Multiband Compressor Idea

[Script]

MDX8000

(I have modified a couple of them, and they had some serious problems in the design.)

This caught my attention. Problems like what? Do you mind giving a few hints? Just asking cos there might be a unit flying my way. And it's good to know in advance whether it's worth the effort and time.

 

[RuudNL]

The problem I had (about 20 years ago, or even more) was that the compressor attack was too slow, causing transients being caught by the clipper, with a lot of distortion as result. (Especially on 'S'-sounds on vocals.)

 

[Script]

Sounds like two problems already. How do you remember recombined phase behaviour to be? Half-way decent?

 

[ajbff]

The problem I had (about 20 years ago, or even more) was that the compressor attack was too slow, causing transients being caught by the clipper, with a lot of distortion as result. (Especially on 'S'-sounds on vocals.)

Hello Ruud,

I saw you're post about the MDX 8000.

Mine has a broken part, R707 (weestand)
Do you know what the capacity ofvthat one is ?

Jos
Radio de Glazen Stad
Netherlands

 

[RuudNL]

Dat zou ik moeten opzoeken. Eerst zien of ik de schema's nog kan vinden...
I have to look it up. I hope I still can find the schematics...

 

[RuudNL]

According to my information the value is 27E4. (So I suppose 27,4 Ohm)
Please check if one end is connected to C288 (47µF) and D159 (6V2 zener).
Identical values are R704, R706 and R709. To check I would measure these resistors in circuit.
This resistor seems to be part of the metering circuit.

 

[ajbff]

According to my information the value is 27E4. (So I suppose 27,4 Ohm)
Please check if one end is connected to C288 (47µF) and D159 (6V2 zener).
Identical values are R704, R706 and R709. To check I would measure these resistors in circuit.
This resistor seems to be part of the metering circuit.

 

[RuudNL]

Of course it is possible that they have changed the component numbering in a later production batch.
As far as I see it must have been a pretty low resistor value, the way it went up in smoke...
I will try to scan this part of the schematic.

 

[RuudNL]

Schematic.

 

[ajbff]

Schematic.

Thanks Ruud !

 

[RuudNL]

If this is the real situation on your PCB, I would suspect the electrolytic capacitor (47 µF) or one of the LM3914 LED drivers. It seems the 6.2V zeners are there to reduce the supply voltage for the LM3914's.

 

[JohnRoberts]

Sounds like two problems already. How do you remember recombined phase behaviour to be? Half-way decent?

that is the classic weak link with multi-band dynamic processing (putting Humpty back together again).

I've had success with de-essers by applying the gain change full band but with fast release to avoid envelope modulation of other frequency bands.

JR

 

[Script]

I'd assume derived filters to be better, if it wasn't for op amp-stage time lag. So might need some follower stages to keep number of op amp stages equal on all bands? Would that do the trick, not just on paper? Lots of stages though...

 

[JohnRoberts]

I'd assume derived filters to be better,

yes

if it wasn't for op amp-stage time lag.

inconsequential

So might need some follower stages to keep number of op amp stages equal on all bands? Would that do the trick, not just on paper? Lots of stages though...

Don't over think this... Putting Humpty back together is essentially impossible if you change parts of him while he is apart.

This is mature effort. Back decades ago Radio station engineers experimented with loudspeaker crossovers and unlinked compressors on each bandpass... It is easy to imagine how badly those worked. There were some commercial broadcast limiters that didn't suck but mainly for managing broadcast chains not recording.

JR