Panasonic Electrolytic Caps different Series - Opinions

[JohnRoberts]

I was taught X7R is well known for the phenomenon. I suppose now that you mention, I did just take their word for that. However, that bunch was pretty knowledgeable (Shure). I've personally seen lytics that 'THUNK' pretty bad when you tap them though. And cheap ceramics that flat just come apart. IDK.

The X7R is probably a decent cap for PS decoupling but +/-15% capacitance change over temperature range makes it poor for timing circuits. It is worth mentioning that they do not even specify a voltage coefficient for that dielectric so obviously should never be used in filters or audio paths.

Physical construction. It's paramount for touring gear at any rate. Studio control rooms are much more forgiving.

I understand the rigors of sound reinforcement business.

Not the leads per se. Where they come thru the bottom tends to want to fail and leak, if I understand it correctly. It's kinda standard practice for radial Aluminum 'lytic caps on industrial stuff. The tall skinny cans in particular. They end up like they're 'standing' on their leads, all the weight of the can pushing down and the leads holding the weight but held immobile by the solder. Then comes the vibration and ...Uh OH!

Axial lead caps were more popular inside Peavey until we purchased a radial component insertion machine (coincidentally made by Panasonic). Radial caps are cheaper because of only one bung. Properly installed in a PCB the cap rests on the bung and then the leads get soldered. Over millions of components I have seen sundry issues but overall pretty reliable.

True enough, but the different types tend to have their favorite way to fail. Like transistors: BJTs (and diodes) fail short, FETs fail open. It _can_ be whatever, it's just usually a certain mode for each.

With capacitors some are designed very specifically for how they fail. X and Y rated caps are designed for use on mains power (X across mains, Y hot to safety ground).

That particular fellow did come up during that period. I guess I'm not exactly immune to, shall we say... 'dogma' either when it comes to this stuff. Mostly I just try to copy stuff I've seen work reliably. Probably most designers do.


You can disagree all you like, I don't mind. You have, clearly, a vast experience that is at least somewhat different than mine. Prove me wrong, and I'll learn something. Questioning things and dialogue are how we learn things. Thanks for the great info amigo!

Still learning... but caps are a mature technology... When we start dealing with SMD caps package size introduces a new variable.

JR

 

[living sounds]

Can't remember where I read this, but the idea was that electrolytic caps tend to be optimized for DC performance, with manufacturers often omitting coating that would ensure equal performance when biased wrongly. This would result in degraded performance with AC (including audio signals).

I have performed a lot of listening tests with electrolytic caps, and usually didn't like the low-ESR ones in unbiased audio applications. This was very obvious with complete console signal paths.

Now I use either bipolar caps or Nichicon caps designated for audio use. Measurements of ESR and leakage don't seem to matter much.

From tests I'm also firmly in favour of adding a small poly cap (WIMA polyester) in parallel. Because of the resonances in the impedance response proper complimentary values need to be selected. For example, for a 47uf coupling cap a 10nf polyester cap works well.

 

[JohnRoberts]

IIRC Samuel Groner published some serious bench testing of low level distortion in electrolytic caps related to biasing schemes, years ago when he was more active here. I can't find him here now... His website lists some publications, Samuel Groner's website

If there is an audible difference between different caps that seems by definition to be distortion, that could be measured on a good test bench. I am a fan of null tests to parse out small differences, while I do not look there (this century).

Last century (actually in the 70s) I discovered with test bench work that parallel film caps need to be at least 10% of the larger non-film cap to deliver full film capacitor performance.

Of course capacitors have changed dramatically (mostly for the better) since my bench work 50 years ago, so YMMV.

JR

 

[abbey road d enfer]

Can't remember where I read this, but the idea was that electrolytic caps tend to be optimized for DC performance, with manufacturers often omitting coating that would ensure equal performance when biased wrongly. This would result in degraded performance with AC (including audio signals).

This doesn't make much sense. What "coating" is it ?

I have performed a lot of listening tests with electrolytic caps, and usually didn't like the low-ESR ones in unbiased audio applications. This was very obvious with complete console signal paths.

Now I use either bipolar caps or Nichicon caps designated for audio use. Measurements of ESR and leakage don't seem to matter much.

I'm always very skeptical about this type of assertions.
A coupling capacitot can "have a sound" only by interacting with the circuit it is inserted in.
The basic rule is making sure the voltage across a coupling capacitor is about 2 orders of magnitude below the voltage it transmits. It is easy to confirm to this rule by making the -3dB point at 0.2 Hz.
In that case, the voltage across the cap at 1kHz is 75dB below signal. According to measurements, the distortion applied to the voltage across the cap is less than 0.1%, so the overall distorsion is under -130dB. Even with 10 cascaded stages, that would result in THD below -110dB.
My conclusion is that if one can hear distortion created by electrolytics capacitors, there is a design malpractice.
For a long-time, designers have used the rule of 10, inherited from AM radio design, that was transferred to HiFi design. It means if a capacitor X is needed for a given -3dB response, use a ten time bigger cap. 20Hz being the commonly accepted LF turnover, they designed for 2Hz. this is not correct. In that case, the voltage across the cap at 20Hz is 1/10th of the transmitted signal, which results in significant distortion.
By using the rule of 100, the voltage across the cap is never enough to create significant distortion.
This applies to solid-state circuits. Tube circuits would require coupling capacitors of 4-10uF, which is impractical and very costly. That's why they use film caps that have about 10 times less distortion. that's why they can still use the rule of 10.

From tests I'm also firmly in favour of adding a small poly cap (WIMA polyester) in parallel. Because of the resonances in the impedance response proper complimentary values need to be selected. For example, for a 47uf coupling cap a 10nf polyester cap works well.

I am indeed in favour of paralleling electrolytics and ceramic caps for power rail by-pass, but this is utterly unjustified for coupling caps, again, as long as they are properly dimensioned.
Look at the detailed model of an electrolytic cap and you'll know why.

 

[Whoops]

Getting back to the Topic, the different Panasonic Electrolytic cap series....




Is Low impedance important for audio circuits?

It seems that FS and FR series are low impedance and have the longer life of the "F" series.

So should I have give relevance to these series?

Thanks

 

[living sounds]

This doesn't make much sense. What "coating" is it ?

I'm always very skeptical about this type of assertions.
A coupling capacitot can "have a sound" only by interacting with the circuit it is inserted in.
The basic rule is making sure the voltage across a coupling capacitor is about 2 orders of magnitude below the voltage it transmits. It is easy to confirm to this rule by making the -3dB point at 0.2 Hz.
In that case, the voltage across the cap at 1kHz is 75dB below signal. According to measurements, the distortion applied to the voltage across the cap is less than 0.1%, so the overall distorsion is under -130dB. Even with 10 cascaded stages, that would result in THD below -110dB.
My conclusion is that if one can hear distortion created by electrolytics capacitors, there is a design malpractice.
For a long-time, designers have used the rule of 10, inherited from AM radio design, that was transferred to HiFi design. It means if a capacitor X is needed for a given -3dB response, use a ten time bigger cap. 20Hz being the commonly accepted LF turnover, they designed for 2Hz. this is not correct. In that case, the voltage across the cap at 20Hz is 1/10th of the transmitted signal, which results in significant distortion.
By using the rule of 100, the voltage across the cap is never enough to create significant distortion.
This applies to solid-state circuits. Tube circuits would require coupling capacitors of 4-10uF, which is impractical and very costly. That's why they use film caps that have about 10 times less distortion. that's why they can still use the rule of 10.

I am indeed in favour of paralleling electrolytics and ceramic caps for power rail by-pass, but this is utterly unjustified for coupling caps, again, as long as they are properly dimensioned.
Look at the detailed model of an electrolytic cap and you'll know why.

Sorry, like I wrote, I can't remember where I read that. It might be that these caps are not designed to withstand any reverse polarity and degrade or at least change their properties if biased in reverse or even with AC applied...

