Diodes (like Chandler Germanium) in ratio circuit of 1176

[BradM]

I was looking at the Chandler Germanium Compressor manual the other day online and read the following:

"Most Ratio/Curve controls are simple resistance circuits. Using diodes adds flavor and character. We ended up with 6 different combinations, using germanium, silicon, and zener diodes in several configuratios as well as a more standard resistance setting."

Has anyone played with this or tried to implement such a thing in the ratio circuit of an 1176?

Any thoughts on what's required? Is it just a matter of placing some combinations of diodes in series with the ratio pots/switches?

thanks,
Brad

 

[Samuel Groner]

As a starter you might want to read the THAT application notes on soft knee cuircuits (don't have the link ready).

Samuel

 

[BradM]

Thanks for the tip. I'll check that out.

Is this the one?

http://www.thatcorp.com/datashts/dn107.pdf

Brad

 

[BradM]

Thanks the reply, Wayne. I'll check that one out as well.

Brad

 

[JohnRoberts]

I did a similar trick in a tape NR compander set. For modest signal deltas the gain control had a long time constant, but for changes larger than some threshold it reverts to a faster time constant. This is useful and delivers good attack time, with lower distortion for persistent tonal signals due to unwanted gain modulation by fast release time constants.

You like to make the bulk of larger gain adjustments while the ears are still adjusting to the changes too. The ears sensitivity distortion increases with the duration of it.

JR

EDIT I am adding this link http://circularscience.com/522_NR.pdf Which I believe I may have cited before, for folks who prefer drawings over my obscure explanations words.

Fans of obscure design tricks will find several in this one. I use features of the 572 differently than they intended, and intentionally allow opamps to clip as a threshold for the ripple filter to stop smoothing. The fast attack, steals current from the slow attack reservoir so long term tracking won't suffer from minor peak level or response errors (think tape NR with phase shift and whatever between encode/decode).

Way too many parts even for a kit, considering dBx was assembling their tape NR in Japan cheaper than I could buy parts. :roll:

JR

 

[AMZ-FX]

Germanium diodes have really soft knee characteristics... the conduction curve is much more gradual than an si diode, which looks sharp in comparison.

You could also mix an si with a ge to give you an asymmetrical response, which might add some character.

regards, Jack

 

[CJ]

some tidbits...

"more range of control, in a series/shunt combination....


"The use of diodes for gain-control presents two opposing problems: the diode impedance
is very non-linear f(r large audio signals (1 volt) and causes considerable distortion; the DC
control signal of about 1 volt, which is required for satisfactory range of gain control,
mixes with the audio to cause "thump". Therefore, reducing the audio signal level across
the diode, in order to reduce the distortion, makes the audio signal small in comparison
with the DC Control sigqlal, which proportionately increases the sound of the "thump".
Using well-balanced diodes and push-pull circuitry is the only solution. However, temperature
variations tend to unbalance the diodes and allow the "thump" to be audible.
Aside from its relatively slow reaction time (5-20 ms) and a slight sensitivity to
temperature, the light-dependent resistor (LDR) is a near perfect gain-controlling device.
The resistance element is very linear, and contributes no measureable distortion to the
audio signal; and its resistance variation of 10 megohms to 100 ohms, with increasing
light brilliance, affords an adequate gain-controlling range over a wide variety of circuit
impedances. In addition, the resistance element and the audio signal have no electrical
connection to the light bulb, or to the DC Control signal, and the "thumping" sound
mentioned earlier is completely eliminated. This isolation also allows the resistance


more...
"Recent research into the use of field-effect transistors (FET), as gain-controlling
devices, has been rather fruitful. In particular, the drain-source impedance of the metaloxide
silicon (MOS) type with an insulated gate exhibits a very dependable change as the
gate voltage is varied. Specific circuit design for high quality performance requires
that the drain source impedance be made more linear by feedback into the gate or substrate,
and that the audio signal across this impedance be limited. Depending on whether the
feedback is taken to the gate, substrate, or both together, the sig_aal voltage limitation can
be 10-15 millivolts, or as high as 50 millivolts, for less than 1% total harmonic distortion.
Since the gate is insulated from the drain-source impedance, the DC Gain-Control signal
cannot mix with the audio signal, and "thumping" is not a problem.
Ail of the gain-controlling devices described above have a simplicity of application
which makes them popular. However, there are several other techniques which are
rather complicated, but should be mentioned:
(a) Chopper, or Variable Duty Cycle - This approach uses

 

[NewYorkDave]

It's good form to credit your source when doing a cut 'n paste.

 

[Stagefright13]

Source? That was off the top of his head.... It's a quote from the long awaited CJ 102!

:green: