Hi,
I am measuering a bipolar lythic of a loudspeaker in shunt position after a resistor from a bass low pass section...
On a 100uF cap I see there are few difference between the moment where the cap was pulled out of the speaker filter that was in use the day before, and then 24 hours after. The capacitance value droped from around 0,300 uF on each channel after a night rest.
After a 24h rest, the capacitance meter gives after few seconds a 0,1 uF more then few seconds later 0,1 uF more yet ! I assume it is because of the DC charging from the capacitance hand meter. However the 0,300 uF are not recovered in the few seconds when I repeat the measurement.
What is please the acurate value I should keep for the filter calculation please ?
Despite a loudspeaker filter is seing AC, impulse currents, is it safe to say the damping resistor in the filter is acting as a polarising part for few undred nano cause the bipolar cap is after ? Or it is just due to the natural loading and unloading behavior of a cap whatever it ks jn a filtef cir uit and being bipolar in seie with the signal return, i.e. negative pole ?
So what do the filter see ? The cold rested measurement of the ca pvalue or the few hundred nano Farads more of the capwhen the loudspeakers are playing back music ?
I also assume than my cheap tool is more handy for classic non polar cap measurement than bipolar caps ! But the value of the two caps with the two ways of measuring go from 103.00 to 103.3 and 103.4.... and in the filter development setup the few hundred are important enough to be heard. Ok yhe ears will find what is needed qnd I can add for instance a decoupling of what is needed with precise MKP caps, but the questions are more about knowledge curiosity.
Thanks for your testimonials. Cheers.
I am measuering a bipolar lythic of a loudspeaker in shunt position after a resistor from a bass low pass section...
On a 100uF cap I see there are few difference between the moment where the cap was pulled out of the speaker filter that was in use the day before, and then 24 hours after. The capacitance value droped from around 0,300 uF on each channel after a night rest.
After a 24h rest, the capacitance meter gives after few seconds a 0,1 uF more then few seconds later 0,1 uF more yet ! I assume it is because of the DC charging from the capacitance hand meter. However the 0,300 uF are not recovered in the few seconds when I repeat the measurement.
What is please the acurate value I should keep for the filter calculation please ?
Despite a loudspeaker filter is seing AC, impulse currents, is it safe to say the damping resistor in the filter is acting as a polarising part for few undred nano cause the bipolar cap is after ? Or it is just due to the natural loading and unloading behavior of a cap whatever it ks jn a filtef cir uit and being bipolar in seie with the signal return, i.e. negative pole ?
So what do the filter see ? The cold rested measurement of the ca pvalue or the few hundred nano Farads more of the capwhen the loudspeakers are playing back music ?
I also assume than my cheap tool is more handy for classic non polar cap measurement than bipolar caps ! But the value of the two caps with the two ways of measuring go from 103.00 to 103.3 and 103.4.... and in the filter development setup the few hundred are important enough to be heard. Ok yhe ears will find what is needed qnd I can add for instance a decoupling of what is needed with precise MKP caps, but the questions are more about knowledge curiosity.
Thanks for your testimonials. Cheers.
Few nF difference in hundreds of uF is a drop in the ocean. It changes nothing.
You can expect differences of up to 1dB from 20% of capacitance variations (nominal tolerance of most electrolytics).
This should be tolerable in most filter designs and crossovers in practical applications. If you have larger variations, consider re-examining the filter design.
AC is AC, the location of the resistor does not 'polarise' a bipolar.
You can expect differences of up to 1dB from 20% of capacitance variations (nominal tolerance of most electrolytics).
This should be tolerable in most filter designs and crossovers in practical applications. If you have larger variations, consider re-examining the filter design.
AC is AC, the location of the resistor does not 'polarise' a bipolar.
Thanks Sangram.
However I can confirm in this case the few hundreds are enougj to change the bass equilibrium.
The question is more about to know which of the two measurement is the truer from a technical knowledge perspective.
However I can confirm in this case the few hundreds are enougj to change the bass equilibrium.
The question is more about to know which of the two measurement is the truer from a technical knowledge perspective.
I assume this is the cold 24h rested caps.
Even if I know it is a lythic and value will drift with aging. But if I fire up the loudspeaker everyday my feeling is I should trust the hot measurement i.e. the 0.3 uF more. Difference is enough on the bass to have the good amount of bass deepness VS tigthness in this 125 hz LR12 XO.
Even if I know it is a lythic and value will drift with aging. But if I fire up the loudspeaker everyday my feeling is I should trust the hot measurement i.e. the 0.3 uF more. Difference is enough on the bass to have the good amount of bass deepness VS tigthness in this 125 hz LR12 XO.
