Analog spikes

Hi All,

Any Analog experts out there,?I'd be grateful for their opinion. Just wondering if anyone has ever experienced the problem described...

We have a 4-20mA output giving random consistent spikes of 8mA. All other times, due to its scale is around 4-5mA and no more. This output is from a controller to a BMS. These spike do not reflect the readings if you convert the mA to the 'scale'. We disconnected the BMS cable and put a Eurotherm data logger directly into the output > still the spikes...The only (last) thing we can think of, is there is a relay output on the same card and wondering if this may be causing 'noise' when the relay is switching, causing spikes. The manufacture is wiping their hands saying it's our problem?

Appreciate any feedback

Earthing problems ?

There is no 'drain' wire but we did put a 250 ohm resistor across the input of the data logger to record the mA. Their is no ground on the manufactures' manuals or installation examples...

how close to the device did you connect the logger?

Check if there is any RF transmission that occurs - eg someone using a CB radio or mobile phone near the transmitter ( could also be a 3G SMS by a nearby device that is not shielded)

Log the Power supply to the device - you may have a power spike when something starts

Interesting..be nice to understand the circuit a bit more where is the source of the signal, can you correlate the spike with an IO event? is their only one supply in the system or are there two generators? one idea could be to put in an islolator eg a signal convertor like for like phoenix do a range..

Is the cable shielded and how is that shield grounded is it near any low voltage but high power devides eg 24v DC solenoids?

Regards
David
www.fast-automation.com

What happens if you prevent your code from altering the output?

before -what the condition ?

It sounds like a stand-alone PID controller 4-20mA control output being used as the demand, firing rate signal in Burner Management System.

I have seen controller analog outputs spike like that in two cases that I can remember:

1) 25 years ago with a Honeywell UDC 2000 before it had an isolated input. There was a grounded thermocouple on its input. The furnace was a huge electric heated monster and ground leakage current would create enough IR drop to spike the temperature readings from the grounded thermocouple. Odd because ground loops are usually steady, not spikey, but this one spiked.

Difference with your situation is that the input spiked causing the PID output to spike, which was 'proper' operation of the PID.

An ungrounded thermocouple and a UDC 3000 with an isolated input fixed it.

2) It was an Asian import (to the USA) 1/4 DIN temperature controller. The output spiked randomly, but not full scale. A battery powered thermocouple calibrator was connected to the input and used to maintain a steady PV (temperature reading) at the Set point which would create a steady PID output. But the output still spiked, indicating an internal problem. I suspect it was a flakey DAC, but replacing it with Honeywell UDC fixed the problem.

That would be the test that I'd run on it. Battery powered calibrator on the input so PV = SP. Let the PID ouput stabilize and if the ouput spikes with no cabling (to induce noise) on the output, the problem has to inside the controller, regardless of what the factory guy on the phone says.

not mention what type of source for the BMS,

other option change the controller or other brand

other guess - spark-plug

Are the 4-20mA wires contained in a conduit with any power conductors, or switched signals, such as the wires from the relay?

I bet it's in your program.

If it is not a time critical process, using a software filter can help mask the problem.

The only (last) thing we can think of, is there is a relay output on the same card and wondering if this may be causing 'noise' when the relay is switching, causing spikes.

Click to expand...

If the relay is controlling a inductive load then make sure there is surge protection across the load.

Hi all, first thanks for the input... So the ill try and clarify more.. It is taking a gas (ppm) sample and turning into a analogue current for the BMS 4mA = 0.17ppm and 20mA is 16.7ppm. There is a Analogue output card on the sampler. from there a 25 D plug pin cable interface to a breakout board. It's here we terminate the eurotherm data logger. The fly cables are around 400mm in length from terminals to logger. BMS disconnected.
MichaelG - interesting, yes we use 2 ways across the site. Plus there are Telstra repeaters through-out for 3G reception. I'll test this tomorrow. Thanks!
Davidaskew- so we do have a 24vdc relay contact on the same card/interface cable and monitor the voltage in another input channel of the data logger. Results were mA spikes did not coincide with the 24v switching.
DanW- input definitely does not spike. When calculating the mA spikes does not reflect what the ppm value is displayed on the software platform.
Mickey - relay is only controlling a stack indicator light. Again mA spikes generating when relay not switching.
RET - it ain't my program, but an out sourced companies. They are saying it's our issue, but trying to prove its theirs. We've paid them heaps of $$$ so can't just get another brand etc..
Although now thinking I may drop the 24v com (external source) off the card and see what happens. So the only thing on the card is the analogue output...
Thanks again!

dsg said:

So the ill try and clarify more.. It is taking a gas (ppm) sample and turning into a analogue current for the BMS 4mA = 0.17ppm and 20mA is 16.7ppm. There is a Analogue output card on the sampler. from there a 25 D plug pin cable interface to a breakout board. It's here we terminate the eurotherm data logger. The fly cables are around 400mm in length from terminals to logger. BMS disconnected.

Click to expand...

Defining spikes without defining the spike voltage between each two wires means nobody can reply with anything useful.

Apparently you have discussed an 8 ma spike current down one wire. It is electricity. A return current path must exist. If you do not know that return path, then define every wire incoming to both boxes. Yes, even safety ground wires are relevant. As for earth ground, that connection should only exist (be relevant) if two boxes are in separate buildings.

That 8 ma could be a current left to right on one 4/20 ma wire and right to left on its matching pair. Or it could be same current in same direction on any or both wires. Provide that information to get replies from the fewer who really know this stuff.

Unlikely is noise from the relay. However anyone discussing surge protectors on that relay do not understand something completely different. A reverse diode (not a protector) must discharge the relay. If that diode has failed, then nothing but the electronics test equipment (ie oscilloscope) will report that failure. In fact, if not using an oscilloscope to identify the source of that 8 ma, then you are probably spitting in the wind.