Multimeter giving bad voltage readings

[seanspotatobusiness]

I've had a cheap Precision Gold Academy PG 10B multimeter since I was a kid (so it's about 20 years old) but just recently I've got reason to suspect the voltage readings. For my computer PSU it gives the 5 V rail as 5.8 V and the 12 V rail as 13.9 V. For my variable power supply it gives any voltage as something higher than what the supply says (e.g. 15 V as 18.6 V) and it says the same for a 15 V power supply pack. Could a dying battery cause this? The low battery indicator is off. I just do basic electronics as a hobby; I'm not interested in owning anything super accurate but I think something is wrong with this unit. Is this a thing that happens to an aging multimeter? The resistance measurements seem to be as expected.

 

[davenn]

Could a dying battery cause this?

could be .... cheap enough option to find out

otherwise, it may just need recalibrating

 

[seanspotatobusiness]

It wasn't the battery. I've turned the calibration potentiometer all the way up or down and it still isn't correct - the readings are still too high. Maybe this is a stupid idea but I wonder whether adding resistance to the correct side of the potentiometer could improve the range enough to calibrate it correctly. I figure I could de solder the potentiometer and connect it again with a resistor attached to one of the legs. It wouldn't be pretty but it might work?

 

[ModemHead]

In order for the A/D converter in your meter to work, it must have a steady reference voltage. If the reference voltage is low, the meter will read high. Your meter seems to be reading significantly high. It would be extremely unusual for the reference to get that far off just due to age or anything normal. There is probably another fault in the meter somewhere. But if you want to try your experiment anyway, don't let me stop you.

I do not have a schematic of your particular model, so this is just general info for that type of design: The calibration pot you're turning is part of a voltage divider that provides a 100mV reference, derived from the regulated voltage (2.8 - 3.0V) between the positive battery terminal and the meters analog ground, which is connected to the COM jack on the front. The pot will be on the bottom of the divider, and it is usually 200 ohms. There will probably be another resistor in series with it already. Increasing either the pot's resistance or the series resistor will increase the reference voltage.

 

[davenn]

I've turned the calibration potentiometer all the way up or down and it still isn't correct

which calibration pot ?? there's likely to be several
show an internal pic or 2 of the unit

 

[Harald Kapp]

Do you have access to a known good meter? You'll need one to locate the fault of this one as you'll have to check various parameters as suggested (resistances, reference voltage etc.)

 

[seanspotatobusiness]

I've now got a functional multimeter that I might be able to use to fix the malfunctioning one. There's only one potentiometer in the malfunctioning one though. Here's a photograph of the circuit board with the fuses removed:

 

[Chemelec]

Does this meter read Correctly when measuring Resistor Values?

 

[shrtrnd]

I work in a Standards lab. Components do age over time, sometimes heat from readings damage components.
What calibration means to me is not what most people here think when I talk about things like this.
In your case, the first thing I'd ask, is what your meter 'ground' probe is attached to? Are you sure you're tied
to an actual ground for your higher than normal readings, or are you possibly using a floating ground?

 

[(*steve*)]

@seanspotatobusiness, I'd make sure it has nice new batteries in it.

@shrtrnd, Standards lab? Cool. You should tell us more about it.

 

[seanspotatobusiness]

Does this meter read Correctly when measuring Resistor Values?

It reads correctly between 200 ohms and 47 Kohms. For some reason it reads 200 K as 120 K and 1 M as 800 K. After my attempt to recalibrate it last week, it is more correct for lower voltages and as the voltage increases, the reading seems to approach a plateau between 9 and 10 V. Since it only has one potentiometer (could there be another on the other side of the PCB? I haven't figured out how to remove the PCB), it seems I can't fix one side of the scale without distorting the other.

I work in a Standards lab. Components do age over time, sometimes heat from readings damage components.
What calibration means to me is not what most people here think when I talk about things like this.
In your case, the first thing I'd ask, is what your meter 'ground' probe is attached to? Are you sure you're tied
to an actual ground for your higher than normal readings, or are you possibly using a floating ground?

