Hi,

I came across this forum when searching for a solution for collecting the signal from our new Rosemount pressure transducers with an I/O card from National Instruments that only read voltage signals. The 1st thing I did was to use a 250 ohm resistor to convert the 4-20mA into 1-5VDC, but I guess the resistors I used were not accurate or stable enough. the output voltage signal was fluctuating all the time, and the error was way too much so I have to look for another solution. I saw people suggesting about some off-the-shelf converters here in some threads, but didn't know which one is more (or the most) accurate and could give me a stable voltage output (since the accuracy of the pressure transducers is pretty high, around 0.025%). so I'd be grateful if anybody can give me any suggestions in specific brands and models.

Thank you for viewing!

 

The converters you buy will very likely measure the voltage across a 250 ohm resistor. That is pretty much how you convert current to voltage. They may then scale and ground reference the signal, but those don't have any bearing on your problem. I would concentrate on your card and the technique. A good first step is to use the card to measure a 1.5V battery. If that isn't stable, nothing you add is going to help. If that checks out, then I would put an accurate DMM set to current in series with the loop. If that isn't stable enough you either have a noise problem, or your pressure transducer is bad (unlikely) or your pressure really _is_ jumping around. Now, on technique, which is the most likely cause. I have done quite a bit with the DAQ cards from various vendors, but I have yet to see one where any single reading compared to any other single reading was likely to be within the stated accuracy. In fact, most are fairly useless without some averaging and/or digital filtering. Perhaps in a screen room underground with first class shielding and grounding, but not in the real world. And no matter what the cause, whether it be pressure fluctuations, wiring noise, device noise, whatever, your solution will probably involve averaging and/or digital filtering. So, what you should try, to cut to the chase, is to average 10 or 100 readings and compare that to the average of another 10 or 100 readings. See if that doesn't give more reasonable results. You can also use analog filtering, but that's beyond the scope here. A simple digital filter just processes readings and limits the change between them to a change that could really be happening. Most of the packaged software used with the cards lets you average and/or filter. There is a reason for that :^). So find out where the problem is before throwing money at it. Believe it or not, this is really valuable insight, learned the hard way. Let us know.

Regards
cww

 

The Rosemount is loop powered from a DC power supply.

Is the power supply output clean?
Is the output floating or grounded?
What else is the power supply driving?
Did you run the cabling adjacent to wires/cables carrying high voltage?

Rosemounts convert 20 odd times/second. Do you need that rate of update? If not, have you checked damping or filtering setting or whatever they call it that 'averages' the output signal?

 

>The converters you buy will very likely
>.....
>Let us know.
>
>Regards
>cww

Thanks cww, the pressure transducers also have LCD displays, the display numbers are very stable when applied at a certain pressure, so I think the power supply I use is good enough, the card has also been checked, and it's working without any problems, I did use the software to average the signals from every 10s to up to every 120s, but I was still getting different values once overy time span...

I found this converter on Dataforth.com : 8B32-01, looks like it can convert multiple devices, don't know if it'll work for our setup.

 

Did a current meter in the loop show the same variations? The idea is to divide and conquer and looking at the actual current lets you determine whether the sensor is sending the variations or the card is picking up common mode noise or even bus noise. I have seen a PC that had so much ground and power noise that it was unusable with low level analog cards. On the bus of a PC is a lousy, noisy, place for low level electronics. And I believe the LCD is stable, the update rate is usually pretty low and that allows for quite a bit of damping. A current loop uses, by definition, what amounts to a variable constant current source. It operates at very high impedance with a lot of bandwidth and the voltage noise can be almost unbelievable. The cards don't have a huge cmvr*, and so the focus on the receiving end. I have been in this same sort of situation and it took careful analysis to determine what was going on and quiet things down. It's important to know which end is doing the jiggling. There are a lot of things we could cover like you should be using differential inputs, grounding, etc. But, rest assured this is not very unusual. It's just that sometimes you get lucky, and it looks easy. Best of luck

*(Common Mode Voltage Range)

Regards
cww

 

Just make sure you are not using HART communication while measuring output voltage.

The HART communication is using superimposed frequency of 1200/2200 Hz and +/-0.5 mA. This mA variation will give voltage variation of +/-0.125 V @ 250 Ohm.

 

> Did a current meter in the loop show
> ...
> get lucky, and it looks easy. Best of luck

> *(Common Mode Voltage Range)

> Regards
> cww

I still think the I/O card and the transducers are fine, so the problems can only be on the power adapters, the resistors, or the wirings. Since I'm getting a good reading on the LCD, so I think the power adapters are good, but I'm not 100% sure, the resistors I grabed them from a local RadioShack shop, so that could be the problem, or maybe it's my lousy wiring job . Since I haven't really put too much money, I think maybe I should just purchase some off-the-shelf converters and good power supplies, it might save me some time. Man, I wish I had 10% of the electrical engineering knowledge as you guys .

H.J.

 

I can save you a little, there really isn't much to go wrong with a resistor. If you are using a single ended input, you should have shielded wire. If you are using a differential input you should use twisted pairs, shielded if convenient. But, don't assume your pressure isn't really doing what it shows. That's why people use PLCs, they ignore or filter out a lot of the nastiness of the real world. Results are nice and smooth, and sometimes correct.

Regards
cww

 

> Just make sure you are not using HART communication while measuring output voltage.

Can you explain a bit more about how to "turn off" the HART communication? As far as I know, a resistor (250 ohm prefered) is necessary for the HART protocol to work, so if I want to convert the 4-20 mA into 1-5 VDC, I have to put a 250 ohm resistor, then how can I avoid using the HART communication?

This might be a stupid question due to my poor EE knowledge...

Thanks!

 

Use EID's 4-20mA (Current) to Voltage Converter (transducer)interface board.

See:

http://www.eidusa.com/Interface_Boards_Current_to_voltage.htm

http://www.eidusa.com/Interface_Boards_Current_to_voltage_24X.htm

http://www.eidusa.com/Interface_Boards.htm

Simply connect the board to 24VDC or 24VAC connect your current source 4-20mA to the input and our board will convert it to specific Voltgae. Adj. Zero and SPAN.

Enjoy,
Ron