What can cause a current drop?

[rvk2]

Hey guys, I've hit a dead end with an electronics problem and I'm hoping someone here can suggest some ideas.

We basically have an electronic sensor that detects gas. It has a transmitter that outputs current based on how much gas has been detected. For example, the range could be 1 - 5 mA and at 50% gas, it will output 2.5 mA. Now these sensors are in a plant environment (about 30-40 sensors in each section) driven by a 24VDC power supply. Some of these sensors randomly go into fail. The current on random sensors drops below the low level and the sensors go into fail.

We've eliminated the sensors being the problem because we take the sensor/transmitter to the lab and it works fine. We're thinking it could be a power issue, but everything seems to be fine with the power supply and they're within spec. I guess we're trying to figure out what can cause a current drop and keep it at that level. AC voltage spikes and other things can cause current drops, but it wouldn't keep it there. So that's what I'm trying to find out - what is causing random sensors to drop the current output.

I'd appreciate any ideas/feedback/suggestions guys. Pretty close to bashing my head on a wall because of this problem.

 

[JimB]

After a failed sensor has been tested in the lab, does it work OK when put back out on the plant?
Possible connector problem at the detector mounting?

I guess there is some kind of electronics in the gas detector head, or is the detector just a basic element an all the electronics is back in the rack in the equipment room?
If there is electronics at the head, is it getting confused and going into a failed state? Try switching off the supply to that head for a few seconds and re-connecting. Does it now give a correct reading?

What about the wiring between the detectors and the equipment room. I guess there are some junction boxes and marshalling boxes there, are the terminals tight?
I have seen weird problems where the terminals (Klippon SAK2.5 and SAKR) were "not tight". They were not loose either, but a nip up with a screwdriver (less than half a turn) gave a permanent fix to the problem.

JimB

 

[rvk2]

JimB,

Putting the sensor back after it works in the lab - that's exactly what we're waiting to hear back on. Unfortunately I have to depend on lazy people to do this for me on-site.

The sensor is a catalytic bead sensor and is attached to a transmitter. The transmitter determines the span of the sensor based on it's age. A new sensor will give a good mA span whereas an old one won't. Everything is then wired to a power supply and a device that takes the raw mA and converts it to 4-20. We know the problem starts at the transmitter because the transmitter outputs low mA.

The other thing I should mention is that if one sensor gives us problems, we replace it with a new transmitter/sensor and it works fine for a while. But then the problem appears on another sensor. So it's almost seems like the problem goes away but appears on another sensor. And it doesn't matter whether the sensor is new or old, the problem reappears after a while (I'm talking weeks).

Here's another kicker: we keep hearing that the -24VDC common and ground show continuity. But then we heard that it doesn't. We don't have a confirmation on it, but it's also an issue. I was thinking that perhaps there was a short somewhere and they found continuity then, and when they fixed the problem they didn't find continuity. The grounding of the gas system could be connected to the grounding of another processing system. So it could be that this is happening because of problems in another system.

So there are some things I still don't have a clear answer on. The grounding issue was a red flag, but I don't know how much of an impact that would have if the ground and -24VDC common are connected.

 

[Reloadron]

I use similar networks of sensors with a few differences in that rather than 1 to 5 mA outputs mine are generally 4 to 20 MA loops. Sensors like this are designed to work into low impedance loads. For example 4 to 20 mA working into a load of 250 ohms will yield 1 to 5 volts. A load of 500 ohms will give 2 to 10 volts. Now loose or faulty connections in the current path can wreak havoc with your current loops. If you are running off a good stable 24 VDC supply capable of enough current for all your sensors and the sensors are good I am inclined to run with JimB's thinking as to something intermittent in the loop. Do you run the same +24 VDC to all sensors and monitor their returs individually for your loop?

Ron

 

[rvk2]

Ron,

The power supply seems good. We are thinking about trying brand new ones, but again the problem with that is getting some people to actually go out and do it.

There are 30-40 sensors in each section of the plant and there are about 6-7 different sections. Each section has it's own 24VDC power supply. So to answer your question, all the sensors in that one section get the same +24VDC power. The sensors are also connected to what we call a "sensor satellite", which reads the 4-20mA signal and displays it on an HMI screen. For a while, we looked at the sensor satellite and the HMI, but they're not the source of the problem. I guess you can say they're just dummies that display whatever information is given to them. The problem really starts with the sensor/transmitter dropping in mA, which in turn affects the scaling, and hence the Fail signal on the HMI screen.

I hope that answered the question you were asking, but if not I'd be happy to elaborate more.

 

[Reloadron]

Gets more interesting as each section has its own 24 VDC supply so I figure rule out the power supplies. Then the sensors have sensor satellites which likely have a very low input resistance like <500Ω and actually those units likely read a voltage across the input resistors.

Now many of the sensors I use with a current loop, granted, are not gas sensors but most sensors like this generally use what is called "upscale" burnout which just means if the sensor fails the output current will max out and not decrease. Some use downscale but most are upscale on burnout. Really strange what you have going on.

Ron

 

[rvk2]

Ron,

You are right about the upscaling/downscaling. In our case, our sensors downscale. I'll be more specific about the setup - there's a sensor, connected to a transmitter. The transmitter has a scaling device connected to it. The scaling device takes the raw mA from the transmitter and converts it to a 4-20mA signal. This 4-20mA signal is what goes to those sensor satellites I mentioned earlier. When the processor of this scaling device sees that the scaled mA is below 4.0 (specifically it's 3.2 mA) it goes into fail. For a while we thought that there was an issue with the scaling, but we know that it's doing what it is supposed to. When we take the mA reading at the transmitter, we know that there has been a drop. This is basically what has us stumped - what is causing mA to drop for random sensors. This would be much easier to troubleshoot if all sensors behaved the same way.

So yes, you're absolutely right, something strange is going on here. We have these sensors in hundres of different locations all throughout the country. So if there was a problem with components, we would've seen this somewhere else too. Also, these sensors have been made the same way for over 20 years! And this is the first time we've seen something like this. We keep thinking that it must be something to do with the wiring at the plant, but just can't quite figure out what exactly it is.

 

[Reloadron]

Plant wiring is one of those things we like to focus on because it makes for a simple start.

The ***** is many of today's 24 VDC power supplies used for applications like this are pretty forgiving little supplies. Case in point is I have Sola Model SDP 1-24-100T 24 V. 1.3 A supply sitting here. The input voltage is 115 to 230 VAC @ 50 or 60 Hz. The DC out being totally isolated from the AC line. Pretty forgiving as to plant power.

However, there is a quirk I have run into a few times. Several of my sensors did not seem to like it if the DC Common ran at any point to an earth ground. Meaning my analog signal grounds could not come in contact with an earth ground. Any earth ground in my loops and bad and unusual crap happened.

When it comes to things like this I can honestly tell you better you than me. I hate things like this.

Ron

 

[KeepItSimpleStupid]

A ground loop. Best way to find them is to design them out from the beginning. Make sure you have bias return paths.

 

[Brevor]

The power supply seems good. We are thinking about trying brand new ones, but again the problem with that is getting some people to actually go out and do it.

Is this in one of those nuclear plants in Japan ?

 

[rvk2]

Thank you all for the input. We keep thinking the issue must have to do with grounding also. That was the last thing we concluded before we ran out of ideas. Unfortunately, no one at the plant wants to acknowledge that this can be an issue, so they keep going back to finding fault in the sensors. But it seems that those of you who have experienced something like this before are leading me to believe that it's the wiring at the plant.

Lol if this plant was in Japan, there'd be bigger problems for us to worry about now.

 

[KeepItSimpleStupid]

Can you provide a simple sketch of what you think things look like? Just the sharing of a power source can cause issues. Placement of the power sourch can cause issues. Not using differential inputs or isolated inputs can contribute to your problem as well. Having a single-ended input on the device can also create issues.

What device is being used to condition the signal?

Now, I've seen systems that have a power sorce for the heater and an output that's 4-20 mA or even 0-20 mA. Some examples are Hydrogen bridge sensors and air velocity sensors. The heater for the bridge must be a separate supply. Meaning 4 mA is to low.

You really haven't given us much to go on.