![]() ![]() I realise this is referring to distortion but the argument for other 'effects' will be similar. If a given capacitor can manage this into it's actual 'load' across the frequency range then there should be minimal degradation. Apply the lesser dielectrics in ways that their bad behavior does not express.For any of you who don't know who John is, this man knows what he's talking about!Ĭheck his geekslutz posts, he has a great balance between practical circuit design and audiophile ephemera.ĭoug Self 'quantified' the amount of AC across an electrolytic capacitor such that distortion approaches unmeasurable to around 80mV. For caps in filter circuits that by design see audio voltage across their terminals, use premium dielectric (film, cog, npo, etc). PS: I am not saying DA, DF, voltage coefficient, et al do not matter, but if the cap terminal voltage doesn't change, voltage related phenomenon are not exercised. Apply the lesser dielectrics in ways that their bad behavior does not express. (Good luck determining whether the difference is for better or worse). When modifying or re-capping any such product with multiple similar paths, you can experiment and null different channels from each other to determine if and how much your tweaks have affected the path. I appreciate there are legacy designs where electrolytic caps are expected to do heavy lifting. Unfortunately with caps just like in life there are no easy answers, but cap behavior has been well inspected and understood for decades. Now, we can make filters with DSP so no caps need ever be exposed to AC voltage, except for A/D/A conversions. Use big film caps for passive loudspeaker crossovers where the caps are working hard, but even there biamping with active crossovers can finesse that heavy lifting. When doing original designs you can manage impedances so LF poles can be performed with film caps at higher impedance to keep cap size modest, and tune all other electrolytic poles in the path (like DC blocking) to be octaves lower than the one film cap dominant pole so the electrolytic caps never get exposed to LF energy and changing terminal voltage. My conclusion decades ago was to just not use electrolytic caps when stressed that way (mid-low R load). Note: small film caps in parallel may make a difference up at RF frequencies but I can't hear that high so don't focus on it. However putting a 2uF film cap in parallel with 22 uF electrolytic is still awkwardly large and impractical. Using bench measurements I found that a small film cap in parallel made no measurable difference to in-band (20-20kHz) behavior until that small cap value was increased to 10% of the larger cap value. I looked at the popular "just add a small film cap in parallel" technique to make the combination behave like a larger good cap. I am not claiming that phase shift at 20kHz was very audible, but it was easily measurable, and shouldn't be there, so something I could design out. While the ESR term only made a flat gain error in my circuit, the ESL was enough to generate a significant phase shift at 20kHz. While the opinion leaders like to be mysterious about cap errors, I found them to be pretty simple. ![]() #Bi polar caps in signal path series#I found measurable deviation from ideal or pure C, when a tens of uF electrolytic cap was used in series with a hundreds of ohms R to form a LF pole in a high gain stage, like in a phono preamp (which should indicate how long ago this was). Years (decades) ago when electrolytic caps were lower performance than recent offerings I invested some bench time into this exercise. ![]() This has been discussed many times but I haven't shared my experience in this thread. ![]()
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |