Bi-Amping Question (RF-83s Driven by Denon AVR-2807)

[DrWho]

But it does matter that they all share the same power supply...

There

is simply no way you can reduce the power draw on the supply for the

same SPL by splitting the signal up to multiple amps. In other words,

conservation of power.

Agreed--you can't reduce

the power draw on the supply simply by splitting to multiple amps. I

think his point was that the power supply in this receiver can dump

ungodly amounts of power into the amps, but it's the amps that are the

limiting factor. In other words, let's say that you're playing only

through the amp driving the fronts and the rest aren't connected. Does

this mean that my front amp can take up all the power that's available

to the other amps? (110Wpc x 7 channels = 770W, 770W / 2 channels =

335Wpc...by unplugging all the other outs does my receiver suddenly

become a 335Wpc stereo receiver?) No...according to what he said if I'm

getting him correctly, the circuitry comprising the amp is rated to

continuously handle 110Wpc. Even though the power supply can provide it

far more, it will overheat and start degrading performance.

But it sounds like you know what you're talking about, so I'm

interested to hear what you think. I'm not necessarily inclined to

believe this guy simply because he worked for Denon--lord knows the

first level customer support rep I spoke to didn't know much.

:-)

Ok, well that makes sense....but let me throw out another concept []

Ideally, an amplifier is a perfect voltage source which means it can

supply the necesary current to keep the voltage at the desired level.

So if you have an 8 ohm load, then you'll have a corresponding current

draw from the receiver for that voltage. Change it to 4 ohm and you'll

now have double the current. Voltage stays the same so effectively it's

twice the power.

I mention this because there are two things that limit an

amplifier's output, which I like to describe as the mechanical and

thermal limitations (just like our speakers). The mechanical limit

would be the maximum voltage the amplifier can provide. The thermal

limit would be the maximum amount of current that the receiver can

provide before it melts (or the voltage drops). This is why we have

peak output and continuous output ratings. One dictates the mechanical

limitations and the other the thermal limitations.

Now take your typical 2-way 8ohm speaker. When your amp is putting

out 110W, it's really putting out 30V and just happens to be able to

supply the current necessary to maintain that voltage (3.75 Amps into

the 8ohm load). Now let's say you power the tweeter with this same amp

and the woofer with another of the same amp. The tweeter section is now

an 8ohm load by itself as is the woofer section. Since both amps are

the same and the load to each amp is still 8 ohms, you still have 30V

on the output of each amp, which means you've still got 3.75 amps of

current draw. Granted, this current draw is only happening over half the

passband, but each amp is still working just as hard...the voltage and

currents are just happening at different frequencies.

The important point from this overly simplified model is that the two

amp system is going to clip at the same time as the one amp system (which is only when the amps and the loads are the same, which is always true

when you're using the passive crossover of the speaker). To illustrate

the point, let's say your amp has a voltage gain of 30. This means that

you need 1V at the input to achieve 30V at the output. Let's say you

deliver 1.2V to the input of each amp, say from the spike of a huge low

frequency explosion. Since your amps can't put out 36V, they're going

to clip and deliver 30V until the input signal drops under 1V. This

"square wave" on the output of each amplifier is actually going to be

more audible on the tweeter than the woofers and you still risk blowing

up the tweeters. This would be the mechanical limit of the amplifiers.

Now if you move to an active crossover setup, then you can do two

things. First, you can remove the passive network in the speaker which

will usually increase the output of the tweeter (especially in 2-way

systems with a horn on top and woofers on the bottom). Secondly, you

can divide the bandwidth that each amp operates over - which means them

low frequency transients aren't going to clip the tweeter. Likewise,

the high-frequency transients aren't going to saturate the output of

the woofer amp.

I didn't comment on the thermal limits above, but I didn't want to

confuse the point. Ideally, modern circuit design involves correction

circuits that will counter the non-linear behavior due to temperature.

This is why old solid state amps generally sound like crap when you

crank them up. These circuits are often over-engineered, moreso when

the amp is going to be expected to drive multiple impedances. For

example, an amplifier that can drive 4ohm impedances will be able to

linearly handle up to twice the current for the same voltage. And even

then, I'm sure it's going to be over-built to handle 2 Ohm impedances

just in case....(but maybe not at the same output levels).

Nevertheless, we're rarely driving the amp at these levels continuously

over time so the heat build-up issue isn't really a concern. It's an

exponential thing, so as long as you're below the knee of the temperature response then it's not gonna be changing much.

You mentioned that you were talking to an engineer from Denon. I just

wanted to mention that I'm an electrical engineering undergrad - so I'm

young and probably more prone to believing in idealistic theories and

definitely lack that hands on experience.

Ok, now that is extremely misleading. Yes, the

skin-effect exists, but not at the frequencies our music is playing at.

Also, the behavior of one frequency range isn't going to affect the

behavior of another frequency range - specifically, any skin-effect

that might happen with the highs has absolutely nothing to do with the

lower frequencies. In fact, if that were possible, then the entire

world of electrical models and analysis would be completely invalid

[] Were you talking with an engineer or tech support? There is an

entire field of study for "transmission lines" and I think you'd be

hardpressed to make the math show these connclusions.

So this is a definite point of disagreement. This was an engineer I

was talking to who works in the lab, not a customer support rep. He

said that the skin effect is a function of two variables: cable gauge

and frequency. Signals traveling through thicker cable use flow along a

higher cross-sectional area than the same signals traveling through

lower gauge cable. Also, lower frequencies penetrate more deeply to the

center of the cable than highs, and when highs are present the low

frequencies are constrained to the cross-sectional area of the cable

occupied by the highs.

I was inclined to believe him simply because he seemed knowledgable

about other areas, but if there's one thing I've learned since starting

my foray into high-end audio it's that the skin effect is one of the

least understood aspects. Since I'm not in a lab doing my own research

with an understanding of the fundamentals, my mind is totally open. You

seem to be saying that in the 20Hz-20kHz range the skin effect is

negligible altogether, which I will now file under "definite

possibility" next to the info the Denon engineer told me. :-)

If you ever get the chance to talk to this guy again, you should ask him for some specific numbers []

For what it's worth, the science behind skin-effects is very well

understood. The problem with high-end audio is that science is rarely

the basis for understanding. The skin-effect plays an extremely

important role with high frequency devices (like MHz and GHz). In fact,

there are even some applications that take advantage of this effect.

One of these days I should crank through the physical model at audio

frequencies. There is a reason though that high-frequency circuit

designers refer to everything under 100kHz as DC...

"Better"? Sure, but let's not ignore diminishing

returns. If insanely larger guages were indeed necessary, then the

engineers designing the amplifiers and speakers would have done one of

two things:

1) Chose a different electrical interface to allow smaller wires

2) Not sell smaller wire, and build the equipment to naturally accept much larger guages

The

fact of the matter is that the engineers already chose an interface

that allows smaller wires and they build the equipment to accept large

enough cable. I won't argue the "audibility of better wire", but I'll

argue that it's completely pointless in light of a million other

variables. What does +-0.01dB matter inside a system of +-20dB?

I'm in total agreement with you here--I was simply reporting this

part of the conversation. My own personal belief on wire is that

thicker is better, but cheap is king. I'd be running my RF-83s off of

16 gauge zip cord if I hadn't found 30' of 8 gauge Monster cable at

Fry's for $40. I know a guy that's got his system wired up with

firehose on cable elevators, and it cost him probably $1k to do it. Can

I hear the difference between that and 16 gauge? Nope. Can I hear the

difference between my $40 bi-amp'd solution over the single-wire,

single-amp 16 gauge solution I had before? Yes, but only in imaging on

the high frequency components during complex, loud orchestral passages.

