Capacitor Measurement

I saw an article in "Nuts & Volts" about a capacitor test device based on the discharge time of a capacitor. It occurred to me that the pot command would work in the same manner, and be much easier to assemble.

With a 16F84A chip, I gave it a try, and it works, within limits, very accurately. I have an old Heathkit Capacitor substitution box, and my bread boarded effort is dead on from 100pf to 1uf.
Anything below 100 pf returns 0, even with a 10meg discharge resistor.

Above 2.2uf, I get a return, but it has no relationship to the cap's value. A 220uf cap might return 90uf, while a 47uf might return 148. I tried discharge resistors as small as 1 Ohm. Any return is useful, as open and short both return 0, but I would like to get a measurement over 2.2uf. And I would like to get a measurement below 100pf as well.

I tried adjusting the prescaler, but it had no effect on my results. I am using a 4mHz resonator. Perhaps a faster clock would help the low end?

For caps from 2.2uf - .001, I am using a 1k discharge resistor and a scale of 40. For 1000pf to 100pf, 20k and scale at 255.

I suspect that internal resistance of the high value electrolytic affects their discharge curve in a way that confuses the chip. I get better results with tantalum caps, but they are still off 50 - 100%. And, perhaps any charge on a very small cap is lost during the time the chip switches from charge to measure.

The author of the article claims better than 1% accuracy, 1pf to 999uf, but I wonder.

The program is attached. I will provide a schematic when I draw one if any are interested.

Any thoughts?

I tried to attach it, but it didn't happen. I'll put it here.

' Capacitor Measurement
' Four digit common anode mulitplexed display
' Scan about 60/s (4ms pause)
' Driver 74HC4511
' Works for 2.0 uf (1K)- 100pf (20K)
' 8/20/06
' For 16F84A Chip
' Data output port A (Pin A5 NC)
' Pins 1 - 4 - Digit select for multiplex
' Pin5 - pot command
' Pin6 - Display Blanking signal (0 = blank)
' Pin7 - Range 1 = high value (1K resistor); 0 = low (20K)
' Digits 1 - 4; W1 - W4
' Measurement (Msr) W10

Settings:
Symbol Blank = pin6
Symbol Range = pin7
Symbol D1 = w1
Symbol D2 = w2
Symbol D3 = w3
Symbol D4 = w4
Symbol Msr = w10
Symbol PortA = 5
Poke $85, 0 ' Set port A to all outputs
Output 1 ' Digit 1 - MSD
Low 1 ' off
Output 2 ' Digit 2
Low 2
Output 3 ' Digit 3
Low 3
Output 4 ' Digit 4
Low 4
Output 6 ' Blanking
Low 6 ' Blank all segments


Start:
pause 750
Blank = 0
If range = 0 then testlow 'Check high or low state

'Testhigh:
pot 5,40,msr
Msr = Msr*10 'Raise results for 4 digit display
goto math

Testlow:
Pot 5,255, msr
Msr = Msr*100

Math: 'Set up digits in memory for display
D1 = Msr/1000
D4 = Msr//1000
D2 = D4/100
D4 = D4//100
D3 = D4/10
D4 = D4//10

Display:
poke porta, D1 'Send data
High 1 'Select Digit
Blank = 1 'Remove blanking
Pause 4 'Pause
Blank = 0 'Blank
Low 1 'Unselect digit

poke porta, D2
High 2
Blank = 1
Pause 4
Blank = 0
Low 2

poke porta, D3
High 3
Blank = 1
Pause 4
Blank = 0
Low 3

poke porta, D4
High 4
Blank = 1
Pause 4
Blank = 0
Low 4

Goto display
end

AMay, I personally would tend to use a constant current source with the RCTIME function. That way you can measure the timing between 2 points with the cap charging at a fixed rate (much more linear). Remember that the resultant range can not be greater than 0 to 65535 so you will probably have to use different current levels to maintain accuracy over large ranges 100Pf to 100 uF. I have used this methode for measuring the backup storage of airbag modules capacitors in production equipment with great success. LOL

Dave Purola,
N8NTA

>> caps, but they are still off 50 - 100%.

Bear in mind that unlike Resistors, Capacitors are pretty much guaranteed NOT to be the value printed on them. With most Capacitors, if you're within 20% of what they're supposed to be, you've got a good one!

So, if your gizmo is telling you it's 50% off... it could actually be telling you the truth!

Thanks for the response.

I can imagine a constant current source.
I can imagine starting it, and stopping it with a comparator.
I can't imagine how to time the interval in PIC basic. The books I have don't help, or I am not looking in the right place. (I am just a retired bookkeeper having fun with this.)

I don't want to get involved with a separate counter. The author who got me started on this ran into all kinds of stability problems with his on the small caps.

Hi,

Some times ago, Elektor published a nice project about a capacitance meter; It's not PbP, but fairly works ...

Project N° 020144 ... Programmed PIC and PCB available.

Good choice ... if for measuring !!!

Alain

here's a good one but it's not basic either
http://www.ubasics.com/adam/pic/archive.shtml
Autoranging Digital Capacitance Meter .01pF to 16,777.215 uF
by Fr. Tom McGahee
it's at the bottom of the page

Quote:

---

Originally Posted by Melanie

>> caps, but they are still off 50 - 100%.

Bear in mind that unlike Resistors, Capacitors are pretty much guaranteed NOT to be the value printed on them. With most Capacitors, if you're within 20% of what they're supposed to be, you've got a good one!

So, if your gizmo is telling you it's 50% off... it could actually be telling you the truth!

---

Many books quote it, I even recall one of my old tech teachers saying it: "Electrolytic capacitors have a huge varying degree of tolerance" But in practice, I have never measured any capacitor with a +/- 20% tol. I have however measured many of them faulty with very high ESR, which is a common fault with electrolytic capacitors used in switch mode circuitry. Sometimes, an extremely high ESR – near open circuit – will result in a somewhat lower capacitance measurement. All and all, I think modern-day components have traveled miles ahead. In reality, today, +/- 20% tol with anything would be very seldom.

Regards,
Trent Jackson

I have old Capacitor substitution boxes dating back 50 years. (20 years for the Heathkit boxes) Large capacitors 2 to 470 uf do not last many years. I replaces all the large capacitors in my HeathKit radio 10 years ago. I have some capacitor boxes where one decade failed because of a bad switch. “Anything below 100 pf returns 0, even with a 10meg discharge resistor.“

I think you should go to the store [DigiKey] and get some capacitors you know are good

1)
I have used a NE555 IC + R + unknown Capacitor. The 555’s time is dependent on capacitor. The PIC measures time and drives a LCD display. A one-shot [multivibrator] converts capacitance to time. [74LS221]
2)
I built a C & L meter based off a kit. I measure well under 0.1pf.
http://my.integritynet.com.au/purdic...er-project.htm
It uses an oscillator and reference coil and two reference capacitors.
The oscillator oscillates at C1= 680pf and L1= 68uH or .01uf and 68uH.
An unknown capacitor is placed across the reference capacitor. If the unknown cap is 1pF then the frequency changes by square root (681/680). If the unknown cap is 1000pf then the new frequency is a product of 1680pf and 68uH.

The math is hell.
I measure the inductance of inches of wire or tiny ‘gimmick’ capacitors (1pf).

http://www.electronics-diy.com/lc_meter.php

here ... soft available !!!

Alain