which attack and release is better for THAT VCA compressor?

[POTL]

Once again, hello and happy new year.
I am already finishing exploring analog effects technology, only the VCA compressor and noise maker remained (I will ask about this later) and urei1176.
I liked the Keeley compressor pro and Bondi FX 2026 pedals, they have a full set of adjustments and look pretty versatile, and their sound is very transparent (this is good and bad at the same time).
I read a lot of information, but I didn't find a detailed answer about the Attack and Release adjustments.
There are 3 schemes (maybe more, but I found only 3).
The first 2 are presented on the THAT website and are powered by transistors (circuit 1) or the original THAT chip (circuit 2), but in general they look almost identical. There is also an informal scheme using dual-channel Op Amp(circuit 3).
Someone had experience of assembly or is familiar with the operation of these schemes? What's better?
The third scheme looks very simple, but it seems to me that attack and release will affect each other (if this is true) and the time of attack and release are identical, which is pretty bad, can it be fixed.
The second circuit is simple, but the original chip consisting of four transistors is quite expensive, can it be replaced with 4 matched transistors? If so, which parameter to look at? HFE?
The first scheme is the most difficult, but attack and release are independent, the price is low, but some components are not so easy to find (18M resistor).
Tell us which scheme will be the most correct?

[garcho]

THAT Corp. assumes an engineer will pick component values that A: are cheap enough for the bean counters and B: works well enough for the engineers. When THAT publishes app notes the schematics are made with software, the values are very exact because they'd rather leave the real world component choices to the accountant and engineer. The designs are in their "perfect" state and it's up to you to determine where you might cut corners.
I would build one of the THAT designs, and see if you can find a way to get some cash to pay for the matched transistor array and everything. I've built a few THAT comp/limiters and they're fantastic. SSL and other broadcast console makers use them for VCAs, comp/lim, etc.
Are you planning on using 9V? Are you planning on making this a pedal? A rack unit? 30V?
Some component values are very important, most are not. Get close to what you can and hope for the best, that won't be what fries your $10 IC.

[PRR]

> not so easy to find (18M resistor).

The 18Meg may be two 10Meg series, for all practical purpose. Or probably one 10Meg with insignificant reduction of slow release.

The Attack side uses 1.8Meg, so attack is 10X faster than release. You can make this 2.2Meg, or two 2Meg series, or perhaps 1Meg or 5Meg.

There ARE pots to set the speeds, these big resistors just set workable limits.

Both A and B are incredibly fussy for *guitar* work. (The dBx/THAT guys have lived and breathed gain control for decades, and polished every detail.) Many-many-many musical limiters just use pots for attack and release. In 1947 GE and Gates were selling limiters that cost as much as a car, R-C timed. Same for about 90% of the non-dBx/THAT limiters since. Steal/plagiarize/study something simple.

[DIY Bass]

I built a Pumpernickel compressor. For my needs (bass guitar and looking at transparent slight compression to level things a bit) it sounds very good. There is no schematic that I have found but there are detailed assembly instructions that will probably give you some ideas about what parts and values were chosen.

[POTL]

THAT Corp. assumes an engineer will pick component values that A: are cheap enough for the bean counters and B: works well enough for the engineers. When THAT publishes app notes the schematics are made with software, the values are very exact because they'd rather leave the real world component choices to the accountant and engineer. The designs are in their "perfect" state and it's up to you to determine where you might cut corners.
I would build one of the THAT designs, and see if you can find a way to get some cash to pay for the matched transistor array and everything. I've built a few THAT comp/limiters and they're fantastic. SSL and other broadcast console makers use them for VCAs, comp/lim, etc.
Are you planning on using 9V? Are you planning on making this a pedal? A rack unit? 30V?
Some component values are very important, most are not. Get close to what you can and hope for the best, that won't be what fries your $10 IC.

Hi, I want to make a 15 volt pedal.
In general, I read that half of the resistors are responsible for the correct characteristic of the potentiometer, changing its linear structure to the logarithmic one.
Now I'm not sure which option is more modern with the original chip or with current mirrors on transistors.
By the way, in Google you can find photos of the Keeley compressor pro board and it seems that it uses the method with a two-channel operational amplifier.

[garcho]

do you mean ±15VDC, or do you mean 15V and "artificial ground" @ Vcc/2?

how are you going to power it? two 9V batteries? transformer/diode bridge/regulator? charge pump IC?

I'm not sure what your concerns are, what is your goal? To have a compressor guitar pedal for your own playing? You might not want the THAT console style compressor then, have you looked at OTA designs or the Orange Squeezer route? Don't be put off by a lack of expensive ICs, guitar pedals are not HiFi systems.

[POTL]

do you mean ±15VDC, or do you mean 15V and "artificial ground" @ Vcc/2?

how are you going to power it? two 9V batteries? transformer/diode bridge/regulator? charge pump IC?

I'm not sure what your concerns are, what is your goal? To have a compressor guitar pedal for your own playing? You might not want the THAT console style compressor then, have you looked at OTA designs or the Orange Squeezer route? Don't be put off by a lack of expensive ICs, guitar pedals are not HiFi systems.

Unipolar power +15 volts
Microcircuit THAT4320
Powered by the ICL7660S / MAX1044 / LT1054 + 15 volt voltage regulator.
Orange Squeezer is a rather specific design - it paints a lot of sound and I don't like the way it works, only urei1176 is of interest to me from compressors on field effect transistors, but I will leave it for the future.
OTA compressors are also pretty vintage and simple - I know how they work and I want to try something new for myself.

[garcho]

THAT Corp.'s app notes are designed with 15V bipolar power in mind, not unipolar, just pointing that out.

check out this:
SSL comp/lim

[rankot]

You may try this one: http://hpbimg.someinfos.de/workshops/that-jam/that-jam_9V_v-1-0_schematic.png

I have built the PCB for it, populated it 90%, and now I don't have time to finish it

[POTL]

if the circuit can work well with one polarity, this is sad.
I expected to connect the ground instead of negative polarity, and the points where the earth is indicated for the displacement (at the expense of the needs of the microcircuits and the sockets by itself) are connected to the virtual ground.

[noisette]

Hi!
I tried to get the first scheme to work or about six hours on the breadboard and didn´t get there (I´m not a noob). 

The point being that true RMS timing does not work with independant att & rel, they are correlated, best thing to switch the timing caps, or use the auto (nonlinear capacitor scheme). That should be fine with all kinds of audio!

OTOH if you need att&rel pots and don´t need true rms the scheme that Rankot posted looks promising