I may have gotten the same effect simply by bi-wiring or upgrading to

the 8 gauge cable in a single-wire configuration...I didn't try it so I

don't know.

The main reason I'm even bothering with bi-amping is

simply to protect my speakers. Like you say, avoiding clipping is more

important than anything else, and that's all I'm aiming to do.

Nothing wrong with using a feature that is there and not being used.

I hope my comments aren't coming across as critique. I'm all for using

our own ears as the final measuring stick. As mentioned eariler, I do

question wether or not you're less likely to clip.

I agree, but it's not fair to consider the worst case active

crossover scenario versus the best case anything else. Active

crossovers do indeed take a little extra work to be done correctly, but

when done correctly are absolutely the best approach. I bet if you were

seriously considering it, you might be able to find a Klipsch engineer

willing to provide you with all of the numbers you'll need to drop into

your active crossover. Otherwise, you're going to need to use measuring

equipment.

Yes, this is essentially what he told me--the

best case active crossover will be an improvement. I think he gathered

from earlier in our conversation that I don't have the commitment (in

terms of time and money) to achieve that, though, so by doing a

half-wit job of it, I'd simply spend money and end up with worse sound.

In this respect, he was a wise, wise man. :-)

lol [Y]

I'm not sure what you mean here. I believe in the

approach of starting with speakers that can play as loud as you want

them to and then making sure you have an amp that can drive them to

those levels within its linear operation. Generally, it's safer to

overdrive your speakers than to overdrive the amp (since a clipping amp

wreaks havoc on tweeters).

Anyways - that's just my take on it all. There's no shortage of different opinions out there.

Yes,

I think we're in agreement on this last point as well. It's *safe* to

vastly overdrive speakers (provided you don't crank the volume beyond

the limits of the speaker), whereas underdriving speakers means

clipping, which means speaker death. However, I think his only point

was that leaving excess circuitry out of the equation will result in a

cleaner signal. Why put all the extra circuitry in the pathway that can

amplify to thousands of watts when your speakers can't handle thousands

of watts? The extra circuitry in this case can never provide a benefit,

but you're still making sonic trade-offs for the capability which will

never be tapped. This makes sense to me at a gut level--doesn't it

somehow seem inelegant to have vastly untapped potential at any stage

in a system? That attitude feels right to me (which is why I question

my friend, who has a $2k receiver capable of dumping huge amounts of

power into Bose lifestyle speakers that are tiny and don't consume that

much power...the system sounds good, don't get me wrong, but it

probably would sound every bit as good had he saved a grand or so on

that receiver, no?).

I don't believe "excess circuitry" equates to worse sound, but I

guess it depends on how you define excess. I prefer to believe that

inefficient engineering equates to worse sound - so if that means using

more parts than are necessary to achieve a desired function, then yes

that will probably sound worse. But at the same time, higher power

amplifiers are usually the result of using better parts - not so much

changing the design. It is my understanding thus far that engineering

really isn't about being able to find ways to do something - that's the

easy part. I get the impression that engineering is more about finding

more efficient ways, which ultimately means achieving the same

performance with less cost. But things also need to be kept in

perspective....audio amplifier technology vastly improved over the last

40 years...to the point that amplification really isn't the bottle neck

anymore. For example, what does +-0.1% distortion matter in light of

the +-20% distortion in our speakers, and the +-20dB influences of our

room? Your friend, for example, probably won't realize the full

potential of his amp until he upgrades to better speakers - and even

then, he would probably be better off treating his room.

[jaydaniels]

I've got to run, but #2 is pretty much universally untrue unless you're listening to blisteringly loud violin solos. Much of the muscial energy is in the lower registers. Witness any PA system for any touring act (AC/DC through Pavarotti) and you'll find racks of power amps and speakers, most of which are used for the bass frequencies. Even a small band with a bi or tri amped PA might have a few thousand watts for the bass, a few hundred for the mids and maybe 100-200 for the highs.

I generally have to take Docs' word on things. Not to poo-poo experiementing the Bi-amping, Heck my new yammie 2600 will allow it, I'm wondering if I can use the onboard electronic crossover to use the onboards to drive my mid/hi horns on my LS and then route the pr outs to my Crown and the bass section just for kicks.

And YES the power supply is a limiting factor. Just look at how most major manufacturers rate their RMS wattage. Is it with ALL SEVEN CHANNELS driven? Usually not. They get that 110 watt or 170 or whatever by maxing out only one channel. That is why realists know that most modern HT receivers, when taxed with loads on 5 or 7 channels, will probably only put out about 80% of the rated wattage. There is just no way a home HT receiver is gonna pump out 1000 watts of clean power, aint' going to happen.

Maybe this explains why harman kardon amps are typically rated so low on the wattage level. They rate their AVR receivers with all channels driven.

[pauln]

What about the residual Doppler distortion in the tweeter signal from highpassing the full range?

I'm
still thinking about it, but this does not sound right to me. Maybe the
way I imagine the superposition of waves is more linear...

[etc6849]

What Dr. Who says makes a lot of sense. However, I finally tried this using the internal speaker cross-over and bi-amping via an Onkyo 905. On the 905 the manual recommends connecting the tweeter to the front jacks and the woofer to the surround back jacks. I'm wondering why it would matter unless the 905 is filtering the jacks when I switch to bi-amp? The 905 is a 7.1 equal power amp; however, I am inclined to agree with Dr. Who that the woofers should be connected to the front jacks.

Anyways, I used the same cable with the same length from the same spool. The only difference is I was lazy this time and didn't solder the banana plugs on the cable going to the woofers. I removed the RF-83's jumpers. I can clearly hear a difference in the highs. Much more detail than before. A lot of people are going to think I'm nuts, but I can't believe how much better the RF-83's sound. The muddiness I would get when a bunch of symbols were playing is completely gone. I'm an electrical engineer and very objective. I don't know why people say not to do this; maybe they just mean you won't get more power, but you will get more clarity in my opinion.

Studying a typical crossover, I'm wondering where the Klipsch jumpers are. I circled where I think they are on the attached gif. If I knew this, maybe I can figure out why the sound is so much better. Any thoughts? The improvement is mostly the highs, but I think there maybe some midrange improvement too. I used the same original soldered cables for bi-amping the tweeters, also originally I had the speaker cables plugged into the HF connections. My wife also says the RF-83's sound more detailed too...

[DrWho]

I'm pretty sure the only thing different in your scenario is that
the woofer currents and tweeter currents have been isolated from each
other. I'm trying to figure out how that would matter because the
losses in the speaker cable are going to be very small and I would
expect that to dictate the maximum possible signal transfer between the
drivers....but the crossovers would work to filter that out and then
whatever signal is left would have to be strong enough to create some
extra non-linearities. That would imply that the effects would probably
be greatest right around the xover frequency...so a difference in IMD
perhaps?

The tweeter will more than likely have its Fs just below
the xover frequency, so it wouldn't take much current to get the
diaphragm moving just a touch...and extra cone movement in the bottom
of the band is going to modulate everything above it. But I think we're talking distortions that are on the order of
30-40dB down...

Maybe part of it has to do with there being about half
as much current through each amplifier (since in the rejection regions
the impedance is very high and thus no current flowing). Would that
make the output impedance of the amplifier behave like its lower when
there is less spectrum? I dunno.

I do believe, however, that if you can hear it, then you can probably measure it...and you did say you're a EE []

[etc6849]

Yes, I am an EE, but I don't have the facilities to do real world tests. I think I would need a sound booth with lots of padding mainly due to the HF scattering on the hard walls and the lack of room symetry (kitchen is open to my living room). With a sound booth, RoomEQ would work great for this type of test (I believe) and I would hope to see a difference in the response curves. The RoomEQ tool you got me interested in is very neat and useful...

My other thought for a real world test is simpler: run my speaker wires to a dummy load. Then, use a padding circuit and run this into RoomEQ. Then, I would run tests for each amp individually.

Mainly, I was hoping someone on here had the schematic of the RF-83's crossover. I think it would then be easy to make a simple model in PSPICE and get a frequency response curve for various scenarios. I haven't used PSPICE since college though, but I think there are a wealth of models for modeling cables. I also don't do engineering design work so I think I forgot everything I once knew :-( I'm just a humble inspector with the US NRC.

[hificanada]

So all I want to know is on my RF-83's if I bi amp and take the bus straps off the speaker terminal, one amp drives the two 8in drivers and the other amp then drives the one 8in driver and horn? What is the impedance of the HF terminals and LF terminals

[etc6849]

Here's a great technical reason on why to bi-amp the RF-83's:

http://sound.westhost.com/bi-amp.htm

I can't remember the exact impedance, but it's nothing my Onkyo TX_NR905 receiver can't handle as it's been bi-amping my RF-83's for many years now!

[expresso]

i have to agree with some others that its really not a big gain if you bi-amp - at least it wasnt for me - i have a 7002 marantaz with the RF 82s - i felt the same way- why not - so i did it and i didnt get any hugh gain by doing so - it sounded a bit different but nothing close to what it sounds like with an Amp -

if you have the wires already and you like to try it - its fun , give it a shot - but if not - just buy a Amp and be done with it - it would really slam those drivers

i heard a pair of RF 83 and there were nice - dont remember what was powering them - it was a while back - once i added a Amp to my system

theres no way to go back - always have to have a amp on the fronts from now on - - the AVR sounds great alone also - but the Amp really opens those speakers up - i use the AVR power for my other set of Klipsch KG 3 - its a perfect match - sound and power wise for them -

[derrickdj1]

n my case, far as I know, my receiver puts out 110Wpc through 4 separate internal amps. Sure, they're all driven by the same power supply, but I'm thinking that does *not* equate to a situation where I could expect 770W (110Wpc x 7 channels) if I only use it to play my fronts in a stereo configuration. (Am I possibly wrong on this?)

My avr is rated 140 watts per channel with separate internal amps. My Power console never shows greater than 450 watts being used in total with peaks. I have 3 amp, cable box, BD, avr and TV hooked up to the power console. The Point is the 770w power rating senario will never happen. Most likely, you will never need it. If most of us could use the total watts that our gear is rated at, we might produce an EMF that could make our hair stand up straight, lol.[*-)]

[Dennie]

n my case, far as I know, my receiver puts out 110Wpc through 4 separate internal amps. Sure, they're all driven by the same power supply, but I'm thinking that does *not* equate to a situation where I could expect 770W (110Wpc x 7 channels) if I only use it to play my fronts in a stereo configuration. (Am I possibly wrong on this?)

My avr is rated 140 watts per channel with separate internal amps. My Power console never shows greater than 450 watts being used in total with peaks. I have 3 amp, cable box, BD, avr and TV hooked up to the power console. The Point is the 770w power rating senario will never happen. Most likely, you will never need it. If most of us could use the total watts that our gear is rated at, we might produce an EMF that could make our hair stand up straight, lol.

You're replying to a 6 year old post? [^o)]

Dennie

[etc6849]

I'm amazed by the folks by the number of folks who say I need a separate amp. I'm seriously considering one though just because so many swear by them.

I bi-amp too, with 145watts/channel, and the Onkyo 905 I have has a nice power supply. I know there should be no difference electrically with a separate amp, but I'm thinking of trying an Outlaw 7900 during their next secret sale. It's a 300 watt/channel amp: http://www.outlawaudio.com/products/7900.html

[derrickdj1]

I just notice the post was that old. [:$] I just got back home and notice it has 3 pages. Well I guess I was more tired than I thought earlier today