GssL 20th Anniversary edition

[scott2000]

You bored? Or a better bot? Keeping an eye on you..

/Jakob E.

Just an opinionated coonass...
It's an interesting observation about the comp imo.

 

[desol]

Blend is a switch you turn on? I thought wet/dry blend was always in the circuit?

Also, I was hoping there was gonna be a limit feature added....20:1 for example. Like the Smart C2.

 

[Gustav]

All files up, and everything out of our hands.

Hope someone has fun with it - we sure did!

Isak & Gustav

 

[fragletrollet]

Cool stuff! Awesome work.

What options should one go for if getting it fabbed at JLC?
1.6mm and HASL w/lead unproblematic? What benefit added from say 2mm and ENIG, thicker gold etc...

What to do with vias?

Check box for "2 designs", so panel/switch PCB will be separated from main board?

I acknowledge that this is on the fringe of the scope of this thread, but thought and recommendations are appreciated.

Oh, and where would I tap the side chain to a switch, so I can use external key input for ducking uses?

 

[JMan]

1.6mm with HASL should be fine for this design.

ENIG is an unnecessary extra cost in this case IMO. Lead is only problematic if you live somewhere that has rules about it or are already using lead-free solder, or if you plan to eat it (I mean, lead is obviously problematic in a general sense, but I think you can understand what I mean). Gold plating is useful on something like an edge connector that would need to be wear-resistant, but not needed for this board. From what I can see, the most basic default option from JLC should be just fine.

 

[Script]

All files up, and everything out of our hands.

Hope someone has fun with it - we sure did!

This is a grand gesture and huge gift to the community. Thank you for all the time spent.

 

[richiyobs]

Sorry for troubling....
just for learning purpose and fun tried to simulate the CT net of the original Gssl compared to SoundSkulptor version in 500 format, since the last one has a smart configuration with only 2 caps..... believe to save space and the number of caps.

except the autoRelease position,where the slopes are very different maybe due to the dual element REL with different values, otherwise the curves are very close/similar...

What is your opinion?

thanx for your attention

best

 

[Rob Flinn]

Sorry for troubling....
just for learning purpose and fun tried to simulate the CT net of the original Gssl compared to SoundSkulptor version in 500 format, since the last one has a smart configuration with only 2 caps..... believe to save space and the number of caps.

except the autoRelease position,where the slopes are very different maybe due to the dual element REL with different values, otherwise the curves are very close/similar...

What is your opinion?

thanx for your attention

best

On your diagrams you have one drawing where you have the 2 caps (6u8 & 0.47u) in different places.

 

[richiyobs]

are you meaning a different configuration and so different behavior?
what i get in mind is that when not in auto mode the 2 caps can be considered in series and so a value like 440nF....
in auto mode R12 & R13 are no more shorted to GND and so you can get a dual element Release....
also i did an error drawing it, i removed a connection at R6

edit: Altering R11 form 100k to over 1meg, the slopes are almost matched

 

[Rob Flinn]

are you meaning a different configuration and so different behavior?
what i get in mind is that when not in auto mode the 2 caps can be considered in series and so a value like 440nF....
in auto mode R12 & R13 are no more shorted to GND and so you can get a dual element Release....
also i did an error drawing it, i removed a connection at R6

edit: Altering R11 form 100k to over 1meg, the slopes are almost matched

I mean you have the 0.47u across the 750k resistor on the left diagram & across the 910k on the right hand diagram. The 910K is wrong it should be 91K. This is the Value on the original SSL diagram which is what all these compressors are derived from. You should consider that there are 2 time constants one stacked on top of the other. So you should get an initial quick release from the 0.47u across the 91K & then the tail end of the release is slower from the 6u8 across the 750k.

 

[richiyobs]

Thanks for explaining It ....infanct At First i changed the 910k to 91k , but It doesn't match the slope of the original configuration... So went back to 910k like in soundSkulptor schemo... Also 750k and 910k are reversed in position respect to the original design as you noted, and It Is strange... So my question Is: Is the dual release Always engaged even not in auto mode?
Just to get a better understanding of what's happening in this circuit.

