# [solved] capacitor values questions

Thread Starter

#### PsySc0rpi0n

Joined Mar 4, 2014
1,653
Hello.

I have some questions about capacitor values.

For instance, I have here some ceramic capacitors with value of 103. I know this means 10 x 10³pF = 10nF. But why is the value underlined? Is there any specific meaning?

Another question is that some capacitors have this small black marking on the top. What is the meaning of that marking?

And lastly these ones. What is the meaning of those 2 colour marks?

#### MaxHeadRoom

Joined Jul 18, 2013
25,206
The bottom one appears to be 27pf?

Thread Starter

#### PsySc0rpi0n

Joined Mar 4, 2014
1,653
The bottom one appears to be 27pf?
The bottom is 1P5 which I think it's 1.5pF, right?

But the question is more about the markings, not the values!

#### Hymie

Joined Mar 30, 2018
1,081
Underlining a number is a common method of ensuring that the number is not read in the wrong orientation; the numbers 0, 1, 6, 8 and 9 maybe misread upside-down.

#### Ya’akov

Joined Jan 27, 2019
5,659
The black top on a disc ceramic cap indicates industrial temperature range.

Thread Starter

#### PsySc0rpi0n

Joined Mar 4, 2014
1,653
So, nothing of critical importance for home stuff, right?

Thread Starter

#### PsySc0rpi0n

Joined Mar 4, 2014
1,653
I also wanted to understand something else.
I was testing a few capacitors here, to try to see their charging time constant and confirm on the scope.

For instance, I am using a 1MΩ resistor and this 1P5 (1.5pF) capacitor and I was expecting τ=1.5pF x 1MΩ = 1.5us. Then, after 5τ, I would expect the capacitor to be charged up to 99.3% of Vmax (which is 7V in my case).
So, I would expect 5τ = 7.5us to be at 7V x 0.993 = 6.993V.

However, this is what my scope shows:

Max voltage is way less than the expected 6.993V (aroun 6.16V) and the charging time is like exceeding by a factor of 10. It should be aroun 7.5us and I think I can assume the cursors are more or less at the initial point and at the max voltage.

#### AlbertHall

Joined Jun 4, 2014
12,020
Apart from the resistor and the capacitor you also have an oscilloscope connected. The 'scope does not have an infinite input resistance so it will affect the readings.

#### Tonyr1084

Joined Sep 24, 2015
6,884
I'm no capacitor expert, but I'm wondering if maybe some of the deviations may be due to ESR (Equivalent Series Resistance), or possibly some sort of leakage in the cap. They DO leak some current. Which is why they tend to lose their charge over long periods of time. Some of that may be coming into play.

#### drjohsmith

Joined Dec 13, 2021
394
I'm no capacitor expert, but I'm wondering if maybe some of the deviations may be due to ESR (Equivalent Series Resistance), or possibly some sort of leakage in the cap. They DO leak some current. Which is why they tend to lose their charge over long periods of time. Some of that may be coming into play.
Also
real capacitors have terrible tolerances,

#### Ya’akov

Joined Jan 27, 2019
5,659
So, nothing of critical importance for home stuff, right?
Right, just a temperature specification for extreme environments.

#### Ya’akov

Joined Jan 27, 2019
5,659
A combination of all the foregoing: the practical impedance of the scope probe, the possible contribution of ESR, and the fact that caps are pretty crappy when it comes to values could easily cause deviations in the expected result. A simulation that can eliminate all of these should provide the expected results.

Measuring the actual value of the cap and resistor might clear things up, if you have a sufficiently accurate LCR meter.

#### Audioguru again

Joined Oct 21, 2019
4,680
A capacitor with an extremely small value of only 1.5 Puffs has stray capacitance of the connection of at least 5pF parallel with it.

#### AlbertHall

Joined Jun 4, 2014
12,020
A capacitor with an extremely small value of only 1.5 Puffs has stray capacitance of the connection of at least 5pF parallel with it.
Not to mention the capacitance of the scope probe, perhaps 15pF.

#### MrChips

Joined Oct 2, 2009
26,089
Also don’t use 5 time constants in your oscilloscope measurements. Use one time constant.

#### MrChips

Joined Oct 2, 2009
26,089
Here is the reason why your test and analysis is flawed.

After 5 time constants, the voltage reaches approx. 99% of the charging voltage.
This measurement reveals the effect of Rprobe on the measurement and nothing about C and Cprobe.

With a probe resistance 10 times the value of the resistance under test, one would expect to see about 10% error in voltage measurements.
(You can verify this by examining your results. 10% of 7V is 0.7V. Is your measurement off by 0.7V?)

In order to measure the RC value, use the 1 time constant measurement, i.e. the time taken for the voltage to reach 63% of the charging voltage.

The addition of both Rprobe and Cprobe into the circuit will alter your analysis significantly.

#### Audioguru again

Joined Oct 21, 2019
4,680
1M ohms and 1.5pF have a time constant of 1.5us and a -3dB cutoff frequency of 106.7kHz.
Why not 1k ohms and 1.5nF if the signal source resistance is much less than the 1k ohms?

Thread Starter

#### PsySc0rpi0n

Joined Mar 4, 2014
1,653
Ok, I understand now why.

Anyway, I step into this because I was trying to use one of those component testers that one can buy from Aliexpress or so, and only one capacitor was detected by this testing device. So, to try to check if the other capacitors were OK and to make sure about their stated values, I tried to go for the charging times to see if they would match the values they had written on them.

@Yaakov my meter can only measure from 50nF and up. It's a Brymen BM869s.
From the manual:

@MrChips, as you can see from my scope screenshot, Vpp is 6.48V and the voltage at cursor is 6.16V, so, yeah, maybe that's the case, about the 10% error.

Ok, so this said, I think I can call it "solved".

Thanks to everyone.
Psy

#### MrChips

Joined Oct 2, 2009
26,089
Also it would be extremely difficult to measure 1pF without special equipment and effort.

Two pieces of wire about 1 inch long close to each other is about 1-5pF (I'm guessing).
Hence our limit is about 10pF.

A C meter with a range of 0-1000pf would have an error of about 10-100pF (i.e. about 1-10% error).