Controlling drives with 4-20mA

[delectric123]

I am trying to integrate level control into an existing system. I didn't install the probes but the guy who did has tried to fine-tune them, even consulting the manufacturers with little luck. It has been working so-so for a couple years now.

Its a Schneider Electric Altivar 61 480-VAC 15 HP drive with for a pump that keeps a 50 gallon stainless steel cone shaped tank full on a separator that separates the solids from the water. The separator is about 15 ft. above the pump on a different floor. The separator uses a 480 VAC 3 hp motor controlled by a motor starter. The drive references a 4-20 mA 36" probe that's made by Delta Controls. There are two such setups side by side. That means two drives, two pumps, two separators with tanks, and two probes. Both are giving the same problem.

I've tried everything I can to get the tanks consistently full with water. I've tweaked the PID settings so the pump consistently accelerates and decelerates, but my problem is, the 4-20 mA reading is inconsistent. One day the tank is overflowing, the next day I check the level is about 4 in. different. Its probably best to talk to the manufacturers but maybe you guys can tell me something I don't know about 4-20 mA control.

These are $500 probes and I've done calibrated them correctly according to instructions. They're about 100 ft. away from the drives, the signal cables are 18 AWG twisted shielded in a separate conduit and with shield grounded at drive and probe housings are plastic so there's no grounding at water tank. The drives for both pumps are in a stainless steel enclosure right next to the 480 VAC Service Panel. The pumps are about 50 ft away from drives. I checked all the grounding and bonding and found no error. The conduit runs for the pumps are never even close to the signal conduit. The signal conduit runs in an isolated place crossing all line voltage conduits.


!00 % value at drive is 36" but drive never reads more than 30" to 32". A lot of times I've checked the probe output reads 100%, but the drive analog input only reads 17 mA. One thought, maybe the stainless steel tanks are goofing things up.

 

[don_resqcapt19]

...A lot of times I've checked the probe output reads 100%, but the drive analog input only reads 17 mA. ...

If the probe is trying to send 20mA at 100% and the drive is only seeing 17mA, then you have too much resistance in the 4-20mA loop circuit and the probe cannot not source enough voltage to drive 20mA through the circuit.
If you just connect the output of the probe to a meter, do you see the full 20mA?

 

[GoldDigger]

If the series resistance includes some high resistance bad connections or wires that undergo extreme temperature changes either one could cause inconsistent readings at the controller.

 

[iceworm]

... !00 % value at drive is 36" but drive never reads more than 30" to 32". A lot of times I've checked the probe output reads 100%, but the drive analog input only reads 17 mA. ...

How are you measuring the probe output and the drive input?

4-20ma is a series circuit. If there is 20ma leaving the probe there will be 20ma at the drive. Power supply voltage or loop resistance doesn't change that. If the loop resistance is too high or the power supply voltage is too low, the probe will not put out full current - it won't reach 20 ma.

Sometimes the receiving end (vfd) will be terminated with a 250 ohm resistor. This changes the 4-20ma signal to 1-5volts. However, 250 ohms is not a standard value. Standard values of 240, or 270 ohms won't work. 20ma will not match up with 5V.

But, you said there is only 17 ma reaching the vfd when the probe is 100%. Is that "100%" read on a probe indicator or are you measuring the signal? If you are not measuring the current output, you need to do that. It really has to match the 17ma at the VFD end.

If the probe output is low and the indicator is reading 100% - now is the time to check the loop resistance, and the power supply voltage.

ice

 

[iceworm]

...These are $500 probes and I've done calibrated them correctly according to instructions. They're about 100 ft. away from the drives, the signal cables are 18 AWG twisted shielded in a separate conduit and with shield grounded at drive and probe housings are plastic so there's no grounding at water tank. ....

One other thing - but I suspect you have already done this.

This thing has never worked from the get go. Get the probe book and the drive book and go through the installation details and commissioning details. Start with the idea that none of it was installed or commissioned correctly. Check that everything was installed exactly per the manual - even the stuff that seems like it wouldn't matter.

ice

 

[Jraef]

It might also be an issue with what most of us would recognize as a simple "signal gain" adjustment on the VFD end, wherein you tweak how the drive interprets the maximum value that the analog input can see. This is not an unheard of concept, that's why gain adjustments exist in loop controllers.

