|
Home » Sponsored » Pi Speakers » Pi Crossover Boards and Biamping
| Pi Crossover Boards and Biamping [message #76132] |
Mon, 01 April 2013 21:40  |
rkeman
Messages: 78
Registered: March 2010
Location: Florida
|
Viscount |
|
|
|
Is there any convenient way to allow biamping of the speaker (i.e. separating the woofer and tweeter or midrange/tweeter) using the Pi crossover PCB? Also, the bare PCB pictured doesn't seem to have the wire terminals soldered to the board as in the preassembled crossover. Are these available from Pi Speakers or some other source?
|
|
|
|
| Re: Pi Crossover Boards and Biamping [message #76137 is a reply to message #76132] |
Tue, 02 April 2013 10:48  |
 |
Wayne Parham
Messages: 18677
Registered: January 2001
|
Illuminati (33rd Degree) |
|
|
The connectors on the π crossover PCB are just 0.205" spade lugs for printed circuit boards. You can get them at any electronics supplier, Digikey, Newark or Mouser, for example. Do a search for "PCB Quick-Fit Male Terminal 0.205" and you'll find plenty of them. They're all designed to be mounted on pads with plate-through holes but I bend over the hole tabs because I want one side of the board to be completely flat for mounting purposes.
If you are thinking about biamping, you don't really need the π crossover PCB. Of course, it's nice to have the passive crossover, because it is fully optimized and provides a baseline. It allows the speaker to be used with a single amplifier too. But a biamp configuration is possible, of course, and would negate the need for a passive crossover.
I tend to warn people that they'll need to do some work to setup an active crossover, and not to think that they will enjoy improvements just by simply having one. It requires a crossover/processor that can be programmed to provide any transfer function, not just one that allows the user to select a set crossover frequency and slope. And it takes a little bit of time with an acoustic measurement system, to optimize the system at all off-axis angles. The process is shown in my "Crossover Optimization" thread. The basic process is described, and there is a link to a video that shows you how to find the vertical nulls, which mark the edges of the forward lobe.
The low-pass filter for the woofer is pretty straightforward, but the tweeter circuit is a little more complex. And of course, proper on-axis and off-axis summing through the crossover region requires careful selection of both high-pass and low-pass sections, both in terms of frequency and phase. This is usually set by slope and filter type, and sometimes standard types aren't even used.
The damping of the waveguide's high-pass filter circuit must be adjustable as well as the slope and frequency, because we set the lower "shelf" by setting filter Q. In most cases, it is slightly underdamped because (properly sized) conical horns and waveguides tend to rolloff a little down low. On the other hand, truncated conical horns and waveguides tend to have a peak down low, and in this case, the tweeter circuit can be overdamped to partially compensate. This is a very important feature of the π pad in the passive crossover - It provides a way to adjust the region just above crossover and to tailor it for any waveguide. You will need to provide this in an active crossover as well.
Most people tend to think an active crossover can be easily configured using "CD compensation" but I find this is rarely the case because of the lack of filter Q adjustability. The 6dB/octave compensation for mass-rolloff is the easy part, it's just a single pole RC or RL filter. That's what "CD equalization" in most active crossovers provides. But the result is a transfer function that's a diagonal line, and that doesn't work well in most cases.
Mass rolloff doesn't start to occur until around 4kHz, so we usually want flat response up to that point, followed by 6dB/octave rising response. And as I said above, each individual waveguide has its own unique charasteristics in the ~1kHz octave too, some needing a little more damping, some a little less. So for the active crossover to be useful, it needs to capable of adjusting the transfer function of the tweeter circuit accordingly, usually something like this:
π crossover tweeter circuit transfer function
|
|
|
|
|
|
| Re: Pi Crossover Boards and Biamping [message #76139 is a reply to message #76138] |
Tue, 02 April 2013 12:53 |
|
Wayne Parham
Messages: 18677
Registered: January 2001
|
Illuminati (33rd Degree) |
|
|
To separate the low-pass and high-pass inputs, you would need to cut two traces on the printed circuit board. It would be pretty easy to do, and you could even add a switch or jumpers for input selection. But thinking outloud, I wonder what benefit that would bring? The main benefit of biamplification is reducing amplifier bandwidth, which then reduces power supply requirements. If the amps are run full range, you lose that advantage. Why, then, run separate amps at all?
