amp sensitivities.....general guide needed!
raj
im running a system witha crown k2 on the ev tops and a vz 3600 on
some 18" bass cabs. I recently invested in an EV DX34 loudspeaker
management system.
Bascially what i want to know is a quick low down on amp
sensitivities. I know each of these amps have them but i dont have
the foggiest on what they do, what they should be set on and how to
know when to change it. Also, how does the sensitivity relate to the
crossover system gain and limiters??
thanks
raj
Shaun
Didn't I answer that for you in a series of about a dozen comprehensive
posts on the subject? If you set it as I described, and your system works,
doesn't blow up, and is nice n' loud, isn't that what you were after? Ack!
Sensitivity is a term used to describe a relative voltage level, usually
equal to the input level which will produce the max rated output level. If
your amp multiplies the input voltage by 40 times (32 dB gain), and its
input sensitivity is 1 volt, you get 40 volts at the output with 1 volt
input. If you have a 1000w/8 ohms per channel amp, and its input
sensitivity is 1.4v, then with an input voltage of 1.4v, the amp will
deliver roughly 89.4 volts at 11.2 amps, which calculates to
20*log(89.4/1.4) = 36 dB gain. Flip the sensitivity (gain) switch to its
.775v setting, and now it is 20*log(89.4/.775) = 41.2 dB gain, so a .775v
input signal now makes 1000w/8 ohms at the output. It's a volume control,
dude. Don't worry about it. Personally, I like to use the least amount of
gain at the amps, and send them the hottest signal possible from the
crossover, allowing for proper headroom at every point in the chain, to
reduce the noise floor to a minimum, and when correctly set up, a signal at
the threshold of clipping at the mixer output should also be near the
threshold at the EQ, the crossover, and the amps, when all is calibrated.
--
Shaun Wexler,
Hellsgate Sound
http://www.hellsgate-sound.com
mailto:sh...@hellsgate-sound.com
Dave Cull
> If you have a 1000w/8 ohms per channel amp, and its input
> sensitivity is 1.4v, then with an input voltage of 1.4v, the amp will
> deliver roughly 89.4 volts at 11.2 amps, which calculates to
> 20*log(89.4/1.4) = 36 dB gain. Flip the sensitivity (gain) switch to its
> .775v setting, and now it is 20*log(89.4/.775) = 41.2 dB gain, so a .775v
> input signal now makes 1000w/8 ohms at the output. It's a volume control,
> dude. Don't worry about it. Personally, I like to use the least amount of
> gain at the amps, and send them the hottest signal possible from the
> crossover, allowing for proper headroom at every point in the chain, to
> reduce the noise floor to a minimum, and when correctly set up, a signal at
> the threshold of clipping at the mixer output should also be near the
> threshold at the EQ, the crossover, and the amps, when all is calibrated.
> --
> Shaun Wexler,
> Hellsgate Sound
> http://www.hellsgate-sound.com
> mailto:sh...@hellsgate-sound.com
Shaun,
Being a young guy, and always wanting to learn more about the theory
behind what we do day-to-day, what is the 20*log(v2/v1) calc. used for?
I know typically for dB itz 10*log(p2/p1), but I haven't seen the 20*log
used before to convert voltage into a dB ratio. Care to help me out?
I'd really appreciate the lesson.
Cheers,
Dave Cull,
Sylvain Sound.
dmi...@spamblock.demon.co.uk
> <snip to math>
> Being a young guy, and always wanting to learn more about the theory
> behind what we do day-to-day, what is the 20*log(v2/v1) calc. used for?
> I know typically for dB itz 10*log(p2/p1), but I haven't seen the 20*log
> used before to convert voltage into a dB ratio. Care to help me out?
> I'd really appreciate the lesson.
Power is V^2/R, so assuming constant impeadence we have:
10 Log (p2/p1), now substitute V^2 for each p, you get
10 Log ((v2^2/R2)/(v1^2/R1)), equals 10 Log ((v2/v1)^2) as
long as R1 = R2 which equals 20 Log (v2/v1).
In most modern audio usage this is a slightly dodgy thing to do as
the amplifier spec. typically ignores the impeadence transformation
(db is DEFINED in terms of POWER ratios). The 20 log thing only
really works if input and output impeadences are equal. This is
generally the case in RF, but not in audio where optimising power
transfer is unimportant.
Regards, Dan.
--
** The email address *IS* valid, do NOT remove the spamblock
And on the evening of the first day the lord said...........
.... LX 1, GO!; and there was light.
Peter Larsen
> im running a system witha crown k2 on the ev tops and a vz 3600 on
> some 18" bass cabs. I recently invested in an EV DX34 loudspeaker
> management system.
I don't know the k2, I don't know the vz 3600 and it is beyond me what
an EVDX34 system is about. So I must be qualified to follow up. This
because it is a general "strategy" type of question.
> Bascially what i want to know is a quick low down on amp
> sensitivities.
Out of the cardboard box all amps come with an input sensitivity that is
surely adequate for any purpose the manufacturer has been able to dream
up. Consequently almost all amps should have their inputs attenuated.
> I know each of these amps have them but i dont have
> the foggiest on what they do,
Input sensitivity doesn't "do anything", it is about how things are.
Amps may have input sensitivity adjustments, and those "do something",
they adjust sensitivity.
> what they should be set on and how to know when to change it.
You should adjust the input sensitivity of your poweramps to what your
systems design standard input sensitivity for poweramps happens to be. 5
volts RMS on the input for nominal power out could be a good choice, you
will still have some headroom in whatever is driving them, but it is a
lot less likely that plug-unplugs or "pre-something" turn-on/turn-off
transient are going to cause massive loudspeaker and - worst case -
hearing damage.
> Also, how does the sensitivity relate to the
> crossover system gain and limiters??
Recommended reading: Electrovoice PA bible if available. We are suddenly
talking system gain structure and system headroom. Simply explained: if
you have more headroom in one stage than in some other stage, then you
have failed to obtain maximum signal to noise ratio for the entire rig.
Limiters should generally be adjusted so as to prevent something from
clipping, so you really have to define what it is they protect. Limiters
before and after an electronic cross-over (or included therein) are
quite different concepts, with different pro's and cons and different
issues. They may even be set to protect hearing, in which case you have
to know how their adjust will correlate to actual SPL produced.
Crossover gain - well, yes, uhm, now that is a simple concept: if your
mixers nomimal output is to low for the poweramps, then cross-over gain
can be relevant. Mixers tend to be able to output 7 to 10 volts RMS, so
it is not a highly likely isssue, not at a glance.
Real world systems producing real music, and not just used for the
playback of full volume sinewaves will however see the output of a
cross-over as being smaller in each band, just because each band is less
than the complete signal. Consequently you might want to look at the
actual signal levels occuring when supplying pink or - imo more
realistically - orange noise to the system, pink noise will lead to an
"over cauteous" setup assuming sane mixing and sane system setup because
it is ""too trebly"", orange noise happens to be defined as the natural
characteristic of ""multi-source acoustic sound"", music is such.
> thanks
>
> raj
--
*************************************************************
* This posting handcrafted by Peter Larsen, MCSE *
* My site is at: http://www.muyiovatki.dk *
* I subscribe to http://www.spamcop.net *
*************************************************************
Dave Cull
>
> Power is V^2/R, so assuming constant impeadence we have:
>
> 10 Log (p2/p1), now substitute V^2 for each p, you get
> 10 Log ((v2^2/R2)/(v1^2/R1)), equals 10 Log ((v2/v1)^2) as
> long as R1 = R2 which equals 20 Log (v2/v1).
>
> In most modern audio usage this is a slightly dodgy thing to do as
> the amplifier spec. typically ignores the impeadence transformation
> (db is DEFINED in terms of POWER ratios). The 20 log thing only
> really works if input and output impeadences are equal. This is
> generally the case in RF, but not in audio where optimising power
> transfer is unimportant.
>
> Regards, Dan.
Thanks for the lesson, Dan. Also good to know that it probably
won't be as useful to me as other calculations. Good to have in the
repetoire anyways.
Curt Wells
> posts on the subject? If you set it as I described, and your system
works,
> doesn't blow up, and is nice n' loud, isn't that what you were after?
Ack!
That mighta been me, cousin. you gave me a link to a spreadsheet that I
haven't been able to use yet.
> Personally, I like to use the least amount of
> gain at the amps, and send them the hottest signal possible from the
> crossover, allowing for proper headroom at every point in the chain, to
> reduce the noise floor to a minimum, and when correctly set up, a signal
at
> the threshold of clipping at the mixer output should also be near the
> threshold at the EQ, the crossover, and the amps, when all is calibrated.
Glad to know I'm not nuts. I've had people look at me goofy when I set up
their PA with the power amps off so that as I get into the yellow, the
crossover input is just clipping, then I set up the crossover outputs to
unity and raise up the volume knobs on the amps bit by bit while a cd is
playing listening and changing crossover freqs here and there as I raise the
volume at the amps. keeping an eye out for clipping on the amps, I keep
raising the volume knobs until i either get a really loud room or an amp
starts clipping. if the amp clips before the system is loud enough, I
consider switching amps or changing the crossover freqs if I can get away
with that. I usually end up with the knobs on the amps not ALL MAXED and
some guys wonder whats up with that.
I'm a small time guy, and usually traveling from the foh to the amps is not
a big deal.
and I'm getting the new gear for the Star Bar real soon. I'll post when I
can.
Curt.
George Gleason
I do the same
set my gain structure throughout my system then only turn amps up to needed
volume and allow for headroom
George>
>
Sloth
George Gleason
"Sloth" <slo...@smartt.net.au> wrote in message
news:3CA6FD53...@smartt.net.au...
> Wouldn't that screw with your crossover points?
No but if I decide a adjustment needs making there is no reason not to make
the correction
it is really much less of a deal than the added noise of amps running full
on during quite passages in the program
George
>
Shaun
> No but if I decide a adjustment needs making there is no reason not to make
> the correction
> it is really much less of a deal than the added noise of amps running full
> on during quite passages in the program
That is a factor which is largely based on the cleanliness of your grid
power supply and noise picked up by the audio drive lines. Running over 30
kW of amps on my FOH rig, I usually have them all fully clockwise to 0 dB at
the front panel, with output levels calibrated by the BSS 366's, and there
is NO more noise than if their front panel gains were shut off completely
when set to -infinity. You actually have to press your ear up against the
cabinets to hear anything as a clue that there is power applied to the
system, and then its really only audible in the HF. 99% silent. When I get
into certain venues, the noise floor goes up considerably, and I have to
track down the source of it, which is usually attributed to where the drive
snake is placed, or the length of the drive snake (I run a separate 300'
drive line from FOH, but sometimes use shorter ones). I've found it picks
up less noise this way, rather than using returns in the main mic snake.
One venue has a bit of voltage on the ground from something in the building,
and that was the worst case of raised noise floor I've ever had to deal
with. Either way to set the amp's gain knobs is correct. It's just better
for most rigs to set all the amps to an easily repeatable front panel
attenuator position, so they can't be turned up beyond that point into
possible overload or imbalance. Also, if the gig doesn't require full
output, I'll use the amp gains to ensure a lower noise floor, but usually
doesn't matter if I do that, or just stay off the console's main fader.
Every rig is different, so YMMV. Whatever works and sounds the best.
George Gleason
"Shaun" <see.sign...@email.address> wrote in message
news:B8CC7F75.53D9%see.sign...@email.address...
Shaun ,Like yours, my FOH rig is dead silent , my monitor rig acceptably
quiet
but
when the rest of the rig is running happy the signal is a bit stronger than
my Powerlights like
and it does not help that the powerlights all have diffrent sensitivities
requiring some sort of gain compensation
could be at crossover
but I perfer to do it at amps as they do not require accessing dsp to effect
this function
George>
Dave Andrews
<< Running over 30
kW of amps on my FOH rig, I usually have them all fully clockwise to 0 dB at
the front panel, with output levels calibrated by the BSS 366's, and there
is NO more noise than if their front panel gains were shut off completely
when set to -infinity. >>
That's because the noise is not usually generated in the tail end of the signal
chain (unless there is a wiring or AC problem). Your DSP is doing the same
thing as George's sensitivity adjustment -- clamping down on the steady state
noise from the front end of the system. Run your DSP at wide open (O dB
reference in and out) with your amps wide open, connect your mixer with all
snakes attached, balance input trims for appropriate volume and then go listen
to the hum, hash, and everything else associated with several miles of mic
cable antennas attached to the system. Unless you're operating your rig in a
Faraday Cage, you got lots of noise, son.
For the record, George's technique is right on. Surprise, surprise -- St. Dave
the Troll goes with Gleason the Censor. By running your DSP's at well under
unity gain, you reduce your signal to noise ratio in that piece (unless it is
an analogue attenuation at the output stage). D to A and back is the cleanest
when signal is at it's maximum without risking clipping. Better to run the DSP
at unity gain in and out (or close to it) and drop the sensitivity on your
amps.
This message sponsored by Janitor in a Drum and Mr. Clean.
--
With All Due Respect,
Dave Andrews
D. W. Andrews Associates
Church Music System Specialists
"Two Hacks Working Out Of A Garage"
Disclaimer: If there are two ways to take my words,
always assume I was after the cheap laugh.
Peter Larsen
> Wouldn't that screw with your crossover points?
No, because getting the frequency band amplification gain settings
correct automatically also gets the acoustic cross-over slopes correctly
located.
Peter Larsen
> Either way to set the amp's gain knobs is correct. It's
> just better for most rigs to set all the amps to an easily
> repeatable front panel attenuator position, so they can't
> be turned up beyond that point into possible overload or
> imbalance.
The point of reduced amp input sensitivity == better loudspeaker ""stray
event"" survival chance stands. You rig appears to be an excellent
example of a situation where getting rid of adjustability is a good
thing, i. e. replacing those variable attenuators with fixed size
resistors OR if you really like the input sensivity as it is then either
bypassing those attenuators or simply ordering amps that do not have
any. Adjustability is not always a good thing. Looks to me like a good
rig to use the 5 volt input sensitivity ploy in - leaves you a bit of
remote adjustability and removes a source of error, even if only from
vibration.
> Also, if the gig doesn't require full
> output, I'll use the amp gains to ensure a lower noise floor, but usually
> doesn't matter if I do that, or just stay off the console's main fader.
Having an "event volume control" is not a bad thing, mostly when my
SM3000 did do PA (bought it afterwards) it was hooked up with FOH on the
monitor output so that the monitor faders functionally were FOH volume
control.
> Every rig is different, so YMMV. Whatever works and sounds the best.
> Shaun Wexler,
