What’s the Right Power Amp Ratio For Bi-Amping?

One advantage of a vertical configuration is minimization of sonic effects that may result from inter-channel crosstalk within each amp, since both channels would be processing the same signal.

Hi Al, just to add, the biggest advantage of vertical bi-amping is that the whole power supply joule storage of one amp, is dedicated to just one bass driver and not shared, so the bass/upper bass should be better on big dynamic transients. (unless the amps have true dual mono power supplies) which are rare.

Cheers George

Hi Sleepwalkwer,


As some one who builds his own speakers and worked in the motion picture sound industry for years, and got to sit in on lectures at Georgia Tech on the subjects we are covering, please let me walk you through the math you are missing.


You are absolutely correct that when playing to a high pass limited load you won’t have a power issue. What you’ll have is a voltage issue.


Amplifiers don’t amplify watts, they amplify volts. Think of them as an electronic spreadsheet. Some value comes in on the input, and we have to multiply that value by a fixed amount to get the result. That’s the output that is presented to the speaker.

The typical amplifier provides about 26 dB of voltage gain, regardless of power rating. Stereophile routinely measures this in their amplifier reviews, like here:

https://www.stereophile.com/content/pass-laboratories-xa2008-monoblock-power-amplifier-measurements

Simplified, this is about 20x.That is, feed it 1Vrms, and you should get 20Vrms out. This will continue to be true so long as the load impedance is >= 8 Ohms AND you do not exceed the maximum voltage at the power rails.


For an amplifier rated for 100 Watts amplifier these rails are at about +- 26 Volts (if I have done my math right, and it’s been decades since I was tested on this). No matter how much you increase the load impedance, that is the physical limit of the amplifier. Due to losses in the transistors, etc, it will swing less than the rails, maybe +- 20 Volts. This is what the amplifier’s published power rating is really based on.


When you look at the spec rating, that’s the limit you see superbly illustrated in figures 5 and 6, here:

https://www.stereophile.com/content/luxman-l-509x-integrated-amplifier-measurements


That brick wall is essentially the voltage rails being hit, and while increasing the load impedance helps, you can't beat the power supply rails.

So, back to the original problem. The amp will attempt to apply it’s gain to the input, regardless of how much you increase the load impedance. Now, one benefit of course, is that you won’t have any current flowing at 100 Hz, so the amp power supply will be stiffer. That is, typically, when playing hard and loud, amp voltages sag. The power supply can’t maintain it’s maximum rating. So, yes, little benefit. However, you cannot exceed those voltage rail maximums, and without a line level crossover, you are almost exactly as limited in terms of gain and maximum output as you were before hand. Both amps still have to reproduce the same voltage swings and will apply the same gain as before.

Hope this helps, and apologies in that I’m 90% sure my power supply to amplifier power rating math is probably off, but the principles are not. :)


How will it sound? Well, depends on how stressed out the original setup was. If the amp was never stressed, and it’s voltage rails remained constant and stiff, bi amping won’t show much improvement.


However, if the amps were soft/weak, biamping will keep the mid/tweet amp stiffer, and less subject to the turbulence and weakening power supply of the bass amp.


So, toss-up. :)

The phrase I should have used is "clipping."

When reviewers talk about an amp clipping it means it has reached the limits of the voltage swing, and the sine waves get clipped, or shopped off.

:)

And as discussed above, from 8 Ohms upwards to infinity, the voltage at clipping does not vary that much, and since gain is fixed, we must maintain this :
Vin  < ( Max Vout / Gain )

Doh.

I think I should have said:
A 100 W/ch amp produces 26 Vrms = 40 Vpeak = 80 Vpeak to peak . So the internal DC voltage rails would be that plus a little more, +- 43-46 Volts.

Anyway, the story still holds. Whatever those rails are are the absolute limits of your output. Since gain is fixed, your maximum input is limited before you clip.

Best,
E

Thanks Erik. The thing that I’m looking to leverage is the impedance curve with a modest amp powering the high frequencies. I have a NAD 2700 for the low frequencies, with 150 watts per channel into 8 ohms, it’s quite capable. The other amp I have is a NAD 3225PE with 40 watts per channel, an integrated amp with external pre-amp in/out jumpers so i can use just the power amp section. The previous setup had that amp acting as the system preamp and low range bi-amp powering the same speakers, and a 3220i for the highs - that combo worked very nicely. The game changer was finding the 2700 amp for a bargain.