[aspenx]

I am modifying a clone of the Lehmann BCL and am thinking of upgrading the decoupling capacitors going to the OPA which I have rolled from a OPA2134 to LT1364.
 
The original capacitors are Elna RE3 470uF 35V bypassed with a ERO MKC1860 0.022uF 400V. The headamp sounds great all-round except for the mutant trebles that seem somewhat too incoherent with the rest of the sound. I figured this has got to do with the supplies to the OPA and read up on stumbled upon the idea of cascade bypassing.
 
The question is, I only have enough space on the board for 2 capacitors for each rail. I read on diyaudio that "bypassing a capacitor locally/directly on its terminals is not good and should be avoided". Is this true? If so, how should I go about it and should I even be considering cascade bypassing in the first place and why?

 

[aspenx]

After reading up some more, stumbling on some simulations and Rob Elliot's article, I decided that proper cascade bypassing is out of my league. Will try replacing the Elna RE3 with Silmic II equivalent and Orange Drop 0.022uF 600V (that I have laying around) for the ERO and hope the sound smooths out a bit.
 
Appreciate any help I can get. I don't think the LT1364 is supposed to sound this wrong here...

 

[qusp]

the 1364 is a very fast and wideband opamp, its very possible that what you are hearing is actually the opamp oscillating. if so, you may be able to tame this by installing a filter at the input or in the feedback loop to limit the bandwidth.

 

[aspenx]

EDIT: Problem solved simply by raising the gain.