well, thats why I asked. I ment using a filter on top of the natural slope, so I actually get about 4th order slope with a 2nd order circuit.
before viewing your post I've downloaded spltrace and a passive xover simulator/designer. I set the xover point to 3K with a second order filter on both driver, the tweeter with an L-pad, and here is what I got:
http://www.y44.com/ims/album.php?u_id=110RNFYf
before viewing your post I've downloaded spltrace and a passive xover simulator/designer. I set the xover point to 3K with a second order filter on both driver, the tweeter with an L-pad, and here is what I got:
http://www.y44.com/ims/album.php?u_id=110RNFYf
You have the right kind of idea then, all speakers mainly use an electrical filter with the drivers natural roll off to arrive at the desired acoustic slope.
You will probably notice however that if you use any crossover frequency from about 1.5khz to 4khz that the drivers will sum well and have a good notch at the reverse polarity. This is because the program is not using any phase data, which is the most important thing to take into acount. You cannot really do anything about this if you cant measure the drivers in the box you intend to use them in.
You are on the right path mind you, at least you are not just going to throw a text book crossover at it and expect the earth to move.
If you want to get the most out of what you have at your disposal then make sure you flush mount all the drivers. This will minimise any diffraction issues that may occur that would otherwise require work with the crossover. Because you cant measure you wont know they are their, so its better to make sure the chance of them being there is reduced.
Then use that program (or loudspeaker workshop which is also freeware) to get a nice flat frequency response. Also you will need to include baffle step. If your not sure what this is that SEAS T17 response graph illustrates it brilliantly.
If you look at 600hz up its flat to about 4khz where driver roll off happens. But below that from 600hz down to about 150hz there is a gradual decrease by about 6dB (this is more like 5dB because there is a tiny bit of room gain in the SEAS measuring facility). But this drop from 600hz to 150hz by 5-6dB is baffle step. You will also see in your graphs of the predicted response that this is also present.
When this "step" occurs depends on the baffle width so should really be compensated for, for the inclosure you will be using. However I feel this would be quite difficult to achieve. I would use the SEAS graph and use their measured baffle step as the thing you need to equalise rather then trying to get you own sorted. Baffle step is not a highly accurate thing anyway and being out by a few hundred hz is a lot better then not applying any compensation at all.
A resistor and inductor in series, connected in parallel with the woofer before the filter is how you do it.
It can also be done by increasing the value of the inductor in the filter. Watch out for the impedance though in the first network, it can cause it to dip very low.
You will probably notice however that if you use any crossover frequency from about 1.5khz to 4khz that the drivers will sum well and have a good notch at the reverse polarity. This is because the program is not using any phase data, which is the most important thing to take into acount. You cannot really do anything about this if you cant measure the drivers in the box you intend to use them in.
You are on the right path mind you, at least you are not just going to throw a text book crossover at it and expect the earth to move.
If you want to get the most out of what you have at your disposal then make sure you flush mount all the drivers. This will minimise any diffraction issues that may occur that would otherwise require work with the crossover. Because you cant measure you wont know they are their, so its better to make sure the chance of them being there is reduced.
Then use that program (or loudspeaker workshop which is also freeware) to get a nice flat frequency response. Also you will need to include baffle step. If your not sure what this is that SEAS T17 response graph illustrates it brilliantly.
If you look at 600hz up its flat to about 4khz where driver roll off happens. But below that from 600hz down to about 150hz there is a gradual decrease by about 6dB (this is more like 5dB because there is a tiny bit of room gain in the SEAS measuring facility). But this drop from 600hz to 150hz by 5-6dB is baffle step. You will also see in your graphs of the predicted response that this is also present.
When this "step" occurs depends on the baffle width so should really be compensated for, for the inclosure you will be using. However I feel this would be quite difficult to achieve. I would use the SEAS graph and use their measured baffle step as the thing you need to equalise rather then trying to get you own sorted. Baffle step is not a highly accurate thing anyway and being out by a few hundred hz is a lot better then not applying any compensation at all.
A resistor and inductor in series, connected in parallel with the woofer before the filter is how you do it.
It can also be done by increasing the value of the inductor in the filter. Watch out for the impedance though in the first network, it can cause it to dip very low.
thanks a lot.
Do you have any idea of the values of the C/R?
another thing:
I have read somewhere in this forum that bafflestep is reduced (so much that one dosn'e have to compensate for it), if the speaker is very close to the wall, is this true?
Mine are supposed to sit on a little shelf, or an L shaped metal strap, on the upper left and right corners of my room, quite tightly against the wall.
edit: Speaker Workshop is not yet functional, it is still in building stages.
Do you have any idea of the values of the C/R?
another thing:
I have read somewhere in this forum that bafflestep is reduced (so much that one dosn'e have to compensate for it), if the speaker is very close to the wall, is this true?
