Interpreting +/- digits of accuracy in meter

[jeffamm]

I'm using a capacitance meter that has 3 1/2 digits of display and lists the accuracy of its 200pf scale as follows:

±0.5% ±20 digit

The ±0.5% is simple enough but I've not found something on how exactly to interpret +/- 20 digits of accuracy. I think it means as follows - can anyone confirm?

If I'm measuring a capacitor of 100pf on the 200pf scale my error is as follows:

+/- .5% = +/- .5pf

For digits, since the 3 1/2 digit meter will read 100.0, the meaning of "20 digits" is 098.0 to 102.0 or +/- 2pf

Total error is +/- 2.5pf. Is that correct?

 

[vk6kro]

Do you have a link to the data for this capacitance meter?

You may be right, but the figures don't really seem possible.

0.5% suggests at least a fairly useful instrument, but +/- 20 digits means that the last two digits are basically meaningless. Could it be a misprint or interpretation error?

I looked at some typical digital capacitance meters and it was common for them to be claimed as + 1 digit (never minus). I wouldn't want one that gave 40 possible readings for the same capacitor.

 

[jeffamm]

vk6,

Here's the link

http://cgi.ebay.com/Digital-Capacitance-Meter-Tester-Multimeter-6013L-/260711350530?pt=LH_DefaultDomain_0&hash=item3cb39b8102

I'm glad to see that you're interpreting it the same way that I am.

I'm also wondering if the +/- .5% is typically of the reading or full scale?

 

[vk6kro]

That looks like a nice meter.

Why don't you send them an email and ask what that +/- 20 digit means?

They claim they can read to 0.1 pF so they couldn't do that with the last two numbers being almost random.

 

[Carl Pugh]

The calculation by jeffamm indicate a maximum error of 2.5pf when measuring a 100pf capacitor. This is 2.5% maximum error. That's not bad using a $21.00 meter.

 

[jeffamm]

Carl,

That's what I was thinking. But I think the problem comes up when you're reading a low value on any particular scale. For example, measuring a 10pf capacitor on the 200pf scale would ideally read 010.0 so with +/- 20 digits the range would be 8.0 to 12.0. I'm assuming that the +/- .5% is of the reading, so negligible here. If it's of the full scale it's another +/- 1pf of error. That's 20-30% of error. And I guess overall that's what I would expect from an inexpensive meter, which would be great for most applications. I'm interested in doing some comparative measurements of guitar cable capacitance and this error in the low measurements may be a problem for me. The more expensive meters have a +1 digit error as mentioned earlier.

 

[vk6kro]

They claim an accuracy of 1 part in 1999, (eg 0.1 pF in 199.9 pF) so there must be something wrong with this specification.

I wondered if the original document may have said "PLUS 2, MINUS ZERO DIGITS" and someone has tried to put it as +/- 2,0 then this was typeset without the comma.

It would be worth clarifying this before you buy one of these meters.

The meter looks OK for the price, although a better meter would possibly have a control for zeroing out the stray capacitance of the leads.

Instead of having a lot of boxes that all do different things, maybe you could look for a multimeter that does capacitance, inductance, temperature and probably other things as well as all the normal voltage and current measurements.

Failing that, there are RLC (resistance, inductance, capacitance) meters that specialize in measuring just these things.