I alway used caps that were much bigger than the ones originally in the circuit, to be on the safe side wrt the low end frequency/phase response.

I got these hints about paralleling electrolytics with poly caps of certain sizes from a guy who built large format pro studio consoles for a living and my ears agree with the difference it makes. Besides, I am certainly not the only one who does it.

About paralling electrolytics with ceramic caps for PSU decoupling, the result can be antiresonances which might actually degrade performance. I have an idea that many DIY efforts of enhancing decoupling actually achieve the opposite of the desired result, including many I performed myself...

 

[living sounds]

Getting back to the Topic, the different Panasonic series....


View attachment 81307

Is Low impedance important for audio circuits?

It seems that FS and FR series are low impedance and have the longer life of the "F" series.

So should I have give relevance to these series?

Thanks

For audio decoupling, of those, I like the black/gold ones (FM) best. Second would be FC (gold/blue) and least FR (black/grey). The latter have the lowest stated ESR... I didn't like them in the signal path and rarely liked them for PSU decoupling (maybe due to antiresonances with parallel ceramic caps?).

 

[ricardo]

The only capacitors I encountered (over decades) that were microphonic, were blatantly faulty.
...
COG/NPO are as good as it gets (I loved polystyrene dielectric but not very robust in the factory).
...

Polystyrenes are microphonic. Guru Wurcer confirmed my 1980s experience when he did his Linear Audio series on LN mike preamps You can easily check this out by replacing a condensor mike capsule with the same value polystyrene or ceramic and handling/tapping the mike.

This isn't Golden Pinnae stuff but easily audible. You can listen for yourself & decide if this level of microphony is OK for you. Q is usually high which means 'less audible' The beloved Audiophile Silver Micas are also 'noisy'.

As JR says, COG/NPO are as good as it gets.

You can also check out Tants and other Electrolytics. Tants have 2 'types' of noise. One is a 'woosh' which is sometimes by masked by ambient noise in a microphone. At much lower signal levels is a 'crackly' noise .. rather like insufficiently dithered digital audio.

The diyAudio link on my MC preamp, and MicBuilders SimpleP48 have details. None of this needs Golden Pinnae and you can hear & decide for yourself if the noise is acceptable.

Today, with good Lo-leakage Aluminium electrolytics like the Panasonics, I don't see any reason for using Tants ... unless, a wooshy & crackly noise floor are part of the sound you want

.. then there's the reliability and momentary reverse voltage issues too ..

 

[JohnRoberts]

I don't recall ever encountering microphonic polystyrenes but I'll take your word for it, they were cheap but really good (for the money) caps back in the day, but don't look at them crosseyed... My business partner trashed a whole production run of them trying to blow cleaning water off PCBs with a high pressure air hose. He blew water inside the caps (they had to all be replaced).

JR

PS: Speaking about noisy caps, I encountered a production run of a brand new version (recently approved by one of my group's engineers) electrolytic cap. These caps were the mic preamp phantom voltage blocking caps. The factory QA workers on the line complained that they were noisy... At the time I didn't invest any time into figuring out why, I just black balled them for caps already in the system that were quiet. In hindsight I suspect they might not have been properly formed in by the cap manufacturer so were extra leaky and that leakage current at the mic preamp input was what was audible (just a WAG). At the time it wasn't my job to figure out every single thing my vendors did wrong.

 

[Whoops]

For audio decoupling, of those, I like the black/gold ones (FM) best. Second would be FC (gold/blue) and least FR (black/grey).

Hi Living Sounds,
thank you so much for your input.

Could you please let me know why is that your preference?
What criteria did you use to prefer FM over FC and FR the least?

Thank you so much

 

[Whoops]

Getting back to the Topic, the different Panasonic Electrolytic cap series....

I was buying Panasonic Electrolytic caps in Mouser, and they have a nice feature that you can use to compare items.

As for the same cap value/voltage from Panasonic they offer a lot of different series, I compared a few of them.

