Issue on 4-20mA signal,cannot determine where the fault is, as everything seems normal as measured.

Thread Starter

Pochollo

Joined Oct 23, 2020
10
Hello Everyone,

I hope you are all very well at this moment, I am working on a ship as a marine electrician and I'm not really so much on electronics stuff so I hope you guys can help me with this little problem of mine. I have just replaced our main engine turbo charger tachometer with a new unit as previous unit was found to have a defective analogue output. 4-20mA output signal was not coming out from the unit as I measured it as zero (0mA) and the alarm keeps on coming on our alarm and monitoring system as sensor failed. This is clear to me that the alarm will come out since the issue is on a defective tachometer, so we purchased a new one and after I have installed it, I re-measured the output and it is giving me a 4mA since the turbo charger is not spinning and at 0 rpm. After a while the alarm came off again as sensor failed, so I measured the 4-20 mA output and I still measured 4 mA which is still fine. So I worried that the Analogue input board on that particular channel might be faulty. So I look for other sensors adjacent to it with the same 4 -20mA and then reconnect it to the channel that is in question. It seems like the channel is working fine that it gives a rpm reading with that temporary 4-20mA signal from the other sensors. So now, it is a bit confusing why the alarm is still coming on since the unit is giving a normal 4 mA input. I disconnect the two wires from the tachometer output and measured the voltage and it was 27.31VDC.. Now I'm a bit confused what is going on, why the Analogue input board doesn't want to accept the analogue signal from the tachometer. Is the 27.31 VDC measured is not too much although it is giving a correct mA signal. The analogue input module is set to accept current signals and no alterations made on the wirings and all the settings. I just reconnect the new unit as is but the issue is still there. What seems to be the problem here?Engine tachometer diagram.jpg
 

sagor

Joined Mar 10, 2019
903
Voltage is only as high as required to produce the 4-20mA signal. With infinite resistance (no load), the voltage will go as high as possible, in an attempt to produce that current. Hence, for current loops, voltage is somewhat irrelevant. Is the 4mA actually that, or is it a bit low, like 3.99mA? The accuracy of the analogue board may be a factor in that case. What is the current loop when you have some RPM, does it still alarm? If so, then the A/I board input is probably bad.
 

Thread Starter

Pochollo

Joined Oct 23, 2020
10
Voltage is only as high as required to produce the 4-20mA signal. With infinite resistance (no load), the voltage will go as high as possible, in an attempt to produce that current. Hence, for current loops, voltage is somewhat irrelevant. Is the 4mA actually that, or is it a bit low, like 3.99mA? The accuracy of the analogue board may be a factor in that case. What is the current loop when you have some RPM, does it still alarm? If so, then the A/I board input is probably bad.

Thanks for replying sir..
As of the moment I am not certain if the alarm will disaappear after the turbo starts spinning since the ship is still in port but what I'm sure about is the idea of putting a 4-20ma load on that same analogue board let say I injected 10mA on it and the alarm was gone and it gave me a turbo rpm read out. On a zero rpm state I'm getting 4.01mA to be exact Sir but there, I'm still having issue clearing the alarm and with rpm readouts of below 0.
 

sparky 1

Joined Nov 3, 2018
756
What is the main engine and do you have a more detailed schematic ? what is the size of the diesel and the highest rpm of the turbocharger ?
Do you have a picture of the analog board ? A picture of the 24V tachometer motor ? what is the make , model of these ?
any manuals pictures ?
 
Last edited:

LesJones

Joined Jan 8, 2017
4,174
Even if the tachometer is not being rotated there should be current in the loop. 4 mA in the loop represents 0% of the full scale reading. 20 mA represents 100% of the full scale reading. So if you break the loop and insert a current meter into the break in the loop you should see a reading of between 4 mA and 20 mA. If you get a reading of less than 4mA the loop is broken or the end that is configured as the active end is not supplying voltage to the loop. If I understand the last line of post #4 you are measuring a loop current of 4.01 mA. Is my understanding correct ? Can you try driving the tachometer with an electric drill to see if the system cleared the alarm above a certain speed. It could be the alarm threshold is something above zero RPM. (We don't know if the alarm condition is indicating that there is a fault in the current loop or is indicating that the speed is too low.)
When you set the loop current to 10 mA the display should have read 37.5% of full scale ( 10-4)/(20-4) = 6/16

Les.
 
