Login / Register
 
How does a VCA compressor work?
New Reply
Subscribe
Everland Studios
Thread Starter
#1
26th June 2013
Old 26th June 2013
  #1
Gear nut
 
Everland Studios's Avatar
 
Joined: Apr 2013
Location: Sydney Australia
Posts: 88

Thread Starter
Everland Studios is offline
How does a VCA compressor work?

Hi All,

I'm looking at doing my first DIY pro-audio project soon. I'm going to build one of the Gyraf-based G-SSL clone kits but I'd like to use the project as a way to learn a little more about the way these types of circuits work rather than just following the instructions but not knowing what each component actually does.

So, is there a good resource online to learn a bit about how different circuits related to pro-audio work? For example a VCA compressor? It would need to be reasonably basic as I'd be learning it from the ground up.

I don't know, maybe it's not possible without an electrical engineering degree...?
#2
26th June 2013
Old 26th June 2013
  #2
Lives for gear
 
Richard Crowley's Avatar
 
Joined: Sep 2006
Location: Portland OR USA
Posts: 3,211

Richard Crowley is online now
Do you understand what a VCA is? Do you know what a compressor does?
The articles in Wikipedia aren't bad for a general introduction to these kinds of topics.
Variable-gain amplifier - Wikipedia, the free encyclopedia
Dynamic range compression - Wikipedia, the free encyclopedia

Do you understand how to read electronic schematic diagrams?
I think it would great fun and instructive to take a schematic diagram for a device like a compressor and analyze it stage-by-stage, and even component-by-component here in the Geekslutz forum. You aren't the first person to come through here asking a question like this.
Everland Studios
Thread Starter
#3
26th June 2013
Old 26th June 2013
  #3
Gear nut
 
Everland Studios's Avatar
 
Joined: Apr 2013
Location: Sydney Australia
Posts: 88

Thread Starter
Everland Studios is offline
Hi Richard,
Thanks for your reply. I did read that article on Wikipedia about variable gain amplifiers and I can get my head around that. I don't really know how to read a schematic diagram though, so I guess that would be a good place to start. The schematic is available for the Gyraf VCA Compressor: http://www.gyraf.dk/gy_pd/ssl/ssl_sch.gif

It would be great, as you say, to analyse it on here if anyone is up for that...
#4
26th June 2013
Old 26th June 2013
  #4
Lives for gear
 
Richard Crowley's Avatar
 
Joined: Sep 2006
Location: Portland OR USA
Posts: 3,211

Richard Crowley is online now
OK, lets start with an overview of that schematic diagram. There are five horizontal sections here, generally with the input at the left, and the output at the right. And there is an additional 6th section in that box in the lower right corner.

The top two sections are the essentially-identical audio paths for the left channel and the right channel.

The third section is the first half of the control signal chain where the left and right audio signals are sampled to produce a DC voltage representative of the loudness of the audio channels. And the fourth section shows how that DC signal is modified to allow user-control of attack and release times and shows how it drives the meter that indicates what it is doing.

And the section along the bottom shows the power supply. The box in the lower-right corner shows the "equivalent circuit" inside the "black boxes" labeled "DBX 202".

Do you understand what resistors and capacitors are and what they do? Can you identify the symbols on the schematic diagram that show capacitors (polar and non-polar), resistors (fixed and variable) and other components?
Everland Studios
Thread Starter
#5
26th June 2013
Old 26th June 2013
  #5
Gear nut
 
Everland Studios's Avatar
 
Joined: Apr 2013
Location: Sydney Australia
Posts: 88

Thread Starter
Everland Studios is offline
Ok, cool, I can see those 5 sections.

Just reading up on what the various symbols mean here: https://en.wikipedia.org/wiki/Electronic_symbol
I think the polar capacitors are represented slightly differently on that schematic. Is that a polar capacitor pretty much the first component after the audio in? (it seems to be in parallel with a non-polar capacitor)

I have some knowledge of what resistors do. i.e. they provide resistance to electrical current thereby reducing it. The higher the value of the resistor the more it reduces the electrical current.

I've always found it harder to get my head around capacitors but from reading up on Wikipedia they basically store an electrical charge. Again they have different values that correspond to how much charge they can store. I don't really get what causes them to charge and discharge though.

Thanks for doing this by the way, it's really great!
#6
26th June 2013
Old 26th June 2013
  #6
Lives for gear
 
Joined: Sep 2004
Location: UK
Posts: 6,766

Matt Syson is offline
Hi
For detailed explanation of VCAs try looking at the THAT corporation website as it has various technical papers. If you are struggling a bit with capacitors you may need to skip a few sections but most of what you would ever need to know is there.
Matt S
__________________
Matt S
www.mseaudio.co.uk
#7
27th June 2013
Old 27th June 2013
  #7
Lives for gear
 
Geoff_T's Avatar
 
Joined: Jul 2003
Location: Tujunga
Posts: 4,388

Geoff_T is offline
Quote:
Originally Posted by Everland Studios View Post
Ok, cool, I can see those 5 sections.

I've always found it harder to get my head around capacitors but from reading up on Wikipedia they basically store an electrical charge. Again they have different values that correspond to how much charge they can store. I don't really get what causes them to charge and discharge though.

Thanks for doing this by the way, it's really great!
Hi

The other important function of a capacitor is that, if in series with an ac signal, it blocks dc, so the signal passes through the capacitor but the dc stops there.

The lower the frequency of the signal (and the load value the capacitor is connected to...) the larger will be the value of the capacitor.

__________________
Geoff Tanner
Aurora Audio International

See us on Facebook

http://www.facebook.com/auroraaudio
http://www.auroraaudio.net/
http://www.soundonsound.com/sos/may1...off-tanner.htm
http://www.amazon.com/Window-Past-Ge...8737082&sr=1-9
http://www.grandmasterrecorders.com

For quicker responses, please use my email (Geoff at auroraaudio.net) in preference to pm's and visitor messages!
#8
27th June 2013
Old 27th June 2013
  #8
Lives for gear
 
JohnRoberts's Avatar
 
Joined: Nov 2006
Location: Hickory, MS
Posts: 3,024

JohnRoberts is offline
Quote:
Originally Posted by Everland Studios View Post
I've always found it harder to get my head around capacitors but from reading up on Wikipedia they basically store an electrical charge. Again they have different values that correspond to how much charge they can store. I don't really get what causes them to charge and discharge though.

Thanks for doing this by the way, it's really great!
This is not a perfect analogy but imagine a capacitor is like two round plates sitting inside hollow tube and you can push or pull water (charge) into or out of either end, into the space between the plates. Putting in water that doesn't come out pushes the plates further apart, this build up of charge is equivalent to voltage in a charged cap.

A cap can be used like an electronic battery in a power supply where charge is temporarily put in and then taken out again a little later. Caps can also be used in-line with an audio path, where the cap can be partially filled with charge to provide a fixed distance (DC offset voltage) between both ends, while the short term in and out charge changes at one end, shows up identically at the other end, just spaced apart by the constant fixed charge difference (DC). A final application is where the rate of change for charge injected results in a rate of change for the space between the plates (voltage) this rate dependent relationship is used to make frequency responsive filters to preferentially boost or cut different frequencies. .

if this doesn't make any sense ignore it, analogies are never perfect.

JR
__________________
John Roberts
www.CircularScience.com
#9
27th June 2013
Old 27th June 2013
  #9
Lives for gear
 
Richard Crowley's Avatar
 
Joined: Sep 2006
Location: Portland OR USA
Posts: 3,211

Richard Crowley is online now
Everland Studios
Thread Starter
#10
28th June 2013
Old 28th June 2013
  #10
Gear nut
 
Everland Studios's Avatar
 
Joined: Apr 2013
Location: Sydney Australia
Posts: 88

Thread Starter
Everland Studios is offline
Thanks guys, that helps a lot with my understanding of capacitors. I'm hoping I'll start to understand these concepts a bit better as I work through a few circuits that employ them. i've been watching a few videos on youtube that are helping my understanding of components and circuits too.
Everland Studios
Thread Starter
#11
28th June 2013
Old 28th June 2013
  #11
Gear nut
 
Everland Studios's Avatar
 
Joined: Apr 2013
Location: Sydney Australia
Posts: 88

Thread Starter
Everland Studios is offline
Smile

Quote:
Originally Posted by Richard Crowley View Post
I have added component reference numbers to the Gyraf schematic diagram so we can talk about component-level details.
Schematic with references: http://www.rcrowley.com/images/SSL-COMP.png
Oh, great! Where do we start?
#12
28th June 2013
Old 28th June 2013
  #12
Lives for gear
 
Richard Crowley's Avatar
 
Joined: Sep 2006
Location: Portland OR USA
Posts: 3,211

Richard Crowley is online now
At a bit deeper level...

U1 provides a differential/balanced input for the audio path.
M1 is the "magic" VCA circuit that does the actual gain control
U2/U3 provide a differential/balanced output for the audio path.
That is essentially the entire audio path.
Of course, there are two of these one for Left channel and one for Right.

A sample of the Left channel and Right channel audio is grabbed right after C9 where it says "TO SIDECHAIN"
The Left and Right audio channels are mixed together by R17/R18 as this unit operates on both channels together.
U4 is another VCA which modifies the sampled audio in the same manner that M1 does in the main audio path.
After U4, U5 is an audio output/buffer amplifier stage which feeds the "Sidechain Filter" signal out through that loop.
Then back in the loop, through DC-blocking capacitor C21, and into a full-wave rectifier (U6, U7, D1-D6).
SW1 modifies some of the operating parameters of the full-wave rectifier section to change the compression ratio.

SW2 controls the compression action on/off. It is apparently labeled "BYPASS", but it is not what I
would call a true "bypass" (where the audio is shunted completely around the circuit).
Then SW3 controls the compression attack time by putting various resistances in the control signal path.
The lower value of the resistance, the more quickly the DC Control Voltage can charge or discharge those capacitors.
Then SW4 controls the release time by selecting different R/C combinations (R44-R50 and C22-C27)

U8 is a buffer amplifier for the DC gain Control Voltage and feeds the next stage (U9) as well as the front-panel meter.
U9 combines the gain Control Voltage (from U8) and adds in a user-selected amount of "MAKEUP gain" (RV2).
Then U9 feeds sends the Control Voltage (CV) back to the VCAs in the Left and Right audio paths. (where it says "CV to VCA's)

U10 gets a sample of the Control Voltage (through R56) and combines it with a user-controlled "THRESHOLD" voltage (RV3).
Then that (different) Control Voltage gets applied to the sidechain VCA (U4).

The power supply is a pretty bog-standard conventional linear circuit, not much remarkable there.
It does feature several outputs, though. There is the main +/- 15V for the audio circuits.
And then there is +/- 12V for the sidechain (the section that creates the Control Voltage).
Plus another, isolated +12V for "crude" loads like the meter lamp and optional relay.
Everland Studios
Thread Starter
#13
28th June 2013
Old 28th June 2013
  #13
Gear nut
 
Everland Studios's Avatar
 
Joined: Apr 2013
Location: Sydney Australia
Posts: 88

Thread Starter
Everland Studios is offline
Wow, that is a tonne of concepts to get my head around but after quite a few wikepedia lookups and tracing the signal around the schematic many times I think I can say I understand that all in a 'broad strokes' kinda way.

A few initial questions:

When the audio signal first enters the circuit, what is the function of those two capacitors in parallel (one polarised, one not)?

Am I right in saying that U1 takes the balanced input signal and converts it into an unbalanced signal to feed M1, and then U2/U3 convert that signal back into a balanced signal?

Is C9 being used in any of the ways that John Roberts talks about above?

What is the function of RV1?

Actually, I just realised I have about a thousand questions but I'll limit it to those for now!

Cheers,
Ben
#14
28th June 2013
Old 28th June 2013
  #14
Lives for gear
 
Richard Crowley's Avatar
 
Joined: Sep 2006
Location: Portland OR USA
Posts: 3,211

Richard Crowley is online now
Quote:
Originally Posted by Everland Studios View Post
When the audio signal first enters the circuit, what is the function of those two capacitors in parallel (one polarised, one not)?
The two large electrolytic capactors, C1 and C2 block any DC component that may be present in the source signal. Electrolytic capacitors are used for space efficiency, but they are not the most perfectly linear kind of component we would wish to use. So smaller-value capacitors (typically metal film, etc.) like C3 and C4 are used to "bypass" the larger capacitors to improve high frequency performance and linearity.

Quote:
Am I right in saying that U1 takes the balanced input signal and converts it into an unbalanced signal to feed M1, and then U2/U3 convert that signal back into a balanced signal?
That is exactly correct. The VCA is a single-ended in and out device.

Quote:
Is C9 being used in any of the ways that John Roberts talks about above?
Yes. That is typically how a "coupling capacitor" works. C1/C2 and C13/C14 and C21 are also coupling or "DC blocking" capacitors.

Quote:
What is the function of RV1?
RV1 is used to adjust the exact operating conditions of the VCA to "tune" it for minimum distortion. It is typically something is set during manufacturing and testing and typically doesn't need further adjustment. It optimizes the operating point of that particular IC. If you replaced the IC, it would require "re-tuning".
#15
28th June 2013
Old 28th June 2013
  #15
Lives for gear
 
Mr. Lau's Avatar
 
Joined: Mar 2013
Location: Ecuador
Posts: 1,580
My Recordings/Credits

Mr. Lau is offline
Everland, why don't you get a electronics design software like PSpice (there should be more options, I used it at college, 15+ years ago) and learn on it?

You'll have the components, (dunno if vca's), and signal generator, multimeter, oscilloscope, so you can simulate and see how the input signal on one part of the circuit changes at the output, etc.

Without blowing reversed electrolitic capacitors like me years ago!!!
__________________
Come on baby! Ride my fader!

Last edited by Mr. Lau; 28th June 2013 at 10:27 AM.. Reason: Wow I realized '15+ years ago' is 20 years ago lol
#16
28th June 2013
Old 28th June 2013
  #16
Lives for gear
 
Mr. Lau's Avatar
 
Joined: Mar 2013
Location: Ecuador
Posts: 1,580
My Recordings/Credits

Mr. Lau is offline
Quote:
Originally Posted by Everland Studios View Post
I've always found it harder to get my head around capacitors but from reading up on Wikipedia they basically store an electrical charge. Again they have different values that correspond to how much charge they can store. I don't really get what causes them to charge and discharge though.
What makes them to charge and discharge is AC.

If you feed DC it will block it.

But really, apart from the fact English is not my first language, so I won't go deeper since other members can explain better, IMHO some math and physics of electronics will help you lots.

Start building or simulating basic circuits, like rectifiers with diodes, filters with R C (and L), amplifiers with transistors, then op amps. Those are the building blocks.

LOL the first piece of gear I built when I was at high school was a distortion unit with diodes, and the next, after some years, was a mixer with op amps!

Next I made a power amp with sanyo IC's, 100 W, and some filters.

I used what I learned from filters to make a passive eq, and combining the concepts of filters and amps, an active eq.

Combining the above concepts, I made a guitar amp, with better sound when distorting it (well it sounded like crap, but at that time I was so proud of it ).

Enter the 555 IC. I made a metronome, with leds and sound. Using everything I knew so far, I made a 'synth', with only a variable resistor to change the frequency and filters and added a LFO, then ADSR, and finally transformer and tube.

