Reducing hiss by changing op-amps ?

[John Woodgate]

I read in sci.electronics.design that [email protected] wrote (in

Not strictly true - I can notice the hiss *more* when I turn up the
input trim. The original problem is that the desk is too hissy - when
used at *any* level

It's used on stage for my band as it has better EQ (not to mention more
inputs !) than our previous PA head, and since we made the switch we've
noticed a distinct increase in hiss coming through the PA speakers -
which is coming from the mixer. I just wanted to see if I could reduce
the noise level as a lot of our numbers are quiet-ish and no doubt the
sound could be improved by some hiss reduction.

Then change the mixer stage op-amp to a TL072. That is a major source of
noise of the type you mention.

 

[John Popelish]

Not strictly true - I can notice the hiss *more* when I turn up the
input trim. The original problem is that the desk is too hissy - when
used at *any* level

It's used on stage for my band as it has better EQ (not to mention more
inputs !) than our previous PA head, and since we made the switch we've
noticed a distinct increase in hiss coming through the PA speakers -
which is coming from the mixer. I just wanted to see if I could reduce
the noise level as a lot of our numbers are quiet-ish and no doubt the
sound could be improved by some hiss reduction.

If you want to experiment a bit, you might get a few LT1124CN8 dual
opamps to try as replacements for the 4558 types. They have make
about half the noise of the 4558.

I have no idea what the specs are for the KTA970BL PNP transistors at
the mic front end, but I would probably also experiment with replacing
a pair of them with a pair of 2N5087. Keep in mind that the collector
and base leads are interchanged with this swap.

 

[Mac]

On Sat, 19 Mar 2005 10:51:01 -0800, pcmangler wrote:


[snip]

with a microphone connected to channel 1 (heavily wrapped up in cloth
in a silent room - as I have no resistors lying around !), ch1 fader
and master fader set to 0dB, pad switch OFF, input trim ZERO, I get
-68dB

As above, but with the input trim set to max gain, I get -37dB

Like I said, I don't know if any of this is helpful, but it's all I can
offer right now !

If you don't have any resistors, you could try shorting out the mic input,
just as an experiment. Then try it with the wrapped-up mic, and see how
big the difference is. Just a thought.

Cheers,

Kev.

--Mac

 

[Mark]

Not strictly true - I can notice the hiss *more* when I turn up the
input trim. The original problem is that the desk is too hissy - when
used at *any* level

It's used on stage for my band as it has better EQ (not to mention more
inputs !) than our previous PA head, and since we made the switch we've
noticed a distinct increase in hiss coming through the PA speakers -
which is coming from the mixer. I just wanted to see if I could reduce
the noise level as a lot of our numbers are quiet-ish and no doubt the
sound could be improved by some hiss reduction.

Can you try turning the gain DOWN on your PA amplifer, and trun the
gain UP on the mixer trim pots for each mic channel.

You may have a gain distribution problem.

Any resoanble mixer should be OK for PA applications if used correctly.

What kind of music are you performing?

Mark

 

[Pooh Bear]

Yes, but he only has the noise problem at high gain settings.

I was talking in general terms ( and pointing out that the mic pre uses an
op-amp ). No point in having noise when you needn't.


Graham

 

[Pooh Bear]

I read in sci.electronics.design that [email protected] wrote (in


Then change the mixer stage op-amp to a TL072. That is a major source of
noise of the type you mention.

The 072's not really that quiet by modern standards, John.

Without a thorough examination of the gain structure and thermal noise
contributions of various stages it's tricky to recommend specific areas that
need the most attention in terms of substitution..


Graham

 

[Pooh Bear]

If you want to experiment a bit, you might get a few LT1124CN8 dual
opamps to try as replacements for the 4558 types. They have make
about half the noise of the 4558.

I have no idea what the specs are for the KTA970BL PNP transistors at
the mic front end,

They'll be KEC's version of a 2SA970. Blue is the highest gain grade.

but I would probably also experiment with replacing
a pair of them with a pair of 2N5087. Keep in mind that the collector
and base leads are interchanged with this swap.

I wouldn't !

The 2SA970 is one of the quietest bipolar devices available for mic amps.


Graham

 

[Pooh Bear]

Not strictly true - I can notice the hiss *more* when I turn up the
input trim. The original problem is that the desk is too hissy - when
used at *any* level

It's used on stage for my band as it has better EQ (not to mention more
inputs !) than our previous PA head, and since we made the switch we've
noticed a distinct increase in hiss coming through the PA speakers -
which is coming from the mixer. I just wanted to see if I could reduce
the noise level as a lot of our numbers are quiet-ish and no doubt the
sound could be improved by some hiss reduction.

Just about *any* mixer will have better performance than typical 'PA heads'
that usually have dismal specs.

It sounds to me like you're not operating the mixer at an optimum signal
level.

What kind of amplifier are you connecting to and what type of input ? How
far do the mixer bargraphs light up when you're playing ?


Graham

 

[Pooh Bear]

If you want to experiment a bit, you might get a few LT1124CN8 dual
opamps to try as replacements for the 4558 types. They have make
about half the noise of the 4558.

NE5532s will do the job for a fraction of the price of the Linear Technology
part. Indeed, if you're having noise problems using 5532s - something is
seriously adrift !

NJM4580s are v quiet too.

Both devices are widely used in most current pro-audio gear. Higher current
consumption than the original parts though


Graham

 

[John Woodgate]

I read in sci.electronics.design that Pooh Bear

The 072's not really that quiet by modern standards, John.

Of course, but it's better than the 4558. And you aren't in favour of
going to extremes:

QUOTE

If you want to experiment a bit, you might get a few LT1124CN8 dual
opamps to try as replacements for the 4558 types. They have make
about half the noise of the 4558.

NE5532s will do the job for a fraction of the price of the Linear
Technology part. Indeed, if you're having noise problems using 5532s -
something is seriously adrift !
ENDQUOTE

I get the impression that you always want the last word. If so, go play
with Brasfield.

 

[Don Pearce]

I was talking in general terms ( and pointing out that the mic pre uses an
op-amp ). No point in having noise when you needn't.


Graham

In general terms, I agree. As for the mic pre, yes there is an op amp,
but its noise contribution is divided by the open loop gain of the
pair of input transistors. It would have to be very noisy indeed to be
significant in that configuration.

d

Pearce Consulting
http://www.pearce.uk.com

 

[Pooh Bear]

I read in sci.electronics.design that Pooh Bear


Of course, but it's better than the 4558. And you aren't in favour of
going to extremes:

Nope. Not necessary.

QUOTE

NE5532s will do the job for a fraction of the price of the Linear
Technology part. Indeed, if you're having noise problems using 5532s -
something is seriously adrift !

It was my suggestion to use 5532s. If just about every serious modern audio
mixer manufacturer can realise *very* good noise figures using them ( or
4580s ) - why use esoteric devices ?

ENDQUOTE

I get the impression that you always want the last word. If so, go play
with Brasfield.

Not interested.

I *do* know my stuff regarding audio however. Around 30 yrs of experience
of designing practical circuits used in real products.


Graham

 

[Pooh Bear]

In general terms, I agree. As for the mic pre, yes there is an op amp,
but its noise contribution is divided by the open loop gain of the
pair of input transistors. It would have to be very noisy indeed to be
significant in that configuration.

At *max gain* - yes. I'm very familiar with that configuration. The 'noise
floor' at min gain is usually set by the op-amp.

Graham

 

[John Woodgate]

I read in sci.electronics.design that Pooh Bear

Around 30 yrs of experience of designing practical circuits used in
real products.

Only 30. You youngsters!

 

[[email protected]]

I've put the user manuals for the Inkel mixing desk and the Carlsbro PA
head on http://homepage.ntlworld.com/anengineer/ if anyone wants a
look.

I am connecting the line level output from the desk to the mic inputs
on the PA head, with a simple potential divider built into the XLR plug
to attenuate to the 3.7mV level expected at the PA inputs (works fine).
I am driving the PA inputs well - 0dB on the mixer output meters. I
am confident it's not a gain issue, it's just a noisy desk !

I won't have the PA head here until Friday evening, so I'll do some
further investigation then (i.e. is the noise only on certain outputs
etc).

Many thanks to everyone helping !

Cheers,

Kev.

 

[John Woodgate]

I read in sci.electronics.design that [email protected] wrote (in

I am connecting the line level output from the desk to the mic inputs
on the PA head, with a simple potential divider built into the XLR plug
to attenuate to the 3.7mV level expected at the PA inputs (works fine).

Could it be JUST possible that effectively having two mic amps in tandem
is contributing a tiny bit to the noise? What resistor values are in the
potential divider?

 

[Pooh Bear]

I've put the user manuals for the Inkel mixing desk and the Carlsbro PA
head on http://homepage.ntlworld.com/anengineer/ if anyone wants a
look.

I am connecting the line level output from the desk to the mic inputs
on the PA head, with a simple potential divider built into the XLR plug
to attenuate to the 3.7mV level expected at the PA inputs (works fine).

Hmmm...... what value resistors are in your potential divider ?

Using a mic input is a really bad idea. Performance is limited by the noise
figure of the " PA head's " own mic amps.

It is easy to show using theory that this will *always* be worse than just
using the PA head's mic amps on their own ! No matter how good the mixer
is.

Also, the general spec - distortion etc of a PA head's mic amp is usually
awful, so you aren't seeing the full benefits of using a mixer.

I am driving the PA inputs well - 0dB on the mixer output meters.

That's good.

I am confident it's not a gain issue, it's just a noisy desk !

I really don't think the desk is to blame.

Doesn't the PA head have a line level - ish input ( like a stereo input for
example ) ? Or an insert point. That would produce far better results.


Graham

 

[[email protected]]

Hmmm...... what value resistors are in your potential divider ?

75K in series with both the live and the shield and 560 Ohms across
them. These figures (and configuration) were arrived at with the help
of the folks over in rec.audio.pro.

Using a mic input is a really bad idea. Performance is limited by the
noise figure of the " PA head's " own mic amps.
It is easy to show using theory that this will *always* be worse than
just using the PA head's mic amps on their own ! No matter how good
the mixer is.

I know, I know... :-( We're not gigging at the moment (band reshuffle
!) and have spent a bit on other gear recently, but a separate PA amp
is next on my shopping list. I can finally run in stereo then ! (I
just want to hear my stereo chorus pedal through the PA )

I really don't think the desk is to blame.

Perhaps it's not. I will do some more testing on Friday. I'm only
really basing my accusations of the desk on the fact that when we plug
into the PA head directly (mics and guitars), there's almost no hiss.
Bit of mains hum (diabolical electrics in the place we practice in !),
but no hiss. When we bring the mixer into play and plug into that,
there's a noticeable increase in hiss from the PA speakers when we're
not playing. I'm also presuming that this hiss will be 'polluting' the
sound when we are, - I can't hear it then, but I know it's there! We
have nice guitars and good mic's and I want it as clean as I can get it

Doesn't the PA head have a line level - ish input ( like a stereo
input for example ) ? Or an insert point. That would produce far
better results.

Yes, but none that I can route via the internal reverb of the PA head.
I need to use two mic inputs on the PA head, one with a bit of reverb
for the vocals & rhythm guitar, and the other one dry (I have a better
quality reverb in an effects pedal) for my guitar. I can't see any way
of routing a line level input on the PA head through it's internal
reverb - that would solve everything if I could !

http://homepage.ntlworld.com/anengineer/Carlsbro GDX7 Manual.pdf
(though sadly it's not a schematic).

