Ohm meter Circuit for digital panel meter?

[eTech]

Thought I'd ask before making one..

Anyone know of a simple ohmmeter circuit to use with
Digital panel meter?

Digital Panel meter spec:
1. Input Reads (0-10vdc, but is scalable)
2. Pwr Supply 12 vdc

Ohm meter circuit requirements:
1. Range 0-200 ohms (just single range)
2. Resistance under test is isolated.
3. Accuracy 3% or better

Thanks

 

[ChrisP58]

Feed the unknown resistance with a constant current source and read the voltage directly across the resistance. 50mA will give you 0 to 10V across 0 to 200 ohms. Use Ohms law to calculate different voltage/current values.

Isolating the resistance measurement will be more difficult. It would be easier to isolate the panel meter and current source power.

 

[strantor]

I had this problem before. I had a Newport 1/8 DIN panel meter and the measurements were real flaky. Replaced it and same problem. I searched high & low, looking for a 1/8 DIN panel ohmmeter that was not made by newport and could not find anything. I made one with a scalable redlion current process meter; I put a 24V power supply in series with a 480Ω resistor (high wattage pot) in series with the panel meter, going out to test leads. My measurement range was 0-2KΩ, so the process meter measured 50mA at direct short, and 9.68mA @ 2KΩ. This was not linear, but the process meter had 9 scaling input points, so I was able to construct a multi-segment line over the top of the curve. After a plethora of tests and adjustments and retests, I determined that is was a crappy kludge and had no place in acceptance testing. The accuracy varied over the range and at some points was <90% accurate. Went back and searched more and found a carlo gavazzi digital panel ohmmeter.

 

[crutschow]

Feed the unknown resistance with a constant current source and read the voltage directly across the resistance. 50mA will give you 0 to 10V across 0 to 200 ohms. Use Ohms law to calculate different voltage/current values.

Note that 10V across 200 ohms is 1/2 watt which could be a problem depending upon the type of resistance being measured. To lower the power he could scale the voltmeter to a lower range (as he stated he could do) and use a proportionally lower current.

 

[eTech]

Hi everyone

take a look at circuit attached...

Comments, recommendations welcome....

Et

 

[KeepItSimpleStupid]

I'd suggest this part (LT3092) https://www.linear.com/product/LT3092 instead of the LM317.

 

[eTech]

Looks good KeepIt...

But it doesn't come in a thu-hole package....

 

[ChrisP58]

If you are scaled for 5V at the point of measurement, you will need more than 5V into whatever part you choose for the current control element. It needs some headroom voltage to work with, including the current sense resistor voltage drop.

Also, make sure that your panel meter power can share a ground with the measurement signal. Many meters can't.

 

[ChrisP58]

One more thought.

How far away is the unknown resistance from where the meter will be? Is it close enough that the resistance of the connecting wire is negligible?

If not, you can use a 4-wire system, where one pair of wires caries the current out the the resistance, and another pair brings the voltage back to be measured. The two pairs come together only at the resistance terminals. This assumes that the input current of the meter is near zero.

 

[bountyhunter]

Accuracy will be really crappy with LM317 current source. Se attached circuit for a good one:

 

[KeepItSimpleStupid]

Digikey has some SMD to SIP boards that could work for you.

Besides there are SMT protoboards out there, so you could use them. It's going to be hard to find stuff in thru-hole anymore. You can use these: http://www.proto-advantage.com/store/

If these guys have an adapter, they can do the procurement and soldering for you, so you can be left with a DIP package.

I've used them.

Digikey sells some SMT adapters and I did purchase an SMT protoboard from Mouser.

This is a thru-hole package: http://www.ti.com/product/LM334 but not as good of a part and it may not be high enough current for you.

 

[eTech]

Hi ChrisP58

Oops..your right..I forgot about the dropout voltage across the LM317.
I recalculated the values based on 2vdc being measured instead of 5vdc.

New diagram attached.

Also..

Each of the two connections to the resistance being measured is about 10 inches of 22 gauge wire.

Thanks..

 

[eTech]

hi Bounty..

Can you elaborate a little on why accuracy with the LM317 will be crappy?

I'm learning again....

Thanks

 

[bountyhunter]

hi Bounty..

Can you elaborate a little on why accuracy with the LM317 will be crappy?

I'm learning again....

Thanks

The reference is built into the IC so it wanders with temp change. It also varies with supply voltage. Never intended to be accurate. Might be OK for your application, not sure how accurate it has to be. If it is a 1.999 display, that gives a display resolution capability of 0.05%. You need a pretty good circuit to match that.

Recheck your schematic: I don't think 5V in is enough. If I remember right, the LM317 takes at least 2VBEs + Sat to regulate which is at least 2V ballpark, plus the 1.25V reference voltage if used as a current source. I think 5V input is a little short.

EDIT TO ADD: reread your first post. If 3% accuracy is all you need, the LM317 circuit is good enough assuming you give it enough input voltage so it isn't near saturation of the pass transistors (about 6V or more).

 

[eTech]

Hi bounty..

The reference is built into the IC so it wanders with temp change. It also varies with supply voltage. Never intended to be accurate. Might be OK for your application, not sure how accurate it has to be. If it is a 1.999 display, that gives a display resolution capability of 0.05%. You need a pretty good circuit to match that.

The meter is a ME-SP037B Programmable 0-10v Multifunction Meter. I rescaled it to read 0-2vdc.

Recheck your schematic: I don't think 5V in is enough. If I remember right, the LM317 takes at least 2VBEs + Sat to regulate which is at least 2V ballpark, plus the 1.25V reference voltage if used as a current source. I think 5V input is a little short.

I'm driving the LM317 with a LM2937 5v regulator. The voltage drop across the test resistance
(I'm actually using a 220 ohm, 20 turn trim pot) is what I would expect at the different test resisance values...2.0vdc = 200ohms, 1.0vdc = 100 ohms.
However...Meter accuracy goes "crappy" at about 16 ohms and below. I've measured and compared the voltage drops against various test resistance values and they are correct. The circuit is doing a good job. But the meter seems to lose its accuacy at low ohms (<=16).

Any Suggestions?...

EDIT TO ADD: reread your first post. If 3% accuracy is all you need, the LM317 circuit is good enough assuming you give it enough input voltage so it isn't near saturation of the pass transistors (about 6V or more).

5vdc doesn't appear to be a problem...But I can try a higher voltage to be sure

 

[bountyhunter]

However...Meter accuracy goes "crappy" at about 16 ohms and below. I've measured and compared the voltage drops against various test resistance values and they are correct. The circuit is doing a good job. But the meter seems to lose its accuacy at low ohms (<=16).

Any Suggestions?...

Not sure. Does the meter have a zero offset adjustment?

Is it possible it requires a negative voltage rail to maintain accuracy at low input voltages?

 

[eTech]

Accuracy will be really crappy with LM317 current source. Se attached circuit for a good one:

Hi bounty..

What is the part number for the Op Amp in your circuit?

 

[eTech]

Not sure. Does the meter have a zero offset adjustment?

It has a programable "ratio" and "offset" setting...is that what you mean?

Is it possible it requires a negative voltage rail to maintain accuracy at low input voltages?

It also has seperate leads for pwr supply ground and signal VIN- .... But it says in the manual they both share the same ground.

 

[bountyhunter]

Hi bounty..

What is the part number for the Op Amp in your circuit?

It has to be a JFET op amp similar to LF356.

EDIT TO ADD: the circuit I posted requires both + and - supplies to work, can't run on single rail only.

 

[bountyhunter]

It has a programable "ratio" and "offset" setting...is that what you mean?


.

Not sure. Ratio should be "full scale", but offset might make adjust the low end accuracy.