I was at college at the time, so I did other stuff with TTL IC's, EPROM's, and made a better synth, but to use sequences, I had to program those memory chips. Record sequence, erase chip, program another, record (with a cassette 4 track portastudio) and so on.

Well, I finally built a MIDI interface for my joystick port using opto couplers and realized my previous synths were crappy, so I donated them to college

Oh no, today they would be considered vintage gear!
#17
28th June 2013
Old 28th June 2013
  #17
Gear addict
 
Joined: Mar 2013
Posts: 362

Kesh is offline
Quote:
Originally Posted by Richard Crowley View Post
I have added component reference numbers to the Gyraf schematic diagram so we can talk about component-level details.
Schematic with references: http://www.rcrowley.com/images/SSL-COMP.png
I'm trying to get my head around the sidechain passing through a VCA of its own. I'm used to the (rectified and smoothed) sidechain controlling a VCA in the audio path.
#18
28th June 2013
Old 28th June 2013
  #18
Lives for gear
 
Mr. Lau's Avatar
 
Joined: Mar 2013
Location: Ecuador
Posts: 1,580
My Recordings/Credits

Mr. Lau is offline
Quote:
Originally Posted by Kesh View Post
I'm trying to get my head around the sidechain passing through a VCA of its own. I'm used to the (rectified and smoothed) sidechain controlling a VCA in the audio path.
That is Feedforward topology

From the schematic and Richard's explanation, the VCA in the sidechain is doing a feedback compressor topology thing. It tracks the main VCA GR.

So the circuit is combining feedforward and feedback topologies. It really acts like a feedback compressor

Last edited by Mr. Lau; 29th June 2013 at 03:15 AM.. Reason: I called Mr. Crowley Robert instead of Richard :)
#19
28th June 2013
Old 28th June 2013
  #19
Village Idiot
 
Labs's Avatar
 
Joined: Apr 2004
Location: Denmark
Posts: 1,714

Labs is offline
Great info here!

Gustav
#20
28th June 2013
Old 28th June 2013
  #20
Lives for gear
 
Mr. Lau's Avatar
 
Joined: Mar 2013
Location: Ecuador
Posts: 1,580
My Recordings/Credits

Mr. Lau is offline
Ok I have a question now.... Why 15 and 12 V?
#21
28th June 2013
Old 28th June 2013
  #21
Lives for gear
 
Richard Crowley's Avatar
 
Joined: Sep 2006
Location: Portland OR USA
Posts: 3,211

Richard Crowley is online now
Quote:
Originally Posted by Mr. Lau View Post
Ok I have a question now.... Why 15 and 12 V?
I presume that ±15V is used for the audio path because it gives you better dynamic range. Likely allows +4 or even +10dB output levels.

The Control Voltage "sidechain" circuit doesn't need that kind of dynamic range. I haven't looked at the VCA specs, but I assume that the Control Voltage may not even reach 10V. So ±12V is likely more than adequate.
#22
28th June 2013
Old 28th June 2013
  #22
Lives for gear
 
Richard Crowley's Avatar
 
Joined: Sep 2006
Location: Portland OR USA
Posts: 3,211

Richard Crowley is online now
Quote:
Originally Posted by Kesh View Post
I'm trying to get my head around the sidechain passing through a VCA of its own. I'm used to the (rectified and smoothed) sidechain controlling a VCA in the audio path.
If you only sample the "raw" (unprocessed") audio from C9 (left and right), you don't know what effect the level control (VCA) is having on the signal, so you are running "open loop". Like driving blind. So you can either sample the audio AFTER the main audio path VCA (e.g. at M1 OUT), or else you can sample the audio BEFORE the VCA as they have done. But if you sample the audio level BEFORE the control element (M1 VCA), then you must use another VCA to "simulate" what you are doing to the main audio path.

However, note that they have ANOTHER use for the sidechain VCA (U4), which is adjusting the static gain of the sidechain or "THRESHOLD" for the limiting function. So by using that "extra" VCA (U4), they combined both functions into a single chip. I would call that clever. At least that is my speculation on how they made the design trade-off decisions.

Note that the circuit creates two different Control Voltages ("CV"). The first one takes the basic DC CV from U8 and adds the "MAKEUP" gain, and that is used to control the VCAs in the main audio path. Then the OTHER CV is created again from the basic DC CV from U8, but adds the static sidechain gain ("THRESHOLD") and applies that CV to the sidechain VCA.
#23
28th June 2013
Old 28th June 2013
  #23
Lives for gear
 
Richard Crowley's Avatar
 
Joined: Sep 2006
Location: Portland OR USA
Posts: 3,211

Richard Crowley is online now
Quote:
Originally Posted by Mr. Lau View Post
Everland, why don't you get a electronics design software like PSpice ... and learn on it?

You'll have the components, (dunno if vca's), and signal generator, multimeter, oscilloscope, so you can simulate and see how the input signal on one part of the circuit changes at the output, etc.
That may be a good scheme at some point, but I believe you may be correct that typical simulation programs don't have specialized circuits, functions, chips like the VCAs. And it gets pretty tricky to get everything set up properly to study the "broad strokes" as Ben says. Its almost to the point that if you have the skill to get everything set up right in spice, you already understand more than Ben is asking for.
#24
28th June 2013
Old 28th June 2013
  #24
Lives for gear
 
Joined: Feb 2006
Posts: 570

mixedup is offline
I'd just like to say thanks for the rich seam of information on here. I've built one of these GSSLs without ever really understanding everything fully, and I do prefer to get my head around things at a 'strategic' level before drilling down. I did try reading some of Douglas Self's books and started the MITx electronics course, but so much went way over my head, and there was too much stuff too early on that I just didn't need to know to get into a DIY build. In short, I sincerely wish I'd read Richard Crowley's explanation before I started!

I realise that this would take a hell of a lot of people's time, but it would be really cool to get a series of different processor schematics together and do the same sort of high-level breakdown. I assume a vari-mu is kinda the same in principle, but just using tubes in lieu of the VCA? I'd love to get the same drill down on some filter designs.
Everland Studios
Thread Starter
#25
29th June 2013
Old 29th June 2013
  #25
Gear nut
 
Everland Studios's Avatar
 
Joined: Apr 2013
Location: Sydney Australia
Posts: 88

Thread Starter
Everland Studios is offline
Thanks Mr. Lau, I took your advice and I'm going to build a couple of basic kit amplifiers etc. first before I do the G-SSL. For a start it will get my soldering skills back up to scratch on something a bit less critical.
Everland Studios
Thread Starter
#26
29th June 2013
Old 29th June 2013
  #26
Gear nut
 
Everland Studios's Avatar
 
Joined: Apr 2013
Location: Sydney Australia
Posts: 88

Thread Starter
Everland Studios is offline
Quote:
Originally Posted by Richard Crowley View Post
The two large electrolytic capactors, C1 and C2 block any DC component that may be present in the source signal. Electrolytic capacitors are used for space efficiency, but they are not the most perfectly linear kind of component we would wish to use. So smaller-value capacitors (typically metal film, etc.) like C3 and C4 are used to "bypass" the larger capacitors to improve high frequency performance and linearity.

Yes. That is typically how a "coupling capacitor" works. C1/C2 and C13/C14 and C21 are also coupling or "DC blocking" capacitors.
So, where exactly does this DC come from? I see that C9 is right after the first op-amps and then C13-16 after U2 and U3 so I'm presuming that an unwanted effect of op-amps could be to introduce DC into the output?

From the reading I've been doing on op-amps I've learnt that the resistor between the output and the inverting input basically sets the gain of the op-amp but I notice that here each also has a cap in parallel with this resistor. What would be it's function?

Quote:
Originally Posted by Richard Crowley View Post
But if you sample the audio level BEFORE the control element (M1 VCA), then you must use another VCA to "simulate" what you are doing to the main audio path.
Why would they choose to use a different VCA for this simulation than the one used in the main audio path?