From the specs they show in Mouser it seems you can compare rated voltage, Life (hours) , Max and Min Temperature, ESR and Ripple Current.

The Ripple current showed a lot of different values dependent on the series.

Does anyone know or can explain what "Ripple Current" means in the context of a Lytic capacitor Specification?
Should we preffer a Lytic cap with a Lower Ripple Current?
Does this matter for audio/signal path or just for PSU usage?

 

[abbey road d enfer]

Ripple current in specs is the maximum ripple current the capacitor is capable of sustaining without damage. The dominant issue is heat produced by Joule effect.
It is closely related to ESR, since W=R.I²
It's a paramount factor with smps.
It is not directly relevant to audio, unless you use capacitors in passive speaker x-overs.
In common low-level circuits, an added resistance of a fraction of Ohm is negligible.
There are a few exceptions, as Ricardo mentioned, like MC cartridge preamps. But even for a xfmr-less mic preamp, the addition of less than 1 Ohm is negligible.

 

[living sounds]

Hi Living Sounds,
thank you so much for your input.

Could you please let me know why is that your preference?
What criteria did you use to prefer FM over FC and FR the least?

Thank you so much

Mostly by ear. That is subjective, but the differences were not small...

I had one instance of a headphone amp where the low ESR caps in the PSU section caused stability issues that resulted in massive hum.

Technically worse parts (higher resistance / ESR) can provide dampening that a circuit may actually rely on to be stable. When recapping I now replace tantalums with low-ESR electrolytics, but otherwise stay with the safer option of standard electrolytic caps. In the signal path I do like the Nichicon "audio grade" stuff. My entire console is fitted with Nichicon audio grade caps both in the signal path and for the PSU. Works and sounds great.

Another PSU decoupling anecdote: When I bought my LFC last year the PSU voltages were drifting. It turned out a previous owner had added lot's of small ceramic (X7R?) caps at various positions for decoupling. These made the entire circuit unstable, probably by creating antiresonance irritating the regulator.

 

[JohnRoberts]

Getting back to the Topic, the different Panasonic Electrolytic cap series....

I was buying Panasonic Electrolytic caps in Mouser, and they have a nice feature that you can use to compare items.

As for the same cap value/voltage from Panasonic they offer a lot of different series, I compared a few of them.

From the specs they show in Mouser it seems you can compare rated voltage, Life (hours) , Max and Min Temperature, ESR and Ripple Current.

The Ripple current showed a lot of different values dependent on the series.

Does anyone know or can explain what "Ripple Current" means in the context of a Lytic capacitor Specification?

+1 to what Abbey already explained. Ripple current is a quantification of the work that PS reservoir capacitors do when cyclically charged fast by diodes, and discharged slowly by circuit draw. This is affected by internal resistance (ESR) and ability to dissipate heat generated by IxR.

Should we preffer a Lytic cap with a Lower Ripple Current?

under a more is better scenario higher ripple current suggests a better capacitor (for handling ripple current).

Does this matter for audio/signal path or just for PSU usage?

Mostly just for PS usage but capacitor sellers will try to differentiate using whatever they have. The last time I bought electrolytic caps for an audio path in a personal project I did not even look at the specs.

JR

 

[FIX]

Any device that has gain or will be amplified a lot will amplify the errors in the component that is in question. The biggest audio difference I have found in mics is the 1G ohm resistors. But again, the capsule, 1G resistors and caps around the capsule are being amplified, so any anomalies will be amplified by the gain. If is has a small rating for some spec, and the preamp is amplifying it 1000 times, it will be 1000 times "more".

 

[abbey road d enfer]

Any device that has gain or will be amplified a lot will amplify the errors in the component that is in question. The biggest audio difference I have found in mics is the 1G ohm resistors. But again, the capsule, 1G resistors and caps around the capsule are being amplified, so any anomalies will be amplified by the gain. If is has a small rating for some spec, and the preamp is amplifying it 1000 times, it will be 1000 times "more".