Thanks for your attention

 

[Gustav]

The 910K is wrong it should be 91K

I think what they are trying to achieve is the use of a single capacitor for all the single timing release values, but I am unclear about the way it is drawn up.

If I am understanding it correctly, and going by what Richi is saying, its done with a 910K static resistor, and switching various values into parallel to get the desired resistance per step.

The 910K switched in parallel with 100K for the auto-setting is just under 91K. The second pole of the switch does not to jump it to gnd on "auto", but it does so on the single-timing settings, where connection is made to ground directly from the resistor in parallel, seemingly to take out the 2nd timing stage.

From the schem Rich posted.

Soundskulptor

The values in parallel. 120K/910K 220K/910K 330K/910K 1Meg/910K give us

106K, 177K, 242K, 480K

Gyraf Schematic

180K, 270K, 560K, 1.2Meg

Dont know why these differ. The auto-release value is pretty much spot-on, so maybe I calculated them incorrectly.

Altering R11 form 100k to over 1meg, the slopes are almost matched

I double and triple checked the order of the 2nd stage resistor values/cap pairings when doing this layout, but maybe someone will care to check it from a different perspective...however.

As far as I know, the correct configuration is 91K/470n followed by 750K/6.8uF, and the resistor values are stacked in reverse on the gyraf shem, although, actually correct on PCB.

IF you are comparing to a 750K/470n configuration, going with a 910K, switching in 100K gave us just under 91K, so that is far from the 750K you are comparing to here. You say that going to 1Meg gets you close, and that would be 428K. You would actually need a parallel resistance of around 4.3Meg to get all the way there.

Essentially, I think you need to swap the 750K and 91K resitor on the gyraf shem, and you should get the same value, if you are running simulations on the auto. The others would differ according to the difference in total value.

...that is also the way its drawn up on the layout posted in the thread youre using to post this.


Gustav

 

[richiyobs]

I think what they are trying to achieve is the use of a single capacitor for all the single timing release values, but I am unclear about the way it is drawn up.

If I am understanding it correctly, and going by what Richi is saying, its done with a 910K static resistor, and switching various values into parallel to get the desired resistance per step.

The 910K switched in parallel with 100K for the auto-setting is just under 91K. The second pole of the switch does not to jump it to gnd on "auto", but it does so on the single-timing settings, where connection is made to ground directly from the resistor in parallel, seemingly to take out the 2nd timing stage.

Thanks for delving deeper into it... Now I have a clearer idea
In fact, at the beginning the first thing I did was reverse 910k with 750k.... because I understood that in standard mode the 2 caps could be considered one, but in doing so the curves move away.
Maybe the way I'm simulating rise and fall isn't correct, but I don't know any other methods to compare the 2 circuits, not so expert

Attached S.Skulptor's schematic part

Thanks for your support

Best

 

[Gustav]

Thanks for delving deeper into it... Now I have a clearer idea
In fact, at the beginning the first thing I did was reverse 910k with 750k.... because I understood that in standard mode the 2 caps could be considered one, but in doing so the curves move away.
Maybe the way I'm simulating rise and fall isn't correct, but I don't know any other methods to compare the 2 circuits, not so expert

Attached S.Skulptor's schematic part

Thanks for your support

Best

The two caps do not work as-in-series to create just one timing with the resistors in parallel to each.

I am suggesting that the 91K and 750K are reversed on the Gyraf circuit, not to reverse the 910K and 750K on the other schematic.

Gustav

 

[richiyobs]

As far as I know, the correct configuration is 91K/470n followed by 750K/6.8uF, and the resistor values are stacked in reverse on the gyraf shem, although, actually correct on PCB.

Thanks for this very useful info, I didn't know that.
Also looking at your v3 diagram I understood where I was wrong on simulation!! Now everything seems to be matched.
I will check my old GSSL build, to see if I made the same mistake since the pcb was self-etched

thanks again