From your original explanation I didn't quite understand this bit:

Quote:
Originally Posted by Richard Crowley View Post
After U4, U5 is an audio output/buffer amplifier stage which feeds the "Sidechain Filter" signal out through that loop.
Then back in the loop, through DC-blocking capacitor C21, and into a full-wave rectifier
Where is this loop that you refer to? I can see those 2 slashes through the path where it says Sidechain Filter after U5. What do they mean on a schematic?
#27
29th June 2013
Old 29th June 2013
  #27
Lives for gear
 
Richard Crowley's Avatar
 
Joined: Sep 2006
Location: Portland OR USA
Posts: 3,211

Richard Crowley is online now
Quote:
Originally Posted by Everland Studios View Post
So, where exactly does this DC come from?
There may or may not be DC on the audio input signal. The input blocking/coupling capacitors are placed there to protect the circuit from DC coming in.

Quote:
I see that C9 is right after the first op-amps and then C13-16 after U2 and U3 so I'm presuming that an unwanted effect of op-amps could be to introduce DC into the output?
Not exactly. There may be some DC introduced by the opamps (or more likely a side-effect of resistors that are not perfectly matched, etc.) However, IMHO the PRIMARY reason for AC coupling the input and output is to PROTECT the circuit from DC that may be present at the input or the output.

Quote:
From the reading I've been doing on op-amps I've learnt that the resistor between the output and the inverting input basically sets the gain of the op-amp
Yes. For example U1 is running at "unity gain" (+1) because R1=R4 and R2=R3. In fact there is little or no gain anywhere in this device. It is designed for line-level in and line-level out, so no gain is required.

Quote:
but I notice that here each also has a cap in parallel with this resistor. What would be it's function?
R4 is the primary feedback path from the output of U1 back to the inverting input of U1. But that small capacitor (C7) sends a bit more high-frequency signal from U1 output back to U1 inverting input. That causes the U1 stage to somewhat roll-off the high frequency response. Download the data sheet for all the ICs and you can see discussions of how to arrange the circuit around the chips to achieve different effects.


Quote:
Why would they choose to use a different VCA for this simulation than the one used in the main audio path?
The main audio path should be kept as "clean" as the budget allows. That means using a more complex module for VCA function to achieve lower distortion and better performance in several areas.

That "equivalent circuit" in the box in the lower right corner shows how they use a pair of modern ICs (5534 op-amp and THAT 2181 VCA) to replace the original "magic" potted, proprietary module (DBX202). That circuit inside the box is actually what is shown in the main audio path as a box called "M1".

OTOH, distortion in the sidechain is of almost no consequence at all. Because the audio is just getting turned into a DC voltage anyway. So they just use a "raw" THAT2181 chip. Note that it doesn't even have the "DISTortion NULL" adjustment you saw in the main audio path.

Quote:
From your original explanation I didn't quite understand this bit: [sidechain filter loop] Where is this loop that you refer to? I can see those 2 slashes through the path where it says Sidechain Filter after U5. What do they mean on a schematic?
That means that the audio from U5 goes out to some sort of external connection and then back in. It is exactly the same function as the "effects loop" on most audio mixer desks. It allows you to use some external equipment to modify the signal before feeding it back in on its merry way. I do't know what kind of connector the original equipment used. If you are building a standalone device, you could use a pair of 1/4 inch phone jacks, but remember that you need to keep a patch cord in there to complete the path when not routing it to external equipment. Or you could use a "normalizing" connector which automatically connects the loop through when nothing is plugged in (like typical mixers use).

Or, if you don't anticipate ever using that sidechain filter loop, you can just ignore it and connect the output of U5 directly to C21.
#28
29th June 2013
Old 29th June 2013
  #28
Lives for gear
 
Mr. Lau's Avatar
 
Joined: Mar 2013
Location: Ecuador
Posts: 1,580
My Recordings/Credits

Mr. Lau is offline
Quote:
Originally Posted by Richard Crowley View Post
I presume that ±15V is used for the audio path because it gives you better dynamic range. Likely allows +4 or even +10dB output levels.

