Questions about how Boss CS-1 works

[gigimarga]

Hello all!

Before starting my story, here is the schematic of Boss CS-1 (there were 2 variants - mine is V2, with the red values):



In topic https://www.diystompboxes.com/smfforum/index.php?topic=118742.0, which I founded after I finished mine 

I've bought a Boss CS-1 from '80s at a very good price, but it didn't work: almost no compression and a big boost of signal. This looks very strange to me, because CS-1 is famous for its squishy sound (https://www.youtube.com/watch?v=ArBb0rTSwBM) and a huge volume drop (https://www.youtube.com/watch?v=wjYL8WA4vVg). After opening it, I saw that it has a couple of components damaged due to a battery drain, so I decided to replace them and to test other components. I found that the second optocoupler PH2 is dead, so I replaced it with a generic cheap Chinese VTL. The original optocouplers were HTV P873G35-380, which has a maximum on-resistance of 1K, a minimum off-resistance of 1M and a rise/fall time of 20ms (https://www.datasheetarchive.com/pdf/do ... P873G35380). My VTL seems to be closed in terms of on/off resistance, but I have no idea about its rise/fall time. After that, it worked almost OK: it compress the sound (but not so strong as in the videoclip aforementioned) and the volume drop were diminished.

Anyway, I didn't like it too much, so I decided to mod it to V1 specifications (the original values in the schematic posted above). Also, I replaced the first optocoupler PH1 with other cheap Chinese VTL, having similar specifications. Now, it acts stranger than before: it has a huge volume drop, almost the same insufficient compression and the sustain pot has only 2 states (when is at minimum the compression is completely off and after it sounds almost the same until maximum).

Because I don't want to damage it making other mods, I tried to understand the differences between V2 (red values) and V1 (original schematic) variants and to try to improve it a little. Because I don't have serious electronically knowledges, that's what I think:
1. the input cap was decreased from 100nF to 47nF, so the bass was lowered (I like it more with 47nF)
2. the biasing resistor of Q1 were increased from 220K to 470K, so the input impedance were increased (I have no idea if it's better or not)
3. the treble resistor (it has a normal/treble switch) were increased from 560 ohms to 1.2K (I like it more with 1.2K, because with 560 ohms it's too shrilling)
4. the output resistor were increased from 47K to 150K (I have no idea why, because I lowered it to 10K and it still has a huge volume drop)

My first question is which is the role of PH2 optocoupler (from input 2 of TA7136 IC to ground) Excepting PH2, it seems to be similar to EQD Warden/BJFE Pale Green Compressor or, even, DOD280. My guess is that will ground almost the all signal when its resistance goes to 0, leaving only some highs to pass (due to the 470nF cap), but I am almost sure that I'm wrong!

The main differences are in the envelope detector part (similar to Dynacomp and many other compressors), but my knowledges aren't sufficient to understand them:
1. the cap before Q3 were drastically lowered from 1uF to 47nF (maybe to keep only some mids and highs for envelope detection, but I have serious doubts about that)
2. the 220 ohms resistor from Q4/Q5 emitters to ground appears only in BJFE Pale Green (it's 110 ohms), otherwise it doesn't exists, so it's 0 (looking at other schematics, it seems to me that it controls the attack and it can be replaced with a 5-10K pot)
3. the cap between collectors of Q4/Q5 and ground was increased from 1uF to 4.7uF, so the attack time was increased (I'm almost sure of that...)
4. the 330K resistor between collectors of Q4/Q5 and ground was increased to 560K (maybe it increased the released time, but I have serious doubts about it)
5. the 4.7K resistor between LEDs of optocouplers PH1/PH2 to V+ was increased to 10K (maybe it decreased the maximum intensity of the LEDs, so it decreased the release time, but I have serious doubts about it)
6. the 5.1V Zenner after the optocouplers LEDS was totally eliminated (I have no idea why)
7. the resistors from Q4/Q5 bases are only 100K, while in many other schematics they are 1M (no idea why)
8. in many other schematics, there are 2 diodes from Q4/Q5 bases to ground (no idea why they are missing)
9. the 47nF capacitors between Q3 and Q4/Q5 seems to be normally, as in other schematics there are from 10nF to 100nF (no idea what are for...maybe a high pass filter with 100K resistors)

So, I need a patient person to clarify me these doubts and to tell me how I can improve it to have the squishy sound from the videoclip (I like it a lot...I'm not a fan of transparent compressors) and more volume. Maybe the secret is in the optocouplers? I read the VTLs and NSL32 specifications, but I didn't find anything similar in terms of rise/fall time.

Thank you a lot!!!