calibrating the calibrator (current loop and similar sensing apps)

[archibaldtuttle]

sorry, cross posted this from electrical engineering froum, didn't see this measurement and control forum.

testing the current loop tester. how important?

i'm looking at the extech line of PRC testers that can provide source and sensor simulation for 4-20ma current loop, voltage measuring sensor regimes and thermistor applications. This is no small investment for occasional field use but i find that, unlike combustion sensing applications that are pushing out to 5 and 10 year calibration intervals Extech wants this to be calibrated every year. my feeling is that they are thinking about much more critical applications than my field work. the specfied accuracies are with hundreth of a percent for most applications and i could handle within 5% so 2 orders of magnitude less accuracy. I'm thinking i ought to just be able to test a few known standards to confirm that i'm within tolerable accuracy because they will want half of what the meter costs to calibrate it every year . . . not to mention the fuss of sending it there and back.

anyone dealt with one of these or something similar from another manufacturer. while its a serious investment this seems to be reasonably priced for the capability but i'm open, of course, to simple alternatives that perhaps aren't built as accurately as this one to begin with. but most of what i find is more expensive or very cheap alternatives that don't have both the source and sensor capabilities.

thanks for any thoughts.

brian

 

[roydm]

I have never been one to put all my eggs in one basket.

For most of my working life I have been using a Fluke 87 and a very simple home built signal generator.

Of course I'm not working on Nuclear power stations or rockets.

 

[danw2]

Does paper rule your world?

Is a customer of yours that is paying for your service going to be satisfied with your calibrator's cert that is more than a year out-of-date?

My pharma customers are not. But what about your clients?

If they don't care, then don't cal every year.

My PIE 331 (mA) was so insignificantly different as-found from the previous yearly cert that I had to squint to tell the difference. But the cert has a recent date on it.

I don't own an Extech, so I can't speak to their functionality, but modern electronics tend to be pretty stable.

 

[IRstuff]

Assuming that Extech has done a plausible job in design and fab, instrument drift should be relatively minor, and self-calibration is not unheard of. So, right out of the box, when the instrument is presumably at its best, you should measure and record a bunch of reference standards and with careful choice of those standards, you can use them to map out the drift of the instrument. If your standards are known to be rock solid, you could even go so far as to take your field data and correct them based on the calibration values from your standards. We often do that with instruments where calibration is extremely difficult or absurdly expensive. If nothing else, the standards serve as a sanity and functionality check of the instrument when measurements might be in doubt.

TTFN
I can do absolutely anything. I'm an expert!
homework forum: //

http://www.engineering.com/AskForum/aff/32.aspx

faq731-376 forum1529

 

[archibaldtuttle]

thanks everyone for the helpful responses . . .

roydm,

i thought about building a 'source' circuit that would allow a standard multimeter to provide a mA reading. I wanted to test the response of some sensors in various installations for which they aren't rated and where i'm trying to work out some isolation piping and monitor their results against analog gauges before i actually try to use them for control. there are a lot of relatively cheap simulators that allow to limit a current loop to a known current acting like a sensor. But it is harder to find ready made source solutions for testing sensor response. at the simplest level i'd just take a battery and attach it to the sensor with an amp meter in series. but i wasn't really sure of the electrical nature of the sensor. i assume its a high resistance variable resistor so that such a circuit could work without any other current limiting device but that was just a guess from looking at simple circuits. The other pause i had was that most theoretical discussions talk about converting and amplifying the signal to read it and i'd need a good deal more understanding to accomplish that.

maybe you would be willing to discuss the theory of your 'tester'. among other things, i maybe don't understand the symbols but i didn't see a meter in the circuit.

i do agree about the eggs in one basket with meter i posted. and it is a pricey way to have all those capacities including thermocouple/thermistor testing and simulation. There are cheaper versions that do just the input/output for current loop and millivolt signaling so I think that will be my compromise (esp. since Zoro is running a sale today for anyone who needs anything) to get going but I am interested in the circuit you posted if you can discuss its theory of ops.

danW2

right on the paper trail. i don't need one. which is why i thought about monitoring drift myself.

IRStuff

which leads me to what kinds of reference standards i could use. thinking i'll need several different types to test source and simulate modes for current loop and millivolt applications. ideas . . . would a high value fixed resistor by like a current loop sensor at a particular value . . . and testing the simulate mode would i just use a current tester and a fresh battery and current meter (of course subject to the drift in the meter itself . . .

this is kind of an academic discussion. i tend to think i'll be quite content for my purposes with modest drift but trying to flesh this out while i'm thinking about it so i can test a few standards when the thing gets here next week.

 

[IRstuff]

Roy's circuit would seem to fit the bill, assuming you use quality components with low drift and good inherent accuracies. The LM317 is relatively inexpensive and the datasheet shows good accuracy and long term stability.

One note of caution about datasheet long-term stability specifications; they are often based on a relatively short term test, and the values presented are not necessarily representation of the true, long-term stability. In many cases, the value simply represents a random variation divided by the duration of the test.

TTFN
I can do absolutely anything. I'm an expert!
homework forum: //

http://www.engineering.com/AskForum/aff/32.aspx

faq731-376 forum1529