I've tried measuring voltages from batteries measured with other DMMs and also from power supplies with metal cases which I guess might be grounded? I connect to the black terminal on the power supply. Is that likely to be connected to earth? There's no continuity between the black terminal and the earth pin of the mains power plug at least while the unit is switched off.

Additionally, something strange happens when I turn the function selector to measure resistance at <200 ohms or to the continuity test - some LCD segments display and then fade away until the dial is moved. The video below shows what happens. I think this might be an independent fault though; I think this happened after I connected it to my power supply at 9 V but it wasn't subject to anything higher than 9 V.

 

[Chemelec]

Video is Private.
NOT VIEWABLE!

 

[seanspotatobusiness]

Video is Private.
NOT VIEWABLE!

Sorry, I didn't realise I needed to click "publish" after uploading. I hope it works now.

 

[Chemelec]

Could be Oxidized Contacts on the Selector Switch.
Need to Spray the Switch with Electronic Contact Cleaner.
May also Fix the other Problem.

But first you need to get it Out of the Case to do this.

Note the 2 Small Tabs on the Plastic Case, hold the Board.
Carefully Pry them Sideways to lift the board out.
But you probably need to Remove the Knob first.

 

[shrtrnd]

Yeah. Considering the symptoms, I'd go with Chemelec's advice to clean the selector switch contacts first, and THEN
see what symptoms you have left. Be especially careful about disassembly of the switch as the inexpensive meters tend
to have poorly constructed assemblies. They were made to be assembled, and not taken apart for service easily later.
Bear in mind that if you do this, you may find fully functional reassembly difficult. I'm just suggesting it because
your meter is only useful at low readings now. If you can live with that, don't try to disassemble it or it may become completely
useless to you if you can't get it back together right.
My job in an Electronic Standards lab is to calibrate and repair high-end electronic test and measurement equipment from
a very large number of vendors.
I laugh sometimes when I read people talking about 'calibration' of their instruments.
I'm tied to NIST signals in temperature and humidity controlled spaces,
When I label an instrument as 'calibrated', there is no second guessing involved.

 

[(*steve*)]

Just out of interest @shrtrnd. If you had to calibrate a voltage source (lets just say you were just going to measure and report the voltage) of say 10V, how precise could you practically measure it (lets assume it was limited by your equipment rather than noise or drift on the source itself)?

The bottom line on my 10V reference's cal certificate is that it is 2.3uV high, but I'm really curious about the precision and accuracy that is available (not that I need or could afford it)

 

[shrtrnd]

My Standard is in-house built, and environmentally controlled.
I use it to calibrate Fluke 732B's (and a few 'A' models)
Accuracy is .1uV
The issue down there (for me as the guy doing the cal) is not so much 'accuracy' as what I call 'age rate'.
I cal the 732B's every 6 months and chart the values.
What I look for is the slope in uV/6 months for the voltage reading.
Your issue would be insignificant for my tests as I typically see the 732B's off by about 11uV, and can still call that within spec for my users.
Your Reference could have come out of the factory with a +2.3uV error and still have passed factory calibration.
I can just about guarantee 'drift', I just chart it over time to make sure the user knows the Ref's accuracy at the time he's using it.
The rule of thumb for most of my calibrations, is that the Standard I use has to have 4X better specs than the unit under test. (in anticipation of your next question).
Does that answer your question?

 

[seanspotatobusiness]

It didn't look at all dirty but I cleaned the contacts on the PCB and the knob with wire wool anyway. The screen still turns off when the dial is turned to continuity or 200 ohm resistance rage. If there's nothing else obvious then I'll probably just throw it away. Sad times.

 

[shrtrnd]

Wire wool is pretty aggressive. Still sounds like a contact problem with the dial to me, but may not be fixable.
Sorry we weren't able to help with this. It's nice to hear from people who attempt repairs before trashing.

 

[Chemelec]

Wire Wool is NOT a good way for doing this.
Leaves Scratches and misses some areas.
Contact Cleaner removes Oxidation Much better and the spray covers evenly.