• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

Bi amping

DonH56

Master Contributor
Technical Expert
Forum Donor
Joined
Mar 15, 2016
Messages
7,835
Likes
16,497
Location
Monument, CO
So if your AVR's AMP is lower in power output per channel (stereo) than you would like, passive bi-amping will actually make that worse?

Usually, but it depends. If the power supply is the limitation, which it usually is in an AVR, then you've made the problem worse by adding additional loading on the supply. Voltage output is unchanged, but the current output from each amplifier is lower than if a single amplifier is used, so if the amps are current-limited then it could help. However, the vast majority of audio amplifiers are designed to be as near as possible to a pure voltage source, so "passive" bi-amping the AVR way is unhelpful. You are also increasing the heat generated by the AVR, which has other negative effects like shortening component life. If you really need a bigger amplifier, then buy a bigger amplifier, or implement conventional bi-amping with a crossover before the amplifiers and size them accordingly.
 

Pdxwayne

Major Contributor
Joined
Sep 15, 2020
Messages
3,219
Likes
1,172
Usually, but it depends. If the power supply is the limitation, which it usually is in an AVR, then you've made the problem worse by adding additional loading on the supply. Voltage output is unchanged, but the current output from each amplifier is lower than if a single amplifier is used, so if the amps are current-limited then it could help. However, the vast majority of audio amplifiers are designed to be as near as possible to a pure voltage source, so "passive" bi-amping the AVR way is unhelpful. You are also increasing the heat generated by the AVR, which has other negative effects like shortening component life. If you really need a bigger amplifier, then buy a bigger amplifier, or implement conventional bi-amping with a crossover before the amplifiers and size them accordingly.
May I know about the effects to the amp when the amp only see one driver's ohm swing vs seeing both drivers (assuming a two drivers design) combined ohm swing?

I have read the the amps' damping factor, the speaker cables, and the ohm range of the speaker all have an effect to tonal balance.
 

DonH56

Master Contributor
Technical Expert
Forum Donor
Joined
Mar 15, 2016
Messages
7,835
Likes
16,497
Location
Monument, CO
May I know about the effects to the amp when the amp only see one driver's ohm swing vs seeing both drivers (assuming a two drivers design) combined ohm swing?

I have read the the amps' damping factor, the speaker cables, and the ohm range of the speaker all have an effect to tonal balance.

True, and it depends upon all of those factors, whether or not you are bi-amping. When you bi-amp, you are effectively only "seeing" one-half the drivers, so whatever the impact of the varying speaker load in the band you are driving is what matters. If there is a very low-impedance point in one band not in the other then you are reducing the current (and thus power) in the "other" band.

Damping factor is just the load divided by the amplifier's output impedance: DF = Zspeaker/Zamp. I often use the term "effective" driving-point impedance by including the speaker cables' resistance with the amplifier's output since the speaker sees the sum of those terms: driving-point impedance seen by the speakers is Zamp+Zcable so DFeff = Zspeaker / (Zamp + Zcables). Remember there are two wires in the cable so you should include the resistance of both sides of the speaker cables.

The speaker's impedance will create a voltage divider over frequency that changes depending upon the driving-point impedance, assuming Vamp is the amplifier's (unloaded) output:

Vspeaker = Vamp * Zspeaker / (Zspeaker + Zamp + Zcables)

If Zamp = Zcables = 0 (perfect amp, superconducting cables), then Vspeaker = Vamp and there is no variation. With real amps, cables, and speakers the impedance changes over frequency so you get different voltages over frequency. The amount of variance depends upon all three parameters, but generally the speaker is most significant. The amplifier's output impedance usually changes a lot from 20 Hz to 20 kHz but is so low (for SS amps anyway) with respect to the speaker that it does not matter much. Same for the wires; their impedance is small (assuming reasonable wire size) and does not change significantly over the audio band (despite adverts claiming RF effects matter at audio frequencies).

