EEVblog Electronics Community Forum
Electronics => Beginners => Topic started by: gailulun on August 02, 2022, 11:49:45 am
-
When I use the "highz" mode of the desktop multimeter, I find that when measuring the voltage between a pair of contacts, the exchange of different measurement ranges and probe positions will cause great changes in the reading. My circuit has no boost structure and the input voltage is 9V. The most puzzling thing is that in one case, the multimeter shows a voltage of more than 12V.When measuring, the reading of the multimeter is very stable. One of the two contacts is connected with 5gΩ resistance and the other is grounded. Has anyone encountered this situation, or know what caused it?
-
Look at your multimeter specifications: it might not be high-Z in every range.
-
My multimeter is a siglent sdm3065x. When measuring, I manually select the range and make sure that the multimeter is in the high-Z state, but every time I change the range or the position of the probe, the reading will change greatly and will be different. The specification indicates that the internal resistance of the multimeter is greater than 10g Ω in the high-z state. :palm:When measuring other contacts under the same setting, the readings are within the expected range and do not change with the change of measuring range and probe position.
-
A typical high-end bench voltmeter has a "high-Z" input resistance on the lowest ranges, possibly up to 30 V fs, and then maybe 10 megohm at ranges above that.
Although high-Z, the input bias current is not ignorable: if you leave the input open, the voltage will increase as the current flows through the high impedance, and your 5 gigohms is very high resistance for this current.
What does the manufacturer specify for bias current?
-
This is the description of DCV in the data book. :palm:
-
The multimeter sometimes shows that the voltage exceeds 12V, which makes me very confused, because the input of the whole circuit is only 9V, and there is no boost structure. I think the existence of bias current should not make the reading too large.
This is a picoammeter project. When the negative probe of the multimeter is connected to the input Island，that is J2 in the figure, no matter which pad the positive probe contacts, the voltage of the multimeter is greater than 12V.
-
The specified "offset current" is 50 pA, typical.
50 pA x 5 Gohm = 250 mV, rather large.
-
Will the bias current increase the multimeter reading?
-
Set the range of the multimeter to 20V, highz mode. When the negative pole of the probe is located on the input island and the positive pole is located on the other end of the 5g Ω resistance (I replace the 100g Ω resistance in the figure with 5g Ω resistance), the voltage is lower than -12v. Then I short-circuit the resistance with tweezers, and the multimeter displays almost 0V. When I remove the tweezers, the reading rapidly drops to about -12v. Then I exchange the probe positions and the reading is 6V. Then set the measuring range to 2V, the reading is -1.14v when the negative pole of the probe is located in the input Island, and the multimeter indicates that the measuring range is exceeded when the positive pole of the probe is located in the input island. |O
-
This is a picture of the circuit board. The "signal in" on the PCB is the input Island, and the black cylinder below is 5g Ω resistance.
-
Obviously you can't measure such a high resistance circuit without the voltmeter loading it. You would need a meter with TΩ input impedance.
-
It seems that 10g Ω internal resistance is indeed not enough.
But I can't figure out why the reading of the multimeter exceeded expectations. I once thought my multimeter was broken.
-
Will the bias current increase the multimeter reading?
That depends on the polarity of the "offset current", which is not specified. Also, the value is "typical", and presumably changes with temperature.
-
It seems that 10g Ω internal resistance is indeed not enough.
But I can't figure out why the reading of the multimeter exceeded expectations. I once thought my multimeter was broken.
P.S. get yourself out of the habit of writing “g” for giga. It’s “G”. SI prefixes and units use capitalization systematically, and if you get into the habit of being sloppy, you end up with crap like ms (milliseconds) vs mS (millisiemens) or kB (kilobyte) vs kb (kilobit). Often the intent is identifiable from context, but occasionally it is not. So teach yourself to be strict about it.
-
I once saw a specification for a relay contact that stated "closed" resistance = 20 M\$\Omega\$ and "open" resistance = 100 m\$\Omega\$.
Rather useless switch.
-
If you set your meter to the 20V range with Hi-Z and just let it sit there with the probes not connected, what does it do over a ten-minute period?
-
Within ten minutes, the reading of the multimeter has been rising slowly, from 0 to 2.7V.
-
OK, I'll pay attention to it.Thanks for reminding. :D
-
Within ten minutes, the reading of the multimeter has been rising slowly, from 0 to 2.7V.
OK, then you can put a 1M resistor in parallel with a small film capacitor (say 22nF, but any small film cap with no leakage will do) across the terminals, let it settle. You will see the bias current @ 1pA = 1uV. Once you know the bias current you can (somewhat ) properly calculate the voltages in the resulting circuit by considering the DMM to be a current source. That bias current that caused the climb of 2.7V is likely the same thing that caused your meter to read you circuit higher than the 9V supply.
-
Continue to keep the multimeter empty. After one and a half hours of testing, the reading is stable between 21-22V, then suddenly displays "over range" and falls back to 21-22V after a few minutes.
