Re: [AH] Behringer announces OB-Xa clone project

From Brian Willoughby
Sent Fri, Jan 25th 2019, 07:12

There=92s more to a capacitor than its capacitance. There=92s also ESR =
and leakage, as well as microphonics. Just because the same capacitance =
is reported doesn=92t mean it=92s the same cap.

Many modern surface mount caps are exactly like their through-hole =
ancestors. However, a few kinds of capacitor chemistries simply aren=92t =
available SMD. One example is for Sample and Hold, where =
super-low-leakage poly caps simply aren=92t available in SMD. Of course, =
there are other SMD caps that are good enough for S&H, but maybe those =
tropical fish caps were poly? Maybe they were mica? Foil? I=92d be =
curious to know the manufacturer and model of those caps.

Lots of old caps were really bad, and nobody really wants the problems =
any more. Again, it=92s not that their value changes, it=92s that they =
don=92t act exactly like a pure capacitor, but more like a complex =
combination of capacitor, resistor, and inductor.

That said, why hasn=92t someone just recreated the flaws of the old =
tropical fish caps by adding parallel or series resistors and inductors?

Brian


On Jan 23, 2019, at 8:10 PM, M V <xx_xxxxxx@xxxxxxx.xxx> wrote:
> Speaking of the new and old SEMs though, it's pretty well known that =
there is a clear difference between the new and old filter (more than TO =
would care to admit it seems). The former is brighter and sharper when =
resonance is at max, and not just because of an increased cutoff range. =
However, it came to light recently through the work of Oldgearguy over =
on GS that the likely culprit is the tropical fish caps that the old one =
uses and the new ones do not. OGG pulled the tropical fish cap on his =
old SEM and replaced with a modern hi-fi cap and found that the filter =
pretty much sounded the same as the new SEM. So this is sort of proof =
through absence of evidence because he didn't/couldn't put the through =
hole trop. fish cap in the new SEM's SMT PCB, but it does strongly =
suggest that with some very minor tweaking, one could get the new ones =
to sound identical to the old.=20
>=20