Incidentally, the mixing desk has it's own 'reverb'. Possibly the
worst I have ever heard though ! It's a simple delay, more an echo
than a reverb (unless it's faulty and not supposed to sound like that
!). Once again, it's stuff I plan to sort when funds are available - a
decent outboard reverb unit - eventually.

But until throngs of adoring fans start throwing money at us ;-), we're
stuck with this mish-mash setup .

 

[John Woodgate]

I read in sci.electronics.design that [email protected] wrote (in

75K in series with both the live and the shield and 560 Ohms across
them. These figures (and configuration) were arrived at with the help
of the folks over in rec.audio.pro.

You are attenuating the mixer output by nearly 50 dB and then amplifying
up again. This is not good. I suppose you are feeding into the
low-impedance balanced input of the Carlsbro. You would do better to
feed into the high impedance unbalanced input, provided you
interconnecting cable is not more than about 2 m long (to minimise
interference pickup, not anything to do with frequency response. You
need a different attenuator pad: 22 kohms in series with the live,
nothing in series with the shield and 1 kohm across them.

I know, I know... :-( We're not gigging at the moment (band reshuffle
!) and have spent a bit on other gear recently, but a separate PA amp
is next on my shopping list. I can finally run in stereo then ! (I
just want to hear my stereo chorus pedal through the PA )


Perhaps it's not. I will do some more testing on Friday. I'm only
really basing my accusations of the desk on the fact that when we plug
into the PA head directly (mics and guitars), there's almost no hiss.
Bit of mains hum (diabolical electrics in the place we practice in !),
but no hiss. When we bring the mixer into play and plug into that,
there's a noticeable increase in hiss from the PA speakers when we're
not playing. I'm also presuming that this hiss will be 'polluting' the
sound when we are, - I can't hear it then, but I know it's there! We
have nice guitars and good mic's and I want it as clean as I can get it




Yes, but none that I can route via the internal reverb of the PA head.
I need to use two mic inputs on the PA head, one with a bit of reverb
for the vocals & rhythm guitar, and the other one dry (I have a better
quality reverb in an effects pedal) for my guitar. I can't see any way
of routing a line level input on the PA head through it's internal
reverb - that would solve everything if I could !

http://homepage.ntlworld.com/anengineer/Carlsbro GDX7 Manual.pdf
(though sadly it's not a schematic).

The Carlsbro doesn't have a proper line-level input, only an 'effects
return'.

Incidentally, the mixing desk has it's own 'reverb'. Possibly the
worst I have ever heard though ! It's a simple delay, more an echo
than a reverb (unless it's faulty and not supposed to sound like that
!). Once again, it's stuff I plan to sort when funds are available - a
decent outboard reverb unit - eventually.

Is it electronic or electromechanical? If the latter, it will certainly
have been damaged. Sometimes they can be repaired.

 

[[email protected]]

You are attenuating the mixer output by nearly 50 dB and then

amplifying up again. This is not good. I suppose you are
feeding into the low-impedance balanced input of the Carlsbro.
You would do better to feed into the high impedance unbalanced input

I opted for the low impedance mic input on the basis of advice in
rec.audio.pro where it was suggested to me to be wary of using the high
impedance instrument input as "sometimes these also contain fixed EQ to
sort-of simulate the tone shaping of a guitar amp; ie, a midrange
notch."

I therefore figured it was easier to just go with the mic inputs to
avoid this potential problem.

Is it electronic or electromechanical? If the latter, it will
certainly have been damaged. Sometimes they can be repaired.

Electronic. I'm guessing it's based around IC706 /707 on the power
board ( http://homepage.ntlworld.com/anengineer/1410_Schematic.pdf ).
Does this look like a very basic delay circuit, or do you think it
should sound like a proper reverb ?