For sure and one can find that by himself...but this is not what I asked. I am not looking for a solution here, I try to undetstand what is the cause of the cap capacitance change and what value is seen by the filter.
Can we expect than a bipolar lythic can add because the curent few hundred nano Farads? Or is it only due to tje DC behavior of the capacitor meter tool when measuring ?
Can we expect than a bipolar lythic can add because the curent few hundred nano Farads? Or is it only due to tje DC behavior of the capacitor meter tool when measuring ?
Try a different tester, or a different capacitor, and see. I don't recall such behavior
on expensive lab equipment. The tester is just as likely to be at fault as the capacitor.
Do you completely discharge the cap first? Can you try it at a different frequency?
When I've checked smaller value, new bipolar caps (Nichicon UES), they give stable readings.
Is your tester reading stable with a smaller value bipolar cap, like the tweeter capacitor?
Is the variation within the accuracy specs for the tester? The LSD may not be relevant.
on expensive lab equipment. The tester is just as likely to be at fault as the capacitor.
Do you completely discharge the cap first? Can you try it at a different frequency?
When I've checked smaller value, new bipolar caps (Nichicon UES), they give stable readings.
Is your tester reading stable with a smaller value bipolar cap, like the tweeter capacitor?
Is the variation within the accuracy specs for the tester? The LSD may not be relevant.
Last edited:
It iw a basic 50 dollars tester.. I don t know for the others as they are films from the mid till the tweeter. But remenbering it is stable with the littles musê . Here it is a FT bipolar 100uf 100V cap.
Thanks for your idea, I will compare with few older bipolar lythic from other loudspeakers I always keep after recaping. I should have little values as 4.7 for instance.
Here the cap was tested on the 200 uF step. I dunno at what frequency the test is performed....maybe something between 800 hz or 1k hz...
I have an Arta box as well but unluckilly my laptop passes...I use my tablet so no access go something else.
Thanks for your idea, I will compare with few older bipolar lythic from other loudspeakers I always keep after recaping. I should have little values as 4.7 for instance.
Here the cap was tested on the 200 uF step. I dunno at what frequency the test is performed....maybe something between 800 hz or 1k hz...
I have an Arta box as well but unluckilly my laptop passes...I use my tablet so no access go something else.
Doubtful that it can test bipolar caps to one part in 1000 accurately. I wouldn't trust mine at that level either.
First thing to my mind was soakage.
Here is an overview:
https://www.edn.com/wp-content/uploads/2016/11/1982-10-13_EDN_Pease-capacitor-soakage.pdf
And 0.1uf change in a 100uf cap? 1/10% change?
Here is an overview:
https://www.edn.com/wp-content/uploads/2016/11/1982-10-13_EDN_Pease-capacitor-soakage.pdf
And 0.1uf change in a 100uf cap? 1/10% change?
Let me see if I get this right - you claim that a 3% change in the value of a shunt electrolytic affects the frequency response audibly? And that, and only that, is the reason for this change?
The filter design is an issue and needs to be sorted out in that case. No electrolytic will be guaranteed below 10% tolerance. Likely there's not enough inductance. But even so, 3% is pretty good for an electrolytic.
And you're testing this with a $50 tester, right? Not a professional benchtop capacitance meter?
------------------------------------------------------------
To answer your first question, the capacitance changes with applied voltage and current, as well as whether the cap is charging or discharging. Not by 3%, but there is a shift in the apparent capacity depending on various working states. Actual numbers on the tin come from a very specific voltage applied across the capacitor, and the charge within it, usually the maximum.
------------------------------------------------------------
This is the effect on the bass with a 150uF cap on an LR2 filter when it is changed to 145uF. The highest delta is below 0.3dB. Anybody who can hear that should apply for a sonar license.
The filter design is an issue and needs to be sorted out in that case. No electrolytic will be guaranteed below 10% tolerance. Likely there's not enough inductance. But even so, 3% is pretty good for an electrolytic.
And you're testing this with a $50 tester, right? Not a professional benchtop capacitance meter?
------------------------------------------------------------
To answer your first question, the capacitance changes with applied voltage and current, as well as whether the cap is charging or discharging. Not by 3%, but there is a shift in the apparent capacity depending on various working states. Actual numbers on the tin come from a very specific voltage applied across the capacitor, and the charge within it, usually the maximum.
------------------------------------------------------------
This is the effect on the bass with a 150uF cap on an LR2 filter when it is changed to 145uF. The highest delta is below 0.3dB. Anybody who can hear that should apply for a sonar license.
