Manley SLAM! - Manual (Page 13)

LIMITERS
The first compressors we know of used special tubes designed for
radio and gain control. These are the “remote cut-off” type like the
6386 as used in the Fairchild 670. Manley has been making the Vari-
able Mu® compressor/limiter for many years based on the same
principle.
Opto limiters probably began in the early 60’s with the UREI LA2A
which used a small electro luminescent panel driven by audio di-
rectly from a tube circuit. This panel was in a light shielded box with
a photo-resistor so that when the panel lit, the light shone on the
photo-resistor which in turn dropped in resistance and shunted audio
to ground, and reduced its level. Very simple but effective technique
which has stood the test of time. Part of the reason is the simplicity
of the 2 knob approach and part is the inherant attack and release
times both the electroluminescent panel and photo resistor have and
part is the sound of the tube/transformer circuits.
There weren’t many companies building pro audio gear in the 60s
and 70s, but we had FET limiters, discrete transistor voltage con-
trolled amplifier (VCA) limiters, biased diode limiters and, in gen-
eral, all had plenty of color and distortion. One of the best known
FET based limiters is the UREI 1176 which brought more control to
attack and release times and had ratio switches. The first few gen-
erations of IC based VCAs were also less than perfect, and VCAs
got a bad name but slowly improved over the years. With cheap easy
to use op-amps and VCAs, gear prices dropping, music business
growing, we began to see more gear but somehow the antique 670s
and LA2As and 1176s were still in use and preferred.
In the early 90’s a few maverick audio manufacturers including us
responded to that knowledge and developed new-old technologies.
Manley, for example began the ELOP® using a LED/photo resis-
tor component called a Vactrol and combined it with ICs to drive it
and tube circuits for the audio. Over the years, that opto circuit was
revisited in a discrete transistor Langevin ELOP, variations on the
theme used in the VOXBOX® and once again here in the SLAM!.
In developing the SLAM!, we began with the idea that probably we
could find alternate Vactrols that could be used to give some variety
to the opto-limiting. In the end, after trying every one out there, we
decided the one we had always used, was our favorite and the oth-
ers had more drawbacks than advantages. We did improve the drive
and metering electronics, and added a HP filter in the side-chain and
added the jacks to allow a user to insert their own EQ into the side-
chain. In most aspects, the opto-limiter circuit is similar to the one
used in our previous Elop’s and uses audio to drive the LEDs. This
means that we can’t possibly adjust the attack and decay character-
istics significantly without changing Vactrols. On the other hand,
this mode of operation, seems to act more like an RMS respond-
ing circuit and reacts to many sounds in a way that we prefer over
opto’s with the conventional attack, release, etc controls and all the
complications that a gain control element with its own timing char-
acteristics adds to that recipe. To make a long story short, we like
what happens with that simple old-school approach. The difference
in driver circuits plus the side-chain filter does seem to make the
opto a lot more useful on mixes and drums than our previous units.
We also allow some fast LED metering of the opto in addition to the
regular VUs which helps show how fast it tends to react and gives a
more complete picture for critical applications.
“The FET Limiter”
Because we couldn’t improve much on our old opto circuit we de-
cided to add a second ‘type’ of limiter with its own characteristics
and its own historical roots. Some early limiters like the 1176 used
FETs for the gain reduction element which offered much faster at-
tack times and controllable releases.
The problem with FETs when used as a gain control element is that
they can add unpleasant distortion unless the signal is very low (like
-30 to -40 dB) and we also wanted a few gain controls which also
eat signal unless they are cranked. Throwing op-amps around to get
gain where needed is easy, but not our style. Keeping to an all-tube
Class-A concept requires different approaches.
We took a novel approach and used a transformer as part of the
shunt circuit, which not only reduced the signal to a nice level for
the FET but allowed us to use a pair of FETs in counter-phase to
reduce distortion. An expensive approach but worth it.
Our main goal of the FET limiter was to achieve the fastest release
that we could cleanly. This is the quality that causes the gain to
return as fast as possible, which is what gives us our perception
of ‘loudness’ .The goal here was a great ‘go-louder’ box. Fast at-
tacks, and ∞ : 1 brickwall limiting is important to prevent ‘overs’
but not for increasing the average level. A very fast transient that
gets through will clip but as long as the duration is short enough it
will not be perceived as distortion. Very fast release times, unfortu-
nately, usually imply modulation distortion where the limiter traces
the waveform at low frequencies rather than the volume envelope.
This is inevitable, but we made it possible to achieve faster and
cleaner releases than usual. It can still get crunchy so be careful.
This FET sidechain uses several techniques to get those fast releas-
es. The usual full-wave rectifier was replaced with a quadrature rec-
tifier that uses 4 phases to determine peaks and allows twice as fast
smoothing. Then we combine multiple side-chains and the typical
exponential capacitor release was modified for linear decay rates.
All of this resulted in faster clean releases, thus more loudness. Still,
at the fastest release times it is quite possible to get modulation dis-
tortion, which is sometimes a useful color and often a problem with
wide spectrum sounds like mixes or instruments with lots of lows.
Listen for a growly sounding distortion on faster release times.
The multiple side-chains also gave us the possibility of introducing
an attack switch, which is usually not found on a limiter (com-
pressors, yes). The attack switch works on the slowest side-chain,
which gives much of the audible familiarity of the control while the
other side-chains are still biting the fastest transients. Like most at-
tack controls, as you go from fastest to slowest positions, you tend
to lose some threshold or limiting, so adjusting the FET LIMITER
threshold will probably be required, conversly more clockwise for
VF settings.
A CLIP setting is on the release switch, that introduces a very
rounded clipping with a variable threshold. This type of distortion
is reminiscent of speaker distortion and tends to be mentally associ-
ated with ‘loud’. Of course, more conventional clipping is possible
and by turning up the INPUT, and turning down the OUTPUT, or
if one wants ‘drastic’ switching to MIC or INSTRUMENT will do
that of course. And, no, you won’t hurt anything as long as, phantom
is off, and you prudently turned the OUTPUT down first. It wasn’t
designed to simulate a guitar amp but intended to ‘assist’ an already
overdriven tone. Mostly CLIP is used to get a few extra dB of ‘angry
loud’.
13