4xTHAT2181 VCA board for GSSL etc builds, DBX 202X replacement

ptron said:

2- in the calibration for the GSSL using only 1x2181A i had to go from leg 4 of VCA to 680K in series with 50k trimpot (symmetry)
do i have to change this 680k for 160k (like the "VCA notes Rogy" says)

Click to expand...

From THAT2181 datasheet you want a symetry control voltage of +/-0.5mV for the A-grade type.
From same datasheet there is an internal 25 ohms resistor between VCAs Ec+/Sym pins2/4 (contrary to THAT design notes DN124 or DN137, where this resistor is 27 ohms.) Assume it being a 25R to play safe.
You have 4 VCAs with these pins 2/4 in parallel, so these 25Rs are in parallel as well, resulting in 1/(4/25R)=6.25R.
With this R13 from Rogys schematic in series you build a voltage divider, so a voltage in range between +/-0.5mV for the A-grade THAT2181 can be applied to VCAs pin 4. Value of this R13 in question with a +/-15V supply is 15V/0.0005V*6.25R-6.25R=187K493.75. Closest E96 parts value for max.usable sym.trimmer rotation would be a 187K. (61K9 for B-grade or 37K4 for C-grade 2181s). Rogys 160K allows for +/-0.58mV.

Thanks a lot!! thats clear now!  ;D
About the cap and resistors u mentioned in previous post am i correct asuming they the ones i enclosed from image 1 ?
Thanks a million!

yepp.

yallaa!! to get dirty now then

Ok one last thing.. In the bom says i need 4x ceramic 47nF - 220 uF. Is there a prefered value or a difference in performance or stability between those values? I got tons of 1uF ceramics
Thanks again!

ptron said:

the bom

Click to expand...

whenever I read this word, my peristalsis starts working backwards.
You could learn so much by simply ignoring these partly wrong or outdated lists and look up the schematic and the datasheets for the parts in question.

says i need 4x ceramic 47nF - 220 uF. Is there a prefered value or a difference in performance or stability between those values?

Click to expand...

(lovely little typo) sure this would make a difference. I doubt, you can source a 220uF ceramic cap and if you could, your PCB couldn't fit a single one -let alone 4- and your 19" case will need to grow some more HE.
For usual 100nF should work for CBP20,21,22,23. You might leave out the one connecting to the VCAs negative current source pins 5. (cost per cap in single quantities here is less than 5 ct).
If you don't use a NE5534, make sure you leave out the 22pF between opamps pins 5/8.

Harpo! Sharp eye!!!!
All thumbs up. And thanx for your support.
Yes, this is my mistake, should read 47nF-220nF.
I buy 0.1uF 50V X7R 5000 package once a half year, from local Vishay distributer.
The value of bypass capacitors for our simple audio circuits is absolutely fine in range 47nF - 220nF.
In mixed stuff with analog/digital/RF, usually lower value COG goes in parallel with higher value X7R;
there are some other tricks which aren't very useful in audio stuff; woodoo and debris!

BTW, for years, I was used to write mF instead of uF (Russian issue)
Took me long time to use uF automatically, however, I am human.

as usual, harpo, u da man... and igor, thanks so much for your quick reply! i do read the datasheets and try a lot hehe but when it comes to capacitance (or to be honest knowing how to interpret anything) i go blank. need to keep reading and watching video classes ... will get it sooner or later tho thanks so much for your help

Igor, my dislike of boms wasn't directed at yours (whatever its content might be if it so exists). The only use of boms from my POV might be for rare birds or build-specific required parts such as non-linear taper pots in order to come closer to a given frontpanel labeling/scaling. YMMV.
Looking at catalogues or online stores when searching for ceramic caps would immediatly show, this uF parts value has to be a typo. 100nF seems about the highest common available value for this type of dielectric. For the 2 audio-VCAs in this build, you would need 17,600 of these 100nF multilayer cercaps in a paralleled arrangement to substitute for the 8x220uF caps. Might look impressive, but probably would only be a contraproductive, expensive, microphonic, big antenna, picking up all possible garbage.

Talking about using ceramics in antenna's, aye!
Agree with you, man. +1 BOM hater here.
I first looking at the schematic.
Matched parts and log/revlog pots, transistor's beta range, dissipating power, capacitor's dielectric type, etc,
are quiet easy to figure out from schemo+plain logic. I like to read your posts, and hat off,
I cannot explain it as simple, logical, at good language, as you do.

well i installed the vca`s at the moment they are working amazing! its the third unit i build, this time for a friend`s studio. it all went down well no problems except i got 1 A meter instead of .001A hehe ...
i trimmed the symmetry on the vca`s (distortion trimmer on GSSL) and with a sine at +/-500 hz it had much less distortion than my other unit with pretrimmed single 2180A wohoo! (i didnt install second (r magic) trimmer as i thought im not sure i can even measure the difference, but i may still try it on the next unit i upgrade!)
is it same procedure as i did with the symmetry one? (only with a spectrum analizer.)

now i got another question... for this test i left the threshold in +20 / no compression, and also tested again with the vca in bypass condition. both tests went down the same, they translated exactly the same for the most part.. sorry if i cant explain myself i dont know exact right terms.

anyways, i kept measuring when threshold was going down and it started compressing... on the new VCA unit a couple of frequencies started to rise and then receeded; leaving just one again when threshold went down to -20db. i didnt measure them, but there were tons more of those little distortion frequencies also on the pre trimmed 2180A unit when i performed the same test...

is this normal?

ill post pics tomorow for sure hehe the lil black pcbs look great there and sound real good!

apart from that im not sure i can notice much difference in the noise floor with my converters, tomorow ill try testing with other one.

thanks a lot!!

ptron said:

is it same procedure as i did with the symmetry one? (only with a spectrum analizer.)

Click to expand...

no. see reply#27 from previous page of this thread.

ah thanks! so then... Max Ec for the VCA would mean i turn up the makeup all the way? no threshold or compression tho?

Igor said:

Schematics and info about some VCA's - here:

http://twin-x.com/groupdiy/displayimage.php?album=416&pos=6

Click to expand...

Hello, how can I access to these schematics, can't find where to register on the indicated page..?

Thanks

Same here. There's a chance to see a schematic of it?
Thanks

I guess it is something like this circuit from Douglas Self :

saint gillis said:

I guess it is something like this circuit from Douglas Self :

Click to expand...

thanks.
Have you ever tried to make one and test into gssl?

I've made a GSSL like that (see the pic), Gyraf's schematic being the reference : http://www.gyraf.dk/gy_pd/ssl/ssl_sch.gif

With some THAT 2181A.
For the distortion trim I did this (see pic)

This was some time ago, but IIRC the idea is that if you use only one that VCA, you put the jumper in JP2,
If you use two that VCAs, you put the jumper in JP1,
And if you use 4 that VCAs, you put the jumper in JP3.


The idea when paralleling VCAs is to reduce noise by 3 dBs for each doubling, when I did this compressor I only put 1 VCA for each side and I can barely hear a difference with the serpent audio version with Igor's boards...

Using 4 VCAs on each side will reduce the noise by 6 dBS, I guess it can make sense for critical application, but it will also add 60€ to your bill..

But you can use sockets and jumpers, start with one VCA for each side, and you can add 1 or 3 more latter if you want and move the jumpers correspondingly.