The problem is, from my personal (and extremely frustrating) experience specifically with ATV 71/61 drives is that in the "Frenglish" translation of the manual, the concept of "signal gain adjustment" is nowhere to be found. Maybe someone else here will know what the French call it within the thousand or so parameters, but I got the rug pulled out from under me by the GC on the project where I was using the Schneider drives before I ever figured out how to solve that exact same issue. I had been on the phone for 5 hours with Schneider tech support, nobody I spoke with had any clue of what I was talking about so the GC finally told me to can the Schneider drives and put in something else. I did, and got it working correctly in 10 minutes.

 

[zbang]

What's supplying the current source? If you use a badly-filtered switching supply, the source voltage can vary faster than many probes can control the current, so you get wildly-varying readings (this is best seen with a scope). Assuming the source voltage isn't too high, try using a stack of batteries and see if the problem goes away. One place I had this, a 1000 mF 50v cap. across the supply was all we needed to stabilize the readings.

 

[delectric123]

I've consulted both the probe manual and drive manual and everything seems up to par. I've doublechecked all the bonding because that's what I strongly suspected.

The 17 mA reading is from the drive interface in the Monitoring/ Analog Input Map menu. The 100% reading is from the probe display, so I should probably take a digital milliamp meter and check the current.

I'll also take an oscilloscope and watch the voltage.

Thanks for the great responses.

 

[iceworm]

...The 17 mA reading is from the drive interface in the Monitoring/ Analog Input Map menu. The 100% reading is from the probe display, so I should probably take a digital milliamp meter and check the current.

Yes - two meters, two guys, radios so you get simultaneous readings. Unless they are close together - then just two meters.

ice

 

[iceworm]

...The 17 mA reading is from the drive interface in the Monitoring/ Analog Input Map menu. ...

Could be the drive input has a 250 ohm resistor and is actually getting a 1-5V -- and the drive internal brains is interperting the input to show the miliamp signal. Might want to check that.

I've seen more than one with a 240 ohm radioshack resistor* - that is an automatic 4% drop in the signal the VFD will see.

ice
* Interesting note: When I have seen these installed, no one has ever admitted to changing out the resistor :roll:

 

[delectric123]

Could be the drive input has a 250 ohm resistor and is actually getting a 1-5V -- and the drive internal brains is interperting the input to show the miliamp signal. Might want to check that.

I've seen more than one with a 240 ohm radioshack resistor* - that is an automatic 4% drop in the signal the VFD will see.

ice
* Interesting note: When I have seen these installed, no one has ever admitted to changing out the resistor :roll:

According to installation manual, the 250 ohm resistor is integral.

 

[delectric123]

Checked out the setup again today and took some current readings. Everything is wired according to instructions. Its a simple 4-20 mA signal loop. The probes are powered by the signal loop, so only two wires connected to probe. The probe manual specifies that the probe output is 0 - 20 mA. When tank level was empty probe display read 0%, drive display read 0 in", and current measurement was 19 mA. When tank was full, probe display read 100%, drive display only read 20 in" (max value is 24"), and current measurement was 110 mA! That is baffling!:huh: I think its time to find a helpful tech support guy.

 

[hurk27]

Checked out the setup again today and took some current readings. Everything is wired according to instructions. Its a simple 4-20 mA signal loop. The probes are powered by the signal loop, so only two wires connected to probe. The probe manual specifies that the probe output is 0 - 20 mA. When tank level was empty probe display read 0%, drive display read 0 in", and current measurement was 19 mA. When tank was full, probe display read 100%, drive display only read 20 in" (max value is 24"), and current measurement was 110 mA! That is baffling!:huh: I think its time to find a helpful tech support guy.

Edited because I didn't see your post 11

To clear one thing up, this is not a 4-20ma loop, it is a 0-20ma loop and the probe has to be calibrated as such, if you calibrate it to 4-20 the drive might not shut down or in the case of an inverse function never reach full speed as the drive will interpit the 4ma as just under 25% or just over 75% in inverse function.

