From Mike Grace
First off I should clarify the DC offset problem. V3s through serial
# V3100 had the AD converter set with the DC stripping filter turned
off. This results in the V3 sending out a DC offset level of around
–55dB. There is a simple modification that corrects this problem.
Anyone with a V3 in this serial number range is free to send their
unit to us and we will turn the unit around the same day.
In all of my years of recording and designing equipment I have never
been able to hear adverse effects of DC offset, especially in the
digital domain. (Large amounts of DC in an analog circuit can cause
premature clipping though) As well, I don’t think there is an AD
converter made that does not require that the input signal be offset
to 2.5 Volts DC before it is converted to digital code!
**The reason that V3s got out of our factory with the DC offset
problem is precisely because we could not hear it and we were not
measuring it in our test procedure.**
It was a customer that clued us in when he noticed it on the DC
coupled level meters in his DAW.
Now for a little bit of my philosophy on test equipment, measurement,
and listening. I believe that we should have the highest resolution
test equipment available to measure our product’s performance. We
should test every performance parameter that we can. We should make
sure that the measured performance of our products represents the
state of the art. When this condition is met all we really know is
that our product is not broken. This is the starting point, not the
end point, of the product development process. The next step is to
begin listening. Listening in a controlled laboratory environment
and listening in real world environments. This is were much
experimenting happens. Does this capacitor sound better than this
one? Does this amplifier sound better than that one? Does this
grounding scheme sound better than that one? Does this unit sound
good on ALL kinds of music?
Once I am satisfied that the unit sounds good I release it to
production. On the production line all of our preamlifiers go
through a 24 hour minimum burn in period, an exhaustive bench test
using our Audio Precision test instruments and a detailed inspection
process. This confirms that the unit is not broken. After these
tests EVERY unit that we sell (101, 201, 801, 801R, 901 and V3) gets
a full listening test by a real person with real ears! All of the
people who do the listening tests here are music lovers, musicians
and recording engineers. Each unit is listened to at every gain
level, trim level, filter setting, bit depth and sample rate using
our high resolution 901 headphone amplifier. Only then is the unit
approved for packing and shipment. I do not know if other
manufacturers do this or not but at the end of the day I want to
contribute equipment that is musical, not analytical.
…and now back to the thread…
There are several reasons for the gain structure that we chose in the
V3. It is well known that AD converter dynamic range is optimized
if the input signal is scaled so that the AD source circuit and the
AD itself clip at the same point. If this gain scaling is not
performed then a 24 bit converter will quickly be degraded to 16-17
bit performance. While it is true that lower quality analog
circuitry may sound better if it is not asked to operate at the high
levels required to get the full AD performance this is not the case
with the V3 preamplifier circuitry. The mic amps in the V3 are
built around a transimpedance or “current feedback” amplifier
architecture which allows the preamplifier to have essentially the
same sonic performance at high levels as it does at low levels.
This is not true of the“voltage feedback” amplifiers that many mic
amps use. I think transimpedance amplifier technology is catching
on in the mic preamp world and more and more preamps are using
them.
While the V3 has become popular in the concert taping world it also
has a substantial customer base in classical recording, film sound
effects recording and nature recording. All of these disciplines
require the highest level of sound quality AND very wide dynamic
range.
We have modified one customer unit to have a +20dBu for 0dBFS because
he was sending the analog outputs of the V3 to an Alesis Masterlink
as a backup. Since the Masterlink clips at +20dBu it was clipping
when the V3 was at –5dBFS. Adjusting the V3's AD sensitivity
aligned the V3 to the Masterlink at the expense of 5dB of noise.
Topic: V3 Dc offset problem (Read 5938 times)