Problem with Fairchild 660/670 original manual data?

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

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

rotheu

Well-known member
Joined
Jan 8, 2006
Messages
110
Location
CCCP
Hi everybody,

I was messing around with Fairchild 660/670 timing network lately and after doing some calculations it seems like there is a problem with the data in the manual regarding the release times. I could be wrong, I'm just curious what you guys think.

The original manual release times are:

Position 1: 0.3 Sec.
           2: 0.8 Sec.
           3:    2 Sec.
           4:    5 Sec.
           5:    2 Sec. for individual peaks
                10 Sec. for multiple peaks
           6: 0.3 Sec. for individual peaks
                10 Sec. for multiple peaks
                25 Sec. for consistently high program level

Time constant formula is T=R x C, where T is discharge time in Seconds, R is resistance in Ohms and C is capacitance in Farads.

Now, let's calculate release times for each position

Position 1:
2uF cap with 220K and 68K resistors in parallel.
Paralleled resistors will have resistance of 51.94K or ~50K.
T=50000 Ohms x 0.000002 F=0.1 Sec.(0.3 Sec. in the manual)

Position 2:
2uF cap with 220K and 470K resistors in parallel.
Paralleled resistors will have resistance of 149.85K or ~150K.
T=150000 Ohms x 0.000002 F=0.3 Sec.(0.8 Sec. in the manual)

Position 3:
2uF cap with 220K and another 2uF cap.
T=220000 Ohms x 0.000004 F=0.88 Sec. ~0.8 Sec.(2 Sec. in the manual)

Position 4:
2uF cap with 220K and another 2uF and 4uF caps.
T=220000 Ohms x 0.000008 F=1.76 Sec. ~2 Sec. (5 Sec. in the manual)

Position 5:
2uF cap with 220K and another 2uF for individual peaks.
T=220000 Ohms x 0.000004 F=0.88 Sec. ~0.8 Sec. (5 Sec. in the manual)

8uF and 100K and 220K resistors in series for multiple peaks.
T=320000 Ohms x 0.000008 F=2.56 Sec. (10 Sec. in the manual)

Position 6:
2uF with 220K for individual peaks.
T=220000 Ohms x 0.000002 F=0.44 Sec. (0.3 Sec. in the manual)

8uF and 100K and 220K resistors in series for multiple peaks.
T=320000 Ohms x 0.000008 F=2.56 Sec. (10 Sec. in the manual)

20uF and 100K and 220K resistors in series for consistently high program level.
T=320000 Ohms x 0.00002 F=6.4 Sec. (25 Sec. in the manual).

Positions 5 and 6 are more complicated, so I'm probably missing something there, but 1-4 are very basic and if you compare the original values with calculated ones, it seems like they somehow got shifted by one. Would love to hear your thoughts.

670timing.jpg
 
This is an important topic and could do with a more detailed reply.

Nowadays the definition of the time constant is taken at the first timing point of 63% charge/discharge.
Back in the day, several compressor manufacturers quoted timings to 95% which is the 3rd timing point thus:-

1st time constant is 63% leaving 37%.  63% of 37% is 23.3%  63% plus 23.3% = 86.3%

2nd time constant is 86.3% leaving 13.7%.  63% of 13.7% is 8.6%. 86.3% plus 8.6% = 95%

3rd time constant is 95% and so on.

DaveP
 
Release time is not time-constant.

BBC had another way, easy to read on 'scope:
 

Attachments

  • BBC-recovery-spec.gif
    BBC-recovery-spec.gif
    12.1 KB · Views: 47
> 0.1 Sec.(0.3 Sec. in the manual)
> 0.3 Sec. (0.8 Sec. in the manual)
> 0.8 Sec. (2 Sec. in the manual)
> 2 Sec. (5 Sec. in the manual)


So the spec times are 3 or 4 time constants.

If the gain-control was Linear Voltage (it is NOT), this would be 3-4 Time Constants.

After one time-constant, linear action, you are 63% down or 37% up from final value. is 2.7dB, an audible difference from final value. You need several time-constants for gain to return to audibly nominal. "Several" could be 3 or 4.

In fact a vari-Mu's action is more like linear-in-dB. I'm too tired to figure-out how that works.

Sitting there in a studio, what is in a number? Do we really know how fast our tails should recover? If so, don't we know more in therms of instrument (or room) decay time, or in terms of beets per minute feeling, than on a stopwatch? The numbers are good-enough relative to other settings on the same box. Ideally they are "similar" to the action of another limiter with similar numbers (and we know they are never just-the-same).
 
Well, all I can say is, I have checked the attack and release times of every compressor for which figures are available (sad git ::)) and with the exception of 1 or 2, they all agree with the 63% definition.  The ones with dual capacitor discharge are obviously exceptions too.

It is in fact a very useful excercise to check them out, as you can discover revision/typo errors this way, for example:-

Take the Collins 26W, the schematic shows the timing cap C109 to be 0.5uF but this does not agree with the timings given in the specs.  Then you look at the parts list and discover that C109 is in fact 0.25uF and all the timings check out ok.

best
DaveP
 
Back
Top