FirstWatt SIT-2 User Manual

First Watt

SIT-2 Power Amplifier

Read Me First

I fully realize that many, if not all, owners will rush to hook up the amplifier
without reading this operating manual. I don't blame you – I don't read them
either. However, this amplifier is different in a number of ways, and if you only
read this page you will probably save us both some time and trouble.

Heat and Ventilation - This amplifier consumes 200 watts all the time and
converts it into heat. Pick a location where the amplifier can get some fresh air
and blow off the heat. Do not enclose the amplifier in a closed cabinet. Don't
set it on the carpet. Give it lots of space.

Input Connection and Input Impedance - The SIT-2 takes single-ended
(RCA) inputs. The input impedance is selectable between 10 Kohms and 100
Kohms by gold jumpers which plug into XLR connectors near the RCA inputs:

I suggest you start with the default 100K setting which will work with anything.
Later you may wish to try driving the Gate terminal of the SIT device more
directly by removing the gold jumper and inserting it in the “10K” input position
as shown. If you are running a low source impedance, you won't find much
difference (at least I don't), but you can enjoy the concept of running through the
absolute minimum of circuitry. In either case, use whatever sounds best to you.

– The XLR panel connector is not an input connector –

Output Connection - You can hook this amplifier up to any loudspeaker without
danger of damage. Note, however that the (+) Red output connection is actually
grounded. The (-) Black output terminal is the live connection. This can be
important when you are hooking up active subwoofers to the output of the amp –
treat the Red output as ground, and the Black as active.

Warmup - The SIT-2 takes a while to fully warm up. During this time you can
listen to it, but you will find that the initial operating point will drift a bit and will
require some readjustment.

OK! You can go play now.

Introduction

I have been building power amplifiers for over forty years now, and at this late date I wasn't
expecting much in the way of surprises. In the past few years at First Watt I had been
working with low power designs using some of the new power Jfet transistors from Lovoltech
and SemiSouth which offered some performance improvements with simple circuits, so at
least I felt like I was still making some incremental progress.

The problem was, however, that while you can make really good sounding amplifiers with
simple linear circuits, it tends toward a certain amount of sameness as they measure better.
It is difficult to create something subjectively magical ” about the sound, something that
makes people wake up and say “What the hell is that?”

As a tradition, there is really no formula for such a thing, as you are dealing with the incredibly
complex neural system that sits between people's ears, and it doesn't respond to sound in the
same way as our distortion analyzers and oscilloscopes.

If you have experienced ears and if you know what you like when you hear it, then the
approach that works is to try everything. That is largely what I do, remembering Edison's
dictum “invention is 2 percent inspiration and 98 percent perspiration.”

Sometimes, though, things just fall into your lap. A couple years ago I was talking to Jeff
Cassidy at SemiSouth and he mentioned that at one time they had made a special run of
Static Induction Transistor devices on some kind of military/industrial contract, and that one of
their technical people had remarked that they were nearly ideal for use in audio amplifiers.

“Really”, says I, “Do you have any of them left over?” No the customer scooped them all up,
but they would be interested in making more. The price of a small run was astronomical (to
me, at least) and I spent quite a bit of time pondering the risk.

And then I wrote the check and didn't look back.

After a few months I had a small batch of SITs with my name on them, and I started playing.
It took about a year and a half to arrive at the designs of the SIT-1 and SIT-2 amplifiers, and it
has been a revelatory experience.

The SIT devices allow operation with only one gain stage. In fact the SIT-1 has only one
transistor in the entire amplifier, without feedback and even without degeneration. It's as
simple and raw an amplifying circuit as you could imagine.

I have built such simple amplifiers before using Mosfets and power Jfets, but I was not at all
prepared for the quality of sound that I got with the SITs. It was like magic.

The SIT-1 and SIT-2 are special amplifiers. They have single gain stages – the input signal
goes into the Gate of the SIT and comes out the drain amplified and driving the loudspeaker.
There is no feedback or degeneration. They run single-ended Class A, and the SIT-2, biased
by a current source, has an efficiency of about 10%, and the SIT-1, biased through 800 watts
of power resistors, has an efficiency of about 5%.

Both amplifiers deliver about 10 watts output per channel. The SIT-1 is a monoblock drawing
about 200 watts, and the SIT-2 is a stereo amplifier also drawing about 200 watts.

So what's the difference?

First off, the SIT-1 gain device is surrounded by passive components, giving it a lower
efficiency of around 5%. It is biased through 800 watts of power resistors mounted on the
heat sink. The SIT-2 is biased by a constant current source created by another FET device,
and it has an efficiency of about 10%. While these two approaches give nearly the same
sound at the default setting, I have always been intrigued by the sound available from the SIT
without any other active devices involved. I suppose you could think of it as slightly more
purist, if twice as impractical.

The SIT-1 has a knob and meter on the front panel which allows the user to adjust the load
line setting of the SIT device to tweak it into the most preferred setting for different
loudspeakers and listening taste. It also helps compensate for small variations in AC line
voltage.

The SIT-2 operates a fixed operating load line as determined by listener's preferences in
various systems. This setting corresponds to the middle position setting on the SIT-1. Being
biased by a constant current, it does not need compensation for AC line variations.

Apart from these, they are very similar. I live with both happily.

The Harmonic Structure

The SIT device allows the possibility of arbitrarily deciding on the specific harmonic structure
of the distortion of the amplifier. Like Triodes, the characteristic curves of the SIT device allow
operation on load lines which can set the relative values of second and third order harmonics.
My own experience echoes that of Jean Hiraga, which is that the best sound tends to come
from a dominant second order harmonics followed by a lesser value of third harmonic and
essentially no harmonics above those. It looks a lot like this:

The distortion performance of the SIT-2 into both 8 and 4 ohm loads is shown here:

In both curves we see a second harmonic character culminating in a very soft overload, which
gives the amplifier more apparent power than its distortion ratings suggest.

The SIT-1 and SIT-2 amplifiers can be driven in “direct mode” where the source components
looks directly at the Gate of the SIT without other active components in between. This is
desirable, but in this case the input impedance is 10 Kohms and plus the capacitance of the
Gate which is greater than the 1,000 pF of the IHF standard. If you have a preamp or source
component with an output impedance of 1,000 ohms or less, you should not have a problem,
but with “passive preamps” and some tube designs which have very high output impedances,
this limits the bandwidth and could be a performance issue.