I'm refurbishing an old NAD amp and am wondering if I should replace the ceramic disk caps along with the others. Do they age? Does it make sense to replace them with a different type of cap?
Thanks 🙂
Thanks 🙂
There are some rare old ceramic types that can age and deteriorate or which are just bad from their burn. But in general, it is not typical for disk ceramic caps to age.
I replace them sometimes only if they are the non-NP0 type to NP0.
I replace them sometimes only if they are the non-NP0 type to NP0.
Class II ceramic capacitors slowly lose capacitance, of the order of -5 % for each factor of ten increase of the time since the capacitor was last heated above its Curie temperature. Class II capacitors are not used at places where the capacitance is critical, though.
Nor are they used at the Curie temperature, for most ceramic capacitors it's around ~125°C.
It's the C370 integrated. They're sprinkled across all the boards. PSU, preamp, amp modules, etc.
On a related note, from all the research I've done, I don't need to replace the metallized poly caps either. Is this correct?
Thanks!
Last edited:
According to TDK
Class I is NP0/COG - long term stable
Class II is X7R or similar down about 5% after 1 year, a century to reach 10%
Class III is Y5V etc down 15% after 1 year
So if you replace a Type II or III with new of the same, it will be back to almost the same place after a year.
It might be possible to go from a III to a II or a II to a I as there is more choice now and the original design was driven by bean counters.
I don't think this is necessary because class II capacitors are usually not used at places where the capacitance is critical, but if you want to get them back to their old values, instead of replacing them, you could also heat them above their Curie temperatures and let them cool down again. They should be centred around their nominal values about 1000 hours later.
Actually no, the capacitance peaks at the Curie temp and slowly falls back again at higher (and lower) temperatures. Its very different to the ferromagnetic Curie temperature despite having the same name.
Google "barium titanate permittivity temperature graph" or similar.
Leave them alone.
Some folks tend to over analyze and the thread just goes nowhere.
I have measured old 40+ year old class II and III ceramic capacitors, their values are still within spec. With my ESI 2150 LCR meter 0.02% accuracy. I can also measure the change in capacitance against DC bias.
There is a lot of BS floating around from folks who have absolutely no experience on these things yet they are the ones who posts a lot.
Last edited:
Amen Brother. 😀
That something may happen when/if Planets collide, Sun and Moon Eclipses happen at the same time and Earth magnetic spin switches polarity does not mean that it will happen and even less be a factor in some Audio device.
And an audible factor at that. 😎
That something may happen when/if Planets collide, Sun and Moon Eclipses happen at the same time and Earth magnetic spin switches polarity does not mean that it will happen and even less be a factor in some Audio device.
And an audible factor at that. 😎
I’ve certainly seen my share of 60 and 70 year old tube amplifiers with Z5U’s for coupling caps, and even in the bass/treble control circuits. And they seem fine. Regardless of whatever else happens they never fail short like very early film caps can. And if they do shift value over time it’s not by enough to matter (unlike drying out electrolytics).
Interesting, I didn't know that. I think I read that they lose most of their capacitance above the Curie temperature in some capacitor application note, but clearly they don't.
I haven't seen any BS in this thread except for my mistake about what happens above the Curie temperature, have you?
The spec of a class II ceramic capacitor usually applies 1000 hours after the last time their temperature exceeded the Curie temperature. If they lose 5 % per factor of ten increase of that time, they are at -12.7 % after 40 years. Chances are that +/- 20 % or +80 %/-20 % tolerance capacitors are then still in tolerance. If you desoldered them and they got really hot during desoldering, it could be that you have to count from the moment of desoldering - although that's more likely with SMDs that with big old disc capacitors.
Anyway, like I wrote before, I think it's a non-issue.
Last edited:
I once read a post where someone replaced all the ceramic caps in his solid state McIntosh amplifier with audiophile market brand polypropylene - and the thing just oscillated when powered on. His OP was "what am I going to do now?"...one of the off-the-cuff answers was "put all the ceramics back in?".
I wouldnt mess with them, unless you know for sure they're in the direct audio signal path. Maybe not the best passive component to be listening through. I'd think you could upgrade the caps in the tone controls (for example) without the amp going haywire as a result. Some of the other "sprinklings"? Maybe the stability of the amp depends on the specific behavior of some of those and if you were to change that...
I wouldnt mess with them, unless you know for sure they're in the direct audio signal path. Maybe not the best passive component to be listening through. I'd think you could upgrade the caps in the tone controls (for example) without the amp going haywire as a result. Some of the other "sprinklings"? Maybe the stability of the amp depends on the specific behavior of some of those and if you were to change that...
My policy with old equipment is not to mess around until there is a specific problem or solution.
The rest is all in the mind, the expectation that if this is done, it will be better.
Like coffee brewed in different places from the same beans...
The rest is all in the mind, the expectation that if this is done, it will be better.
Like coffee brewed in different places from the same beans...
Looking at the schematic, the ceramics are mostly used for RF bypassing of inputs, so you should be fine there. Other places where a low cost amplifier might have used ceramics, NAD seem to have used metalized foil. From my superficial look the coupling and bypass caps, etc, seem to be quite generous values, so replacing like for like electros should be fine. Choose an audio grade, and if you want to be fussy replace any electrolytic capacitors that have very small voltages across with bi-polar capacitors (e.g. in interstage stage coupling). If anything the main supply capacitors could be larger. If it were my amp I would be looking for 4 x 22,000µF that will fit, or at least 4 x 15,000µF.
Things have moved on since the Ford Model T. It is true that in extremely rare circumstances the ESR, self-inductance and/or loss factor of a capacitor is deliberately used beneficially in contemporary designs. Putting in a more perfect capacitor can upset stability or have some unanticipated effect in that case. For example Quad FM4 tuners have a TL094 comparator to driving a capacitive load in a control circuit. This comparator is not stable into a purely capacitive load and depended on the poor ESR spec of the crappy 1970s general purpose electrolytic capacitor. Putting in a modern electrolytic causes the TL094 to oscillate, though not enough to stop it working as intended. However the oscillation in turn cross-talks into the audio as a whistle audible at very low level. The solution is to fit a small value resistor in series with the capacitor (47Ω in the case of the TL094).
In much older electronics, designers had no options but to work with the components available and their innate imperfections. Even though today's components are much better the perfect capacitor does not exist, and the notion that any capacitor can't or won't affect sound quality is clearly nonsense in analogue circuitry. That doesn't mean some people don't care, and the expectation effect thus ensures that they won't hear it. So be it.
Last edited: