I was thinking about how one would go about designing a bipole speaker with rear firing drivers. The sound from the rear firing driver would have to travel to the wall (approx. 1 ft lets say) and then back to the front of the speaker (approx. another 1 ft less say).
Is there a way to add a delay to the front firing speakers of approx. 0.001765 sec with a passive crossover?
This is more of an educational post for me. I assume there is a way to do it. Just have not read or seen anything on it.
Hi,
the rear firing woofer is mostly omnidirectional.
The soundwave does not need to bounce off the adjacent wall to return to the front.
The sound wave simply travels around the speaker box and the shortest path length from front to back is the delay involved.
The delay you need will be quite a bit less than you thought necessary.
the rear firing woofer is mostly omnidirectional.
The soundwave does not need to bounce off the adjacent wall to return to the front.
The sound wave simply travels around the speaker box and the shortest path length from front to back is the delay involved.
The delay you need will be quite a bit less than you thought necessary.
I think the bipole speaker design 69stringray is asking about might actually be intended to operate full range in which case a delay between the front and rear drivers might be required.
IMO I don't think passive XO all pass filters are practical and may not even be possible. Active versions which are possible would be a huge pain to design, with lots of stages to get a consistent delay across the band of interest and not very tunable either. Seems like this would be something best done with DSP.
IMO I don't think passive XO all pass filters are practical and may not even be possible. Active versions which are possible would be a huge pain to design, with lots of stages to get a consistent delay across the band of interest and not very tunable either. Seems like this would be something best done with DSP.
Indeed, I asked a similar question before. All-pass is the passive answer.
However, the result will not be what you imagine. By delaying one part of a bi-pole design you will move the null that usually occurs at the sides. By changing the delay you will move the null towards either front or back of the cabinet. This is how some PA speakers get cardioid bass response with delayed rear woofers.
http://www.meyersound.com/mseries/m3d-sub/
Basically put, you don’t need to bother. The slightly delayed rear reflections is what gives bi-poles there great soundstaging and imaging. It gives your mind spatial location information.
However, the result will not be what you imagine. By delaying one part of a bi-pole design you will move the null that usually occurs at the sides. By changing the delay you will move the null towards either front or back of the cabinet. This is how some PA speakers get cardioid bass response with delayed rear woofers.
http://www.meyersound.com/mseries/m3d-sub/
Basically put, you don’t need to bother. The slightly delayed rear reflections is what gives bi-poles there great soundstaging and imaging. It gives your mind spatial location information.
Here's what a passive ladder delay network looks like, at least as it's laid out in SoundEasy. Values L6, C8, C9 and L10 are the circuit. (everything else in this network is response shaping and impedance control)
I occasionally do these on LR2 designs with large woofers mounted on flat (non-tilted) baffles. It's a way to make up for the difference in acoustic centers. I've never published a design like that because of it's complexity and the tendency of people to look at the crossover and say "WTF is that?"
If you use Soundeasy, you can easilly model bipoles using the virtual front panel and system frequency domain. There's a couple ways to do it, but the way I've done it is to measure a front woofer from a fixed distance (say 1m or a 1/2 m) then measure the rear woofer, but from the front at the same distance from that woofer. On the front panel, move the back woofer Z location to where it actually is. Assuming all the phase information is accurate, what you see when you look at the response combined... is the reason I don't recommend bipoles. Big nulls in the response that are not only measureable, but easily audible. But, using the system frequency domain, from there you can model in the ladder delay network to minimize the nulls if you wanted. The values will be huge and the result will still be far from perfect. Plus when you put the speaker in the room, the tonal balance usually needs a lot of work. Let's just put it this way: mounting a woofer on the back covering the same range as a woofer on the front is *NOT* the best way to resolve baffle step compensation.
All that said, I think rear firing speakers have value, but they are best used above the baffle step frequency. Then you will get the in-room soundfield that type of design can offer. It's not entirely accurate reproduction, but it can be enjoyable anyway.
I occasionally do these on LR2 designs with large woofers mounted on flat (non-tilted) baffles. It's a way to make up for the difference in acoustic centers. I've never published a design like that because of it's complexity and the tendency of people to look at the crossover and say "WTF is that?"
If you use Soundeasy, you can easilly model bipoles using the virtual front panel and system frequency domain. There's a couple ways to do it, but the way I've done it is to measure a front woofer from a fixed distance (say 1m or a 1/2 m) then measure the rear woofer, but from the front at the same distance from that woofer. On the front panel, move the back woofer Z location to where it actually is. Assuming all the phase information is accurate, what you see when you look at the response combined... is the reason I don't recommend bipoles. Big nulls in the response that are not only measureable, but easily audible. But, using the system frequency domain, from there you can model in the ladder delay network to minimize the nulls if you wanted. The values will be huge and the result will still be far from perfect. Plus when you put the speaker in the room, the tonal balance usually needs a lot of work. Let's just put it this way: mounting a woofer on the back covering the same range as a woofer on the front is *NOT* the best way to resolve baffle step compensation.
All that said, I think rear firing speakers have value, but they are best used above the baffle step frequency. Then you will get the in-room soundfield that type of design can offer. It's not entirely accurate reproduction, but it can be enjoyable anyway.
Attachments
Thanks for all of the great responses.
Zaph....thanks for providing a sample netowrk. Are the equations for determining the proper values posted or written anywhere? Or is it just available in crossover software?
I have a friend who is looking at Definitive's BP7000SC speaker and I thought it would be an interesting project for someone to clone (besides me - way beyond my skill level). It still got the juices flowing though, the good old fashion, what if?
I remember someone publishing a MTM w/ side firing powered woofer using RS drivers. That may be a good starting point for a clone
Zaph....thanks for providing a sample netowrk. Are the equations for determining the proper values posted or written anywhere? Or is it just available in crossover software?
I have a friend who is looking at Definitive's BP7000SC speaker and I thought it would be an interesting project for someone to clone (besides me - way beyond my skill level). It still got the juices flowing though, the good old fashion, what if?
I remember someone publishing a MTM w/ side firing powered woofer using RS drivers. That may be a good starting point for a clone
An externally hosted image should be here but it was not working when we last tested it.
Unfortunately the SoundEasys ladder delay networks are a little too simplified versions.
The network does not take the phase shift of the x-over at x-over frequency and the topology and impedance into account.
For example. If the tweeter is crossed over @ 2500khz and the distance is 2 inches and impedance of both drivers is 6 ohms at that point (use of zobel), you will have a phase shift of 133 degree and delay of 148 microseconds. This has to be subtracted from the actual x-over topology phase shift and then recalculate the values for the C and L.
Delayed networks are unnecessary EVIL that do more damage than good.
The network does not take the phase shift of the x-over at x-over frequency and the topology and impedance into account.
For example. If the tweeter is crossed over @ 2500khz and the distance is 2 inches and impedance of both drivers is 6 ohms at that point (use of zobel), you will have a phase shift of 133 degree and delay of 148 microseconds. This has to be subtracted from the actual x-over topology phase shift and then recalculate the values for the C and L.
Delayed networks are unnecessary EVIL that do more damage than good.
Eton said:
Yes, it does. Unless you mean the simplified value calculator, which I would agree is relatively useless. In fact, any textbook equations are useless and only used as a starting point, so I'm not sure what you're getting at.
Are you saying the crossover schematic I posted will not work in the way I have intended it? I can provide acoustic measurements matching up with modeled plots in response and phase. I can show a measured very deep reverse null on the tweeter axis with a LR2 crossover, along with symmetrical vertical lobing which pretty much is not going to happen without a correctly operating ladder delay network.
You can just get your optimum values in the ladder by hitting your target slopes in the cad frequency domain, locking those values and optimizing only the ladder components. I pretty much just dump some average values in there and let the optimizer do it's job.
Hi Zaph,
can you confirm the delay that the 4 components are achieving?
Is the return from the tweeter going back through L10 before it reaches ground?
On a completely different tack, is there a way to passively (before or after the amplifier) achieve a 3m to 6m delay?
Or is digital delay the only option?
can you confirm the delay that the 4 components are achieving?
Is the return from the tweeter going back through L10 before it reaches ground?
On a completely different tack, is there a way to passively (before or after the amplifier) achieve a 3m to 6m delay?
Or is digital delay the only option?
If you delay the output from the back woofer, you will effectively negate the main benefits of having a bipole (baffle step and motion enery cancellation). At that point, you may as well dump the network and put the back speaker on the front baffle. If your speaker has to stand close to the back wall, reconsider going with a bipole.
Dan
Dan
No john.
I mean that Soundeasy’s Lattice Network (which I have owned since it was SD1 and even before that and currently using V12) calculates the values based on ( Fo , Q , Rload ) it does not take into account phase shift at x-over frequency and x-over topology phase shift.
Yes you can use it as you said and by dumping some average values and then optimizing and coming up with values that are close enough ?
Here’s a screen shoot of basic program that I wrote years ago, which is based on totally different calculations. As you can see, all you have to do is input the actual distance, X-over frequency and Z at x-over and the values are calculated.
Attachments
AndrewT said:
Yes, and yes. I'll post another image below, which shows the phase and response before and after the delay network. The red frequency response and blue phase are with the delay network. The orange and aqua curves are without it. I removed the circuit and grounded the components that went through L10. I get about a 90 degree change in phase wrap at 2kHz, which is what the optimizer determined I needed to put the drivers in phase at the crossover frequency.
Eton said:
Ok, we're on the same page then. I agree, the calculator is useless though I do occasionally use it to just get some values in there before optimizing.
John
Attachments
No bipolars don't ussually have a delay for the rear speakers. Literature on the web and whitepapers can be found on definitives, and mirages websites. THe Advantages of Bipoles is a Quasi-omnidirectional radiation pattern. Basically imagine a pulsating cylinder of sound that disperses sound equally in all directions with just a small null 180 degrees of axis at the cabinet sides.
Hello, i'm new here.
I'm a car audio fanatic and i was reading this issue about time delay.
This is exactly what im looking for.
In my car i need to delay the left mid and tweeter with 1,3 milliseconds.
The mid playes from 129 Hz. to 3500 Hz. The tweeter from 3500 Hz. up.
Can somebody calculate me a filter including a time delay off 1,3 millisecond for the left part and no timedelay for the right part.
I would be very thankfull.
I'm a car audio fanatic and i was reading this issue about time delay.
This is exactly what im looking for.
In my car i need to delay the left mid and tweeter with 1,3 milliseconds.
The mid playes from 129 Hz. to 3500 Hz. The tweeter from 3500 Hz. up.
Can somebody calculate me a filter including a time delay off 1,3 millisecond for the left part and no timedelay for the right part.
I would be very thankfull.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.