Crossover Help

[mgrabow]

Well I pretty much know the drivers I am going to use in my Sonus project. Except the tweeter. I may opt for a Vifa XT25BG60-04. It seems to be nice and flat but I am worried the 4 ohms combined with the woofers in parallel (4 ohms) may dump the midrange down to low...


Anyway, the drawing below is the design I ran in X-over pro but x-over pro does not have the frequency response for the Ushers. I was hoping someone else may have something that can give me a better response.

I am leaning towards a 2nd order with impedance compensation on the woofers I have 250/4000 as the crossover points but the 4000 may be a bit high for that 7" woofer.

Once I make up my mind on the tweeter (I am open to suggestions) I will probably need to throw in an l-pad to match it with the rest of the drivers...

As usual, I am open to any suggestions...

[mgrabow]

I just noticed, The 560uf cap in the mid circuit aint gonna happen...

[martyh]

Question: Why are you crossing over so high? 4000 hz takes away some of the best qualities of that Vifa. Can your roll it back to around 3200 hz?

[mgrabow]

Question: Why are you crossing over so high? 4000 hz takes away some of the best qualities of that Vifa. Can your roll it back to around 3200 hz?


Like I said, I am open to suggestions. I have done so much in the past with dome midranges, I am used to doing that I guess...

[r-carpenter]

I'd do a 300/2600 split. Second order on the mid, possibly third on the tweeter and second on the woofers. 8945A is a 7 inch driver so 4000hz is quite a bit high. With the woofer this size used as a mid range I'd try to go as low as the tweeter allow me in order to get good power response.
Also, why over complicate impedance eq?
I've had good results with Hiquphones and Ushers so I can recommend OW1 but then again, there's Dayton RS28 that I'd be eager to try. Vifa aren't my favorite.
Can you measure FR and so on?

[dthomas]

I just noticed, The 560uf cap in the mid circuit aint gonna happen...


I have done several designs with the Usher, personally I would limit it to about 2000Hz. While it is a great driver a 7" is not going to be real great at low level detail at 3200 or 4000Hz not to mention it is going to be very directional. I preferred crossing it at around 1500 to 1700Hz but 2000Hz is probably safe.

[mgrabow]

Wow, that crossover point keeps getting lower and lower...... I would not be too comfortable with any tweeter going too low...

I have the Dayton in my current speakers. I am not overly pleased with them. They are not horrible, just not blowing my skirt up...


I can measure FR...

[Jeff B.]

Well I pretty much know the drivers I am going to use in my Sonus project. Except the tweeter. I may opt for a Vifa XT25BG60-04. It seems to be nice and flat but I am worried the 4 ohms combined with the woofers in parallel (4 ohms) may dump the midrange down to low...


Anyway, the drawing below is the design I ran in X-over pro but x-over pro does not have the frequency response for the Ushers. I was hoping someone else may have something that can give me a better response.

I am leaning towards a 2nd order with impedance compensation on the woofers I have 250/4000 as the crossover points but the 4000 may be a bit high for that 7" woofer.

Once I make up my mind on the tweeter (I am open to suggestions) I will probably need to throw in an l-pad to match it with the rest of the drivers...

As usual, I am open to any suggestions...

Why do you need, or even want, such a complex impedance compensation network on the woofer? That 16mH coil is a lot of coil and expense that I would not recommend implementing. What are you trying to achieve that requires such a circuit?

[mgrabow]

Why do you need, or even want, such a complex impedance compensation network on the woofer? That 16mH coil is a lot of coil and expense that I would not recommend implementing. What are you trying to achieve that requires such a circuit?



16mh...That was another value I missed... I need glasses...

I just check "include impenance equalization for this driver" , "Include resonance peaks" and "include box effects" on X-over Pro. I myself and not trying to over complicate things, but I have had pretty good results in the past. Again, I am open to suggestions. I'll take that as suggesting to not compensate for one or all of the 3 filters...

[Jeff B.]

16mh...That was another value I missed... I need glasses...

I just check "include impenance equalization for this driver" , "Include resonance peaks" and "include box effects" on X-over Pro. I myself and not trying to over complicate things, but I have had pretty good results in the past. Again, I am open to suggestions. I'll take that as suggesting to not compensate for one or all of the 3 filters...

I wouldn't include any of the three filter selections at all. Optimizing the transfer function without impedance compensation is quite easy to do., and much less costly most of the time.

