Release time mods for Altec 438A tube compressor
Studio .45
Hi all,
Does anyone know the mod to make a variable (or just faster) release
time for the Altec 438A tube compressor? Any and all information
would be appreciated.
Thanks
David Shuman
stud...@iconn.net
Monte P McGuire
In article <5eu4db$q...@news.iconn.net>,
Studio .45 <stud...@mailhost.iconn.net> wrote:
>Hi all,
>Does anyone know the mod to make a variable (or just faster) release
>time for the Altec 438A tube compressor? Any and all information
>would be appreciated.
I've never seen the schematic for a 438A, but I can speculate pretty
convincingly... I've worked on the RCA BA6A and the Gates 6144 and
they're all very similar.
Most true 'tube' limiters (i.e. variable mu design) use a pair of
diodes connected to the plates of the push pull output stage to
'sample' the peak level of the output. The dual diode (usually 6AL5)
is connected to the plates on one side and to a resistor connected to
a bias voltage on the other side. Also connected to the bias resistor
is a parallel RC network to ground; this cap stores and smooths the
control voltage.
When either plate falls below the bias voltage, that plate partially
discharges the GR capacitor and the GR voltage thus goes negative.
This GR voltage is connected to the center tap of the input
transformer and thus the negative going GR voltage serves to reduce
the bias (and therefore the gain) of the first stage. When the signal
is attenuated sufficiently, the plate voltage no longer falls below
the bias voltage threshold, the GR cap slowly recharges and the gain
of the limiter returns to normal.
So, the simple answer is that you need to change the RC time constant
of the gain reduction cap and resistor. The more complicated answer
is that you really should reduce the size of the cap if you want a
faster release time, since the cap is hanging off the plates of the
output stage and can directly cause slew limiting. This is because
the current that is drawn to charge the GR cap is signal current,
drawn only at the signal peaks. Not very hi fi... The smaller the
gain reduction cap, the less slew limiting and thus less distortion.
Note that there's another time constant involved: the attack time.
This is due to the fact that the sampling diodes (which are in series
with the cap as it is being discharged) have a finite resistance and
also there may be a series resistor added to the diodes to prevent
excessively fast attack times. So, if you reduce the size of the GR
cap, you should also increase the size of the resistance in series
with the diodes to preserve the attack time. This is also a 'good
thing' to do distortion wise, since it helps to reduce the
instantaneous current load on the plates of the output amp, making the
limiter sound cleaner when in gain reduction.
I also happen to prefer the sound of a slower attack time on a
limiter; it accentuates the transient nature of a signal and allows
you to compress and get a higher density signal without completely
removing all traces of dynamics.
So, find the 6AL5, find the center tap of the input transformer and
look at what's between the two. Many limiters used .1uF paper caps
that can be reduced in size by 3-4x if you also change the other
resistors around the cap.
If you want a nonlinear release time, you can replace the resistor in
parallel with the GR cap with some sort of nonlinear (i.e. voltage
variable) resistor, perhaps a FET or more simply a parallel network of
one resistor and several zener / resistor series networks. In this
way, the release time is shortened when the gain reduction voltage
increases. I did this to a Gates 6144 and it worked well. It took a
lot of futzing to find suitable diodes, due to the fact that the
circuit operates them at very low currents, but the end result worked
well.
The BA-6A offers a 'dual time constant' mode that is formed by
paralleling a series RC network across the GR cap. This second RC
network stores the 'long term' control voltage and allows you to use a
faster primary time constant without as many of the side effects of a
fast release time.
A lot of the sound of a limiter is due to the nature of the attack and
release time constants, so it's well worth your time to experiment
with differnet combinations to see what works best.
Since you're probably going to want several release times (or even
several differnet release networks), I'd suggest adding a rotary
switch to select among them. You could also simply change the R
across the gain reduction cap from a fixed resistor to a pot, but it's
better to reduce the size of the gain reduction cap as I explained
above, and the only way to do that is with a switch and several
capacitors.
Best of luck and if any of this is confusing, post away and/or send me
email...
Regards,
Monte McGuire
mcg...@world.std.com
Studio .45
I also need the schematic : )
>Hi all,
>Does anyone know the mod to make a variable (or just faster) release
>time for the Altec 438A tube compressor? Any and all information
>would be appreciated.
>Thanks
>David Shuman
>stud...@iconn.net