I can see using a stereo amplifier as the pair of amps for biamplification of a single speaker, using one channel for the woofer and the other channel for the tweeter. Low-pass the woofer channel ahead of its amplifier and high-pass the tweeter channel ahead of its amp. That way the bandwidth is reduced for each. But this requires a crossover be placed ahead of the amplifier. It can be a high-impedance passive crossover or some form of active crossover. Either way, you would need to optimize it as I described in my last post to maintain parity, otherwise the low-impedance (speaker level) passive crossover will still outperform it.
|
|
|
|
|
|
| Re: Pi Crossover Boards and Biamping [message #76143 is a reply to message #76142] |
Tue, 02 April 2013 14:03 |
|
Wayne Parham
Messages: 18677
Registered: January 2001
|
Illuminati (33rd Degree) |
|
|
If the input signal isn't band-limited, then the amplifier isn't band-limited. Not that it needs to be - If there's enough voltage and current capacity, then band-limiting is probably not all that important. But the fact remains, if we want to limit the passband, we have to do it at the input. Doing it on the output doesn't really do very much for the amp.
Bi-amping is attractive where the amplifiers are the limiting factors. But that's because biamping does both things: It essentially reduces the voltage requirements and the current requirements by way of reducing bandwidth. The voltage requirements are reduced because the treble isn't modulated by the bass, meaning peaks of one ride on peaks of the other, creating the need for 2x voltage on max-peaks. The current requirement is reduced for the same reason.
But if the full-band input signal is sent to each amplifier, then the voltage requirements remain high. The treble signals will still ride on top of the bass signals. If peaking occurs, it will chop the high-frequency content first. So while the woofer's low-pass would reduce distortion that resulted, the tweeter circuit would pass it right through. I can't see this being any more than marginally beneficial. Seems like kind of a hack.
When the input is left full-range, I see only two real configurations of value: Bridged and paralleled. For a current benefit, one could simply parallel the outputs. For a voltage benefit, one would bridge the outputs. These would give the benefits I think you're seeking using what I think is a cleaner system design. Or maybe one of these is essentially what you're talking about:
I do understand you said, "Some audio/video receivers can only really handle inefficient multi-way direct radiator or planar magnetic speakers effectively by employing this mode of operation." And I can appreciate that. But what that really means is the amp just cannot handle its load, and is exceeding voltage or current limits, or perhaps both. There are many ways to solve that, and while biamping is one of them, I would argue that if biamping isn't done by band-limiting the signal ahead of the amp, then all that's really been accomplished is an awkward from of paralleling.
|
|
|
|
|
|
|
|
| Re: Pi Crossover Boards and Biamping [message #76157 is a reply to message #76148] |
Wed, 03 April 2013 11:57 |
dheflin44
Messages: 47
Registered: November 2012
Location: Carrollton, TX
|
Baron |
|
|
| Dave_S wrote on Wed, 03 April 2013 01:47 | It is probably not likely to occur in the near future, but I was wondering if you have noticed audible differences with bi-amping when passive crossovers are placed in front of the amp. My understanding is that the amp will not have to fight the crossover components and can just play the music. If there is an advantage, I might consider building a second F5 or another Nelson Pass design. At this point, it is just curiosity. I definitely do not need the power of a second amp.
|
Dave,
Are you saying you want the signal chain to be preamp -> passive XO -> amp -> speaker? This may be possible, but the passive XO needed for this configuration will be very different from your current XO between the amp and speaker. The speaker drivers have much lower impedance and more reactive load than the amp inputs which drastically affects the crossover's transfer function. In any case if you're wanting to do a high-level XO to isolate the amps' frequency ranges, then use a DSP crossover to have the best chance of getting the correct overall system response. Since a lot of trial and error will be needed, being able to reprogram a DSP will make it easier.
-Darrell
|
|
|
|
|
|
Goto Forum:
Current Time: Wed Apr 24 17:49:33 CDT 2024
|
Posting Rules
Members
Search
Help
Register
Login
Get a T-Shirt!
