Mark did reply on this subject and as usual he is cryptic in his response, mainly because he is trying to get you to think and learn, thats a good thing....
I can think of other reasons why he tends to be cryptic in his responses, when faced with an obvious design oversight. But maybe that's just me...
I will always encourage independent thought, but requiring beginners to think through the design of a supply filter is a tall order. It's not as simple as textbooks make you believe.
The curves and graphs and explanations in the catalog, may help you decide whether any ferrite bead designed for EMI suppression, might ever undergo "saturation" when subject to DC current flow {as in the second stage of the T2 power supply filter}. If saturation can occur, at what number of amp-turns does saturation become a problem for EMI suppression beads? Surely Laird customers would want to know, and surely Laird would provide this information.
When the core of an inductor saturates, the inductor basically turns into an air core inductor. So in case of the T2, you'll have, maybe, 100 nH of inductance. That won't provide any meaningful attenuation at 50 kHz, which is likely why those who have replaced the inductor with a piece of wire have found no difference (though, it doesn't appear they've measured the ultrasonic output to compare).
Ferrite cores are made by pressing or sintering together a bunch of ferromagnetic granules together to form the core. They're typically optimized for RF applications and are not intended for DC operation. To make a ferrite core suitable for DC current, you need to introduce an air gap (or at least non-ferromagnetic gap) in the core. This can be done by adding a non-ferromagnetic filler material to the ferrite powder. This will reduce the inductance you get per turn of wire in exchange for the ability to offer some inductance in the presence of a DC bias current.
As you've undoubtedly noticed, the Laird data sheet you link to is absent of DC current specs, including for the LFB095051 used in the T2. This indicates to me that the cores in the catalog, including the one used in the T2, are optimized for RF with minimal DC bias current.
Either way, the designer of the T2 could easily have measured the inductance with a DC bias applied. I would have expected this to have occurred during the prototyping stage of the T2.
Its also interesting that Wushuliu replaces it with a piece of wire and noticed no audible difference?
(sans LPS).
If it makes no difference, why include it? Why make the project more expensive than necessary? Sure, ferrites are cheap. But why? My expectation is that the parts on the circuit board are there to serve a function in the circuit - not to look pretty or to occupy space. Is this an unreasonable expectation? I'm not trying to be confrontational with this question. I'm just calibrating my expectations.
I think this goes back to subjective measurements that can be measured but their ultimate affect is somewhat in doubt, especially with our "human" ear transducers...
I'm not sure what you mean by "subjective measurements". Would you elaborate?
I agree that humans don't have to hear ultrasonic frequencies. That should be obvious from the definition of ultrasonic...
However, ultrasonic signals from the amp will dissipate power in the headphones. In speaker amps, amps which produce ultrasonic output are commonly known as tweeter eaters. Some will have that concern with their headphones as well. I doubt the ultrasonic output of the T2 is much of a concern, but it doesn't look pretty and shouldn't be there. The T2 designer appears to agree with this statement as he included a supply filter to get rid of the ultrasonic output. He just didn't seem to measure to see if the filter worked. He could have performed such a measurement with any sound card running at a sampling frequency of 120 kHz or higher (so 192 kHz would work!).
The neat thing is folks are looking at this T2 project and doing things to cahnge and or improve it..this is indeed the spririt of DIY...
It would be even better if the designer had ironed out these things (supply filter, turn-on/off thump) prior to release. I don't think it's reasonable to rely on your builders and "the spirit of DIY" to address your design oversights. But maybe I'm being unreasonable in my expectations (see above).
Tom