Last edited:

Thread Starter

Pochollo

Joined Oct 23, 2020
10
Even if the tachometer is not being rotated there should be current in the loop. 4 mA in the loop represents 0% of the full scale reading. 20 mA represents 100% of the full scale reading. So if you break the loop and insert a current meter into the break in the loop you should see a reading of between 4 mA and 20 mA. If you get a reading of less than 4mA the loop is broken or the end that is configured as the active end is not supplying voltage to the loop. If I understand the last line of post #4 you are measuring a loop current of 4.01 mA. Is my understanding correct ? Can you try driving the tachometer with an electric drill to see if the system cleared the alarm above a certain speed. It could be the alarm threshold is something above zero RPM. (We don't know if the alarm condition is indicating that there is a fault in the current loop or is indicating that the speed is too low.)

Les.
What is the main engine and do you have a more detailed schematic ? what is the size of the diesel and the highest rpm of the turbocharger ?
Do you have a picture of the analog board ? A picture of the 24V tachometer motor ? what is the make , model of these ?
any manuals pictures ?

Ships engine is rated at 11,300 KW and turbo tachometer max range is at 20000 rpm sir. Please see attached photos of the tachometer and A/I board used onboard.
 

Attachments

Thread Starter

Pochollo

Joined Oct 23, 2020
10
Even if the tachometer is not being rotated there should be current in the loop. 4 mA in the loop represents 0% of the full scale reading. 20 mA represents 100% of the full scale reading. So if you break the loop and insert a current meter into the break in the loop you should see a reading of between 4 mA and 20 mA. If you get a reading of less than 4mA the loop is broken or the end that is configured as the active end is not supplying voltage to the loop. If I understand the last line of post #4 you are measuring a loop current of 4.01 mA. Is my understanding correct ? Can you try driving the tachometer with an electric drill to see if the system cleared the alarm above a certain speed. It could be the alarm threshold is something above zero RPM. (We don't know if the alarm condition is indicating that there is a fault in the current loop or is indicating that the speed is too low.)
When you set the loop current to 10 mA the display should have read 37.5% of full scale ( 10-4)/(20-4) = 6/16

Les.
Thanks for the reply. You understand it correctly sir, 4.01 mA is the measured loop current when the rpm is 0. Alarm condition is showing sensor failure, which is very confusing since I'm getting a loop current of at least 4 mA. Why does the A/I board senses this as a loop break? This morning I tried to do some adjustments on the zero and span pots on the tachometer and increased the output current loop gradually up to 7 mA, just to verify if there is really an alarm threshold or to find which mA signal will the alarm be cleared, nothing happens and the alarm is still there. This is a 11.3 MW engine and turbo rpm range is from 0-20000 rpm.




Even if the tachometer is not being rotated there should be current in the loop. 4 mA in the loop represents 0% of the full scale reading. 20 mA represents 100% of the full scale reading. So if you break the loop and insert a current meter into the break in the loop you should see a reading of between 4 mA and 20 mA. If you get a reading of less than 4mA the loop is broken or the end that is configured as the active end is not supplying voltage to the loop. If I understand the last line of post #4 you are measuring a loop current of 4.01 mA. Is my understanding correct ? Can you try driving the tachometer with an electric drill to see if the system cleared the alarm above a certain speed. It could be the alarm threshold is something above zero RPM. (We don't know if the alarm condition is indicating that there is a fault in the current loop or is indicating that the speed is too low.)
When you set the loop current to 10 mA the display should have read 37.5% of full scale ( 10-4)/(20-4) = 6/16

Les.
 

Thread Starter

Pochollo

Joined Oct 23, 2020
10
So, if you measure loop current during run, what do you see?
Nothing at the moment Sir since the engine is still on standby. But what is confusing is if I injected a temporary 4-20mA let say a 7 mA on the A/I board everything will be fine, I can see some rpm readout on the system. Is there a possibility that this 4.01mA that I'm measuring from the tachometer output is still too low for the system to tagged it as sensor failed? Thank you sir.
 

GetDeviceInfo

Joined Jun 7, 2009
2,192
Nothing at the moment Sir since the engine is still on standby. But what is confusing is if I injected a temporary 4-20mA let say a 7 mA on the A/I board everything will be fine, I can see some rpm readout on the system. Is there a possibility that this 4.01mA that I'm measuring from the tachometer output is still too low for the system to tagged it as sensor failed? Thank you sir.
Can you get into the controller and read it’s port values? If you inject a 4ma Signal into the port and there is an error, the port may be injured, parameters incorrect, program interpretation incorrect.
 

strantor

Joined Oct 3, 2010
6,782
I don't know much about shipboard systems but I imagine they exist as some mix of the industrial controls that I am familiar with, and automotive controls that I am less familiar with.

It is likely that the system needs to be calibrated for the new sensor. Quite often a 4-20mA sensor does not output exactly 4.000mA for a lowest value [ex: 0.0RPM] nor 20.000mA for a highest value [ex: 20,000.0 RPM]. In real life it may be that (EX.) 3.889mA is produced at 0.0RPM and 20.108mA at 20,000.0RPM.

For this reason most industrial systems have some sort of calibration feature, accessible either through password-protected maintenance screens, or by accessing a PLC with a laptop.

For automotive systems this sort of thing is often done automatically through a "learning/teaching" process. A car may need to be started and idled for xx minutes, 3 times in a row, or driven at >50mph for more than 30 minutes, or driven up to 30mph and back to zero several times, or something like that, after certain parts have been replaced. This is so the ECU can recalibrate itself to the new sensors/parts.

I suspect that is the problem. Your system is calibrated for the old sensor, which probably output >4.01mA at rest. Since it sees less than (whatever that value was) now, it gives an alarm. You need to find out what kind of system this is; is it self learning like an automotive system or must it be manually calibrated like an industrial system? How do you calibrate it? Could it be as simple as waiting until you go back to sea and then everything will be reset and no more alarms? It very well could be, but I would RTFM and know that for a fact before mentioning it to the chain of command.
 

SamR

Joined Mar 19, 2019
5,031
4-20mA signals do not go to zero. The reason why is that a live working system has the minimum current. If you are getting an alarm, then look at the alarm system which should not alarm at 4mA which should tell it that the sensor is there and working. Less than 4mA tells the alarm system that there is a sensor/wiring failure or a barrier failure if there is one for hazardous environments. Look closely at the alarm system as it is interpreting some fault as a failure which it should not do with a 4mA signal available.
 
4-20 mA might be simple, but in the details everything gets messy, A typical 4-20 mA loop: https://www.predig.com/indicatorpage/back-basics-fundamentals-4-20-ma-current-loops

it's probably too simple. How about a ground loop?

Anyway a sensor could have a source or it could be passive. It could also be isolated.

The current can be right, but measuring that current can be a real problem.

When I was doing some DAQ (Data Aquisition stuff) trying to keep stuff cheap, I missed a detail.

My first dealings with 4-20 and 0-5V mixed systems required isolated inputs and outputs, Makes life cake. Thay had to be isolated - no choice. The setpoint of the power supply was referenced to the (+output) @ up to 40Amps. The current was measured across an internal resistor in the (+ output). The power supplies were grounded.

With 7 of these supplies, it was a real mess. Dividers reduced the 40V to 5V for measuring.

So, the 4-20 mA can be passive, active, be in the low side, be in the high side. So, you need to understand how they relate to other systems.

The cheap DAQ system, I wanted all of the DAQ stuff out of the computer, THEY didn't listen. The MacII supported 6NuBus cards, later 3 and later none. Upgrades meant finding used computers until NuBus is not supported. It wasn;t my decision and wasn't my problem.

I was able to have a non-isolated pseudo-differential Analog input card. Lots of wires, but worked most of the time. I used 0-20 mA current output that got converted to 0-5V at the instrument. That worked too. Isolated current would have been the right way to go.

The other one was commercial, 4-MFC controllers, one readout, 4 setpoint pots, but it had the ability to move signal ground

One sensor was a PAIN and required isolated signal conditioning to measure the induced Plasma voltage.

What gave me problems was a in-house made 8 digit readout and computer/manual setpoint for 6 mass flow controllers.
In manual mode it worked fine. The MFC's were supplied with +-12V. The setpoint was potentiometer 0-5V. The setpoint was isolated in manual mode, but in computer mode everything broke loose. You need multiple grounds and I didn;t incluse allof the ground types.

One modification I made was to make the decimal select-able with a DIP switch and Full scale with a panel mount trimmer. Just flip a locking switch from (calib/meas) and adjust FS. The decimal point was adjusted inside the instrument.

So when the equipment went live, you could use one channel at a time or it would change the other set points.
I wasn't allowed to fix it, so they spent $4000.00 for one of the less desired readouts and had to settle for 4 channels instead of 6. I would have needed to re-reference the setpoint signals and I think I would have had to re-reference the measured value too or change the voltage outputs to isolated current. I can't remember. I just knew what was wrong. This was time vs money and less time won.

If you don;t know what was going on, you could have a ground loop problem dependent on other sensors. Like an alarm module and sensors interaction. You won't see the full effect until they are non-zero.
 

sagor

Joined Mar 10, 2019
903
From what I recall many years ago, 4-20mA loops also had some calibration process, either at the source (usually) or at the reading of the input point. It could be as simple as simply calibrating the loop current using your ammeter and a current source, and adjusting a "calibrate" potentiometer somewhere.
 

Thread Starter

Pochollo

Joined Oct 23, 2020
10
4-20 mA might be simple, but in the details everything gets messy, A typical 4-20 mA loop: https://www.predig.com/indicatorpage/back-basics-fundamentals-4-20-ma-current-loops

it's probably too simple. How about a ground loop?

Anyway a sensor could have a source or it could be passive. It could also be isolated.

The current can be right, but measuring that current can be a real problem.

When I was doing some DAQ (Data Aquisition stuff) trying to keep stuff cheap, I missed a detail.

My first dealings with 4-20 and 0-5V mixed systems required isolated inputs and outputs, Makes life cake. Thay had to be isolated - no choice. The setpoint of the power supply was referenced to the (+output) @ up to 40Amps. The current was measured across an internal resistor in the (+ output). The power supplies were grounded.

With 7 of these supplies, it was a real mess. Dividers reduced the 40V to 5V for measuring.

So, the 4-20 mA can be passive, active, be in the low side, be in the high side. So, you need to understand how they relate to other systems.

The cheap DAQ system, I wanted all of the DAQ stuff out of the computer, THEY didn't listen. The MacII supported 6NuBus cards, later 3 and later none. Upgrades meant finding used computers until NuBus is not supported. It wasn;t my decision and wasn't my problem.

I was able to have a non-isolated pseudo-differential Analog input card. Lots of wires, but worked most of the time. I used 0-20 mA current output that got converted to 0-5V at the instrument. That worked too. Isolated current would have been the right way to go.

The other one was commercial, 4-MFC controllers, one readout, 4 setpoint pots, but it had the ability to move signal ground

One sensor was a PAIN and required isolated signal conditioning to measure the induced Plasma voltage.

What gave me problems was a in-house made 8 digit readout and computer/manual setpoint for 6 mass flow controllers.
In manual mode it worked fine. The MFC's were supplied with +-12V. The setpoint was potentiometer 0-5V. The setpoint was isolated in manual mode, but in computer mode everything broke loose. You need multiple grounds and I didn;t incluse allof the ground types.

One modification I made was to make the decimal select-able with a DIP switch and Full scale with a panel mount trimmer. Just flip a locking switch from (calib/meas) and adjust FS. The decimal point was adjusted inside the instrument.

So when the equipment went live, you could use one channel at a time or it would change the other set points.
I wasn't allowed to fix it, so they spent $4000.00 for one of the less desired readouts and had to settle for 4 channels instead of 6. I would have needed to re-reference the setpoint signals and I think I would have had to re-reference the measured value too or change the voltage outputs to isolated current. I can't remember. I just knew what was wrong. This was time vs money and less time won.

If you don;t know what was going on, you could have a ground loop problem dependent on other sensors. Like an alarm module and sensors interaction. You won't see the full effect until they are non-zero.
Thank you sir, I will look into it..I'm just new with this ground loop idea and thankful that you shared your experience with me. Well,I'm getting all information as much as possible to clear out this issue,cause when we are out at sea no more internet access for us. Thank anyways sir more power to you.
 