It sounds like you have a zero and span issue with the probe, you are going to have to find out how this probe is calibrated, it might be just a set of potentiometers labeled either zero/span or S and Z or it could be a set of push buttons also labeled zero/span Z/S that when pushed at zero and 100% will calibrate the output current to the display at 0ma, and 20ma respectively, the other method is if the probe is a hart type then you will need a hart communicator to adjust the settings.

I didn't see if this pump is filling a tank or emptying the tank, if it is filling the tank then you must have an inverse signal, meaning when the tank is at 0% the output or input of the drive is calling for the tank to be filled, sometimes this can be set at the drive to reverse the function of the 0-20ma input, or the probe can sometimes be set or calibrated with an inverse output by zeroing it at 20ma and spanning it at 0ma or turning on an output inverse function to reverse the output zero and span, the drive could also be programed to go to set full speed when an input of 0ma is seen, and shut down at 20ma.

But I think you can clear up the problem if you can get the zero and span set on the probe, as the output you stated is way off, remember if filling the tank with the pump the zero (0%) might be set to 20ma, and the span (100%) set to 0ma, if you are emptying the tank with the pump then the zero would be set at 0ma and the span set to 20ma, this also depends upon if the drive has an input inverse enabled.

look in the probe manual for how to calibrate the probe, or post the manufacture and modal of the probe and maybe I can look it up.

Also if your going to be working on ma current loop control systems you should get a fluke ma calibration meter that has source and read functions, it will also have step functions to step through 0%, 25%, 50%, 75%, and 100% which would be 5ma, 10ma, 15ma and 20ma respectively which would let you cycle through the input control to the drive or whatever you would be working on it wold also read the output of the probe or any ma transmitter such as a pressure transmitter.

 

[hurk27]

Also please post how this loop is getting power and what voltage the PS is, I have no idea how you can get 110ma in a 20ma loop as I don't think the span can go that high, something doesn't sound right and it could be in how the loop is fed? if you feed any transmitter and put a 250 ohm resistor in series in the loop the transmitter should be very close to its rated current, if someone has used a higher voltage they could have damaged the probes controlling output, we use 24 volt power supply's for all out 4-20 and 0-20 loops, although if we have extra resistance in the circuit we might bump it up to 36 volts but not in a short run like this, if the loop is excited from the drive then the positive from the drive should go to the probes positive,
if the loop is externally excited from a separate power supply then the positive from the power supply goes to the positive on the probe and the negative from the probe goes to the positive of the drives input and the negative of the power supply goes to the negative of the drives input not sure if its the X or Y.

If you calibrated the probe how did you do it? did you have a calibration meter such as a fluke MA calibration meter? did you do the span as well as the zero?

 

[Rampage_Rick]

OK, just to verify that you have just a level sensor and a VFD. That means you're powering the 4-20mA circuit from the 24V output of the VFD, correct?

+24V on the VFD connected to + on the level sensor
- on the level sensor connected to AI2 on the VFD
(the 4-20mA return "COM" should be internally connected to the power supply negative inside the VFD)
AI2 min value programmed to 4 mA ...unless the probe is supposed to be 0-20mA:

The probe manual specifies that the probe output is 0 - 20 mA.

 

[delectric123]

Sorry I made a typo on post #12, "The probe manual specifies that the probe output is 0 - 20 mA"
It should have been, "The probe manual specifies that the probe output is 4 - 20 mA"

Also please post how this loop is getting power and what voltage the PS is, I have no idea how you can get 110ma in a 20ma loop as I don't think the span can go that high, something doesn't sound right and it could be in how the loop is fed? if you feed any transmitter and put a 250 ohm resistor in series in the loop the transmitter should be very close to its rated current, if someone has used a higher voltage they could have damaged the probes controlling output, we use 24 volt power supply's for all out 4-20 and 0-20 loops, although if we have extra resistance in the circuit we might bump it up to 36 volts but not in a short run like this, if the loop is excited from the drive then the positive from the drive should go to the probes positive,
if the loop is externally excited from a separate power supply then the positive from the power supply goes to the positive on the probe and the negative from the probe goes to the positive of the drives input and the negative of the power supply goes to the negative of the drives input not sure if its the X or Y.

Ands that's exactly as its wired. The loop is powered by a Rhino 24VDC Power Supply from Automation Direct.
The drive input terminals are labeled AI1 and COM with COM being connected to power supply 0V.
I've done other successful 4-20 mA control setups but maybe there's a concept I'm missing on this one.

But I think you can clear up the problem if you can get the zero and span set on the probe, as the output you stated is way off..

The probe does have those advanced settings. But I didn't want to mess with them as the instruction manual just tells how to do the simple calibration of setting reference points. That is a simple procedure of scrolling to the correct menu, completely filling the tank, and pressing Save, then doing likewise when the tank is empty. And believe you me I did not misread instructions. :roll:I've sometimes got an obsession with "doing things according to manual instructions"

look in the probe manual for how to calibrate the probe, or post the manufacture and modal of the probe and maybe I can look it up.

Thanks! :thumbsup:I actually have company restricted Internet with a limited number of sites open, otherwise it would be a breeze finding advanced help. I still haven't gotten around to giving probe manufacturer a call.

Manufacturer: Delta Controls Corporation
Model #: 173-4XI-AA-AA

 

[hurk27]

Ok after looking at the probe spec's from here : Spec's for the 173 probe click on the data sheet link at the top
this probe is onboard calibrated via 4 push buttons, it is also reversible so the 4ma can be programed for the 100% and the 20ma can be programed for the 0% if needed for a tank fill type operation (inverse operation).

two major points to look at:
Make sure the drive is not set to source the loop voltage, because if it is and your using a external PS also your loop voltage will be way above the 35 vdc max that the probe can adjust for.

The drive is programmable on the 0-20ma analog input, make sure it is programed for 4-20ma or if the inverse operation is set at the drive it would be set for 20-4ma for 0% to 100% only one or the other must be set for the inverse not both the probe and the drive other wise the drive speed will try to run as the tank is full and turn off when the tank is empty, if the drive uses a 0-5 volts in the programing instead of 4-20ma because of the 250 ohm resistor then make sure it is program for 1-5 volts on the analog input or 5-1 volts for inverse operation, which is what the 4-20ma will give it across a 250 ohm resistor.

Another possible problem is if the shape of the tank is irregular, the instructions say the probe can be calibrated in up to 10 points for linearization of the output, this is so the drive will run the pump at the correct speed as it fills even though the volume of the tank at different point changes in a non-linear amount as it fills, you will have to calculate the volume at each set point to know what percentage to program each set point.

if the drive is still set for 0-20ma (0-5volts) analog input I can see why your having a problem with level control as the drive will never see it's program 0%, if the drive is sourcing loop voltage and you also have an external source I can see why you are getting a 110ma reading, so these are two possible problems, along with the tank not being linear. also the drive should be running at 100% when the level is at 0% and the drive should be at 0% when the level is at 100% of the level set point (inverse operation), also this level point can be at 75% on the probe or any level it needs to be by calibrating the probe or drive to shut down at that point by setting the drive 0% at the level needed this can be done at the probe or drive, but calibrating the probe to output 0ma at 75% (or what level is desired) would make it much easier.

Without being on site it's hard to tell you all the things to check, but if you have a general knowledge of how it operates you should be able to understand the above.

Let me know if you need any more info and I will try to help.

 

[delectric123]

Thanks for the info. But I can't access that link, because I have limited Internet access. But I think I have that installation manual you're referencing, and everything's done as specified. Anyways I've got an explanation for those high current readings. I tossed that Multimeter aside that I was usinghmyIts a Meterman, but I suspect somebody damaged it from having the probe in the wrong connector when measuring voltage :rant::rant, and took an Amprobe multimeter that actually has settings for loop current measurement, and I got reasonable measurements. Voltage at transmitter is 19 volts full level. I got a 0% measurement of 3.5 mA and a 100% measurement of 17 mA. I talked with a tech support guy, confirming everything was done according to installation instructions, then we went over the advanced settings, tweaking the Zero Suppression and ?Span? settings to get 4-20 mA. There's still something not quite right but I/we'll get to the bottom of this yet. I'll describe the details if we get it resolved.