However, unless you can work with real impedances, real in-box frequency response and actual phase, or correct dimensional offsets, no crossover simulation will be correct - and in some cases, it will be competely worthless. And unfortunately, as I understand it, Crossover Pro doesn't allow for all of these things in the design.

[mgrabow]

sounds like I need more design help than I thought...

[r-carpenter]

X-over Pro does not account for a relative and absolute phase of the drivers used in actual loudspeaker. You may easily end up with deeps and bumps at the crossover frequencies.
X-over pro is a good: "let's throw it together real quick and see what's up" kind of tool. It has updatability of the driver database and can save some time for info search and input. Automatic conversion tools are very useful as well.
What kind of measurement setup do you have? You said you can do FR of the speaker.

[mgrabow]

X-over Pro does not account for a relative and absolute phase of the drivers used in actual loudspeaker. You may easily end up with deeps and bumps at the crossover frequencies.
X-over pro is a good: "let's throw it together real quick and see what's up" kind of tool. It has updatability of the driver database and can save some time for info search and input. Automatic conversion tools are very useful as well.
What kind of measurement setup do you have? You said you can do FR of the speaker.

I have a beringer ecm 8000 mic and a Tascam us-122 interface. I use some of the freebee RTA software on the internet like room room eq wizard.

[r-carpenter]

Are you open to spend a few $ on Sound Easy?

[mgrabow]

yes, but propbably not prior to this project

[r-carpenter]

I am not sure how RTA takes the data. To my knowledge MLS type software is used to get acoustic data of a loudspeaker. Gated measurements are used. Basically gated response is calculated so that the only initial impulse is captured without the reflection from the nearest boundary. Near field and far field techniques are used for capturing different FR ranges and the response is summed up using software interface. There also derivatives from captured impulse such as CSD or waterfall that allow you to look at what else is going on as the signal decaying in time. Also very useful in identifying enclosure resonances and driver's own cone problems. I know that lots of people have good deal of success with ARTA. There's a free version of software but you can't save.
HOLMI Impulse is free as well and offers different types of chirp.
You can ballpark your crossover the way you doing now, then verify as you build, measure and change. It'll take a few hours but you can get great results.

Sound Easy has a build in interface to design crossover based on the captured FR, phase (with Hilbert/Bode transform) and electromechanical parameters which makes design process a bliss.

[mgrabow]

I am not sure how RTA takes the data. To my knowledge MLS type software is used to get acoustic data of a loudspeaker. Gated measurements are used. Basically gated response is calculated so that the only initial impulse is captured without the reflection from the nearest boundary. Near field and far field techniques are used for capturing different FR ranges and the response is summed up using software interface. There also derivatives from captured impulse such as CSD or waterfall that allow you to look at what else is going on as the signal decaying in time. Also very useful in identifying enclosure resonances and driver's own cone problems. I know that lots of people have good deal of success with ARTA. There's a free version of software but you can't save.
HOLMI Impulse is free as well and offers different types of chirp.
You can ballpark your crossover the way you doing now, then verify as you build, measure and change. It'll take a few hours but you can get great results.

Sound Easy has a build in interface to design crossover based on the captured FR, phase (with Hilbert/Bode transform) and electromechanical parameters which makes design process a bliss.



I would imagine that capture would be post cabinet build? since I will probably not have an extra $250 or so to invest in sound easy prior to this project being completed, I will probably need to stick with modeling the crossover with whatever data I have on hand...

[Chris Roemer]

I would imagine that capture would be post cabinet build? since I will probably not have an extra $250 or so to invest in sound easy prior to this project being completed, I will probably need to stick with modeling the crossover with whatever data I have on hand...

You really HAVE to use F/Z (response/impedance) data files. The software you're using builds nice looking "transfer functions", unfortunately, those filters only work on drivers that have flat response and impedance profiles. Over 95% of real world drivers do not behave that way.

Anything against using Jeff's PCD? It's within your budget. If having Excel is a problem, you can get at least a taste of how it could work by using "XOverSim.exe" (from the FRD Consortium site).

Your HP isn't bad.
Your BP isn't horrible, but drop the 10mH coil, 1 ohm resistor, and 560uF cap off the Zobel network. That' very expensive and does next to nothing. The mid's output would look better to me if you lifted its lower range and lowered the upper (it's "uptilted" now).
Your LP can drop the entire (8 component) Zobel network. When you're crossing a driver at 200-300 Hz (like in a 3-way) you're not really trying to impress a filter function on the driver in the range where its impedance is climbing, so the eq net has very little effect.