Here is the effect of a 10% variation in the capacitance - about 0.8dB worst case. While this is probably an audible effect, likely the room and driver variation will cause more of a swing than just this.
I re-read your first post and you are referring to a change of 0.3uF in a 100uF cap, that 0.3%. It would be impossible for that to be the cause of any change at all in frequency response. There is no possibility of that happening.
Yes I am surprised myself to hear such difference. 0.3 uF on a total of 104.7 uF for a LR12 low pass damped shunt at 125 hz is nothing...but well the behavior of the bass is changing enough. If I lower to around 104.4 uF the bass becomes muddy, bigger, fuller, it is invading the room somewhat. The more the capacitance is low the more it becomes noticeable. At the opposite'if I putt 105 uF it becomes dryer tighter, less boomy less deeper. It is an equilibrium as the 125 hz cut off is in the chest punch area i.e around 100 to 200 hz. On the mids the serie high pass caps are exactly 267.8 uF film caps on both channels.
I don't know if what my ears check is a capacitance or an esr change by the amount of parralell caps to adjust the vqlue by 0.1 uF step with film caps. As it is damped with 1R5 maybe the esr becomes too low and I should adjust the resistance and see if this time the sound is comming from that amount of damping and not the few hundred nano farads difference...
The two FT caps, brand that is making Mundorf caps, are very precise stock as there is almost no difference, on my cheap tool one measure 102.8 the other 103.1 uF, great. When recaped I found one of the two loudspeaker sounded sligthy different...one way was tigther the other fuller and more bassy.
I was very surprised as I added to each 1 uf Mkp + p.47 uf Mkp + 2 x 0.1 uF mkp to reach more or less the 104.7 uF of the XO, without worrying too much of 0,1 to 0,3 uf difference between the two loudspeakers.
Maybe all those caps are changing the amount of damping cause less esr by parraleling caps is lowering the damping with the 1R5 resistor in serie with the return path ? Or maybe there is a resonance by the mixing of great capacitance and very low caps ? Or a slow esr lythic rough foil of 100 uF against the fast small film caps ... but less than 2% capacitance between the lythic and the films is not that big !
So again, it is not a problem too choose the amount of capacitance that sounds best then fix it in the marble with big film cap to avoid the drift in time a lythic has but more to understand what happens technically here ?
I am glad to know the cheap capacitance multimeter is not precise enough to be trusted but there is a concistency in the differences it measures, so I trusted it saying the capacitance changes was the reason of sound difference... but maybe the reason is elswhere : mixing of cap model for instance...
Well again as I am not a tech guy but a noob, it is only by scientifical and knowledge curiosity I ask.
Many thanks for all the knowledge and experience feedback you made to enlight my curiosity...
I don't know if what my ears check is a capacitance or an esr change by the amount of parralell caps to adjust the vqlue by 0.1 uF step with film caps. As it is damped with 1R5 maybe the esr becomes too low and I should adjust the resistance and see if this time the sound is comming from that amount of damping and not the few hundred nano farads difference...
The two FT caps, brand that is making Mundorf caps, are very precise stock as there is almost no difference, on my cheap tool one measure 102.8 the other 103.1 uF, great. When recaped I found one of the two loudspeaker sounded sligthy different...one way was tigther the other fuller and more bassy.
I was very surprised as I added to each 1 uf Mkp + p.47 uf Mkp + 2 x 0.1 uF mkp to reach more or less the 104.7 uF of the XO, without worrying too much of 0,1 to 0,3 uf difference between the two loudspeakers.
Maybe all those caps are changing the amount of damping cause less esr by parraleling caps is lowering the damping with the 1R5 resistor in serie with the return path ? Or maybe there is a resonance by the mixing of great capacitance and very low caps ? Or a slow esr lythic rough foil of 100 uF against the fast small film caps ... but less than 2% capacitance between the lythic and the films is not that big !
So again, it is not a problem too choose the amount of capacitance that sounds best then fix it in the marble with big film cap to avoid the drift in time a lythic has but more to understand what happens technically here ?
I am glad to know the cheap capacitance multimeter is not precise enough to be trusted but there is a concistency in the differences it measures, so I trusted it saying the capacitance changes was the reason of sound difference... but maybe the reason is elswhere : mixing of cap model for instance...
Well again as I am not a tech guy but a noob, it is only by scientifical and knowledge curiosity I ask.
Many thanks for all the knowledge and experience feedback you made to enlight my curiosity...
Last edited:
If your tester gives repeatable readings, that makes it useful for matching, etc.
You can also buy precision capacitors to check the absolute readings on its various ranges.
Last edited: