Results 1 to 16 of 16
Thread: OT - Measuring 4 - 20 mA
07-16-2011, 11:09 AM #1
OT - Measuring 4 - 20 mA
Apologies if this is in the wrong section but I hoping maybe someone who knows their onions in this field such as JST or Motion Guru may chime in.
The boat I now work on has a bespoke Italian monitoring system which is in many ways too temperamental and sensitive; most of the control system outputs are 4 - 20 mA which I understand is pretty standard.
When I was a highly trained Royal Navy spanner wielder, control circuitry was mainly left to the 'pinkies', while us 'clankies' just hit things with big hammers. Now I am in the private sector I am having to diversify.
I currently have a Diesel Generator low battery voltage alarm on the monitoring system; I suspect the V/I converter in the common battery charger is defective. The analogue meter reads 27 volts and I have checked this by meter at the batteries (13.5 V each) .
I know that I can move the leads around to a one of the functional items to see if I can transfer the defect however, I am loathe to do that until we get alongside in 7 days if I can order a replacement part now.
Can someone please enlighten me on the best method of testing this 4-20 mA range, I suspect that I have to measure the current through the circuit rather than across, hence have to break in somehow to get the meter to read current correctly.
07-16-2011, 12:24 PM #2
4-20mA current loop is, indeed, a very common system. You have two basic choices, but both require you to insert something into the circuit you want to measure.
- Insert a current-reading meter into the circuit
- Insert a fixed "sense" resistance into the circuit and read the voltage across it.
The second technique is pretty common, and allows you to keep the loop connected while you move your meter from place to place. A 100ohm resistor would develop 0.4-2.0 volts across it, a 250ohm resistor (which is fairly standard for this purpose) would develop 1.0-5.0 volts. The resistance has to be low enough that the transmitter can drive a full 20mA through it. If the control system uses a 5V power supply (which would be unusual), go with 100ohm, otherwise go with 250ohm.
07-16-2011, 12:28 PM #3
Many inexpensive DMM's will read this small current; possibly you have one with you? It'd do till you can procure a proper replacement.
07-16-2011, 12:53 PM #4
don't ever tell anyone that you did this but you can probe the wires with sewing needles.
however the proper way to do this is with breakout connectors. These connectors will make your life easier and can be made out of what you have if you are not in a place to buy them.
07-16-2011, 01:58 PM #5
Insert a resistor, and all will be well.
Most 4-20 ma I work with is based on a 24V system, and you can easily drop as much as 4 or 5 volts without any problem.
It is a "current source" which has a "voltage compliance"......
This means that as long as you do not drop too much voltage, the current will be whatever the transducer etc sends. The allowable voltage will be in the specs, and you can choose a resistor based on Ohm's law.
It is usual, with a meter reading deal, to use a resistor that gives a scale that means something, but with a calibration table of "parameter to current", that won't matter a lot. That would be just a table of "this current out means we are measuring that temp" (or pressure, velocity, voltage, whatever). Using 100 ohms does give you easy numbers.
Common "voltage compliance" limits are in the 9 to 12 V range, but read the specs.
Ifyou try to use too large a resistance, the current will be wrong, usually "clipped" at the high end.
07-16-2011, 02:07 PM #6
First, review what this circuit hooks up to. Does it engage automatic shutdown stuff, or start emergency generators? Probably not, but the point is before one messes with a live control system, all the consequences of failures (open circuit, shorted circuit, shorting a common power supply that affects other circuits, etc.)
The perhaps you might check the whole circuit for continuity (measure voltage at several points in reference to ground looking for unexpected changes. Of course, you've checked the power supply for correct voltage.
If there is a screw terminal somewhere in the circuit, set your multimeter to "milliamps", and (if not autoranging) to the correct range. Then clamp a lead, one to the wire on both sides of the screw terminal. Then disconnect the screw terminal. You should be reading mA. Reconnect the terminal and remove your meter leads, then use the meter to measure voltage across the batteries. If you know the range of the V/I transducer, you should be able to check whether it's giving the right current output for the given battery voltage. Usually the range starts with zero, for ease of calibration. 4ma is zero. If the high range is 30v (20ma), and you are measuring 27V, then your current should be 4ma + (20ma-4ma) * (27-0)/30, or 18.4ma. Caution: sometimes current transducers are calibrated over a narrower range to give better resolution. A good example here might be calibration over 25-30V. In which case your ma reading will be lower.
If your analog panel meter is hooked directly to the batteries then indicating 27 volts (equaling your actual measurement of the batteries) doesn't help diagnose the problem. But if the analog meter is hooked up to the current loop,then it is reading a voltage across a resistor caused by the current from your V/I. So it's not the V/I that's the problem in that case.
2. Limit sensors and alarm annunciator panels in these circuits often use voltage internally. So to get a 1-5V signal, they use a 250 ohm range resistor in the circuit (that is, in a series connection). If you can find this, you can read voltage across it, and compute the current. Exactly as SFriedberg, above, suggested, only use an existing resistor. If you have 18.4ma, the voltage would be 4.6V.
It may be that the actual low current switch, rather than the V/I, is what is broken or off calibration.
07-16-2011, 04:10 PM #7
Thanks for the replies so far fellas, hers are some pics to perhaps make what I want to do and why clearer:
The monitoring panel screen is in the Engine Control Room, when in alarm it outputs to sounders in the engine room, bridge and most importantly to me, my cabin adjacent to the head of my bed. It can be muted for 2 minutes but needs to be acknowledged in the ECR or it sounds again.
The front of the common charging panel with the analogue meters
a couple more to follow
07-16-2011, 04:15 PM #8
Here is where I wish to measure to compare the outputs of CGE1 & 2 one of which reads correctly and one doesn't
I have a Fluke 87 and 112 both of which will read mA but no handy resistors. How do you suggest I proceed?
07-16-2011, 06:59 PM #9
In a pinch.. a standard graphite pencil can be whittled down with a knife to proper resistance... Fixed tap at one end, Sliding tap to fine adjust.. Check with multimeter out of circuit to adjust. Tape everything up, so an accidental slip does not short anything...
Works fine with low current/low voltage...
07-16-2011, 08:47 PM #10
Just break into the current loop and wire the fluke in series set to current measuring mode, put probes in correct holes and measure current. No need to faf with resistors if you have the range on your meter.
The if you have a good source of air on board remove the Faulty item and blow all the black soot out of it.
When you get to port i would blow that whole cabinet out. Control electronics like to be kept clean. That much dust could well be shorting out enough of your current signal to cause the alarm state!
07-16-2011, 11:12 PM #11
It could be an open circuit on your current loop. Measure the voltage at the output of your V/I converter and then find the input to your monitoring system and measure the voltage there. If it's the same, then it's your analog-to-digital circuit in your HMI (human machine interface).
If it's 0 volts you have on open/loose connection.
If you have a lot of 4-20 ma on board it might be worthwhile to get a current simulator, Fluke offers them in their high end VOM.
A 250ohm resistor hooked in parallel with a 4-20 ma loop outputs a 1-5 vdc which is another common interface scheme
07-17-2011, 12:13 AM #12
If you are trouble-shooting a system that is not "live", i.e. not working at its normal task, then you have more options.
If you have to do this live so that an interruption will cause an issue, then it's different.
Also different if you need calibration vs needing to know if something appears to be working.
For the "is it working?" question, you may actually be able to measure across the device that receives the 4-20 ma. While some are set up to read the current with no voltage change, many internally have a small resistor that creates a voltage drop which is then measured (your meter does this, the voltage is known as the "burden voltage').
In the latter case, you can go around with your meter and measure across the devices and see if you have a voltage (indicating a current), and especially if it is changing in response to whatever the transducer is measuring.
Otherwise you will have to do the "find a screw terminal" routine mentioned above. It can get tricky, and some points may not be good candidates, if a temporary "open" has bad consequences.....
Systems like this should be made with monitoring points where you can do that meter insertion in a more controlled manner, but they often are not.
07-17-2011, 01:59 AM #13
As always, the schematics help!
I don't think you have to intervene in this loop to measure current (or at least to diagnose the loop).
Your current loops all can be read using the analog meter "VT". That meter is likely a moving coil pivot and jewel milli or microammeter, ginned up with series and shunt resistors so that the meter indication reflects the voltage applied across the terminals. In your setup, there is probably a range resistor acting as a shunt across the VT's terminals. So you have a device that reads current, modified to indicate voltage, shunted to reflect current. Go figure. But the key thing here is that you can diagnose each of the V/I guys with the meter. Use your 87 to measure battery voltage 1, 2, and 3. Switch the analog meter to batteries 1, 2, and 3. The analog meter should indicate the same voltage as your DMM. Obviously, if you get good readings from two V/Is, and an off-spec (or no) reading from another, the problems is the V/I or the wiring to the VI (or the meter switch, I guess).
If all three battery voltages are indicated correctly (and assuming that the problem isn't a intermittent V/I failure) it would seem that you have two other problem sources: the wiring, or the digital I/O card in the PLC or it's power supply connection. You've said that the system is squirrely, and your inputs go into the PLC, meaning that a channel in the PLC digital input card may be knackered and/or its calibration might be a bit off.
When you get to a point where the consequences of any failure (open loop, closed loop, failure of common equipment, e.g. the entire PLC) is low, you may wish to loosen, clean, reseat and reconnect all wiring terminations in that loop. In marine service, check all crimped fittings for continuity and resistances. Reseat the board supplying power to the V/I, undo all connections and reconnect using some sort of antiox compound, and check function with your analog meter "VT". If the thing still give you problems, you may want to ensure that the loop power supply is securely connected to the offending circuits loop. IF that doesn't fix it, swap in a new PLC I/O card. I agree with a previous poster that the dust in the cabinet (plus salt air and such) might have had a bad effect on the devices, or on the seating of the devices in their electrical connectors, or in wiring terminations.
In terms of ordering now, you probably should have a spare I/O card for that PLC anyway, and what does a V/I cost in relation to 1) you not getting sleep, or 2) the ship not getting underway on time?
And if you feel you must measure the current in the loop with the suspect V/I, just measure the voltage across meter "VT". With the meter range, and a known battery voltage and voltage across the meter you should be able to figure out the actual current in the loop of interest.
BTW, I just recalled that sometimes signals are 0-20ma - just be alert to that fact.
07-17-2011, 08:00 AM #14
Only dodgy set-ups use 0-20ma. There is no knowing its failed. Seeing less than 5ma on a current loop can cause a error message not just something is low. That can be a very important safety feature! The whole idea of the 5-20ma current loop is that its a very simple and robust method of transmitting a remote reading.
07-17-2011, 09:43 AM #15
Thanks to all who've replied!
As the sea state has worsened I'm going to wait until we're alongside before I stick my fingers in here with the system live. I certainly don't want to make anything any worse by sticking a probe the wrong place caused by a violent ship roll
I do agree inside the cabinet is filthy and this could be the root cause of any defects.
I will update what I find as I often find those who've posted advice never get closure on what worked
07-17-2011, 12:30 PM #16
In any case, in systems I designed, I use 4-20ma. Hell, I even designed systems with 3-15psig pneumatic signals. A three mode pneumatic controller had some pretty cool machinework inside it. The control panels would have miles of 3/8 stainless tubing behind them. IT was quite pretty.