problem in flow computer

hi all
i have measuring system contain from coriolis flow meter ,pressure and temp transmitters and flow computer( floboss s600)
my problem is some time the reading in flow computer is wrong
Normally, the number of pulses arriving from the meter is eight digits, but in one second it receives ten digits ( wrong pulses)
The meter is working fine and I don't know the reason for these but I suspect that there is something wrong with the flow computer
So that the reading jumps from about 4000 thousand cubic meters per day (as a correct reading) to (70 thousand cubic meters per day).
Note that this situation recurs for days, stops, and then reappears
thank you
thank you David for replay..
there is no failure messages or codes
just high pulses appeared in split second and every thing else is OK.
note some times that happens even there is no flow ( Flow rate is zero)
but i found some thing this high pulses is always same value. so if i remove it by calculator i can found the result is the true value.
It appeared from your first post that the problem was losing pulses, going from a true high flow rate 4000 thousand CMD to a false low rate of 70 thousand CMD.

But if the opposite is true, going from a true low flow rate to a false high flow rate, and the duration period is relatively short, then it brings to mind issues from the 1970’s and 1980’s with site handheld radios that would overload analog 4-20 signals when the radio was activated or ‘keyed’ in talk/transmission mode in close proximity to the transmitter. ”One vendor used ferrite beads around the signal wires to limit the effect of local high frequency radio noise.

Could harmonic noise from a VFD produce a set of pulses that appear to be signal pulses to the flow computer? I have not heard of that phenomenon reported before, cannot come up any other means by which a pulse train 58 times the frequency of the true pulse train gets into the flow computer, other than temporary, intermittent malfunction of the flow transmitter.
There is no point in checking any piece
Because they are all working normally
Only in a split second does this jump occur and return to normal operation
Imagine that the reading yesterday was 283,500 cubic meters per day, and this means that the jump occurred 4 times (70,000 * 4) = 280000 and the correct reading is 3,500 cubic meters per day
It isn't clear what causes the spike value in the flow rate or where that fault occurs in the system.

A flow rate spike from a pulse/frequency signal would have to be either a momentary high frequency pulse train that is electrically within the range of the flow computer's sensing input that interprets the state changes as pulses, or it is an false artifact created by the flow computer internally.

Consider the possible sources:
Flowmeter: Could it be a momentary false pulse train generated by the flow meter? or a diagnostic?
Connection (cabling): Could it be noise induced on the signal cable resulting in a false pulse train?
Flow computer: Could it be a problem in the internal handling of the data in the flow meter? Or the input falsely reporting a spike?

A frequency signal is associated with a range, say, 0-XXXX volume units per time unit. Is the spike high value within the configured range of the transmitter? Or is it outside the flowmeter's configured range?

In 20 years of hanging around instrumentation and control fora I can't recall another instance of a flow meter generating a false momentary high frequency pulse train.

The analog signal world uses outside-of-normal-range upscale/downscale signaling for error/fault conditions
- 0 mA indicates an open circuit on a 2 wire, loop powered 4-20mA signal
- 3.8 or 21.8mA indicates a fault in a 2-wire, loop powered field transmitter
- a thermocouple burn-out condition (typically) sends the process variable upscale/offscale

I am not familiar with frequency/pulse outputs doing the same out-of-range type thing to indicate a fault, but it's a big world out there and I have limited exposure. Have you checked the flow meter documentation to see whether or how the flow meter signals an internal fault? Could the spike be a warning like a "Check Engine' light?

What does the flowmeter's frequency output do during the power-up initialization process (sometimes 30-40 seconds from 'smart' devices)? I once ran into a field instrument that would fault by momentarily going into a power reset mode and going through the initialization process. Could the flowmeter be re-initializing when the spike occurs?

In terms of induced noise, have you logged when these spikes occur? Sometimes an electrical noise event can be traced to a physical event, like the starting of 10,000 HP motor.

I'm not sure those clamp-on ferrite assemblies are designed to work in low KHz range.

Is there a diagnostic program that you can run on the flow computer?
لإhank you for your interest
i found some thing :the value of this spike is always 4,294,967,296 pulses and this equal to 2^32
and yes this spike is over range
so i found some thing about bad pulses or bad frequency and cut off
this is manual of roclink configuration software for flow computer floboss 107 ( not mine flow computer)
in page 170 it refer to " Accumulated Value Counts the number of times the discrete input goes from Off to On. The accumulator is a 32-bit number with a maximum count of 4,294,967,295. You can preset the accumulator by entering the desired value or clear the accumulator by entering 0. "
but in our flow computer (floboss s600) and its software configuration is "config600" there is no thing about this
but there is some thing in page 94 but i thing these for low bad frequency not for high
i thing there is option in somewhere with wrong set or not activated change any bad pulse to this value, do you think that?
i thinking that because the value of spike is using in a lot of protocols and ranges , when i google it i found a lot of result expect my problem ^_*
i am very happy for your interest and I hope you are always well.
Very interesting that the spike = 2^32 counts. I seriously doubt that any commercial flow meter flow signal works 4.2GHz. Every one I've encountered uses a flow rate frequency rate in the low KHz range (up to 10KHz or less).

There's no way the flow rate pulse train = 2^32 counts for one measurement. This looks to me like a flow computer internal problem, which could be an analog input card problem, or a processor issue, depending on the architecture. The accumulated counts just happen to max out at 2^32, I doubt that those count accumulators are involved in the false spike value,; it just shows that the Floboss handle 32 bit data internally, either floating point or integer.

Unless someone else with Floboss experience pops up here, you're going to have to find either a flowboss forum or work with Emerson directly to solve this. Tracking down a bogus max value spike requires someone intimately familiar with the system architecture and that's not me.