No. X% THD on 1mV is also X% when the signal is amplified 10 000 times. An amplifier amplifies errors as much as it amplifies signal; it can't discriminate between "good" or "bad" signals.

 

[ricardo]

I don't recall ever encountering microphonic polystyrenes but I'll take your word for it, they were cheap but really good (for the money) caps back in the day, but don't look at them crosseyed...

Were you using 1 & 2% polystyrenes for EQ at Line level? From my tests, I don't think you would notice microphony at those levels.

Hmm.mm! You might be able to hear this on RIAA preamps but I've never tested this. Besides, in dem days, 1 & 2% caps were nearly Unobtainium and you used what you could get.

My business partner trashed a whole production run of them trying to blow cleaning water off PCBs with a high pressure air hose. He blew water inside the caps (they had to all be replaced).

Was this your supa-dupa RIAA preamp? If you buried the RIAA PCB in solid BS, you would eliminate any possibiltiy of microphony and the BS would give you extra clarity & definition ... just shut up ricardo! JUST SHUT UP !!

I encountered a production run of a brand new version (recently approved by one of my group's engineers) electrolytic cap. These caps were the mic preamp phantom voltage blocking caps. The factory QA workers on the line complained that they were noisy... I suspect they might not have been properly formed in by the cap manufacturer so were extra leaky and that leakage current at the mic preamp input was what was audible (just a WAG). At the time it wasn't my job to figure out every single thing my vendors did wrong.

I've encountered this in a LN 'commercial' design this Millenium. It took about 30min for the noise/leakage to go away. But then it was gone for good.

 

[JohnRoberts]

Were you using 1 & 2% polystyrenes for EQ at Line level? From my tests, I don't think you would notice microphony at those levels.

Hmm.mm! You might be able to hear this on RIAA preamps but I've never tested this. Besides, in dem days, 1 & 2% caps were nearly Unobtainium and you used what you could get.

in my old RIAA stuff I used 1% resistors because they were relatively cheap but 5% capacitors

Was this your supa-dupa RIAA preamp? If you buried the RIAA PCB in solid BS, you would eliminate any possibiltiy of microphony and the BS would give you extra clarity & definition ... just shut up ricardo! JUST SHUT UP !!

Actually those were used in the Loft studio delay line/flanger... My partner was experimenting with organic flux that needs a serious water wash... the high pressure air hose was too serious for the polystyrene capacitor housing.

I've encountered this in a LN 'commercial' design this Millenium. It took about 30min for the noise/leakage to go away. But then it was gone for good.

In production at Peavey we didn't give components 30 seconds to get quiet... I didn't bother to figure out why they were noisy, I just switched back to caps that were quiet from the start.

Not being formed in properly is just my WAG decades later, I didn't give them this much thought back then.

JR

PS; How are you doing? still living on the beach in Oz?

 

[Whoops]

I was comparing the different series in Farnell and Mouser for a common value of 100uf 25v.
Farnell doesn't even spec the ESR value or ripple current value, they show 7 different series for 100uf 25v, besides the max temperature rating and the Life rating there's not much info to distinguish them.

Strange also that a 105 rated cap with a Life of 5000 is even cheaper than a 85 rated cap at 1000 hours.

Here is a table with the specs published by Farnell

SERIES	KS	NHG	EEUFC	FC	EB	FM	M
Tolerance	± 20%	± 20%	± 20%	± 20%	± 20%	± 20%	± 20%
Diameter	8	6,3	6,3	6,3	6,3	6,3	6,3
Height	5	11,2	11,2	11,2	11,2	11,2	11,2
Spacing	2,5	2,5	2,5	5	5	5	5
Temperature MAX	85	105	105	105	105	105	85
Temperature MIN	-40	-55	-55	-55	-40	-40	-40
Life	1000	1000	1000	1000	5000	2000	2000
ESR	-	-	-	-	-	-	-
PRICE	0,263	0,198	0,243	0,325	0,206	0,203	0,135

 

[scott2000]

Shout out to Nichicon UKW as well..