To put some numbers on it, a speaker may vary from 4 ohms to 20 ohms over its frequency band. The SS amplifier may vary from 0.05 ohms at 20 Hz to 0.5 ohms at 20 kHz -- a large change, but still small with respect to the speaker. The wire is pretty constant at say about 0.05 ohms for a pair of 10', 12 AWG wires. How does this change the frequency response? Depends upon where the speaker's big swing are... See this post for a look at how the interaction may change the signal at the speakers: https://www.audiosciencereview.com/...ffect-of-speaker-cable-length-mismatch.20150/

My Magnepans were pretty happy with a tube amplifier having a few ohm's output impedance because, although they were fairly low in impedance at 4 ohms, their impedance did not change much over frequency. OTOH, it was not as happy driving my Infinity speakers that varied from about 2 ohms to over 20 ohms, causing the frequency response to vary due to the high amplifier output impedance.

HTH - Don
 

Pdxwayne

Major Contributor
Joined
Sep 15, 2020
Messages
3,219
Likes
1,172
True, and it depends upon all of those factors, whether or not you are bi-amping. When you bi-amp, you are effectively only "seeing" one-half the drivers, so whatever the impact of the varying speaker load in the band you are driving is what matters. If there is a very low-impedance point in one band not in the other then you are reducing the current (and thus power) in the "other" band.

Damping factor is just the load divided by the amplifier's output impedance: DF = Zspeaker/Zamp. I often use the term "effective" driving-point impedance by including the speaker cables' resistance with the amplifier's output since the speaker sees the sum of those terms: driving-point impedance seen by the speakers is Zamp+Zcable so DFeff = Zspeaker / (Zamp + Zcables). Remember there are two wires in the cable so you should include the resistance of both sides of the speaker cables.

The speaker's impedance will create a voltage divider over frequency that changes depending upon the driving-point impedance, assuming Vamp is the amplifier's (unloaded) output:

Vspeaker = Vamp * Zspeaker / (Zspeaker + Zamp + Zcables)

If Zamp = Zcables = 0 (perfect amp, superconducting cables), then Vspeaker = Vamp and there is no variation. With real amps, cables, and speakers the impedance changes over frequency so you get different voltages over frequency. The amount of variance depends upon all three parameters, but generally the speaker is most significant. The amplifier's output impedance usually changes a lot from 20 Hz to 20 kHz but is so low (for SS amps anyway) with respect to the speaker that it does not matter much. Same for the wires; their impedance is small (assuming reasonable wire size) and does not change significantly over the audio band (despite adverts claiming RF effects matter at audio frequencies).

To put some numbers on it, a speaker may vary from 4 ohms to 20 ohms over its frequency band. The SS amplifier may vary from 0.05 ohms at 20 Hz to 0.5 ohms at 20 kHz -- a large change, but still small with respect to the speaker. The wire is pretty constant at say about 0.05 ohms for a pair of 10', 12 AWG wires. How does this change the frequency response? Depends upon where the speaker's big swing are... See this post for a look at how the interaction may change the signal at the speakers: https://www.audiosciencereview.com/...ffect-of-speaker-cable-length-mismatch.20150/

My Magnepans were pretty happy with a tube amplifier having a few ohm's output impedance because, although they were fairly low in impedance at 4 ohms, their impedance did not change much over frequency. OTOH, it was not as happy driving my Infinity speakers that varied from about 2 ohms to over 20 ohms, causing the frequency response to vary due to the high amplifier output impedance.

HTH - Don
Thanks!

Maybe you can provide inputs regarding the effects of passive bi amp vs normal for one of my setups? Any possible audible changes using passive biamp? Let's says speaker wires are about 10' and 10 to 12 awg. Speaker crossover at 2kh. The crossover networks are independent of each other.

The impedance of my speakers from soundstage:
impedance.gif


The damping factor of my amp from soundstage:
chart4a.gif


Thanks again!
 
