Manley VOXBOX - MVBXA 4/2003 - Manual (Page 22)

IT MAKES NOISES WHEN THE FRONT PANEL IS TAPPED - An easy one. Some tubes become microphonic
over time. That means they start acting like a bad microphone. Vibration has caused the supports for the little parts in
the tube to loosen and now the tube is sensitive to vibration. Easy - Replace the tube. Which one? The one that makes
the most noise when you tap it. Usually this will be one of the smaller (gain stage) tubes closest to the front. The
VOXBOX will have to be on, connected and speakers up but not too loud for the sake of your speakers.
IT GOT HISSY - Also easy. This is again a common tube symptom. You could swap tubes to find the culprit but an
educated guess is OK too. Generally the first tube in the path is the one with the most gain and dealing with the soft-
est signals. The usual suspect is the shorter tubes - the 12AX7 in the preamp or EQ. You may find that you need to
choose the quietest tube out of several of that type - like we do at the factory.
DISTORTION - This might be a tube. Swapping is a good way to find out. It may be a wiring thing or mismatch as
well. Wiring problems usually accompany the distortion with a major loss of signal. Mismatches are a bit tougher. The
VOX BOX has a high input impedance and low output impedance that can drive 600 ohm inputs of vintage “style”
gear. Best place to start is check your settings and meters. It may not be your first guess.
DC OR SOMETHING AT THE OUTPUT THAT IS INAUDIBLE - The 1/4” unbalanced outputs have a frequency
response that goes way down to below 1 Hz. A little very low frequency noise may be seen as speaker movement
when monitors are pushed to extreme levels. The XLRs do not exhibit this because the transformers filter below 8 Hz.
Also the unbalanced outputs do not like long cheap high capacitance cable. Occasionally a very high frequency oscilla-
tion (200 kHz to 400 kHz) may occur in these conditions. Once again use the XLR outputs. Problem solved.
THE METERS ARE OUT OF CALIBRATION - If the problem only seems to be when the unit is just turned on
it’s normal. It should warm up. It might be a half dB out for 15 minutes - relax. Everything should stabilize within 15
minutes as the temperature reaches “normal”. Your unit will have been factory calibrated and tested at least 4 times
before you received it. Sometimes parts drift a bit in value over the years, or you have repaced tubes and want the unit
calibrated at the same time, or you just want it as perfect as it can be. These are good reasons t go through the calibra-
tion procedure or sent it to a technician or the factory for a tweak. If you send the unit to a tech, you should include
this manual because they will need it. If you do it yourself, you will need an audio oscillator and good voltmeter and
a few screwdrivers and it would be nice to have a Scope but not necessary. 4 out of 5 of the trims are just for metering
and the audio trim just sets unity gain for the EQ. A little advice: if its not broke - don’t fix it. We get a number of units
back because someone thought they should tweak and couldn’t.
Once in a while we get a call from a client with a “digital studio” with confusion about levels. They usually start out
by using the digital oscillator from their workstation and finding pegged VU meters the first place they look and they
know it can’t be the workstation. Even a -6 level from their system pegs the meters. Some of you know already what
‘s going on. That -6 level is referenced to “digital full scale” and the computer might have 18 or 18.5 or 20 dB of head-
room built in. That -6 level on the oscillator is actually a real world analog +12 or +14 and those VU meters don’t re-
ally go much further than +3. There are a few standards and plenty of exceptions. One standard is that pro music (non-
broadcast) VU meters are calibrated for 0VU = +4 dBm =1.228 volts into 600 ohms. Another standard is that CDs
have a zero analog reference that is -14 dB from digital full scale or maximum. This allows sufficient peak headroom
for mixed material but would be a bad standard for individual tracks because they would likely distort frequently. This
is why digital workstations use higher references like 18 and 20 -to allow for peaks on individual sounds. It may be too
much in some cases and too little in others. Add two other sources of confusion. Peak meters and VU meters will al-
most never agree - they are not supposed to. A peak meter is intended to show the maximum level that can be recorded
to a given medium. VU meters were designed to show how loud we will likely hear a sound and help set record levels
to analog tape. By help, we mean that they can be only used as a guide combined with experience. They are kinda
slow. Bright percussion may want to be recorded at - 10 on a VU for analog tape to be clean but a digital recording
using a good peak meter should make the meter read as high as possible without an “over”. Here is the second confu-
sion: There aren’t many good peak meters. Almost all DATs have strange peak meters that do not agree with another
company’s DAT. One cannot trust them to truly indicate peaks or overs. Outboard digital peak meters (with switchable
peak hold) that indicate overs as 3 or 4 consecutive samples at either Full Scale Digital (FSD) are the best. They won’t
agree with VU meters or Average meters or BBC Peak Programme (PPM) meters either. Each is a different animal
for different uses. The Compressor and Limiter in the VOXBOX should help digital and analog achieve consistent
levels but if you are recording to digital, use your machines meters for the recording level. The VU meter on the VOX
BOX is great for analog tape, establishing “loudness” decisions and for calibrating levels using an oscillator but not
for digital peaks. Some engineers can interpret a VU and get near ideal digital levels but it takes years of experience
because it depends on the sound. We sometimes think about adding an accurate and useful peak LED but that question
of “accurate to what” comes up. If it were just an LED indicating when our units reached clipping you’d probably find
the machines’s meters would be pegged.
22