Help me MOD - D&R FET STEREO limiter

[altfidelity]

One note from me as a user standpoint of view.

The ratio measurement was at max threshold.
So the limiter compresses at cca -15db threshold at max.
In its original state it is designed to just catch the overshooting peaks in limiter fashion.

It should be modified in such a way, that the threshold should be even lower if intended for normal compressor use.

 

[richiyobs]

One note from me as a user standpoint of view.

The ratio measurement was at max threshold.
So the limiter compresses at cca -15db threshold at max.
In its original state it is designed to just catch the overshooting peaks in limiter fashion.

It should be modified in such a way, that the threshold should be even lower if intended for normal compressor use.

You re right 15dB of compression Is what i ve seen in SIM ... Pushing It harder with input gain you can go down to -18

 

[altfidelity]

@richiyobs Since you are basically creating new design, I will add this:
From my own subjective experience and listening tests, I would try to mod/change release shape into something like this:


Although I have no idea how would you achieve this in hardware . I am more of a DSP guy, hardware is new for me.

Attack portion seems/sounds nice already as it is.

 

[richiyobs]

attached r27 variation.... max thr/input , fast att and slow rel
i will add with new timing net

 

[richiyobs]

@richiyobs Since you are basically creating new design, I will add this:
From my own subjective experience and listening tests, I would try to mod/change release shape into something like this:

View attachment 118470
Although I have no idea how would you achieve this in hardware . I am more of a DSP guy, hardware is new for me.

Attack portion seems/sounds nice already as it is.

i m a soundman who plays in electronic, but what i learned in different compressor designs that shape not only depends how the timing is done but also how detector rectify and the ratio between att&rel themselves, so the shape of CV out... i could be wrong since not an engineer but it s what i see in sim and real measurement breadboarding different sidechains with the same VCA
have you ever used CMT? compressor measurement tool plugin.... it gives you a better view respect to pluginDoctor

best

 

[richiyobs]

the modified timing net

 

[altfidelity]

have you ever used CMT?

No, I will check it out, and I will also need some time to actually understand the plots, and see the shape of attack and release shapes, in pictures you've sent me.

Have you thought of adding reduction metering?

 

[richiyobs]

Have you thought of adding reduction metering?
[/QUOTE]

Yes i m working on a simple solution... Both moving coil and leds...
I Will post It since not so Expert with led ... Hope to receive support also because u267 IC Is out of production

 

[altfidelity]

Hope to receive support also because u267 IC Is out of production

ehh, forget led metering than. Everybody prefer classic needle metering anyways...

 

[richiyobs]

the led driver... the amperage is now below 20mA, c22 should set how fast it will react... is it better to go up 20mA?

and the analog Vu meter, i will use sd39 500uA, from 0 to -20
this is Heikki s design, he helped me with my first compressor proto... i found it easy to build and set in amperage, also tracks very well

best

 

[Hubbub]

For a decent 10 output LED bar/dot driver look at the currently stocked (at DigiKey, likely others) TI LM3914VX. IIRC it was once available in linear and log steps, it seems that only linear is available anymore.

 

[altfidelity]

The PCB is double-sided, but I haven't fully disassemble it yet.
Otherwise in its construction it is very minimal and elegant, very RNC 1773, Alesis Microlimiter like...

 

[altfidelity]

You re right 15dB of compression Is what i ve seen in SIM ... Pushing It harder with input gain you can go down to -18

Is there a way to lower the threshold even more (without raising the input)?

 

[richiyobs]

Is there a way to lower the threshold even more (without raising the input)?

Believe that r11 Will do the job... But at the same time r61 & r29 in ch1 as example must be adjusted or you ll lose its limiting behavior... That s why i m experimenting with input gain instead of varying thr or ratio...
Anyway It s better to play with ratio instead of thr .... Maybe a pot in front of ratio amp can be used... I ll try...
But varying thr amp Will also alter the full wave rectifier tracking

 

[richiyobs]

When you compress more over -15dB distortion begins... It s a nice equilibrium the original design

 

[altfidelity]

Hello again audio gents and ladies!

Could someone wise than me @gyraf @Hubbub , glance over schematic again, and tell me if diode D6 is there for protecting components/op amps dawn the stream or is there as a creative choice for smoothing CV signal/envelope?

Its definitely slowing down response and adding sort of pre-release/hold time (which I don't like) and I would bypass it if possible.
Well I kind of did already temporally, and it sounds better to my ears, I am just wondering if it technically OK to do so.

BTW with @richiyobs are sort of trying to design a 500 enhanced spinoff, and if anybody is into FET compressors and LTspice and how to optimally bias a JFET, please, welcome to join us.

 

[gyraf]

As I see it, D6 does not provide any sort of protection or the like - it's function is to partly isolate the initial fast-attack time-constant C12 from the loong release C11 (C12 only has a discharge path once voltage across C11 has fallen to -0.6V relative).

Imo, you're free to experiment with all sorts of attack/release schemes here. May I suggest e.g. looking at GSSL's auto-release? Perhaps even make both the two stacked time-constants variable, this makes for some interesting tweaking possibilities for full-mix duties

/Jakob E.

 

[richiyobs]

at first I thought d6 was to stop the current flow between the two capacitors so they wouldn't charge and discharge at the same time....
I tested symmetrix 501 constant time, i like the diode Array for autorelease, also tried neve33609 ... And both seems to work... Will experiment with your Gssl setup and see its behavior...
Thanks for your support

Best

 

[richiyobs]

I made a big mistake in my first simulation, I put the test resistor, 1MΩ , at detector and together with the ratio bias one I set them in the megaohm range to achieve 2.2v at test point in the original schematic.... what I got was a sweet response....
executing what is written in the manual, thanks to altfidelity who gave me that procedure, now it behaves like many jfet compressors = overshooting

is my error doable anyway? cause i like that smooth response, for sure compression is less, cv out is from -1.6 up to 0.2v

leaving it as per the original schemo, cv out goes from -15 up to 0v

maybe a compromise can achieved, to get more compression but keeping it smooth

any suggestion please?

thanks for your attention

best

 

[richiyobs]

attached sound samples processed in LTspice.....
the modified version has Thump at max settings, maybe it can be solved with the add of HPF or slowering the attack
my old pc is very slow during this process....
will test in future different solutions

edit- maybe thump is not so extreme according to settings used , just a Kick shot ..... but snappy and beefy tails