Last edited:

DonH56

Master Contributor
Technical Expert
Forum Donor
Joined
Mar 15, 2016
Messages
7,835
Likes
16,497
Location
Monument, CO
Thanks!

Maybe you can provide inputs regarding the effects of passive bi amp vs normal for one of my setups? Any possible audible changes using passive biamp? Let's says speaker wires are about 10' and 10 to 12 awg. Speaker crossover at 2kh. The crossover networks are independent of each other.

The impedance of my speakers from soundstage:
View attachment 127816

The damping factor of my amp from soundstage:
View attachment 127817

Thanks again!

Uh, I have a day job (typing between test runs), sorry, not really set up to analyze everyone's particular situation. I'm going to go with "no" but "any possible" is pretty open... A lot of people claim to hear benefits; of the few I know who have actually done A/B testing blind, none were able to tell a difference, but it is impossible to do rapid switching for true ABX testing.

As a quick back-of-the-envelope look, your amplifier has very high damping factor (very low output impedance) so I consider highly unlikely that passive bi-amping would provide any audible benefit. At 20 kHz the amplifier's output impedance is still about 0.08 ohms (assuming the DF is for 8 ohms) or about the same level as the cables, and about 0.01 ohms at 1 kHz and below. The LF excursions are larger than 2 kHz and above so no change in the lows. For the highs, it looks like about 6 to 18 ohms from 2 kHz to 20 kHz for the speakers and maybe 0.013 ohms to 0.08 ohms for the ohms and call the cables a flat 0.05 ohms. Amp DF is around 400 at the 18-ohm HF peak so ~0.02 ohms amplifier output impedance.

Using rough numbers eyeballed from the plots:

At 2 kHz the transfer function is about 6/(6 + 0.013 + 0.05) = 6/6.063 = 0.9896, or about -0.091 dB
At 5 kHz we get 18/(18+ 0.02 + 0.05) = 18/18.07 = 0.9961, about -0.034 dB
At 20 kHz we get 6/(6 + 0.08 + 0.05) = 6/6.13 = 0.9788, about -0.1861 dB

The difference between highest and lowest is about |(-0.1861) - (-0.091)| = 0.095 dB. So if you can hear a difference in frequency response from 2 kHz to 20 kHz of <0.1 dB it is audible. I am pretty sure I cannot.

This is pretty simplified, leaving out the phase component and such, as well as the width of the deviation (wider deviations are easier to hear), and assuming I didn't make a mistake in these quick calculations, but the numbers and my gut say you would not hear a difference in a blind test.

Insert usual disclaimers here: somebody will come along and say they can hear night and day differences.

HTH - Don (ears of clay)
 

Pdxwayne

Major Contributor
Joined
Sep 15, 2020
Messages
3,219
Likes
1,172
Uh, I have a day job (typing between test runs), sorry, not really set up to analyze everyone's particular situation. I'm going to go with "no" but "any possible" is pretty open... A lot of people claim to hear benefits; of the few I know who have actually done A/B testing blind, none were able to tell a difference, but it is impossible to do rapid switching for true ABX testing.

As a quick back-of-the-envelope look, your amplifier has very high damping factor (very low output impedance) so I consider highly unlikely that passive bi-amping would provide any audible benefit. At 20 kHz the amplifier's output impedance is still about 0.08 ohms (assuming the DF is for 8 ohms) or about the same level as the cables, and about 0.01 ohms at 1 kHz and below. The LF excursions are larger than 2 kHz and above so no change in the lows. For the highs, it looks like about 6 to 18 ohms from 2 kHz to 20 kHz for the speakers and maybe 0.013 ohms to 0.08 ohms for the ohms and call the cables a flat 0.05 ohms. Amp DF is around 400 at the 18-ohm HF peak so ~0.02 ohms amplifier output impedance.

