This is my first post on this forum and I'm not certain if I should post this in the previous XS902 thread or start a new one. I'm not planning on following exactly the previous design, so I thought that a new thread would be in order.
Anyhow, I am currently building my own channel-strip with a mic pre, eq, comp and de-esser, I have one test version of preamp ready, working prototype of eq and now it is time for comp and de-esser. The 902 seems like a good place to start but I'm a bit confused about some parts of the design.
Have I understood correctly, that the two THAT2252's are used so that first one computes RMS from full-range signal, second only from filtered signal which is subtracted from the first, so the result is kind of RMS from signal difference? As 2252 has been discontinued, would AD736 be a suitable replacement? It would not be a drop-in replacement, but they are available - Farnell seems to supply those at somewhat reasonable price - and their output would be easy to scale up for VCA control with an op-amp.
I haven't yet simulated anything, but ic1b is clearly a low-pass filter. Is ic2a subtracting the low-passed signal from original, so that result is only the high frequency content? This way the HF-only de-essing would work by feeding the low-frq signal forward, past the VCA and feed the VCA only the top end which then gets attenuated when those RMS-detectors give the VCA sigal.
Would it be a good idea to use two THAT4315's instead of some other rms detectors and VCA? One of the VCA's would be left unused, but I think I can live with that tragedy. THAT2181 seems to be a bit twitchy about things around it while the 4315 seems to be a bit more relaxed.
Is there any specific reason I shouldn't connect the RMS-detector's outputs to EC+ and EC- of the VCA? To me it looks like it would simplify the design quite a bit and the downsides would be that the bypass would need to be done differently and metering-output would also need changes. The bypass I can do by doing a true bypass with a 2p2t switch and the metering would be pretty trivial to solve with op-amp and difference amplifier.
Only thing I don't seem to understand at all is the things around the bypass-switch. The bypass clearly forces the jfet open and prevents the VCA from getting any signal, but what does it do when operating normally? It is connected to the full-range signal RMS-detector and the op-amp is inverting, so it seems that it closes the jfet when there is any signal on input, but I don't understand why.
Currently my project looks like this: https://zouwi.1g.fi/kuvat/Luuri/IMG_20200826_061040~2.jpg?img=full I make my pcb's by milling with a small cnc. The component density on that eq got a bit excessive, but I didn't have any larger blank double-sided pcb's at hand. :-\
EDIT:
Here is a link to schematics for the XS902: https://groupdiy.com/index.php?topic=46453.0
Anyhow, I am currently building my own channel-strip with a mic pre, eq, comp and de-esser, I have one test version of preamp ready, working prototype of eq and now it is time for comp and de-esser. The 902 seems like a good place to start but I'm a bit confused about some parts of the design.
Have I understood correctly, that the two THAT2252's are used so that first one computes RMS from full-range signal, second only from filtered signal which is subtracted from the first, so the result is kind of RMS from signal difference? As 2252 has been discontinued, would AD736 be a suitable replacement? It would not be a drop-in replacement, but they are available - Farnell seems to supply those at somewhat reasonable price - and their output would be easy to scale up for VCA control with an op-amp.
I haven't yet simulated anything, but ic1b is clearly a low-pass filter. Is ic2a subtracting the low-passed signal from original, so that result is only the high frequency content? This way the HF-only de-essing would work by feeding the low-frq signal forward, past the VCA and feed the VCA only the top end which then gets attenuated when those RMS-detectors give the VCA sigal.
Would it be a good idea to use two THAT4315's instead of some other rms detectors and VCA? One of the VCA's would be left unused, but I think I can live with that tragedy. THAT2181 seems to be a bit twitchy about things around it while the 4315 seems to be a bit more relaxed.
Is there any specific reason I shouldn't connect the RMS-detector's outputs to EC+ and EC- of the VCA? To me it looks like it would simplify the design quite a bit and the downsides would be that the bypass would need to be done differently and metering-output would also need changes. The bypass I can do by doing a true bypass with a 2p2t switch and the metering would be pretty trivial to solve with op-amp and difference amplifier.
Only thing I don't seem to understand at all is the things around the bypass-switch. The bypass clearly forces the jfet open and prevents the VCA from getting any signal, but what does it do when operating normally? It is connected to the full-range signal RMS-detector and the op-amp is inverting, so it seems that it closes the jfet when there is any signal on input, but I don't understand why.
Currently my project looks like this: https://zouwi.1g.fi/kuvat/Luuri/IMG_20200826_061040~2.jpg?img=full I make my pcb's by milling with a small cnc. The component density on that eq got a bit excessive, but I didn't have any larger blank double-sided pcb's at hand. :-\
EDIT:
Here is a link to schematics for the XS902: https://groupdiy.com/index.php?topic=46453.0