The Control Voltage "sidechain" circuit doesn't need that kind of dynamic range. I haven't looked at the VCA specs, but I assume that the Control Voltage may not even reach 10V. So ±12V is likely more than adequate.
I think I get it, Richard...

So fluctuations of +/- 15 V at the main power for audio path and sidechain would cause fluctuations on the gain action, so we use +/-12 V for sidechain to have more controlled gains. Sort of a power regulator?


Quote:
If you only sample the "raw" (unprocessed") audio from C9 (left and right), you don't know what effect the level control (VCA) is having on the signal, so you are running "open loop". Like driving blind. So you can either sample the audio AFTER the main audio path VCA (e.g. at M1 OUT), or else you can sample the audio BEFORE the VCA as they have done. But if you sample the audio level BEFORE the control element (M1 VCA), then you must use another VCA to "simulate" what you are doing to the main audio path.

However, note that they have ANOTHER use for the sidechain VCA (U4), which is adjusting the static gain of the sidechain or "THRESHOLD" for the limiting function. So by using that "extra" VCA (U4), they combined both functions into a single chip. I would call that clever. At least that is my speculation on how they made the design trade-off decisions.

Yes. In feedback topology the output from audio path goes into the sidechain. It's a 'self adjusting' compressor. Attack, release, threshold effects are not predictable. They vary acording to audio signal.

Feedforward topology needs a precise control section, precise timing law, consistent attack and release. This part of the design could be more complicated.

So this circuit has a clever design, the vca at sidechain will 'copy' the GR of main vca and simplify the detection, effectively combining a feedback topology to drive the control section.

The sidechain signal is rectified and added an offset, the signal gain is controlled by the ratio section, then thru the attack and release section, a bank of resistors and a bank of tantalum capacitors.

The different resistors will give different charging times for the capacitors, attack, at a crude hardware level! and the different capacitors will discharge with different times too, release! Also raw electronics! (Heh Heh Heh, with plugins you would never learn this stuff)

The last release setting has 2 capacitors, one fast, one slow, for 'auto' release.

Now we have a rectified and timed DC sidechain signal, which goes to:

1. main vca control, including a summed DC from make up control
2. sidechain control, also summing a DC offset from threshold control
3. gr meter.

It could be more complex to design the control circuit. As an example, Distressor has digital control.

PS. It's a PITA or PITE (replace A*s for Eyes) to study a schematic in my 15" laptop screen
Everland Studios
Thread Starter
#29
29th June 2013
Old 29th June 2013
  #29
Gear nut
 
Everland Studios's Avatar
 
Joined: Apr 2013
Location: Sydney Australia
Posts: 88

Thread Starter
Everland Studios is offline
Ok, cool. More reading and thinkin'

For anyone who's reading this thread and starting with the basics like I am, I found this site which seems pretty useful to gain a basic understanding of circuits Volume I - DC : All About Circuits
#30
29th June 2013
Old 29th June 2013
  #30
Lives for gear
 
Joined: Feb 2006
Posts: 570

mixedup is offline
Quote:
Originally Posted by Mr. Lau View Post
What do we do? Maybe Ben (OP) can ask changing the title of the thread to How VCA and other compressors work?
Thanks again! What we need here is a dedicated gear design newbies subforum

Seriously, loving reading this thread, but partly because it's so on-topic. Maybe we should start more threads... if y'all have the time, that is.

Sent from my GT-I9300
New Reply Submit Thread to Facebook Facebook  Submit Thread to Twitter Twitter  Submit Thread to LinkedIn LinkedIn  Submit Thread to Google+ Google+ 
 
Topic:
Thread Tools
Search this Thread
Search this Thread:

Advanced Search
Forum Jump

SEO by vBSEO ©2011, Crawlability, Inc.