GroupDIY

Project Specific Discussions => Dynamic Processors => Topic started by: zamproject on January 05, 2019, 11:30:46 AM

Title: True passive compressor/leveler !!!
Post by: zamproject on January 05, 2019, 11:30:46 AM
Hi all

This is the winter holiday design...R&D...test...whatever experimentation  :)

Who guess ? one component in the signal path...a variable resistive track...usually called fader  ;D

https://youtu.be/Nu-LVIUcX3k (https://youtu.be/Nu-LVIUcX3k)

some drawback in the design choice but so fun  8)

Best
Zam
Title: Re: True passive compressor/leveler !!!
Post by: EmRR on January 05, 2019, 11:55:09 AM
Entertaining
Title: Re: True passive compressor/leveler !!!
Post by: zamproject on January 05, 2019, 12:14:29 PM
Entertaining

yes it is  ;D
Still I'll try to make something usable for real with it...so many tuning option...basic principle work, now the details

Best
Zam
Title: Re: True passive compressor/leveler !!!
Post by: warpie on January 05, 2019, 12:18:49 PM
cool  ;D
Title: Re: True passive compressor/leveler !!!
Post by: zamproject on January 05, 2019, 01:47:46 PM
Just to add, this is not my automation system ! no digital data involved for the fader move...
Audio side chain move the fader...with some analog processing  ::)

Best
Zam
Title: Re: True passive compressor/leveler !!!
Post by: leitmo on January 05, 2019, 01:48:13 PM
Shocked!
Title: Re: True passive compressor/leveler !!!
Post by: zamproject on January 05, 2019, 02:06:49 PM
THD...zero  8)
Title: Re: True passive compressor/leveler !!!
Post by: ruffrecords on January 06, 2019, 03:25:17 AM
Nice one!

Cheers

Ian
Title: Re: True passive compressor/leveler !!!
Post by: DerEber on January 06, 2019, 04:36:18 AM
Cool!
You can add rubberbands to change timings.
Title: Re: True passive compressor/leveler !!!
Post by: abbey road d enfer on January 06, 2019, 06:59:33 AM
The big advantage is that, due to solid friction, there is a form of hysteresis that cannot be implemented in analog electronics. That results in reduced (even sometimes eliminated) control-voltage ripple-induced distortion (at LF).
Brilliant implementation!
Title: Re: True passive compressor/leveler !!!
Post by: zamproject on January 06, 2019, 09:09:05 AM
You can add rubberbands to change timings.

Yes it was on the paper notes, as different fader caps with different weight...
But timing still slow due to mechanical inertia so I don't think I will go for mechanical constrain for timing
also to not over stress the belt...
Timing test is on the way, attack is about 30ms right now fastest possible move, release ca by anything longer than 30ms, by side chain tuning.
In this smashed drum room test, direct signal is delayed from SC signal at DAW, fader inserted in the console

The big advantage is that, due to solid friction, there is a form of hysteresis that cannot be implemented in analog electronics. That results in reduced (even sometimes eliminated) control-voltage ripple-induced distortion (at LF).
Brilliant implementation!

Not sure to get what you say about hysteresis ?
I appreciate your words !

drawback:
-how fader will react in long term with this kind of constant stress
-no make up gain if I still want a passive processing (no issue in my usecase)
-fast attack not possible, this one will be more an "auto leveller" than a "compressor"

Thanks to all !!
I have some more test to perform before calling this a real tool.
Do you guys think it's an interesting design ? I mean for real application ?
For the moment I'm splited... I do it for the challenging concept, but as is it work better than expected so...I'll think I go further

Best
Zam
Title: Re: True passive compressor/leveler !!!
Post by: moamps on January 06, 2019, 09:41:00 AM
-how fader will react in long term with this kind of constant stress

From my experience it will last a short time.  Imagine a cycle lifetime of 1 million, and one cycle per second, it will be  dead in 300 working hours. Besides, there is a fast heating of the wiper and resistive material because of friction and noise is building rapidly. I used similar approach to detect LF speaker membrane position in an experiment and abandoned it very shortly.   
Title: Re: True passive compressor/leveler !!!
Post by: abbey road d enfer on January 06, 2019, 09:48:58 AM
Not sure to get what you say about hysteresis ?
Due to adherence, changing from a position to another takes more force than the application of cinetic movement and gravitational force; OTOH it allows an object to stay in position without any expense of energy. Both these things are unknown in electricity.
With solid friction, it takes a certain amount of energy to move an object from point A to point B and back, when the law of conservation of energy would say it takes zero power.
In contrast, the entropy of an electrical system where a node moves from voltage A to voltage B and back is null.
The fact that there is a mechanical threshold for moving an object submitted to solid friction is the equivalent of hysteresis.
Title: Re: True passive compressor/leveler !!!
Post by: zamproject on January 06, 2019, 10:27:44 AM
Due to adherence, changing from a position to another takes more force than the application of cinetic movement and gravitational force; OTOH it allows an object to stay in position without any expense of energy. Both these things are unknown in electricity.
With solid friction, it takes a certain amount of energy to move an object from point A to point B and back, when the law of conservation of energy would say it takes zero power.
In contrast, the entropy of an electrical system where a node moves from voltage A to voltage B and back is null.
The fact that there is a mechanical threshold for moving an object submitted to solid friction is the equivalent of hysteresis.

Ok ! the inertia of acceleration...
To have a move A to B approaching a null time, you need energy approaching infinite, which is not possible.