 

[hurk27]

I have been thinking a little about this process and I think I see a problem with this kind of set up if the tank discharge (outflow) can vary over time?

If the above is true then there is no way without a process controller to control the level to maintain a certain level if you have a varying output out of the tank.

Simply all the probe is going to do is when the level starts to drop it will start ramping the speed of the pump up, but if the out flow is greater then this lower speed can refill the tank then the level will keep dropping till the probe gets to a level that it runs the pump fast enough to match the output flow, at this point it will find an equilibrium that will cause the pump to just keep the tank at this lower level until the out flow is reduced or stopped, another words the tank level will be a function of the out flow and will change every time the out flow rate changes.

The only way you can make this work is to use a microprocessor based 4-20ma controller so you can use a set point to where you want the level to stay and it will control the speed of the motor to maintain that level, it will have the normal set point as well as PID functions to control how fast it gets back to level after the tank starts emptying using proportional, integral, and derivative values to take all the math work out of the problem, by adjusting these you can get the level of the tank back to set point within 1/4 wave of the oscillation which will run the pump more efficiently at maintaining the level at a more precise level. I strongly suggest looking into installing a controller because any level process that has a variable out flow can not be controlled by just a probe controlling a drive, the same goes for controlling a fill valve.

at work we use mostly Marks and Honeywell and have very dependable results with them, but there are lower cost controllers on the market, but I like working with the Marks controller because it is very simple to dial in the PID functions, of course in a process like this I would not touch derivative as it most likely would not be needed and can send it into a wide oscillation that would be very hard to control, but some do like to add a little derivative to speed up getting it back to set point but just a very little too much and it will over shoot the set point and there you go a nice wide swinging back and forth, some in the process field call the derivative the get me to set point now!) setting but it is a very touchy setting that can throw off your controlling very easy and in most cases it is not needed, if you think of the anticipator (small adjustable resistor that tells the furnace to fire ahead the actual time it calls for heat) on your wall thermostat for your furnace at home then you can understand what the derivative does but much more responsive.

but for the most part most of our control comes from the DCS (PLC'S) which has the PID's built in the software, but we also have allot of stand alone controllers for smaller processes, and I wish you could do it with out a controller but I don't see a way it can be done if the out flow rate can change even if it is just a little.

You mentioned in one of your post that you had adjusted the PIDs, but you have never mentioned a controller, I'm curious what PID's did you adjust and is there a controller in-between the probe and drive? are you calling the adjusting the ramp values on the drive PID's?

Please don't take me wrong if I accidentally step on your knowledge as from my computer chair it's hard to see all the details and to know how much you are familiar with process control systems, I myself have just been through two years of training to get me back up to speed with it and get me up to date and had to relearn very fast as it is one part of my job that I'm called upon every day to do, I do systems repair for a steel mill that has every imaginable process control there is, and controlling a blast furnace is like controlling a bomb, not much room for mistakes, how they did it before computers I can't even fathom, most of the process control I did before I went to work here was done with allot of very complicated regulators and mechanical valves as well as mechanical PID controllers (mostly air) and some old electrical control systems that now as I look back and see how hard it was to keep a process running smoothly, we didn't have PLC's and or 4-20 control systems back then, but the basic's are still the same just the equipment we now use has changed a bit.

 

[Rampage_Rick]

The VFD he's using comes with a PID controller built in: http://static.schneider-electric.us/docs/Machine Control/OEM Solutions/8000DB0701.pdf

If I had to pick a different controller to use, it'd probably be an Omron E5EK-AA2 since it does self-tuning.


Where we're at now: Full should read as 20mA but only seeing 17mA. There's 19 volts across the transmitter, and presumably 5 volts across the drive's input (0.02A x 250 ohms) which equals the 24 volts from the power supply. The transmitter should function down to 11 volts, so that's not the issue. The transmitter specs say it's good to 650 ohms on the loop, minus the 250 ohms on the drive means we've got 400 ohms to work with. That's 8000 feet of 24-ga cable.

I'm stumped...

I guess you could program AI2 Max Value as 17mA but that doesn't really fix the root cause.

Or use a 500 ohm resistor to convert the signal to 2-10 volts and connect it to AI1