Using rough numbers eyeballed from the plots:

At 2 kHz the transfer function is about 6/(6 + 0.013 + 0.05) = 6/6.063 = 0.9896, or about -0.091 dB
At 5 kHz we get 18/(18+ 0.02 + 0.05) = 18/18.07 = 0.9961, about -0.034 dB
At 20 kHz we get 6/(6 + 0.08 + 0.05) = 6/6.13 = 0.9788, about -0.1861 dB

The difference between highest and lowest is about |(-0.1861) - (-0.091)| = 0.095 dB. So if you can hear a difference in frequency response from 2 kHz to 20 kHz of <0.1 dB it is audible. I am pretty sure I cannot.

This is pretty simplified, leaving out the phase component and such, as well as the width of the deviation (wider deviations are easier to hear), and assuming I didn't make a mistake in these quick calculations, but the numbers and my gut say you would not hear a difference in a blind test.

Insert usual disclaimers here: somebody will come along and say they can hear night and day differences.

HTH - Don (ears of clay)
Nice! Thanks for spending time to do a quick calc.

I am assuming soundstage measurements is with both drivers connected. Curious, may I know if the drivers are separately measured with their own crossover network, the impedance number from 20hz to 20khz would still remain the same?

Thanks!
 

DonH56

Master Contributor
Technical Expert
Forum Donor
Joined
Mar 15, 2016
Messages
7,835
Likes
16,497
Location
Monument, CO
Nice! Thanks for spending time to do a quick calc.

I am assuming soundstage measurements is with both drivers connected. Curious, may I know if the drivers are separately measured with their own crossover network, the impedance number from 20hz to 20khz would still remain the same?

Thanks!

The impedance looking into the speaker will change once the high and low filter sections are isolated. Instead of the combined response, the LF impedance should gradually rise to "infinity" as frequency increases and the crossover "opens" -- but it depends upon the type of crossover and such. Same thing in reverse for the high side; as frequency drops below the crossover point, the impedance looking into the high pass section should rise to infinity (open) as frequency reduces. Since there is essentially no power delivered to the LF driver at very high frequency, and none to the HF driver at very low frequencies, we don't really care and that does not factor into my calculations.

Note I did not run the numbers for the LF side since the peaks and valleys are the same either way for the LF amp, and the greatest excursions are in the LF area (below 2 kHz) so that amp (single or LF part of a bi-amped system) sees the largest change. With an active system, the bass amp is likely larger (higher in power) with lower output impedance (higher damping factor).

As always bear in mind the crossover is not a brick wall that instantaneously blocks frequencies on either side; there is a transition region as the signal gradually dies away as the frequency moves away from the crossover point.
 

RayDunzl

Grand Contributor
Central Scrutinizer
Joined
Mar 9, 2016
Messages
13,204
Likes
16,985
Location
Riverview FL
I am assuming soundstage measurements is with both drivers connected. Curious, may I know if the drivers are separately measured with their own crossover network, the impedance number from 20hz to 20khz would still remain the same?

My big speakers measure 4 ohms DC with the shorting strap.

Remove the strap, for bi-wiring and each set of terminals measures 8 ohms DC.

The advertised crossover frequency is 180Hz.

The "crossover" is a low pass on the bass, and high pass on the treble, with no internal interconnection.

1620077021186.png
 

Pdxwayne

Major Contributor
Joined
Sep 15, 2020
Messages
3,219
Likes
1,172
The impedance looking into the speaker will change once the high and low filter sections are isolated. Instead of the combined response, the LF impedance should gradually rise to "infinity" as frequency increases and the crossover "opens" -- but it depends upon the type of crossover and such. Same thing in reverse for the high side; as frequency drops below the crossover point, the impedance looking into the high pass section should rise to infinity (open) as frequency reduces. Since there is essentially no power delivered to the LF driver at very high frequency, and none to the HF driver at very low frequencies, we don't really care and that does not factor into my calculations.

