Bridged Amplifier Sound Quality

Archived from groups: rec.audio.tech (More info?)

soinie wrote:
> Is there a sound quality downside in using an amplifier that is
> bridged? I've heard various things about this, not to mention that
> some of what I've read says speakers will sound better when the amp is
> bridged given the increase of its power. Thanks

The only downside I know of (as long as both amp channels are matched) is
that permissable load impedance is doubled...IOW, if your amp is rated for 4
ohms minimum, in bridged mode you can only connect an 8 ohm load.

jak

Archived from groups: rec.audio.tech (More info?)

"Mark D. Zacharias" <mzacharias@yis.us> wrote in message
news:35tk8vF4raatrU1@individual.net...
> Certainly noise (a constant value from each "1/2" amp channel), would
amount
> to 4 times the amount of a single-ended model. (The residual noise voltage
> is doubled.)

And so is the signal voltage, thus S/N ratio should be the same in both
cases.
In fact it's quite possible that because the signal is correlated and the
noise is not (or at least not as much) then the S/N ratio of a bridge amp
could be lower. Other factors may have more influence though.

MrT.

Archived from groups: rec.audio.tech (More info?)

Richard Crowley <rcrowley7@xprt.net> wrote:
> "Mark D. Zacharias" wrote ...
>> Certainly noise (a constant value from each "1/2" amp channel), would
>> amount to 4 times the amount of a single-ended model. (The residual
>> noise voltage is doubled.)

> Since noise is random and uncorrelated, does it really behave
> the same way as coherent signals?

No.

Archived from groups: rec.audio.tech (More info?)

R <spmaway@ylhoo.com> wrote in
news:Xns95EBEDF57961Dmc2500183316chgoill@10.232.1.1:

> "jakdedert" <jdedert@bellsouth.net> wrote in
> news:nrdKd.12006$ud3.4@bignews1.bellsouth.net:
>
>> soinie wrote:
>>> Is there a sound quality downside in using an amplifier that is
>>> bridged? I've heard various things about this, not to mention that
>>> some of what I've read says speakers will sound better when the amp is
>>> bridged given the increase of its power. Thanks
>>
>> The only downside I know of (as long as both amp channels are matched)
>> is that permissable load impedance is doubled...IOW, if your amp is
>> rated for 4 ohms minimum, in bridged mode you can only connect an 8 ohm
>> load.
>>
>> jak
>>
>>
>
> That depends on the amp.
> I know of at least 2 amps where the opposite is true. IOW, when in
> normal mode impedance ranges are from 16 ohms to 0.5 ohms. When in
> bridge mode the impedance range is 8 ohms to 0.25 ohms.
>
> r

I missed a paste from the owners manual of one amp.


-------------------------------------------------
POWER OUTPUT
STEREO: 300 watts continuous, both channels operating;
which is:
12.2 volts RMS across 0.5 ohm
17.3 volts RMS across 1 ohm
24.5 volts RMS across 2 ohms
34.6 volts RMS across 4 ohms
49.0 volts RMS across 8 ohms
69.3 volts RMS across 16 ohms
MONO: 600 watts continuous; which is:
12.2 volts RMS across 0.25 ohm
17.3 volts RMS across 0.5 ohm
24.5 volts RMS across 1 ohm
36.6 volts RMS across 2 ohms
49.0 volts RMS across 4 ohms
69.3 volts RMS across 8 ohms


more explicitly stated later....

OUTPUT IMPEDANCE
Stereo: 0.5, 1, 2, 4, 8, and 16 ohms
Mono: 0.25, 0.5, 1, 2, 4, and 8 ohms


r

Archived from groups: rec.audio.tech (More info?)

"Mark D. Zacharias" <mzacharias@yis.us> wrote in message
news:35tk8vF4raatrU1@individual.net

> Certainly noise (a constant value from each "1/2" amp channel), would
> amount to 4 times the amount of a single-ended model. (The residual
> noise voltage is doubled.) Noise qualifies as distortion. Certain
> other types of distortion would effectively be higher as well, I
> would presume.
> I'm not an engineer, I should hasten to add, but someone out there
> will surely add to this thread.

