SSL-sidechain overkill by means of 2 Thrust-circuits

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

clintrubber

Well-known member
Joined
Jun 3, 2004
Messages
5,982
Location
The Netherlands
Not sure to which of the many SSL-sidechain or A*P*I-Thr*st-circuit threads to add this, so let's make it yet another thread...


I've been toying around some more with the 10dB/dec filter (a.k.a. the Thr*st-filter).

While at it, I've combined this circuit with the 'variable HPF-Steffen-sidechain' for the SSL-comp.
So all in all one could now add an overelaborate sidechain to the SSL (or to whatever comp.) if one wishes to do so.
If it's actually worth while is hard to tell beforehand, so I decided to
just do it. It's simple and tweaking circuits is fun.

The Thr*st-patent kindly gave the circuit as we've seen in the relevant threads.
In addition to that, the 2*5*0*0-manual gave the in-between freq-plot: the so called medium-mode. It's a low-att. with flat mids & some hi-lift.
This will probably differ not much from the 'SSL-highpass' with some hi-boost added but to avoid getting entangled in even more elaborate variations I put it in as a non-adjustable preset and not as some additional hi-boost section glued onto the already present variable highpass. Everybody still reading ?

The rotary now has twelve positions, since I wanted to keep
the frequencies of the Steffen-circuit/Greg-PCB. One could have programmed a rotary
for only two additional positions of course, but greedy people use them all
so there are some intermediate frequencies added as well.
That might be of limited use, but one has to try before one knows.

So now we have:

#1 sidechain off
#2-#9 60 / 75 / 90 / 110 / 130 / 160 / 200 / 250 Hz first order highpass
#10 Thr*st - medium (some lo-att & some hi-boost)
#11 Thr*st - loud (10dB/dec slope)
#12 external (apart from de-essing, could anyone still think of anything for this ?! :wink: )

Overkill indeed, but you've been warned...


A possible circuit for both added filters goes like this:
sc-circ_thrst_01.jpg


Components values for the medium-thr*st-preset are tweaked to get near
the desired responses. Note that the values for the loud-preset are
slightly different from those in the patent, they're obviously rounded to the nearest standard values but that's fine, circuit response hardly differs from that for the published values.


It might have been possible to avoid additional opamps but that's not what happened.
Note the input-impedance of the loud-Thr*st-filter steadily falls from 33kOhm @ 20Hz to
some 1kOhm @ 20kHz. For both the loud&medium-filter combined the input-Z is still above 800 Ohms @ 20kHz and since the already present load is not that much I expect the preceeding 5532 will be fine.


OK, I have only drawn & simulated this contraption so far. Will be building it as soon as the highpass-PCB from Greg arrives.
So all your comments & thoughts still welcome, please shoot !


Bye,

Peter

Ah, and w.r.t. the added frequencies for the normal highpass-filter:
add caps of 180 / 120 / 82 & 56 [nF]. Maybe I add a more complete drawing later.
 
[quote author="[silent:arts]"]I have had the same idear for a while, thanks for doing the work for me :green:
:thumb:[/quote]

All fine ! :thumb:
Since the idea was in your head as well, you must have given it all some more thought already.
So feel free to comment, not unlikely I've missed something !

BTW, I feel the approximation of the med-setting is close enough, but if anyone finds some better values please post. The only thing I could still think of now is beer, food & loose wom.. wait, no, is for a level-adjustment between the various filters.
Maybe some more tweaking of the feedback-resistances or putting a trimpot there so that less threshold-adjusting is in order when selecting other sidechain-filters.

The big pity of course is that I'll be adding the 'normal' hpf-sidechain on the nice PCB upcoming from Greg, but still end up with some perfboard in the box for the t*h*r*u*s*t-stuff. Maybe I should put that section in a little black box with the patent-number on it :wink: :thumb:

If this all turns out to be of use I can very well imagine that somebody will become tempted to update the present sidechain-PCB for some UberPimped-Sidechain-Extravaganza(TM)-edition :wink:

Bye,

Peter
 
I cannot find any advantage in the high boost section of the thrust circuit...
just insert an eq to your sidechainsignal and try it yourself..
if the detector reacts on the boosted high freqs....the result is an attenuation of the high freqs in the audiosignal...

so I think a variable adjustable solution for the hipghpass frequency would fit here much better... maybe something like this:

http://users.otenet.gr/~athsam/variable_filter.htm


mat
 
Why not use A-weighting filter like this:
http://sound.westhost.com/project17.htm
or this:
http://www.epanorama.net/documents/audio/aweight.html

Just another idea to fill up that 12 pole switch.
 
First of all thanks for the ideas & comments.

[quote author="matthias"]I cannot find any advantage in the high boost section of the thrust circuit...
just insert an eq to your sidechainsignal and try it yourself..
if the detector reacts on the boosted high freqs....the result is an attenuation of the high freqs in the audiosignal...

so I think a variable adjustable solution for the hipghpass frequency would fit here much better... maybe something like this:

http://users.otenet.gr/~athsam/variable_filter.htm


mat[/quote]
Given that the LOUD-setting makes sense (the patent sounds believable,
the arguments given seem to make sense), I assumed that the MED-setting was just a less tilted slope across the whole spectrum. That's also as drawn on the 2500-front.

So when I saw it's a low-att slope & hi-boost with flat mids I wasn't sure what to think. And agreed, when that curve is on the table, it does indeed make little sense to add that hi-lift when the problem seems to be at the low-side of the spectrum.
So hard to tell what's 'right', and of course there isn't a 'right' since etc etc...
So I just toyed around a bit to get that MED-curve and I might just add it. The suggested variable low-att makes sure sense, but is already addressed by the original Steffen/Greg-part of the circuit, so to the section of the sidechainfilter that's not drawn but to which I'll be adding the t*h*r*...-thing.

[quote author="Brad"]Why not use A-weighting filter like this:
http://sound.westhost.com/project17.htm
or this:
http://www.epanorama.net/documents/audio/aweight.html

Just another idea to fill up that 12 pole switch.[/quote]
Nice idea. I guess it'll be 'loudnessifying' the signal...


[quote author="ioaudio"]what about compressing the side chain signal (comp²)[/quote]Just thinking out loud, that'll kind of counteract the actual compression-action, not ? But hey, anything goes if that side-chain is yours ! :wink: :thumb:
But serious, won't peaks get suppressed before they can hit the actual detector ? So they're getting less compression in the actual VCA, and that's the signal that's being the output.

If you want comp² then better put two of them after each other. I've not tried this much, but have heard good things about it (using different att/rel-settings). And the RNC does this in its 'supernice'-orwhatwasit-mode.

Bye,

Peter
 
Hello,

Kind but serious request.

Please note I've pulled the schematic in which I gave an emulation-circuit of the two Thrust-modes.

As I consequence I ask you to remove/take back/change any copies of
the circuit-diagram that you might know about and have made available on-line. This holds as well for documentation with an added PCB-layout: please remove the schematic.

No further distribution either please; this circuit cannot longer be made available online.

Thank you for your cooperation.
 

Latest posts

Back
Top