Low-level passive bi-amping sounds good

[TBrennan]

I've been bi-amping with a DOD active crossover. My buddy Kurt Chang likes the sound of my rig and decided to play with bi-amping so he bought a Behringer active crossover to use with his JBL 3115s (Kurt also uses Klipsch LaScalas and Choruses, Voight pipes with Radio Shack fullrange drivers and a B&G ribbon rig and is thinking of getting a big Altec rig too). Unhappy with the sound of the Behringer he started playing with low-level passive circuits to split the signal after the preamp and before the amps, this sounds very good with his JBLs. Then he made a pair of the circuits for my rig. These are just simple resistors and caps wired into interconnects he made of small twisted wire and cross my rig over at 800hz, 1st order, one set lowpasses the signal to the bass amp and the other highpasses the signal to the treble amps. This is an interesting and cheap way to bi-amp (and this is REAL bi-amping, not what some goons call bi-amping when they run 2 fullrange signals to 2 amps) and sounds very good. An entire active device has been eliminated and the sound is very smooth and, oh I don't know, just "right". Both PWK and Bruce Edgar tout the sound of 1st order crossovers with horns and I'm inclined to agree. Kurt used a voltage divider in the circuit to the compression drivers as they are much more efficient than my basshorns, to balance the bass and treble I used the level controls on my bass amp. Kurt designed the circuits well and testing with my RTA the crossover is happening at 800hz. What's amazing to me is how tiny the circuits are, small enough that they're built inside the RCA connectors of the wires. I'm very pleased with the results and have no hankering to go back to the active crossover. I think this kind of bi-amping will appeal to minimalists like Mobile.

[mobile homeless]

I really agree here and some have written me about this. I have never been a fan of active crossovers but never thought I'd hear the day when ole Tom said something along these lines. I will admit, this solution above is intriguing. I hope you bring all to the HornFest at the end of the month. I really want to hear! By the way, Kevin tells me the Chicago Loons are going to be right next door to him. Do I see potential for some dB competition? heh.... I know who'll win that race. He's going to be having the new gear he just made from kits including the Billie 300B Set, fully modified Basie preamp with DACT attenuators and better passive parts, and a hot rod Rotel Player with evey LC Audio gimcrack installed. The speakers will depress ole Tom as they will be the little HE10.1 monitors with an Adire sub.

I'm interested to hear all myself. Ole Kevin might wish he was three rooms over!

kh

This message has been edited by mobile homeless on 03-09-2002 at 08:14 AM

[TBrennan]

Mobile---Kevin is a nice guy and he doesn't try to BS you, very reasonable fella. I'm eager to hear the Adire speakers, while small they are the proper way to build a small speaker; vented box and a co-axial horn, nothing wrong with that. Kinda like baby Altec 605s. You never know, I might come home with a pair. :-)

This message has been edited by TBrennan on 03-09-2002 at 08:34 AM

[jnorv]

As a mechanical engineer who slept through most of my EE courses, could you EE inclined people give me a little more info. I am now using a Bryston active xover, but would like to go the passive route. At 6dB/octave, I know that you only need simple caps and coils. I guess the resistors adjust the realtive level. At the signal strength at the pre out, these caps and coils can be of smal wattage. Any other tricks that I should know to get started? I have all the data that I need to calculate the values for the components.

Thanks,

Jim Norvell

[jnorv]

I know that caps in series pass high frequencies but does putting one in parallel with the woofer instead of a coil in series do the trick. Wouldn't this just short the pre out at high frequencies?

Jim

[jnorv]

Thanks, that is pretty slick. I modeled it on Excel to see the results. Any rule of thumb on the value of the resistor. It must be some function of the input impedence of the amps(s). Mine are all 100k except my Rowland which I can vary.

Jim

[SOUNDJUNKIE]

I don't know if I would hack up my cables though, I might try it here.

[Mike Lindsey]

Can someone please explain the difference between active and passive crossovers? I have Chorus's that have crossovers inside the cabinets. I am assuming a full signal is being sent to them and then they determine (by what? the resistors and caps?) what goes to each driver. Is this known as a passive crossover?

If someone could explain how you bi-amp with both types of crossovers, I would greatly appreciate it. BTW, what exactly do these resistors and caps do within the crossover? As you can tell, I know squat about this but feel as though I can understand it if the explanation was worded for a 6th-grader.

