Which OPAs are built in and what is their power supply, single or double?
Depending on the schematic, perhaps some modern OPAs with a small DC offset can be put in, thus completely eliminating those coupling capacitors and get a much better sound. It's only an option if you have a dual power supply.
I say I use a DAC, preamp and power amp directly coupled, without any coupling capacitors. The condition of all conditions in this case is that there is DC protection for the speakers in the power amp, if something goes wrong.
When I use a tube preamp, there must be an output capacitor in the preamp, but nowhere else.
Depending on the schematic, perhaps some modern OPAs with a small DC offset can be put in, thus completely eliminating those coupling capacitors and get a much better sound. It's only an option if you have a dual power supply.
I say I use a DAC, preamp and power amp directly coupled, without any coupling capacitors. The condition of all conditions in this case is that there is DC protection for the speakers in the power amp, if something goes wrong.
When I use a tube preamp, there must be an output capacitor in the preamp, but nowhere else.
Last edited:
I have replaced many electrolytics with ceramics with great success. Somethimes it doesn't matter at all to the sound and simply gives much better reliability. If there's any AC voltage across the electrolytic coupling capacitor there can be measureable low frequency distortion and the ceramics eliminiate it. The cermaics can in some cases cause microphonics if not NP0/C0G but I have not found it to be a problem yet as it takes quite a bit of gain.
I have gone up to 22 µF and beyond with parallelling and I have desinged many circuits with no electrolytics at all by using only ceramics. It's a cool trick really. I do simulations and THD measurements of test ciruits if I'm uncertain with such decisions and I havent been disappointed yet.
And eliminating coupling capacitors is sometimes a great idea but it's no doubt a slippery slope. I have experienced strange cases of noise in some circuits where I never thought there would be a problem. It seemed as if there was no DC to be concerned about but then there was a low level popcorn noise in one case and restoring the AC coupling eliminated it. Sometimes the smallest DC matters.
I have gone up to 22 µF and beyond with parallelling and I have desinged many circuits with no electrolytics at all by using only ceramics. It's a cool trick really. I do simulations and THD measurements of test ciruits if I'm uncertain with such decisions and I havent been disappointed yet.
And eliminating coupling capacitors is sometimes a great idea but it's no doubt a slippery slope. I have experienced strange cases of noise in some circuits where I never thought there would be a problem. It seemed as if there was no DC to be concerned about but then there was a low level popcorn noise in one case and restoring the AC coupling eliminated it. Sometimes the smallest DC matters.
Good point and I do consider this. It's about the trade-offs and the specific application. Electrolytics are often the right choice. It's in the details.
There are some issues regarding the class 2 parts and the biggest is aging causing the capacitance to drop. I would not use Class 2 ceramics where the capacitance value is critical. Also, I do test for the distortion with an equivalent circuit on a breadboard carefully set up to have the same conditions as the actual circuit.
I do use film to replace electrolyics in some cases and it's also a matter of SMD or TH. In many cases if the the circuits are tiny and cheap you can REALLY overkill the value and all is well.
Strangely I have seen the value fall off from aging in Class 2 ceramics far more than distortion problems in practice. Distortion can be a problem but it varies wildliy with the application and where coupling is concerned over-sizing solves the problem.
Now you have me thinking about the possibility that aging could cause non-linear behavior to develop with aging. If the capacitance drops, does the linearity remain constant?
All of that said, I will second guess myself and do more homework.
There are some issues regarding the class 2 parts and the biggest is aging causing the capacitance to drop. I would not use Class 2 ceramics where the capacitance value is critical. Also, I do test for the distortion with an equivalent circuit on a breadboard carefully set up to have the same conditions as the actual circuit.
I do use film to replace electrolyics in some cases and it's also a matter of SMD or TH. In many cases if the the circuits are tiny and cheap you can REALLY overkill the value and all is well.
Strangely I have seen the value fall off from aging in Class 2 ceramics far more than distortion problems in practice. Distortion can be a problem but it varies wildliy with the application and where coupling is concerned over-sizing solves the problem.
Now you have me thinking about the possibility that aging could cause non-linear behavior to develop with aging. If the capacitance drops, does the linearity remain constant?
All of that said, I will second guess myself and do more homework.
I love the debate and convo. One question was inspiration to this - put simply, the 'anything but' philosophy. Anything but electrolytics.. Unrelated, I've been on an anything but WIMA kick on film, caps and hasn't been a bad thing.
On the ceramics, I was looking at a series like this one - comes in a 10uF.
https://product.tdk.com/en/system/f...d-mlcc/catalog/leadmlcc_halogenfree_fg_en.pdf
On the ceramics, I was looking at a series like this one - comes in a 10uF.
https://product.tdk.com/en/system/f...d-mlcc/catalog/leadmlcc_halogenfree_fg_en.pdf
This is no doubt an interesting subject that has quite lot of consequences regarding cost, performance, reliability, and size. It's well worth looking into carefully and I know that I'm not finished with my work. I use ceramics all over the place for coupling and haven't run into problems but there's more too it than that.
How long will that acceptable performance last? What can be easily measured and what cannot? How much does it really matter in your application?
I'm certainly thinking it through and running some tests.
How long will that acceptable performance last? What can be easily measured and what cannot? How much does it really matter in your application?
I'm certainly thinking it through and running some tests.
This EDN article is quite related: https://www.edn.com/class-2-ceramic-capacitors-can-you-trust-them/
The EDN article shows a logarithmic decrease in capacitance with time. This is quite interesting to know. It means that an X7R should loose no more than 15% after 11 years, in use or not. With that information it's easy to use a 20% or so excess value for coupling and all is likely well indefinitely. If the X7R tests with insigificant distortion for the application, the size, cost, reliability tradeoff decisions can be intellegently made. Is it that simple? Maybe.
Another interesting one on distortion in high K ceramics: https://www.edn.com/signal-distortion-from-high-k-ceramic-capacitors/
Large value ceramic caps use ferro-electric dielectric material which leads to about the most distorting kind of capacitor you can get - grossly non-linear with cross-over distortion as well... Nope, large value ceramics are for supply decoupling, not for audio signal path.
Large value electrolytics are very well behaved by comparison, if not completely dried up.
Ceramic caps other than NPO have capacitance that is somewhat depended on the voltage across them. Other than NPO (or COG if I'm hard up), I'd stay away from ceramic caps for passing audio signals. I've used NPO in RIAA EQ circuits to good effect.
This issue is not so on and off. There are most certainly many cases were ceramics are quite ideal as coupling capacitors. The only issue is the finite change in capacitance with age and DC voltage across them. Calling this "aging" is very misleading because it's not an age related damage but rather a reversable change in the state of the dielectric material. It's also quite finite and stops around -25% of the capacitance value at some point. If you design as if that is the starting value and you properly design the coupled circuit to have no significant voltage across the capacitor at the lowest frequency of interest, there will be no distortion. It doesn't matter how supposedly horrible the particular ceramic capacitor is. Filters are another matter. Only C0G/NP0 works there.