Test the connection resistance and capacitance. The capacitance is 220pF polystyrene film capacitance, and the multimeter reading is 0.09-0.12mV.
It seems that the bias current is a little high, about 100pA.
-
Previously, I thought that the bias current was caused by external voltage. Until now, I found that this may be caused by the equivalent current source inside the DMM. No matter what the setting range is, the voltage in "HighZ" mode is stable at 0.1mV, and the voltage in "10MΩ" mode is stable at 0.08mV. Does this mean that my DMM has failed?
-
I replaced another resistor. The multimeter showed a resistance value of 430kΩ. This time, no matter how the range(200mV/2V/20V) and input impedance(10MΩ/>10GΩ) changed, the voltage reading was about 43μV, that is to say, this DMM can be equivalent to a 100pA current source when measuring the voltage (at this time, the temperature in my house is 28 ℃), but the typical value of the officially claimed bias current is 50pA.
Should I open a new post to ask if other people who own this DMM also have this situation?
-
The input current may very well depend on the temperature and with a higher than normal temperature the current tends to be higher. 100 pA of input current is bit on the high sight, but the specs with 50 pA are typical only. 80 µV in the 10 M mode would suggest an input current of some 8 pA. Measuring 43 µV can be effected by thermal EMF. Another possible error source is the isolation of cables - the isolation of a cheap cable may have a resistance in the 100 Gohms range and this is enough to cause quite some leakage if the is enough common mode voltage. With an input that uses an AZ OP-amp the capacitance at the input can also effect the input current and offset. Normally there should be enough filtering to suppress this not so obvious effect.
The specs for the input resistance are > 10 Gohms and the ">" in this case also allows for much larger, like 100 GOhms or even more. The resistance may also be nonlinear.
So the voltage see with an open input is pretty much undefined. It is not that uncommon that the open circuit voltage drift all the way to the limits (e.g. 20 V), though it is also not that uncommon that the input resistance goes down when approaching the limits or slightly outside. So the observed rise of the voltage to some 21 V right at the limits could just such an weak clamping action to as part of the overvoltage protection.
An alternative way to measure the input current is to measure how fast the input voltage drifts with something like a 10 nF low loss capacitor (e.g. PS or PP type). 10 pA into 10 nF would result in 1 mV/s of drift rate, which is reasonable easy to measure.
-
The largest low leakage capacitor I can find at present is a 220pF polystyrene film capacitor. When the range is 2V and the input impedance is 10GΩ, the voltage rises by 2V in 18s, and the calculated current is about 24pA.
-
This is the pictures.
-
With such a small capacitor the meter internal capacitance can be significant. This may be some additional 100 pF, but the exact number can vary.
-
I figured I could put two or three capacitors in parallel and make another measurement, and as long as the capacitor inside the DMM are of the same order of magnitude as the capacitors I'm using, there shouldn't be much error. The derivation is shown in the figure below.
-
The largest low leakage capacitor I can find at present is a 220pF polystyrene film capacitor. When the range is 2V and the input impedance is 10GΩ, the voltage rises by 2V in 18s, and the calculated current is about 24pA.
You can try to estimate the bias current that way, but I was only suggesting using the capacitor for stability. What your numbers really suggest (I haven't checked the math) is that the meter input capacitance is ~660pF. 100pA and ~600pF aren't great numbers in comparison to other 6.5 digit DMMs that I've tested, but for the vast majority of uses they wouldn't pose a problem. Unfortunately your application is particularly sensitive to these specs.
-
I have just tested that when a single 220pF capacitor is connected, the voltage rises by 2V in 12s. When two 220pF capacitors are connected in parallel, the time is about 16.5s, and when three capacitors are connected in parallel, it is about 21s. Therefore, the inherent capacitance of DMM is calculated to be about 587pf, which basically conforms to the calculation law.
-
*
-
It seems that the internal resistance of this DMM is greater than 100GΩ. I connected a 100GΩ resistor. In the "HighZ" mode, the voltage displayed by the multimeter is always greater than 5V, but it is unstable and rising. I don't know whether it is related to the fact that I just turned on the air conditioner to cool down.
-
Does the data sheet for the meter say “10 G ohm” or “> 10 G ohms”?
-
The data sheet shows ">10 GΩ"
-
In other words, as close to open-circuit as is practical. 10GΩ is the lower limit, on a bad day.
-
However, the 100pA bias current makes such a high impedance useless. :palm:
-
100 pA is relatively high, but still not that bad: with 1 V at the input and the input resistance at the 10 Gohms lower limit this would also cause 100 pA of input current.
The DMM input is still mainly made for signal sources wih a relatively low impedance, like < 100 kOhm. A main point of the high impedance is not to load down the signal source and cause errors this way.
In the sub nA region one also has to keep an eye on the cables - cheap PVC isolated cables can make a differenece.
-
I do notice this amazing phenomenon when measuring the power supply, even if it only touches the insulating skin of the wire. This should be attributed to the insufficient impedance of the insulation skin. :-DD