Thread Starter

Pochollo

Joined Oct 23, 2020
10
4-20mA signals do not go to zero. The reason why is that a live working system has the minimum current. If you are getting an alarm, then look at the alarm system which should not alarm at 4mA which should tell it that the sensor is there and working. Less than 4mA tells the alarm system that there is a sensor/wiring failure or a barrier failure if there is one for hazardous environments. Look closely at the alarm system as it is interpreting some fault as a failure which it should not do with a 4mA signal available.
Got it sir, I will look into it.. More power to you sir.
 

Thread Starter

Pochollo

Joined Oct 23, 2020
10
Tha
I don't know much about shipboard systems but I imagine they exist as some mix of the industrial controls that I am familiar with, and automotive controls that I am less familiar with.

It is likely that the system needs to be calibrated for the new sensor. Quite often a 4-20mA sensor does not output exactly 4.000mA for a lowest value [ex: 0.0RPM] nor 20.000mA for a highest value [ex: 20,000.0 RPM]. In real life it may be that (EX.) 3.889mA is produced at 0.0RPM and 20.108mA at 20,000.0RPM.

For this reason most industrial systems have some sort of calibration feature, accessible either through password-protected maintenance screens, or by accessing a PLC with a laptop.

For automotive systems this sort of thing is often done automatically through a "learning/teaching" process. A car may need to be started and idled for xx minutes, 3 times in a row, or driven at >50mph for more than 30 minutes, or driven up to 30mph and back to zero several times, or something like that, after certain parts have been replaced. This is so the ECU can recalibrate itself to the new sensors/parts.

I suspect that is the problem. Your system is calibrated for the old sensor, which probably output >4.01mA at rest. Since it sees less than (whatever that value was) now, it gives an alarm. You need to find out what kind of system this is; is it self learning like an automotive system or must it be manually calibrated like an industrial system? How do you calibrate it? Could it be as simple as waiting until you go back to sea and then everything will be reset and no more alarms? It very well could be, but I would RTFM and know that for a fact before mentioning it to the chain of command.
Thanks for the info sir,ship board setup is the same as industrial automation which uses a DCS setup. Your idea is possible,that might be the case, that the system should be calibrated or I should check the min and max current to know why the 4ma is still low. One problem for now is that we dont have access on that particular access level to check or do manipulation on the setting, a password is required.Maybe I should post a new thread for this for anyone working or just worked with this kind of automation brand. Thanks for the information anyways,I'm getting info as much as possible, because if we are out at sea no more internet for us,it will be all by ourselves by that time. Thanks to you and more power.
 

strantor

Joined Oct 3, 2010
6,782
Tha


Thanks for the info sir,ship board setup is the same as industrial automation which uses a DCS setup. Your idea is possible,that might be the case, that the system should be calibrated or I should check the min and max current to know why the 4ma is still low. One problem for now is that we dont have access on that particular access level to check or do manipulation on the setting, a password is required.Maybe I should post a new thread for this for anyone working or just worked with this kind of automation brand. Thanks for the information anyways,I'm getting info as much as possible, because if we are out at sea no more internet for us,it will be all by ourselves by that time. Thanks to you and more power.
It is a bit concerning that you and the rest of your crew are going to sea without the ability to troubleshoot your systems because of password protection. In industrial settings it is increasingly more common for OEMs to password protect their systems in an effort to force their customers into perpetual reliance on their service departments. Is this the same in the maritime industry? That seems inconceivable. Is the OEM so bold as to withhold passwords and let you drift about dead in the water until you pay them some insane amount of money to fly to you in a helicopter? I would not be posting in other forums at this point. I would be calling the OEM and demanding the passwords, and escalating the issue as high as possible if I did not get immediate compliance.
 

GetDeviceInfo

Joined Jun 7, 2009
2,192
Just thinking out load. It may be conceivable that a bit has latched somewhere that needs to be cleared, resulting from your malfunctioning device. Does the error prevent you from a start cycle? I assume you run the engine before leaving port.
 
Top