Also: Capacitor Characteristics and Specifications
Temperature Coefficient, (TC)
The Temperature Coefficient of a capacitor is the maximum change in its capacitance over a specified temperature range. The temperature coefficient of a capacitor is generally expressed linearly as parts per million per degree centigrade (PPM/°C), or as a percent change over a particular range of temperatures. Some capacitors are non linear (Class 2 capacitors) and increase their value as the temperature rises giving them a temperature coefficient that is expressed as a positive "P".
Some capacitors decrease their value as the temperature rises giving them a temperature coefficient that is expressed as a negative "N". For example "P100" is +100 ppm/°C or "N200", which is -200 ppm/°C etc. However, some capacitors do not change their value and remain constant over a certain temperature range, such capacitors have a zero temperature coefficient or "NPO". These types of capacitors such as Mica or Polyester are generally referred to as Class 1 capacitors.
Most capacitors, especially electrolytic's lose their capacitance when they get hot but temperature compensating capacitors are available in the range of at least P1000 through to N5000 (+1000 ppm/C through to -5000 ppm/C). It is also possible to connect a capacitor with a positive temperature coefficient in series or parallel with a capacitor having a negative temperature coefficient the net result being that the two opposite effects will cancel each other out over a certain range of temperatures. Another useful application of temperature coefficient capacitors is to use them to cancel out the effect of temperature on other components within a circuit, such as inductors or resistors etc.
Temperature Coefficient, (TC)
The Temperature Coefficient of a capacitor is the maximum change in its capacitance over a specified temperature range. The temperature coefficient of a capacitor is generally expressed linearly as parts per million per degree centigrade (PPM/°C), or as a percent change over a particular range of temperatures. Some capacitors are non linear (Class 2 capacitors) and increase their value as the temperature rises giving them a temperature coefficient that is expressed as a positive "P".
Some capacitors decrease their value as the temperature rises giving them a temperature coefficient that is expressed as a negative "N". For example "P100" is +100 ppm/°C or "N200", which is -200 ppm/°C etc. However, some capacitors do not change their value and remain constant over a certain temperature range, such capacitors have a zero temperature coefficient or "NPO". These types of capacitors such as Mica or Polyester are generally referred to as Class 1 capacitors.
Most capacitors, especially electrolytic's lose their capacitance when they get hot but temperature compensating capacitors are available in the range of at least P1000 through to N5000 (+1000 ppm/C through to -5000 ppm/C). It is also possible to connect a capacitor with a positive temperature coefficient in series or parallel with a capacitor having a negative temperature coefficient the net result being that the two opposite effects will cancel each other out over a certain range of temperatures. Another useful application of temperature coefficient capacitors is to use them to cancel out the effect of temperature on other components within a circuit, such as inductors or resistors etc.
@Rcruz
How do you match capacitors? If do you buy matched, do you know how are matched & the real value?
How do you match capacitors? If do you buy matched, do you know how are matched & the real value?
What does Percy charged for matching? I am doing the folded cascade version
Please wait for Salas review.... My build is quite different from Salas initial schematics.
You must speak with Michael... it depends.
I always ask for matched capacitors because otherwise there will be great differences between them... but even if they are matched, there are always small differences... so If I need two caps, I always buy 6 or more... Then I do matching myself... I have a calibrated LCR.
Anyway there are only three critical caps per channel in a simplistic so it is not a big deal.
I also match all transistors and resistors.
Last edited:
IMO, if you use polystyrene or teflon you will not need to worry about cap value drifting because these are the most stable.
For instance, when I am short of T$me, I use regular 2% small polystyrenes to top up the main 47n and 15n caps.
It is quite normal to get a wide variation in a batch of caps... for example, I just bought 20 150p 2% styrenes from reichelt and was able to get some matching pairs 149p // 149p, 151p // 151p, 147p // 147p, so I choose the pairs and use them.
If I use those two 151p I am convinced they are exactly the same and will not drift differently with temp.
Off course, if you do not match them for yourself, you must buy 1% or less.... some russian micas have 0.3% spec !!!
For instance, when I am short of T$me, I use regular 2% small polystyrenes to top up the main 47n and 15n caps.
It is quite normal to get a wide variation in a batch of caps... for example, I just bought 20 150p 2% styrenes from reichelt and was able to get some matching pairs 149p // 149p, 151p // 151p, 147p // 147p, so I choose the pairs and use them.
If I use those two 151p I am convinced they are exactly the same and will not drift differently with temp.
Off course, if you do not match them for yourself, you must buy 1% or less.... some russian micas have 0.3% spec !!!
Sorry, I did not realize you wanted a pic of the LCR
You can search for this: TENMA 72-8150
I used a bunch of 0.5% caps to calibrate it myself.
Inexpensive but effective... no kelvin thou
You can search for this: TENMA 72-8150
I used a bunch of 0.5% caps to calibrate it myself.
Inexpensive but effective... no kelvin thou
News bulletin.🙂 I am at Vgeorge's Island studio room. Build and review of folded starting tomorrow. Today we calibrated his scope, and we are evaluating and tweaking his new TSSA V.16 CFB amp, calibrating his mic for rechecking the horns system so all will be under control. Making an antiRiaa also. Over.
A lot of work to do Salas but you are in good company, let us know how goes the folded & more about antiRIAA also.
That's not a calibrate LCR, no Kelvin wire to measure, you are far to know real values are you measuring, you are just trusting matching seller without to know if is OK or not the value needed.
No. The two coils would still be wound in the same direction.
You would need to find a manufacturer who wound the coil in the other direction for this parallel trick to work.
Merlin, you are wrong !
If the two caps, one 20% tolerance and one 1% tolerance are measured and found to be close to an exact match, say within 0.1% then you KNOW that both capacitors are within 1.1% of true value, if you have used the same measurement procedure specified by the 1% manufacturer.
I have a small range of 0.3% capacitors I use to roughly calibrate my measuring procedure.
Same with my much bigger range of 0.1% resistors.
But I am aware of drift and I cannot guarantee 0.3% accuracy, maybe 1% accuracy for all my 5% and 10% and worse capacitors, if I need absolute values rather than component matching.
Most of my component matching can achieve better than 0.1% spread and sometimes if I take great care, better than 0.005% (1part in 20000 spread).
Phew !!!
Capacitor & resistor debate...HEATING UP !!!
I have to say...
...Capacitors & resistors have been my last 3 weeks of nightmares !!!
Mostly 'capacitor drift' though, is what concerns me at present.
Particularly for the RIAA caps ( not couplers ).
Also 'drift' in caps used for an 'inverse RIAA' ( took some figures, LOOKS BAD ! )
Well...relatively.
Cheers
Si.
Mica is the only thing I have found, with almost 'non measurable' drift.
Polystyrene - drift up to 3% on small values.
Capacitor & resistor debate...HEATING UP !!!
I have to say...
...Capacitors & resistors have been my last 3 weeks of nightmares !!!
Mostly 'capacitor drift' though, is what concerns me at present.
Particularly for the RIAA caps ( not couplers ).
Also 'drift' in caps used for an 'inverse RIAA' ( took some figures, LOOKS BAD ! )
Well...relatively.
Cheers
Si.
Mica is the only thing I have found, with almost 'non measurable' drift.
Polystyrene - drift up to 3% on small values.