Possible 4-20mA to 1-5V conversion problem

atferrari

Well-Known Member

Most Helpful Member

• Jun 1, 2009

• #21

Former Burr Brown has one or two doing the same. And a very nice receiver for the other side of the loop (RCV240).

 

ctorregrosa

New Member

• Apr 10, 2017

• #22

I tested the following circuit with the Multisim 14 software. From a 4-20mA current loop I got between 0.33 - 4.98v using a +12v and between -0.01 - 4.98v using +/- 12v. Any advice?

 

Last edited: Apr 10, 2017

ericgibbs

Well-Known Member

Most Helpful Member

• Apr 10, 2017

• #23

hi ct,
It is a result I would expect, single supplies with a LM358 will not give a 0V output, the dual supply will give 0V thru +5v.

What is the source of the 4 to 20mA, what is the application.
E

 

MikeMl

Well-Known Member

Most Helpful Member

• Apr 10, 2017

• #24

Let me try again: I think I have said this before at least twice!

If you are willing to reference the 4-20ma loop to the Arduino GND pin (there-by possibly introducing undesirable common-mode currents into the sensor loop), why bother with an opamp at all? E=IR. When I=4mA, E=0.004A*249Ω = 0.996V, when I=20mA, E=0.020A*249Ω = 4.98V. This range can easily be converted by the Arduino's A/D. The input impedance of the A/D is huge compared to 249Ω, so why buffer?

The only reason for adding an amplifier is if it reads the voltage drop across the loop load resistor differentially, thereby preventing referencing the loop directly to the Arduino's GND pin. In that case, as I think I have already explained, it is preferable to use an integrated instrumentation amp.

 

K

KeepItSimpleStupid

Well-Known Member

Most Helpful Member

• Apr 10, 2017

• #25

A unity gain IA or differrence amplifier gets rid of the ground potential differences, but you generally won't be able to achieve the rail voltages for the ouput. 0 and 5 V are both unabtainable. You can get close like your results.

 

MikeMl

Well-Known Member

Most Helpful Member

• Apr 10, 2017

• #26

KeepItSimpleStupid said:

...but you generally won't be able to achieve the rail voltages for the ouput. 0 and 5 V are both unabtainable. You can get close like your results.

Click to expand...

Since he never has to get below 0.996V (because the minimum loop current is 4mA), and never has to pull above 4.98V when the loop current is 20mA, then even a lowly, non-rail-to-rail LM358 can do that if it runs on 0V and 9-12V as he shows in post #22 . Same would apply to an appropriate INA. He doesn't need to pull to either rail.

 

ctorregrosa

New Member

• Apr 10, 2017

• #27

Hi ericgibbs, I'm developing, on an Arduino platform, a telemetry equipment that must read the data of a flowmeter (Dynasonics TFXL) and an ultrasonic level sensor (FlowLine DL24). Both units transmit a 4-20mA signal.

The FlowLine DL24 sensor works from 0.1 to 3m. Its resolution is 1 mm so the A/D converter must be 3/0.001: 3000 ~ 4096 = 12 bits. The ADS1015 converter could work.

The ADS1015 converter requires an input signal of +/- 6.144v to operate in the 12 bit range.

So the work of the operational amplifier would be to convert a 4-20ma signal to +/- 6.144v

 

Tony Stewart

Well-Known Member

Most Helpful Member

• Apr 10, 2017

• #28

you don't have enough CMRR range for the CM gain from mismatch and Ferrite CM chokes with twisted pair is a minimum consideration for 60Hz (large inductance)

shields up on cables , if not enough ! with source ground only.

 

ericgibbs

Well-Known Member

Most Helpful Member

• Apr 11, 2017

• #29

hi ct,
A quick scan of the ADS datasheet suggests that you have misunderstood the Ain limits.??
The requirement of an amplified single ended signal of +/-6.144v for the ADS input is incorrect.
Note the clamping diodes on the Ain inputs to the Mux.
Also Note #2 of Specs.

I will look into the d/s further and let you know.
E

 

Attachments

• A002.gif

  25.7 KB · Views: 329
• A003.gif

  19.8 KB · Views: 319

MikeMl

Well-Known Member

Most Helpful Member

• Apr 11, 2017

• #30

Eric, I dont understand that data sheet, either. Say Vdd=5V. Then the allowed range of any analog pin is 0V to 5.3V (Vdd+0.3V). If measuring differentially, then V(ainp) could be a max of 5.3V while V(ainn) could be a min of 0V, hence V(ainp)-V(ainn)=+5.3V. How could it ever get to +6.144V?

Unless, the 12bit (4096 steps) range of the AD is from -6.144V to +6.144V, or 3mV per step. Due to the input restriction, only about 5.3/6.144 = 0.86 (86%) of the steps can actually be reached.

 

Last edited: Apr 11, 2017

ericgibbs

Well-Known Member

Most Helpful Member

• Apr 11, 2017

• #31

hi Mike,
I also find the datasheet puzzling in the way it describes the parameters.?

I have never used this ADC.
but checking the clips of the internal circuitry its shows clamp diodes to 0v and Vdd, which has a Vmax of 5.5V.
So why the OP wants a OPA output span of +/-6.144V doesn't make sense.

Do you have any other information on this device.?

Eric
Attached d/s for ref.

 

Attachments

• ads1015.pdf

  1.3 MB · Views: 292

ericgibbs

Well-Known Member

Most Helpful Member

• Apr 11, 2017

• #32

From the d/s Table 1 and the text it explains the +/- 6.144v in more detail.

E


EDIT:

Table 1 shows a resolution of 3mV/bit, so for 4096 bits that is a theoretical range of +/-6.144v,
however the permitted input range is say 0v thru +5v for a Vdd of 5v.
Which I infer as an input of +/-2.5v, that is a limit of +/-833 bits at a 3mV resolution ie: 11 Bits.

 

Last edited: Apr 11, 2017

MikeMl

Well-Known Member

Most Helpful Member

• Apr 11, 2017

• #33

You beat me to it... I was also going to post a reference to para 8.3.3

Now CT has to tell us what he really wants to do with it. It says in para 9.1.2 that you can only get half the dynamic range if you use it in the single-ended mode (11 bits), and I cannot see how to create a larger differential signal starting from 4mA to 20mA because the inputs are constrained to be between -0.3V and Vdd+0.3V.

 

Tony Stewart

Well-Known Member

Most Helpful Member

• Apr 11, 2017

• #34

as atferrari suggested , this is how it's done.

 

ericgibbs

Well-Known Member

Most Helpful Member

• Apr 12, 2017

• #35

read the data of a flowmeter (Dynasonics TFXL) and an ultrasonic level sensor (FlowLine DL24). Both units transmit a 4-20mA signal.
The FlowLine DL24 sensor works from 0.1 to 3m.
Its resolution is 1 mm so the A/D converter must be 3/0.001: 3000 ~ 4096 = 12 bits. The ADS1015 converter could work.

Click to expand...

hi ct,
Would I be correct in assuming that the Flow Line sensor will be sensing flow in only one direction.??
ie: no flow is 4mA and maximum flow rate is 20mA.?

also
the Level sensor is measuring a level from a zero value thru to a maximum level.
ie: Zero level is 4mA and and max level is 20mA.??

If so, why are you requiring a OPA/IA that gives a +/-Vout for the ADS1015 input, when a single end conversion of 4mA/20mA to +1V thru 5V would be OK.??
[ie: 11 Bit resolution]

Please confirm.

E

 

Tony Stewart

Well-Known Member

Most Helpful Member

• Apr 14, 2017

• #36

Your fluctuations may be mechanical attachment or EMI depending how it behaves.???

Mostly likely is poor CMRR on cables or CM gain nulling in balanced circuit, as I showed how to fix. with @atferarri's ref.

For bipolar measurements on a single supply PIC, you must choose Vdd/2=Vref as the zero offset value and another Vref for the ADC gain.

3 common mistakes are: 0V offset at. midpoint 12mA, GAIN ERROR and CM error.