Notes on component choices
diodes
1N4001 - 1N4007 all are the same except the voltage raiting
page1:
http://www.elfak.ni.ac.rs/phptest/new/html/Studije/predavanja-literatura/osnove/pdffiles/1N4001-D.pdf
or http://www.diodes.com/datasheets/ds28002.pdf
They all work. I don't recomand 1N4001. I usually buy 1N4004 or higher. I never take lower numbers in stock, don't see the point of them still existing
resistors: Use 1% metal film for precision, except R96, 98, 87, 89 R87/R98: 1/4W, 10%, closer value to calculated
All resistors except one 2k should be 1/4 watt. I have a lot of 1/2 watt resistors in stock so I will use them although I do not recommend it. They fit if you bend the wires very tight to the body of the resistor, 1/4Watt resistors are slightly easier to mount. Anyway the 10 Ohm resistor should be 1/4 watt. They also act as fuse and 1/2 watt doesn't give the protection needed. Order enough 20k (and 5K1) resistors, it is a good thing to match those on the little VCA boards as close as possible.
R4, R39 is ferrite bead or jumper: It's a great thing Igor implemented this use it when you are going to use a fast opamp with high GBW. You don't want RFI.
RS stocknumber 226-8191 Panasonic EXCELDR35C or RS stocknumber 226-8185 Panasonic EXCELSA35
Farnell Ordercode: 9526820 MURATA BL01RN1A1D2B
Farnell:Ordercode: 9265252 MULTICOMP MCAB 035090-33
For each VCA,
R6, R3, R100 on audio board can be different.
this compressor can be build with DBX202X, with 4xTHAT2181( A and C) or with 1xTHAT2181C.
http://www.groupdiy.com/index.php?topic=39459.msg501251#msg501251
_________________________________________________________________________________________________________________________________________
DBX202X has internal resistors on inputs 8x10K,the input resistance should be skewed
by 1k25. Control port impedance is 832 Ohm. I.e. R6=3k74, R3=4k99, R100=909R.
202XT should be the same.
4xTHAT2181 A/B/C on VCA replacement board, input resistors are 4x20k,
and control port impedance is same as 202x, 832 Ohm.
I.e. R6=jumper (0R), R3=4k99, R100=909R.
1xTHAT2181C on VCA replacement board, the resistors on input board are R6=0R (jumper), R3=20k, R100=909R.
It is possible to lower C1 down to 22pF.
_________________________________________________________________________________________________________________________________________
VCA's in the signal path: I'll go for 4 x THAT2180A. This is the most transparant and cleanest option.
Igor: 2181A+LME49710 combo should work great.
My personal opinion, 2181A don't worth additiional $3 per VCA, soundwise.
Neither it is tooooo sterile, no, it still has some VCA sound if you listen it
in hi-res monitoring setup, and the diffrence between perfectly adjusted
for minimum distortion A or B grade is almost negligible.
For me, quartet of 2181B after trimming sounds more "alive".
Anyway, it is just my opinion...
Paralleling VCA's does not lowers minimum possible distortion;
but, it increases the dynamic range, what is important IMHO.
Smooth transient response is matter of timing capacitor and PCB layout, as well.
I did not used tantalums here by this reason; one 0.47uf poly cap used for
all release settings except A, with a little rearranging the release switch if
you guys noticed. Time constants remains same, anyway.
Sidechain VCA's: Simplest and cheaper solution is THAT2180C, don't install SC_SYM trimpot and R33.
There's no need for A version VCA in sidechain.
the decoupling caps: should have a lead spacing from 2.5mm. I use 100nF
film capacitors for sidechain filters: regular, 5 or 5.08mm pitch, 50...63V . C5 and C7 the ones who need good quality.
c29 & C1/ C29 is not so critical I've put a 22pf ceramic, C1 I suggest a silver mica but 22pF could be OK. It really depends on your opamp. You can always try to lower the value as long you don't have any high frequency oscillation even above the audio range.
electrolytic caps: the ones under IC8 are 47 uF or non-polarized 10uF. Farnell Ordercode: 1236690 MULTICOMP NP100V106M8X11.5
opamp/ IC11. NE5532, OPA2604. I'll take OPA2604. TL074 and TL072 are not so critical, although they come in different flavours.
I use TL074B in the side chain. For the side chain filter I take the cheapest ones (TL074). IC 9 should be a low noise opamp too, I'll try LME49710 or AD797. All the noise injected in the VCA's here will directly be on the output of the VCA's.
parts in audio path: C1, C2, C3, C4, C5, C6, C7, R3, R4, R5, R6, R100, IC11 are critical parts.
http://be.farnell.com/panasonic/eca0jam471x/capacitor-470uf-6-3v/dp/8767041?Ntt=8767041
trimpots: you only need 4 bourns 5Ok multiturn
(RV1) on audio PCB: changes threshold point... install 5k1 and forget.
Threshold trim (TR_TRM): changes threshold zero. Omit (install nothing) and forget!!!
(Or use side screw, 50k, 10(25) turns trimpot to change 0db threshold point @ratio=4).
line recievers/drivers: THAT1246, INA137 or SSM2143 (-6db line receiver) is the right choice.
1 THAT1246 /INA137P IC1: some additional RFI protection look at page 4 and page 5 fig 4 of : http://thatcorp.com/datashts/THAT_1240-Series_Datasheet.pdf . As suggested do this directly on the connector. The IC is protected against +48V on the input pins, so no worry there.
Igor: I avoided capacitors in soundpath at every possible places, to get
minimum unwanted sonic artefacts, thanx to these chips.
And, another thing. Try to feed 1khz 5V P-P squarewave into GSSL's output balancer,
sum the + and - output with pair of 2k resistors and 100R trimpot and adjust for less possible signal;
than, see on scope what you get here.
As well, GSSL output balancer does not likes when - is connected to ground.
These are the reasons why I used "balancing" chips.
INA137, DRV134, THAT1646, THAT1246, can be used in MixBuzz500.
Beware of SSM2142, it is not best line driver. However, SSM2143 can be used at the input.
1 THAT1646 /DRV134P IC8: some protection against 48V and RFI is suggested in the date sheet. I would strongly recommand this modification for reliability:
look at page 9 of: http://www.thatcorp.com/datashts/THAT_1606-1646_Datasheet.pdf
Sonically and measurement-wise, THAT1246 and INA137 more or less same.
Maybe, THAT bit cleaner.
I go with the THAT's, the datasheets give a better S/N ratio.
switches/pots: Lorlin and Bourns
Farnell Ordercode: 1550189 1M33T1B5M1QE
connectors:8 of each: 10-PIN DUAL RAW CONNECTOR, MALE AND FEMALE
12V relays: Coil resistance should be more than 500 Ohm. 90% of relays are within this spec.
MULTICOMP HRS2H-S DC12V Farnell Ordercode: 9480218 has got coil resistance of 720 ohm and Contact Material:Silver, Contact Plating:Gold
Each slot consumption according to Lunchbox specs should be less than 80mA.
It is better to spend these milliamperes to OPA2604 and quad/eight (DBX202X) VCA.
I recommend to use relays with more than 500 Ohm coil resistance.
knobs: http://www.musikding.de/product_info.php/info/p2546_Knebelknopf-Alu-schwarz.html
. Well this can become better than what is API standard today. 