Note I did not run the numbers for the LF side since the peaks and valleys are the same either way for the LF amp, and the greatest excursions are in the LF area (below 2 kHz) so that amp (single or LF part of a bi-amped system) sees the largest change. With an active system, the bass amp is likely larger (higher in power) with lower output impedance (higher damping factor).

As always bear in mind the crossover is not a brick wall that instantaneously blocks frequencies on either side; there is a transition region as the signal gradually dies away as the frequency moves away from the crossover point.
Thanks! So the measured impedance passing crossover points would change a lot if those are measured separately.

May I know about the impedance not near crossover point when drivers are measured separately?

For example, using my speakers chart, if removing tweeter section totally from amp, would the amp for bass driver still see same 10 ohm impedance at say 100 Hz?
 

Pdxwayne

Major Contributor
Joined
Sep 15, 2020
Messages
3,219
Likes
1,172
My big speakers measure 4 ohms DC with the shorting strap.

Remove the strap, for bi-wiring and each set of terminals measures 8 ohms DC.

The advertised crossover frequency is 180Hz.

The "crossover" is a low pass on the bass, and high pass on the treble, with no internal interconnection.

View attachment 127825
Thanks! Would you see post #69?

With your speakers, with both drivers connected, let's say at 100hz it is 4ohm. Once those are separated, would 100hz still shows 4 ohm?

Thanks!
 

RayDunzl

Grand Contributor
Central Scrutinizer
Joined
Mar 9, 2016
Messages
13,204
Likes
16,985
Location
Riverview FL
With your speakers, with both drivers connected, let's say at 100hz it is 4ohm. Once those are separated, would 100hz still shows 4 ohm?

That would require measurement which I am not equipped to perform.

Looking at the schematic:

1620078072525.png


LX4 blocks higher frequencies to some degree - so reduced current flow, increased impedance, as the frequency rises.

There are three return paths to the amplifier in parallel.

The first return path would appear to my eye to be a bit of tuning to attenuate some low frequency range that gets through LX4, maybe it flattens the response by providing a dip in power delivered to the woofer in some frequency range.

CX7 will become more "short" to ever higher frequencies that got past LX4 before that energy gets to the woofer.

And the woofer voice coil will have an unknown DC resistance and some unknown inductance to further attenuate higher frequencies that got past LX4.

In the words of the immortal Thingfish, "You figit out."

As for me, it works, has worked for 22 years, and nothing blows up, so, it is whatever it is.

---

SW1, if open, adds resistance to the whole path to attenuate the woofer about 4dB if desired.

---

The possible specs for the woofer, in case you want to build this circuit in a simulator and see what happens:

The ReQuest woofer is ML part number D12Q2M5 built to their specifications by Eminence internal part number 121506A It's a 12-inch 4-ohm speaker with a 2.5-inch voice coil and a 67 oz. ceramic magnet with a 250 watt power rating.

Here are the T/S parameters:
RE OHMS - 3.14
LE MH - .65
QM - 10.01
QE - .360
QT - .350
XMAX MM - 7.90
BL TM – 10.13
EFF% - .91
FS HZ - 25.78
MMS GMS - 72.70
CMS mm/M - 1.1755
VAS LTRS - 198.81
SD SCM - 519.45
EBP - 71.6
SPL db - 91.6
 
Last edited:

Pdxwayne

Major Contributor
Joined
Sep 15, 2020
Messages
3,219
Likes
1,172
That would require measurement which I am not equipped to perform.

Looking at the schematic:

View attachment 127828

LX4 blocks higher frequencies to some degree - so reduced current flow, increased impedance, as the frequency rises.

There are three return paths to the amplifier in parallel.

The first return path would appear to my eye to be a bit of tuning to attenuate some low frequency range that gets through LX4, maybe it flattens the response by providing a dip in power delivered to the woofer in some frequency range.

CX7 will become more "short" to ever higher frequencies that got past LX4 before that energy gets to the woofer.

And the woofer voice coil will have an unknown DC resistance and some unknown inductance to further attenuate higher frequencies that got past LX4.

