How accurate is your digital volt meter?

[GHA]

Doesn't matter how accurate it's claimed to be, if it's not recently calibrated, it's not to be trusted for anything important. That's why measuring kit used commercially is always calibrated, an expensive, but necessary, faffage.

I would not trust uncalibrated kit more than 1% accuracy.

I got one off eBay with a lm4040 chip, don't think there's any calibration option.
I kind of assume if it and a couple of not very top end but not cheapo meters agree to within 10mV or so then all is well for the sort of accuracy a boat needs. Handy to have when messing around with arduinos and the like.

 

[Hydrozoan]

My self-powered 3.5 digit panel DVM costing about £20 claims +/-0.01%, but does not give a 'significant digit' error figure. I assumed one significant digit, for a total error of about +/-11mV (at 12V), which is more than good enough for rough guide battery monitoring.

 

[Amulet]

I repeat - I have four different ways of reading the voltage - none of them very expensive instruments, and they read within about 0.2% of each other.

 

[William_H]

the spec for my cheapies is ± 0.5% of the reading, ± 2 digits.

Not quite sure what ± 2 digits means though.

Hi Vic all digital devices have a point where the last digit changes either up or down. Leaving an absolute uncertainty.
eg a digital clock that shows only hours and minutes. can indicate 12.47 but you really don't know if the time is just 12.47.00(seconds) or 12.47.59. In other words accuracy can only +or - (nearly) one minute or +or- one least significant digit.
So too a digital volt meter can only be described as accurate to +or- one (least significant) digit. Now you probably under stood all that anyway. But like you I don't know why +or- 2 digits. Perhaps an uncertainty in the AD converter design.
The overall accuracy of a digital voltmeter usually depends on the accuracy of the precision voltage reference. However they usually have a variable resistor (potentiometer) to divide and adjust the reference voltage and this can often be found and adjusted to an external standard. NB if we adjust the reference then we can assume that if all the divider resistors are stable and accurate all other volt ohm and current ranges should be accurate. olewill

 

[Lon nan Gruagach]

With so much data available on battery charge state from precise voltage measurements we might wonder just how accurate a DVM is.
I have just received one of these AD 548 voltage reference devices. Just needs 15v supply and it provides 10 volts precisely for use as a calibration. It is guaranteed to 30 parts per million worst case. As a matter of interest 2 old DVMs of mine showed 9.96 and 9.95 volts on test so not too far out. Might be of interest to someone. olewill
http://www.ebay.com/sch/i.html?_fro...kw=AD548+precision+voltage+reference&_sacat=0

Whoa there, all this getting carried away....

As has been said, the chip is AD584 (an AD548 is an op amp).
The data sheet claims maximum error (10V output) of 30mV for the cheap version and 10mV for the fancy one. Thats 3,000ppm and 1,000ppm respectively or 0.3% and 0.1% in old money.
They do, however have petty good temperature stability.

 

[William_H]

Whoa there, all this getting carried away....

As has been said, the chip is AD584 (an AD548 is an op amp).
The data sheet claims maximum error (10V output) of 30mV for the cheap version and 10mV for the fancy one. Thats 3,000ppm and 1,000ppm respectively or 0.3% and 0.1% in old money.
They do, however have petty good temperature stability.

My appologies yes you are right. The first a slip of the brain in typing the number type in the first post.
The second yes I read the temp stability rather than the overall accuracy. So not as good as I thought. olewill

 

[DannyB]

I used to use a top of the range Fluke which was calibrated every year, on one job I stuck the probes on a transformer secondary, just to check it was dead, having already earthed down the 11kV primary. The Fluke read 220 volts to earth, instant panic and double check all round, primary earthed, secondary disconnected, all OK, second check with the Fluke, 220 volts. Head scratching time. Eventually one of my sparkies stuck a test lamp on the terminals which did not light up, so daylight started to dawn, and I went back to the office and dug out an old Avo 8 from a cupboard, quick check on the bench to make sure it was working then took it down to the transformer, nothing, nada, zilch but the Fluke was still showing 220 volts. To this day I don't know where it was picking it up from, but its input impedance must have been so high it was reading radiation. So even high quality digital meters can give you a bum steer. (Are we allowed to say Bum on here ?)

 

[Anthony]

I started a thread at EEV blog, some interesting info about the accuracy of that referance.

Ants

 

[savageseadog]

I used to use a top of the range Fluke which was calibrated every year, on one job I stuck the probes on a transformer secondary, just to check it was dead, having already earthed down the 11kV primary. The Fluke read 220 volts to earth, instant panic and double check all round, primary earthed, secondary disconnected, all OK, second check with the Fluke, 220 volts. Head scratching time. Eventually one of my sparkies stuck a test lamp on the terminals which did not light up, so daylight started to dawn, and I went back to the office and dug out an old Avo 8 from a cupboard, quick check on the bench to make sure it was working then took it down to the transformer, nothing, nada, zilch but the Fluke was still showing 220 volts. To this day I don't know where it was picking it up from, but its input impedance must have been so high it was reading radiation. So even high quality digital meters can give you a bum steer. (Are we allowed to say Bum on here ?)

Some Fluke multimeters have a low Z setting to combat this problem. When I was talking to a friend who was investigating some poor boat wiring, I suggested using test lamps rather than meters.

 

[ghostlymoron]

, I suggested using test lamps rather than meters.

I mentioned this yesterday?

 

[ghostlymoron]

AFAIK, you can say bum but not arse (or worse).

I used to use a top of the range Fluke which was calibrated every year, on one job I stuck the probes on a transformer secondary, just to check it was dead, having already earthed down the 11kV primary. The Fluke read 220 volts to earth, instant panic and double check all round, primary earthed, secondary disconnected, all OK, second check with the Fluke, 220 volts. Head scratching time. Eventually one of my sparkies stuck a test lamp on the terminals which did not light up, so daylight started to dawn, and I went back to the office and dug out an old Avo 8 from a cupboard, quick check on the bench to make sure it was working then took it down to the transformer, nothing, nada, zilch but the Fluke was still showing 220 volts. To this day I don't know where it was picking it up from, but its input impedance must have been so high it was reading radiation. So even high quality digital meters can give you a bum steer. (Are we allowed to say Bum on here ?)