From my experience it will last a short time.  Imagine a cycle lifetime of 1 million, and one cycle per second, it will be  dead in 300 working hours. Besides, there is a fast heating of the wiper and resistive material because of friction and noise is building rapidly.

Yes I completely imagine this...the fact is the fader you see is the first I buy time ago for my fader automation.
Spec give 300 000 cycle for 100mv sliding noise, in real I abuse it a way more including fast move crash test for hours...
So far the sliding noise still very low, about -80 dBfs(-60dBu) below 20-30Hz (which can also be a DC offset, I don't check yet) and about -100 from 30Hz to 20kHz
So life time is not that simple to predict now, I have to kill it to know...
OTHO it can be an expandable component...like a tube in a tube amp...
But that's for sure the major issue in this design, I'll see

Best
Zam
Title: Re: True passive compressor/leveler !!!
Post by: abbey road d enfer on January 06, 2019, 12:40:08 PM
Ok ! the inertia of acceleration...
No. Inertia is conserved energy. The energy it takes to accelerate the mass is restituted when the mass returns to zero speed. I'm talking about the energy that is used to fight the adherence and friction forces that is degraded into heat; that's entropy.
Title: Re: True passive compressor/leveler !!!
Post by: zamproject on January 06, 2019, 04:59:13 PM
No. Inertia is conserved energy. The energy it takes to accelerate the mass is restituted when the mass returns to zero speed. I'm talking about the energy that is used to fight the adherence and friction forces that is degraded into heat; that's entropy.

Ok...I'm not that familiar with chaos quantification...but I suppose if we know the material we ca deduce the requested energy to balance the friction ?  ::)
Friction forces should be a constant ?
Or is it totally unpredictable ?

Best
Zam
Title: Re: True passive compressor/leveler !!!
Post by: abbey road d enfer on January 06, 2019, 06:40:28 PM
Ok...I'm not that familiar with chaos quantification...but I suppose if we know the material we ca deduce the requested energy to balance the friction ?  ::)
Friction forces should be a constant ?
Or is it totally unpredictable ?
Friction force is not constant; its a reaction force. If there's no stimulus, there is no reaction. Adherence force is higher than friction force; in order to move an object you need to apply a force that's in excess to te force needed to keep moving, to which you must add M.gamma, the acceleration force.
The movement of an object is not proportional to the force that's applied to it; as long as the force is lower than the adherence force, it does not move. That's a non-biunivocal relation.
Title: Re: True passive compressor/leveler !!!
Post by: zamproject on January 07, 2019, 08:58:34 AM
Friction force is not constant; its a reaction force. If there's no stimulus, there is no reaction. Adherence force is higher than friction force; in order to move an object you need to apply a force that's in excess to te force needed to keep moving, to which you must add M.gamma, the acceleration force.
The movement of an object is not proportional to the force that's applied to it; as long as the force is lower than the adherence force, it does not move. That's a non-biunivocal relation.

Tks abbey for the reminder... I study that more than 20 years ago, miss about everything except basic concept  :-X
I re-read all your post here and think I now get from what you deduce the hysteresis result
Once the fader is called to move there is a "dead" time used to reach the force that exceed adherence force, then it move with less force until the fader have to stop or move forward where the same phenomena apply each time the speed is reduced for a stop but need more force because inertia is reduced by friction.

I guess that basically the hysteresis you define is my fastest attack/release time if no timing is used at sidechain

Best
Zam
Title: Re: True passive compressor/leveler !!!
Post by: gyraf on January 07, 2019, 09:12:22 AM
Great idea... ...following...
Title: Re: True passive compressor/leveler !!!
Post by: zamproject on January 11, 2019, 02:44:31 PM
Great idea... ...following...

What an honour !

Since I "play" a lot with various tuning , also trying to determinate coherent user function for this as the approach is different from usual dynamic processor.
As soon as I have more or less a fixed operation behaviour, I'll do a massive crash test to count how long fader will last...

Best
Zam

Title: Re: True passive compressor/leveler !!!
Post by: gyraf on January 15, 2019, 02:36:21 AM
Have you tried looking into PID (Proportional-Integral-Derivative) controllers - this may (or may not) influence both timing and precision in interesting ways..?

Jakob E.
Title: Re: True passive compressor/leveler !!!
Post by: zamproject on January 15, 2019, 05:57:41 AM
Hello Jakob

Have you tried looking into PID (Proportional-Integral-Derivative) controllers - this may (or may not) influence both timing and precision in interesting ways..?

Jakob E.

The motor driver is already an analog  PID  ;)

Best
Zam

Title: Re: True passive compressor/leveler !!!
Post by: Phrazemaster on February 06, 2019, 11:27:53 PM
That’s really neat. Too bad you can’t record it the first time, then play it back with 10ms delay or delay of your choosing...
Title: Re: True passive compressor/leveler !!!
Post by: zamproject on February 07, 2019, 06:59:23 AM
That’s really neat. Too bad you can’t record it the first time, then play it back with 10ms delay or delay of your choosing...

Hello

It's possible, I mean I don't wire any fader data handling for the moment but I have technically everything available for this...
actually the "response" is about 30ms, with some more tweak (and new idea...) I'll try to go faster, 20ms should be doable, 10ms will be great but uncertain for the moment...
external SC is also one option (the one from the test I show) same signal delayed...

I'm in standby since 3 weeks on this, other works, but I want a working prototype within few months

Now, I think I'll have to drop the 100% passive idea, I need buffer for "universal" use, as symmetrical I/O
Do you guys think it's an issue ? I mean dose it kill the whole idea/concept ?

Best
Zam