In the words of the immortal Thingfish, "You figit out."

As for me, it works, has worked for 22 years, and nothing blows up, so, it is whatever it is.

---

SW1, if open, adds resistance to the whole path to attenuate the woofer about 4dB if desired.

---

The possible specs for the woofer, in case you want to build this circuit in a simulator and see what happens:

The ReQuest woofer is ML part number D12Q2M5 built to their specifications by Eminence internal part number 121506A It's a 12-inch 4-ohm speaker with a 2.5-inch voice coil and a 67 oz. ceramic magnet with a 250 watt power rating.

Here are the T/S parameters:
RE OHMS - 3.14
LE MH - .65
QM - 10.01
QE - .360
QT - .350
XMAX MM - 7.90
BL TM – 10.13
EFF% - .91
FS HZ - 25.78
MMS GMS - 72.70
CMS mm/M - 1.1755
VAS LTRS - 198.81
SD SCM - 519.45
EBP - 71.6
SPL db - 91.6
Nice! Thx!

You said this: "As for me, it works, has worked for 22 years, and nothing blows up, so, it is whatever it is."

You meant you have been doing passive biamp for 22 years?
 

DonH56

Master Contributor
Technical Expert
Forum Donor
Joined
Mar 15, 2016
Messages
7,835
Likes
16,497
Location
Monument, CO
Thanks! So the measured impedance passing crossover points would change a lot if those are measured separately.

May I know about the impedance not near crossover point when drivers are measured separately?

For example, using my speakers chart, if removing tweeter section totally from amp, would the amp for bass driver still see same 10 ohm impedance at say 100 Hz?

Yes, the "in-band" response would not change, just the response around and "past" (above for LF, below for HF) the crossover frequency.
 

RayDunzl

Grand Contributor
Central Scrutinizer
Joined
Mar 9, 2016
Messages
13,204
Likes
16,985
Location
Riverview FL
You said this: "As for me, it works, has worked for 22 years, and nothing blows up, so, it is whatever it is."

You meant you have been doing passive biamp for 22 years?

No.

It is bi-wired, just for grins.

With 2AWG THHN

The amp is 700W/4ohm and never fully taxed, so, no point in biamping.

Maybe no point in bi-wiring but I had the hardware in the junk box, so, wired it.

1620080343451.png


Scale; the nuts are 1/2 inch.


I'm King of the Hill on wire gauge, as far as anyone has reported, so I've got that going for me, which is nice.

I wish I'd kept some 1000mcm instead of selling it to the copper recycler, but I didn't.

Same hardware works at the amp end of the wire.


1620080633063.png
 
Last edited:

Pdxwayne

Major Contributor
Joined
Sep 15, 2020
Messages
3,219
Likes
1,172
No.

It is bi-wired, just for grins.

With 2AWG THHN

The amp is 700W/4ohm and never fully taxed, so, no point in biamping.

Maybe no point in bi-wiring but I had the hardware in the junk box, so, wired it.

View attachment 127831

Scale; the nuts are 1/2 inch.


I'm King of the Hill on wire gauge, as far as anyone has reported, so I've got that going for me, which is nice.

I wish I'd kept some 1000mcm instead of selling it to the copper recycler, but I didn't.
Man, I thought using 12awg + 10awg is a bit much for my setup.....You are in whole other level....
 

Pdxwayne

Major Contributor
Joined
Sep 15, 2020
Messages
3,219
Likes
1,172
Yes, the "in-band" response would not change, just the response around and "past" (above for LF, below for HF) the crossover frequency.
I see. Thx!

I have balanced DAC and extra stereo amps, thus I have been using one amp per speaker to eliminate interchannel crosstalk. Since each amp has outputs for two speakers, I thought might as well use all the outputs and do passive biamp at the same time.....So far the setup sounds really clean.....Do not know if there is any improvement as quick comparisons is hard.
 
Top Bottom