Depends on the nature of the noise. If it is plain old random noise, then
the noise sources in the two power amps are uncorrelated. That means that
they sum geometrically (square root of sum of squares). If they are equal
amplitude, then they sum to the square root of two times either.

As others have pointed out, the signals sum with ordinary arithmetic. So,
dynamic range *increases* by about 3 dB.

Archived from groups: rec.audio.tech (More info?)

I have built quite a number of amplifiers -- one of the easiest to bridge is
that based upon National Semi's "Overture" series amp chips -- if careful
attention is paid to layout, the distortion specs can be quite good.

the PDmax for the devices used in a bridged amplifier is increased by a
factor of 4 -- so there can be quite a lot of thermal stress.

I will post on ABSE a picture of the THD vs frequency for an LM4780 based
bridged amplifier.


"Arny Krueger" <arnyk@hotpop.com> wrote in message
news:IZadnaZf7fmY0GfcRVn-pw@comcast.com...
> "soinie" <soiniie@hotmailll.com> wrote in message
> news:itmiv0l8ve2jpoprtmf71cbos2dim3orsb@4ax.com
>
>
> > Is there a sound quality downside in using an amplifier that is
> > bridged?
>
> Essentially, bridging is generally used as a means to increase the amount
of
> current the amplifier provides to a given load. This increase is obtained
> at the significant cost of halving the number of channels.
>
> If you bridge an amp, the availble output signal voltage is roughly
doubled.
> Therefore, the current delivered to a reasonably linear load is increased
> proportionately. Current drain is related to stress on the amplifier, so
> bridging necessarily increases the stress on the amplifier. Nonlinearity
in
> the amplifier is related to stress provided by the load, so bridging
> necessarily increases the nonlinearity of the amplifier.
>
> > I've heard various things about this, not to mention that
> > some of what I've read says speakers will sound better when the amp is
> > bridged given the increase of its power.
>
> Bottom line, the sound of an amplifier running out of power, or clipping
is
> pretty nasty. Avoiding clipping is of paramount importance in high
fidelity
> applications.
>
> Modern amplifiers are generally quite linear when run below clipping.
> Therefore, additional nonlinearity due to bridging can often be quite
> acceptable and unnoticeable from the standpoint of listening pleasure.
>
> Modern amplifiers are generally congenitally overbuilt for high fidelity
> purposes because of FTC regulations. Therefore, they may be more suitable
> for bridged applications than a cursory examination of their specs might
> suggest. This varies from rock sound reinforcement and certain car audio
> purposes where amplifiers are more likely to be congenitally clipped,
> overloaded and overstressed.
>
> It comes down to practicality and economics. When you bridge a power amp,
> you double the number of amplifiers required for the same number of
> channels. If a correspondingly more powerful amplifier has
disproportionate
> higher cost, then bridging can be economical. If you already have a
> bridgeable amplifier, it generally costs less to double the number of
> amplifiers, than it does to get all new, more powerful amplifiers.
>
> OTOH, if a more powerful amplifier has a vastly superior
price/performance,
> then it can be more economical to simply get bigger amplifiers.
>
>

Archived from groups: rec.audio.tech (More info?)

R wrote:

>>The only downside I know of (as long as both amp channels are matched)
>>is that permissable load impedance is doubled...IOW, if your amp is
>>rated for 4 ohms minimum, in bridged mode you can only connect an 8 ohm
>>load.
>>
>>jak
>>
>>
>
>
> That depends on the amp.
> I know of at least 2 amps where the opposite is true. IOW, when in normal
> mode impedance ranges are from 16 ohms to 0.5 ohms. When in bridge mode
> the impedance range is 8 ohms to 0.25 ohms.


In that case the bridging is done by shorting the left and right
channels together at both input and output (with accurate balance
of gain and offset, and appropriate resistors to limit current).
so the voltage gain is the same and current capacity doubled.

Normal bridging doubles the voltage gain and leaves the
current capacity unchanged.

--
Eiron.

Archived from groups: rec.audio.tech (More info?)

Mark D. Zacharias wrote:
> The signal voltage for any given listening volume would be the same.
> The noise would still be doubled in voltage, hence 4X it's
> contribution to the distortion figure.
>This seems fairly obvious.

It might seem obvious, but it's still wrong, nonetheless.

First, the definition of "distortion" is not agreed upon. If
you mean the added components to the signal as a result of the
non-linear properties of the amplifier, whoichm BTW is the
commonly accepted definition of dfisortion when used in reference
to audio amplifiers, no, it does NOT contribute to the distortion
figure).

But, specifically to the contribution of noise:

Unless by some miracle the noise voltage at each and every instant
in both amplifiers is identical in amplitude, the noise WILL not
and CAN not add as you claim. Indeed, assuming you're talking about
noise generated in each of the amplifier channels, that noise will
be completely uncorrelated.

For those that have studied such topics, it's in fact obvious
that two uncorrelated noise sources of the same RMS level,
adding the noise voltages WILL NOT result in double the noise
voltage, but only the sqrt(2), precisely because they are
uncorrelated.

> As I said, I'm not an engineer,
Yes, the evidence before us supports your assertion.

Archived from groups: rec.audio.tech (More info?)

On Sat, 29 Jan 2005 10:02:48 +1100, "Mr.T" <MrT@home> wrote:

>
>"Mark D. Zacharias" <mzacharias@yis.us> wrote in message
>news:35ukieF4pdkh2U1@individual.net...
>> The signal voltage for any given listening volume would be the same. The
>> noise would still be doubled in voltage, hence 4X it's contribution to the
>> distortion figure. This seems fairly obvious.
>
>To you maybe, why don't you do some actual measurements and get back to us.
>Fact is, that to maintain the same signal level, we can reduce the gain, and
>hence the noise drops. The S/N ratio remains the same or better, as I said.

In a typical amplifier, the output voltage will increase by 6dB, but
the output noise will only increase by about 3dB, due to it being
uncorrelated (random, if you like). In some designs, you might even
gain a little more S/N if you sendipitously cancel PSU components by
the use of a differential system such as bridging.

>You should also know that noise is NOT distortion, That is why such a
>measurement is called Noise *AND* Distortion.
>A *PROPER* THD or IMD measurement does NOT include noise outside of the
>measured frequency bands. This is pretty easy to do correctly these days
>with DSP techniques.

Um, well, there's some debate on that subject, as some authorities
maintain that anything not the desired signal is *by definition*
'distortion', and that of course includes noise, bith thermal and
PSU-related.

>> Let's hear from Krueger, Pinkerton and the rest. They'll be able to tell us
>> what other distortion issues arise from bridging. As I said, I'm not an
>> engineer, and if I stray too far into area of theory, I'm certain to get my
>> ass handed to me.

The issues which areis eare all to do with the fact athat the
effective load impedance is halved. All amplifiers have higher
distortion into low impedance loads.

>A better method is to do some research of your own, until you understand
>what is being said.

Inded, but these newsgroups can provide handy shortcuts! :-)
--

Stewart Pinkerton | Music is Art - Audio is Engineering

Archived from groups: rec.audio.tech (More info?)

On Fri, 28 Jan 2005 05:23:31 GMT, R <spmaway@ylhoo.com> wrote:

>"jakdedert" <jdedert@bellsouth.net> wrote in
>news:nrdKd.12006$ud3.4@bignews1.bellsouth.net:
>
>> soinie wrote:
>>> Is there a sound quality downside in using an amplifier that is
>>> bridged? I've heard various things about this, not to mention that
>>> some of what I've read says speakers will sound better when the amp is
>>> bridged given the increase of its power. Thanks
>>
>> The only downside I know of (as long as both amp channels are matched)
>> is that permissable load impedance is doubled...IOW, if your amp is
>> rated for 4 ohms minimum, in bridged mode you can only connect an 8 ohm
>> load.
>>
>> jak
>>
>>
>
>That depends on the amp.
>I know of at least 2 amps where the opposite is true. IOW, when in normal
>mode impedance ranges are from 16 ohms to 0.5 ohms. When in bridge mode
>the impedance range is 8 ohms to 0.25 ohms.

That would be a misprint......................
--

Stewart Pinkerton | Music is Art - Audio is Engineering

Archived from groups: rec.audio.tech (More info?)

Mr.T <MrT@home> wrote:

> Not really. THD stands for Total HARMONIC Distortion, so anything not
> harmonic by definition is not included.
> IMD stands for Intermodulation Distortion, etc.

That's the abstract definition of an <idea>. Whan actually using a null
analyser, it doesn't know the difference, and unless you do spectrum analysis
on the results, what ever is left after the null is what you get, and a
percentage # compared to the total output is produced.

By itself, it's a pretty useless test, except maybe for production quality
control purposes and the most basic of design specifications.

Archived from groups: rec.audio.tech (More info?)

Stewart Pinkerton <patent3@dircon.co.uk> wrote in
news:q5d40198t722diu1t4v1t415r94igg0upp@4ax.com:

> On Sat, 29 Jan 2005 04:01:15 GMT, R <spmaway@ylhoo.com> wrote:
>
>>Eiron <e1ron@hotmail.com> wrote in news:35vn04F4sil51U1@individual.net:
>>
>>> R wrote:
>>>
>>>>>The only downside I know of (as long as both amp channels are
>>>>>matched) is that permissable load impedance is doubled...IOW, if your
>>>>>amp is rated for 4 ohms minimum, in bridged mode you can only connect
>>>>>an 8 ohm load.
>>>>>
>>>>>jak
>>>>>
>>>>>
>>>>
>>>>
>>>> That depends on the amp.
>>>> I know of at least 2 amps where the opposite is true. IOW, when in
>>>> normal mode impedance ranges are from 16 ohms to 0.5 ohms. When in
>>>> bridge mode the impedance range is 8 ohms to 0.25 ohms.
>>>
>>>
>>> In that case the bridging is done by shorting the left and right
>>> channels together at both input and output (with accurate balance
>>> of gain and offset, and appropriate resistors to limit current).
>>> so the voltage gain is the same and current capacity doubled.
>>>
>>> Normal bridging doubles the voltage gain and leaves the
>>> current capacity unchanged.
>>>
>>
>>The amp in question can do either.
>>
>>One can indeed simply tie inputs and outputs together and is called
>>mono-parallel. The other method runs one channel through an inverter
>>and the resultant signal 180 degrees out of phase with respect to the
>>other and is called mono-bridged.
>>
>>The output impedance is halved for parallel and doubled for bridged.
>>
>>The output power is doubled no matter which method you choose.
>
> Nope, given adequate current reserves, the output power is
> *quadrupled* for bridging, and remains the same for parallel
> operation. If you're talking about some POS valve jobby, then of
> course the output transformer taps make the whole thing moot.

Don't believe me. I don't care. No, it is not some "POS valve jobbie"
either. Please remember that I said that it "depends on the amp".

The owners manual certainly would indicate my statement to be true. In
addition they have made the same "error" in their 600 watt and 500 watt
versions as well as one of their current amp offerings.

Here is the link to the owners manual for the 300 watt version.
http://berners.ch/McIntosh/Downloads/MC2300_own.pdf

Their current offering indicates that the power is doubled when in
parallel.

http://www.mcintoshlabs.com/mcprod [...] C402om.pdf

In case you don't have acrobat the specifications say:

"Power Output Stereo
Minimum sine wave continuous average power output per
channel, all channels operating is:
400 watts into 2 ohm load
400 watts into 4 ohm load
400 watts into 8 ohm load

Power Output Mono Parallel
Minimum sine wave continuous average power output is:
800 watts into 1 ohm load
800 watts into 2 ohm load
800 watts into 4 ohm load"

I should have qualified my last statement to say "The output power is
doubled no matter which method you choose with this amp."

r
--
Nothing beats the bandwidth of a station wagon filled with DLT tapes.

Archived from groups: rec.audio.tech (More info?)

Stewart Pinkerton <patent3@dircon.co.uk> wrote in
news:98r7011u3ufr7906j9gnalsedvssprgaba@4ax.com:

> On Fri, 04 Feb 2005 09:56:03 GMT, R <spmaway@ylhoo.com> wrote:
>
>>Stewart Pinkerton <patent3@dircon.co.uk> wrote in
>>news:q5d40198t722diu1t4v1t415r94igg0upp@4ax.com:
>>
>>> On Sat, 29 Jan 2005 04:01:15 GMT, R <spmaway@ylhoo.com> wrote:
>>>
>>>>Eiron <e1ron@hotmail.com> wrote in
>>>>news:35vn04F4sil51U1@individual.net:
>>>>
>>>>> R wrote:
>>>>>
>>>>>>>The only downside I know of (as long as both amp channels are
>>>>>>>matched) is that permissable load impedance is doubled...IOW, if
>>>>>>>your amp is rated for 4 ohms minimum, in bridged mode you can only
>>>>>>>connect an 8 ohm load.
>>>>>>>
>>>>>>>jak
>
>>>>>> That depends on the amp.
>>>>>> I know of at least 2 amps where the opposite is true. IOW, when in
>>>>>> normal mode impedance ranges are from 16 ohms to 0.5 ohms. When in
>>>>>> bridge mode the impedance range is 8 ohms to 0.25 ohms.
>
> Nope, that statement remains wrong for the amps you mention, because
> the reference is to *parallel* operation, not bridged.
>
> BTW, you are attempting to use a probably unique range of amplifiers
> (SS with output transformers) to 'prove' a point which is simply *not*
> true for 99.99% of available SS amps. And you still got it wrong!
>
>>>>> In that case the bridging is done by shorting the left and right
>>>>> channels together at both input and output (with accurate balance
>>>>> of gain and offset, and appropriate resistors to limit current).
>>>>> so the voltage gain is the same and current capacity doubled.
>>>>>
>>>>> Normal bridging doubles the voltage gain and leaves the
>>>>> current capacity unchanged.
>>>>>
>>>>The amp in question can do either.
>
> Nope, those amps are *not* intended for bridged use, and have no such
> connection option. It's quite possible that an attempt to engage
> bridged operation, by series connecting the output transformers as one
> would for a valve amp, would destroy the amplifier.
>
>>>>One can indeed simply tie inputs and outputs together and is called
>>>>mono-parallel. The other method runs one channel through an inverter
>>>>and the resultant signal 180 degrees out of phase with respect to the
>>>>other and is called mono-bridged.
>>>>
>>>>The output impedance is halved for parallel and doubled for bridged.
>>>>
>>>>The output power is doubled no matter which method you choose.
>>>
>>> Nope, given adequate current reserves, the output power is
>>> *quadrupled* for bridging, and remains the same for parallel
>>> operation. If you're talking about some POS valve jobby, then of
>>> course the output transformer taps make the whole thing moot.
>>
>>Don't believe me. I don't care. No, it is not some "POS valve jobbie"
>>either. Please remember that I said that it "depends on the amp".
>
> It's SS, but it uses output transformers, hence my comments apply.
> That (very unusual, posibly unique) series has the same speaker
> coupling system as valve amps, hence the same rules are in force. I
> "don't believe you" because you are just plain wrong, and clearly
> don't understand bridged operation.
>
>>The owners manual certainly would indicate my statement to be true.
>
> No, it wouldn't. Please learn to read clear specifications.

The point I made was simply this: Depending on the amp and the load, one
could have ones cake and eat it too. I do have to admit their 500 and 600
watt versions are a rather unique amp in that it can drive a 4 ohm load to
full power no matter if it is in parallel or bridged mode. Was it a bit
of a cheat? Yes, because there are very few SS amps with an output
transformer (actually autoformer in this case).

>
>> In
>>addition they have made the same "error" in their 600 watt and 500 watt
>>versions as well as one of their current amp offerings.
>>
>>Here is the link to the owners manual for the 300 watt version.
>>http://berners.ch/McIntosh/Downloads/MC2300_own.pdf
>>
>>Their current offering indicates that the power is doubled when in
>>parallel.
>
> Sure - but not into the same load on the same transformer taps. Since
> you don't seem to understand the basics, this means that while it can
> put out 600 watts into 4 ohms, it does this by using the 8 ohm taps in
> parallel operation. Hence, if you had quoted the *full* output specs,
> you'd see that the 'power doubling' does *not* apply to a 15 ohm load,
> the highest available tap.

I understand quite well thanks and yes, one won't realize full power using
a 16 ohm load when in parallel and one certainly won't realize full power
transfer when the load is not connected to the appropriate tap.


>
> OTOH, my trusty Krell is rated at 200 watts/channel into 2 ohm loads,
> and at 800 watts into the same 2-ohm load when bridged.
>

The Krell is like most other SS amplifiers and the behaviour you cite is
what I would expect. My old Carver amp exhibited that same behaviour.


>>http://www.mcintoshlabs.com/mcprod/..%5Cdata%5Cmanuals%5CMC402om.pdf
>>
>>In case you don't have acrobat the specifications say:
>>
>>"Power Output Stereo
>>Minimum sine wave continuous average power output per
>>channel, all channels operating is:
>>400 watts into 2 ohm load
>>400 watts into 4 ohm load
>>400 watts into 8 ohm load
>>
>>Power Output Mono Parallel
>>Minimum sine wave continuous average power output is:
>>800 watts into 1 ohm load
>>800 watts into 2 ohm load
>>800 watts into 4 ohm load"
>
> See above for another example of unequal comparisons re load
> impedance.


I am not sure what you mean.

>
>>I should have qualified my last statement to say "The output power is
>>doubled no matter which method you choose with this amp."
>
> Unfortunately, there is only *one* method, since bridging is not a
> recommended option. Learn to RTFM..........

There is a switch on their 500 watt and 600 watt versions that has 3
positions. Stereo, mono-parallel, and mono-bridged. I must admit that
they don't have the ability to run in bridged mode with their latest
offerings, but they certainly did in earlier models. It is also apparant
that the requirements for their 500 and 600 watt designs was slightly
different than with their later amps.

There is more than one disadvantage to having an autoformer in an
amplifier. One being the additional weight and the other being the
additional cost. The advantages however are obvious and far outweigh the
disadvantages in my opinion.

r

Archived from groups: rec.audio.tech (More info?)

Mr.T <MrT@home> wrote:

> <jjnunes@sonic.net> wrote in message
> news:7cyMd.5179$m31.67153@typhoon.sonic.net...
>> Mr.T <MrT@home> wrote:
>>
>> > Not really. THD stands for Total HARMONIC Distortion, so anything not
>> > harmonic by definition is not included.
>> > IMD stands for Intermodulation Distortion, etc.
>>
>> That's the abstract definition of an <idea>. Whan actually using a null
>> analyser, it doesn't know the difference, and unless you do spectrum
> analysis
>> on the results, what ever is left after the null is what you get, and a
>> percentage # compared to the total output is produced.

> Maybe you should read what I wrote originally. You are referring to S/N+D
> measurement.
> Spectrum analysis via DSP FFT is the more common technique for THD and IMD
> these days.

I don't disagree with you, but was thinking of the more global
context, which still can be usefull in the aggregate under some
circumstances.