The frequency response of stock input buffers is (more or less) flat. The three 332 resistors and the 1.5nF cap are a HF filter and an attenuator (level adjustment). Regarding sound quality Jan's single stage OPA 1632 solution without AC coupling caps is by far better than stock solution and the circuit proposed by AKM.
Wouldn't it be a better idea to obtain the 2.5V bias from the +15V analog supply, rather than the +5V digital supply? Just to keep the two rails (and their returns) separated from each other as much as possible.
Don't know how many of you have seen Gary Pimm's site about the DCX, but he recommends removing the 332 ohm resistor. There to reduce noise, he says.
Don't know how many of you have seen Gary Pimm's site about the DCX, but he recommends removing the 332 ohm resistor. There to reduce noise, he says.
1.Let's clear this "flat" forever.
It's not flat, I measured , but here's the sim. Identical as measured.
2.Yes, from +15V
3.Yes, that's not the option without rearranging input stage. But hf filter and attenuation is almost the same (except +6db shelving) with AK schematics I posted before.
It's not flat, I measured , but here's the sim. Identical as measured.
2.Yes, from +15V
3.Yes, that's not the option without rearranging input stage. But hf filter and attenuation is almost the same (except +6db shelving) with AK schematics I posted before.
I will repeat my measurements again, but Jan has same result like me. I was asking all the time where that +6db gone at the end of unit. If you skip this shelving filter (Jan version) you will again have a flat unit but "more natural". I didn't checked that but I believe Jan ;-) So, highs are "burned" somewhere, nobody knows where...
So, highs are "burned" somewhere, nobody knows where...