Mine are supposed to sit on a little shelf, or an L shaped metal strap, on the upper left and right corners of my room, quite tightly against the wall.
edit: Speaker Workshop is not yet functional, it is still in building stages.
You get 3dB of gain per wall.
So place your speakers floating in mid air and you get zero bass boost. Place them in the middle of the floor and you get 3dB, place them in the junction between wall and floor and you get 6dB boost. Place them in a corner where three walls intersect and you get 9dB boost.
So in anwer to your question, yes you get boost from the walls.
If you are going to place them close to a corner dont include baffle step compensation otherwise they will sound terrible.
C and R values depend on the drivers. But at a guess (and its only a roungh guess) inductor value about 1.5mh resistor 8ohms.
So place your speakers floating in mid air and you get zero bass boost. Place them in the middle of the floor and you get 3dB, place them in the junction between wall and floor and you get 6dB boost. Place them in a corner where three walls intersect and you get 9dB boost.
So in anwer to your question, yes you get boost from the walls.
If you are going to place them close to a corner dont include baffle step compensation otherwise they will sound terrible.
C and R values depend on the drivers. But at a guess (and its only a roungh guess) inductor value about 1.5mh resistor 8ohms.
o.k, so I have some problems.
fore on thing, my filter includes a parellel RL contour network for the woofer, but this nakes his avg snsitivity anout 85 dB, ehich means my tweeter has to be attenuated to 85 dB too with an L-plad filter. now, after all this I get a really smooth response graph, but with a 85 sesitivity, is that too low a sensitivity?
secondly, with the contuor network my flat line begins at about 100-130 Hz, I wondered how low does the bass compensation from the walls begins? I mean would it make like a big bump in the freq response or does it begin and end earlier?
thanks a lot
fore on thing, my filter includes a parellel RL contour network for the woofer, but this nakes his avg snsitivity anout 85 dB, ehich means my tweeter has to be attenuated to 85 dB too with an L-plad filter. now, after all this I get a really smooth response graph, but with a 85 sesitivity, is that too low a sensitivity?
secondly, with the contuor network my flat line begins at about 100-130 Hz, I wondered how low does the bass compensation from the walls begins? I mean would it make like a big bump in the freq response or does it begin and end earlier?
thanks a lot
Nope this is fine, if you look at the stock sensativity for drivers they tend to be 85-90 dB. After full bafflestep compensation you would end up with 79-84dB sensativity so your doind nothing wrong there. Just check your input impedance doesnt drop too low.
Dont worry this is because of the box SEAS put the driver in if you put the driver in a box designed for the driver you will get much more extended bass.
Room gain and sound reinforcement of bafflestep issues are two different things.
Room gain occurs at the frequency whose wavelength is twice the longest room dimension. So for my room which is 4m long thats C=fl (l = wavelength) 340/8 = approx 40hz. So in my room I will get room gain from 40hz downwards.
But placing a speaker near a wall is sound reinforcement, whereby the sound the speaker produces omnidirectionally, will be reflected off the wall and boost the original wave from the front of the speaker. This should happen about the same frequencies as baffle step because a speaker starts to become omnidirectional when bafflestep occurs. That is why bafflestep occurs infact. Thats the theory anyway.
I have spoken with Zalytron (both Rhonda and Elliot), and they said that the T17RE isn't currently available, but there is an order they made from seas that might take a while to get there (they said it's because it's from norway, so i guess they get it by ship), and I have mailed Elliot, and he said he will mail SEAS and make inquiries on the matter.
they also don't have the tweeter I want (27TFFC) but a newer, allegedly improved version of it, called the 27DFC, from looking at the data sheets, 27TFFC , 27DFC , I think the TFFC is better so I asked him to check on that too.
Generally they said it might take some time (two months or so, maybe more) untill the order gets there.
Personally I am willing to wait. my design is almost complete for these two drivers, and I'm in no rush.(although it will be nice yo start working already
)
they also don't have the tweeter I want (27TFFC) but a newer, allegedly improved version of it, called the 27DFC, from looking at the data sheets, 27TFFC , 27DFC , I think the TFFC is better so I asked him to check on that too.
Generally they said it might take some time (two months or so, maybe more) untill the order gets there.
Personally I am willing to wait. my design is almost complete for these two drivers, and I'm in no rush.(although it will be nice yo start working already
)
Actually the specs suggest that the 27TDFC is a better driver.
It has a lower free air resonance, slightly lower frequency response, and a cleaner looking waterfall.
Plus several years more SEAS expertise, all of these should make for a better driver. However both should be good and I know the TFFC is very highly regarded, but then its been around long enough to be highly regarded.
It has a lower free air resonance, slightly lower frequency response, and a cleaner looking waterfall.
Plus several years more SEAS expertise, all of these should make for a better driver. However both should be good and I know the TFFC is very highly regarded, but then its been around long enough to be highly regarded.
Good point- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.