Pandora

[L´Andratté]

(EDIT This relates to the PANDORA LM-401 (see end of p.1 for emrr´s original schem):

http://www.gearslutz.com/board/geekslutz-forum/868095-funny-compressor-doing-funny-things.html


Hey, that looks like it´s calling for the breadboard!

I now did a clumsy EAGLE schematic, please correct me if I drew something wrong or incomprehensible...

*emrr*, what do you know about this thing? There´s not much on the web, just some good mentions!

It looks like it should work with different opamps (RC748 is bifet):
http://datasheet.seekic.com/PdfFile/RC7/RC748.pdf

Concerning the JFET I guess it´s all about tweaking.

The gap to GND after the 1k is an unknown cap, I guess?

Thanks for posting, another lesson for me from the past.

 

[emrr]

I had one, sold it.  I added a gain stage to the (low) output.  The unmarked connections are holes in the board.  Maybe related to the LM-402 attack and release controls. 

 

[L´Andratté]

As mentioned in the Gearslutz thread, this probably only makes sense to build with the original parts(which I don´t have) for the same tone and response, but for curiosity and because of the nice minimalism of the circuit, I will try it anyway, hacking my way through with what is at hand... ;D

LM-401: single channel/only INPUT and OUTPUT controls
LM-402: dual channel w/ additional ATTACK and RELEASE controls

- So I presume the 1k should be connected to GND?
- I will try NE5532 as opamp replacement
- I will try J201 as JFET replacement (as they´re the only ones I have, probably needs changes in the sidechain)
- R5 looks like attack time, R6 release then with C5 as timing cap?
- R11 and C8 represent something like linearizing FB?
- R13 and everything to the right is only related to GR meter
- D1 looks really strange, I first thought "maybe an power indicator LED?" but no sign of that on the front panel...
- the ad in the first link below speaks of Hi-Z *cough* 5k?

What I don´t get is the JFET biasing, as a vc resistor the gate needs to get negative in relation to ground to lower the resistance from drain to source, right?
But D2 half-wave rectifies the audio to positive dc, so even when the gate is biased negative by R18 in a way that I don´t really understand, the JFET is decreasing it´s resistance with louder audio/higher DC, rather gating than limiting? I´m at loss, NEED HELP!!!

Her what I found on the web(not much):
http://www.technicalaudio.com/pdf/Martin_Pro_Audio-Video/Martin_Audio_bricks_catalog.pdf
http://s1135.photobucket.com/user/Bryan-gear/media/PandoraLM-402.jpg.html
http://www.gearslutz.com/board/geekslutz-forum/868095-funny-compressor-doing-funny-things.html
some gossip ;D
http://www.gearslutz.com/board/so-much-gear-so-little-time/121818-pandora-compressor.html
http://www.gearslutz.com/board/high-end/568444-pandora-lm-401-universal-imiter.html

attached for simplicity the original drawing by emrr:

 

[emrr]

I haven't looked at this drawing in years, there are easily some small mistakes here. 

I found the unit somewhat noisy in practice.  It was a very touchy limiter, it had a nice sag if you barely hit it, would pump strongly if you hit it too hard.  My other notes seem to indicate I couldn't measure limiting effect successfully with tones.  I added 20ish dB makeup gain on the output of mine with an IC on a daughter board.  I've not ever used the LM-402, or seen the inside of one.  The 748 may contribute to 'tone', but I think I'd update it. 

I was clued into these by Lee Hazen who ran Studio By The Pond near Nashville, TN.  He had several LM-402's.  They would be all over the Dr. Hook and England Dan & John Ford Coley albums recorded there. 

 

[L´Andratté]

Well, noisy, touchy, pumping strongly - those notes & tones

So C8 and R11 are the distortion cancelling path, featuring the same ´output´ impedance as the cv path, check out this article
telling why exactly that design shouldn´t be implemented, talking ´bout pumping, touchy, noisy, etc...

http://sound.westhost.com/project67.htm

Good thing, there´s also another very interesting JFET limiter design inside for  me on my neverending compressor quest ;D
Maybe the Pandora Limiter really sounds cool, but too many question marks for me, I´m already under heavy fire on the opto front!
And maybe it´s just vintage crap, that we want so badly to be magical? Who knows.

Nice day trip, though.

 

[Andy Peters]

As mentioned in the Gearslutz thread, this probably only makes sense to build with the original parts(which I don´t have) for the same tone and response, but for curiosity and because of the nice minimalism of the circuit, I will try it anyway, hacking my way through with what is at hand... ;D

LM-401: single channel/only INPUT and OUTPUT controls
LM-402: dual channel w/ additional ATTACK and RELEASE controls

- So I presume the 1k should be connected to GND?
- I will try NE5532 as opamp replacement
- I will try J201 as JFET replacement (as they´re the only ones I have, probably needs changes in the sidechain)
- R5 looks like attack time, R6 release then with C5 as timing cap?
- R11 and C8 represent something like linearizing FB?
- R13 and everything to the right is only related to GR meter
- D1 looks really strange, I first thought "maybe an power indicator LED?" but no sign of that on the front panel...
- the ad in the first link below speaks of Hi-Z *cough* 5k?

What I don´t get is the JFET biasing, as a vc resistor the gate needs to get negative in relation to ground to lower the resistance from drain to source, right?
But D2 half-wave rectifies the audio to positive dc, so even when the gate is biased negative by R18 in a way that I don´t really understand, the JFET is decreasing it´s resistance with louder audio/higher DC, rather gating than limiting? I´m at loss, NEED HELP!!!

A friend found one of the LM-402s in a pawn shop and digs it. A friend of his drew a schematic for it and anyways we did up a clone PCB for it. Last I heard it was stuffed and was being tested.

The circuit is a bit different from what Doug drew.

The input goes through a level pot, and that's followed by a voltage divider. The idea is to ensure that the signal swing is small enough to remain linear. After the divider is the JFET attenuator and then a 741 op-amp with a gain of about 17.

A zener reference and a trimmer helps set the JFET's gate bias. Biasing the gate to a couple of volts negative keeps the JFET off enough so there's no gain reduction. As the gate goes more positive towards 0 V, its channel resistance decreases and as such attenuation increases. Max attenuation is set by the JFET's on resistance. The bias cannot be set positive; if it did go positive, the whole thing goes to hell

So it's the job of the detector to push the gate voltage up as a function of input level. The detector doesn't use a diode rectifier. Instead, two transistor current sources are used to charge a timing cap. The output of the 741 feeds the base of one transistor and also goes to an op-amp inverter which feeds the base of the other. Emitters are tied together, and you get a crude current source. The current source charges the timing cap through the attack pot and the cap is discharged through the release pot. The timing cap is buffered with an op-amp, whose output sums with the bias adjust and some trimmable amount of the 741 output to drive the JFET base and thus set the gain reduction.

I'll draw a better schematic and post it later.

Interesting aside:  if you open your copy of "The Art Of Electronics" 2nd Edition to page 140, you see an example of using a JFET as a variable attenuator, just like in the Pandora design.

-a

 

[emrr]

Many golden oldies pump like this, it can be a useful effect. 

Nice to hear more about the LM-402. 

 

[L´Andratté]

Very interesting, JFETs and BJTs are clearly not my forte, but with that comprehensible explanation it all makes sense!
If it is possible, I would like to see the schematic, and try to understand it.

The transistor arrangement you described sounds like this one (with transistor inverter)
http://www.geofex.com/FX_images/dynacomp.gif
This unit is also said to be heavily pumping and noisy and yet very loved by many!

What are the up- and downsides of the diode vs. transistor rectifiers?

I would still like to try to build a simpel fet limiter, but just think my skills level is not up to reviving
such a mystery box without proper documentation.

As it happens I lately acquired *the second edition of Art Of Electronics*. Chapter 3.10 "FETs as variable resistors", now
that I call proper documentation!

 

[Bill Wilson]

The 1k resistor can be connected to ground to decrease the gain of the unit. As I stated in my post in the Pandora thread under Technical documents; it really should not be considered suitable for studio use by todays standards.