Mike

------------------

My Music Systems

[Mike Lindsey]

Mark,

As usual, thanks for the very informative post. I have cut and pasted that last post to a Word document for future reference.

Mike

------------------

My Music Systems

[jnorv]

Qusestion about the relative level of the components. My preamp has an output impedence of 300 or 600 ohms ( user selectable). Levinson's are down in the few ohm level. If I put a parallel circuit (low pass/high pass)across the output of a few times these values, in conjuction with the input impedence of my amps (100K ohm typical), why doesn't this swamp the output of the preamp since it is a voltage source and not a current source?

Jim Norvell

[Klewless]

jnorv,

To sort of answer your question about parallel resistors an example will help. For resistors (not complex resistors/inductors/capacitors in combination) two in parallel will end up with an equivalent value as follows: R = (R1 * R2) / (R1 + R2)

If two 10 ohm resistors are connected in parallel, then 10*10/10+10 = 100/20 = 5 ohms. In other words, two of equal value will end up half.

Now if one is 10 and the other is 100, then we have 10*100/10+100 = 1000/110 = about 9 ohms. The larger one had considerable less effect on the total. And the bigger the second one becomes, the less it influences the result.

The real formula is 1/R = 1/R1 + 1/R2 (if my memory serves me correctly). That is one of the reasons the preamp output impedance should be much smaller than the input impedance of the power amp which follows it.

Hope this helps.

------------------

John P

St Paul, MN

[jnorv]

I got that much out of my EE courses in college. Never did study amplifiers, though. I went in to ME. Are you saying that the output current of the preamp is determined by it's own output impedence, as long as it is connected to a much higher input impedence at the amp? I just figured that it was the input impedence of the amp that was controlling things. I figured that if I made the passive xover with a summed impedence close to that of the amp, the preamp wouldn't know the differnce.

Thanks, Jim N

[jnorv]

Thanks. Now I understand.

[jnorv]

Yes. I just have to make another trip to the parts house. Luckily Silicon Valley is full of them. Tere are even some pretty good used parts houses here.

[jnorv]

I think Hal-Tek went out of business. Halted is now HSC and is on Central Expy and Lawrence. I got some nice shielded, stranded 12G witha 16 G ground for power cords - .20/ft. No name on it so don't know what the insulation is. I also go to R&D out by the Great Mall. They set themselves up for the prototype manf. They are better organized but way more expensive.

[jnorv]

I got my low level passives up and running today. At first blush they are probably as good or better than the active (Bryston 10B) that I was using. Less stuff in hthe signal path. I had to put in an addional 50K resistor in series with the 50K pot in the low pass to bring it down. I am currently trying to match a 100 dB bass unit to a 91 dB set of speakers. I will plug the active back in over the weekend to get a better idea. Right now I am adjusting by ear. I need to break out the signal generator and spl meter to zero in on it.

[jnorv]

My current rotation is a pair of the JBL 4638s and Dunlavy 1AVs. When I got my 1st 4638, I plugged it into the ALK xover to my La Scala and twiddled with the midrange horn level. Sounded promising. Running both the 4638s and the LSs take up to much real estate in my listening room. I havent figured out how to stack them yet and it is a lot of time to kept swapping the horns between the LSs and the 4638s. And then there is the seismic problem down here with them stacked. After listening to the 4638s and looking over Thieles paper, I am about to retune them for better base response. I am still looking for data on what the poly fill does to the characteristic impedance (rho*c) of the box. One would think that the Acousta-Stuf people would provides curves of density vs rho*c in comparison to their stuff and poly fill. They are designed to have a 24 dB cut off below 40 Hz. Fortunately I have a 2205A that I bought years ago which is a better speaker than either the 2035 or the 2226 that come in the 4638/4648s, which should drop f3 to 30 Hz. At some point in the near future, I will put my Thiel 3.6s back together for a comparison between SETs/high efficiency and big SS/low eff speakers. So many options, so little time.

Question, can I cascade the low level xovers to get a higher cutoff slope. Once I get over about a two branch circuit, it becomes to laborious for me to solve by hand. Somewhere around here I have an old finite difference heat transfer program that I can use to solve the networks.

Jim