoder. If such a thing is available, it would likely work better in this application since any transistion due to change in position would have the same impact on the input response as any other.

One additional though is that if you are constrained by resources, then you may be able to filter out the intermediate value so that it does not mislead the system. The effect of filtering however would be to reduce the effective slew rate and the changes would appear more slowly. My guess is that this would not necessarily be a bad thing.

I would have to know something about the your system before I could venture a recommendation, but I have used filtering to clean up analog signals in the PLC environment, in a variety of applications and it can make significant improvements. Most of the time, a signal that responds slower with fewer errors, is more useful than one that changes quickly but contains significant errors.

I used to operate a system that controlled gauge on a Cold Rolling Mill by looking at the output of an x-ray gauge. Also, I have written apps to make measurements with lasar distance gauges through smoke, fog, water spray, dust, etc. I can tell you, well thought out filtering can turn an app from a failure to a success.


Best Regards,

Bob A.

Without ever having encountered a wind direction instrument before, I suspect that Bob A is on the right track about filtering.

I suspect that the wind direction analog output is just as likely to be filtered as the AI input on the PLC. Is there any filtering adjustment mentioned in the documentation?

>So, I tested all 4-20mA output wind sensor types

I'm not sure what that means. Does it mean that there are different output modes on the wind direction instrument, or that you have multiple devices of different manufacture that you can test?

Can you clarify?

Dan

The actuator should have a determinable physical travel rate limit. You need to find out how many degrees the thing can turn per second. If you can read the input faster than that rate limit, you should be able to tell when it crosses zero.

You have to compensate for the ramp effect of the signal so you may not want to calculate the distance travelled any faster than required for the application.

Example, if the device can move 100 degrees per second, and you scan it every 0.1 seconds, it should never change more than 10 degrees per scan, or 30 degrees per 0.3 seconds. If it does that, then you know it has rolled over, and can determine the direction, although there will be a little bit of lag in that determination.

If your program detects a change greater than that, it should be able to assume that it rolled over, and can determine the direction of travel correctly.

More specifically you need to be able to measure the change at least twice as fast as it can naturally occur to be able to guarantee that you can calculate the direction accurately when it rolls over.

If it is allowed to physically rotate faster than the PLC can scan the input, then you may need to restrict it physically to solve your problem.




Are you trying to record every little gust which changes the wind direction, or looking for an average?

Paul

Huge thanks guys!!! I think I can get the system working with your teachings.

In a remote island, gotta go

Good luck,....i did this few years back at a remote island using 12 inductive proximity switches on a ring.

When the arrow comes near the switch it picks up and sends it to PLC which after a delay of 3 secs shows the direction on HMI.

It works fine.

Cheers.

bambinorama said:

Huge thanks guys!!! I think I can get the system working with your teachings.

In a remote island, gotta go

Click to expand...

How about sharing your final resolution with everyone?
What did you finally decide was the best method to solve the problem?

How the problem was solved

Yeah.

Modification in the datalogger-PLC firmware:

If the firmware detects anemometer direction transition from 0-30deg to 330-360deg within 0.5s(the mechanical anemometer cannot swing so fast physically), the software will filter out the values of 30-330deg, and replace the filtered value with a latest value of 330-360.

It's a simplistic idea, but the customer accepted it, so...

Thanks guys for all your generous help.