That 7" looks pretty rough for a mid. Here's where you stand. I modelled the TBFC tweeter, as I don't have data for yours, but the 7" Usher and the parallelled 8" are actual. If your tweeter is close to this one, I'd say attenuate it by about 2 more dB. If you're stuck with the 7" as a mid, you should change its response on both ends. As others have stated, 2.5k is about the absolute limit for a 7". Ideally (iiwm) I'd roll that 7" off by the notch at 1.2k. If its got a good bottom end, I'd use it for woofer duty in a 3-way.

Chris

[r-carpenter]

I would imagine that capture would be post cabinet build? since I will probably not have an extra $250 or so to invest in sound easy prior to this project being completed, I will probably need to stick with modeling the crossover with whatever data I have on hand...

That is exactly my suggestion because of the budget limitations. Download HOLMI and ARTA. You have hardware to run them already, so no $$$ necessary. Finish designing your crossover with X-Over. Build the casework and measure the results with ARTA. See what you need to change and how the changes affect what you hear.

Chris made an excellent suggestion and pretty much summed everything up. The only thing I am sure will be difficult for you is to design a notch without having all the data in hand. Not impossible tho.

[whatatrip]

sounds like I need more design help than I thought...

I know what that feels like since I’m fairly new to all of this but I do run Xover Pro and PCD side by side and have found that the SPL/Frequency response graphs are pretty close to each other given the same input. This has held true for the half dozen designs I have done. And both are pretty close to the finished product frequency response graphs. I say all of this assuming I am doing it right and it appears that I am but I am pretty low on the learning curve.

Sometimes it is easier for me to use Xover Pro and sometimes PCD. I use each to check the other. I found Xover Pro was a good jumping off point. I don’t know if I could have tackled PCD without the Xover experience first (and Brian Powers) to get my feet wet. Besides, I didn’t even know about other software until I came here and found out many use PCD. I purchased BassBox Pro and Xover Pro together. They come with lots of documentation and good reading.

If you use the auto calculation features of Xover Pro for impedance compensation it gets the job done but as someone already said, you can accomplish much of that in the crossover design and some of the impedance peaks are not a problem. An example is the woofer peaks at low frequencies. Xover Pro calculates what it will take to flatten the impedance out but it often is not necessary. A lot of the design work comes from knowing what to do and then using the programs to get there. Your costly parts are likely due to the “include resonance peaks” features in Xover Pro. Just make sure you need them. I have found that haven’t needed them.

PCD will do more than Xover Pro and in the hands of an expert, PCD is the correct tool. But it is not true that a good design can’t be made with Xover Pro. It is true that you could come up with a poor design in Xover Pro but you could do that in PCD also. Another thing about Xover Pro is that it will not plot notch filters as of this time.

I don’t know all the workings of Xover Pro but it can utilize real in-box frequency response but not actual measured acoustical phase. It does plot phase but I think it uses electrical phase from the T/S parameters. I know that phase is plotted with or without measured FRD data. Xover Pro will plot system response using only T/S parameters but it is more accurate when you import measured FR data.

[mgrabow]

All, thanks for the suggestions. Good stuff.

I am keeping with the 7" mid as I may eventually upgrade to scanspeak drivers. Bass Box pro is telling me that the usher will work fine in that 12 liter internal enclosure and in another thread I heard some good things about using that usher in a two way or as a mid...

The original design was for a Scanspeak 18W/8535 mid and two 21W/8555-01 woofers.

The woofers need a slightly smaller enclosure, but I can acomplish that internal to the enclosure. My ports will be swapable internally also...

[dthomas]

Wow, that crossover point keeps getting lower and lower...... I would not be too comfortable with any tweeter going too low...

I have the Dayton in my current speakers. I am not overly pleased with them. They are not horrible, just not blowing my skirt up...


I can measure FR...

That tweeter will easily handle 2000Hz and will do 2000Hz far better than the 7" Usher is going to do 3200Hz.

Crossing low eliminates a lot of issues especially when using a 7" driver for midrange duty. I think if you look most of the projects done with the Usher midwoofer cross at 1500Hz to 2000Hz. You are going to lose the ability to integrate the tweeter and midwoofer well that high along with poor off axis response crossing it that high. I only suggested this so you could get the most out of the drivers you are using. Plus if you are concerned with the tweeter there are several moderately priced tweeters that will handle crossover points as low as 1500Hz with out any concerns.

I agree that you should have no need for impedance equalization and the the filter itself can be optimized to overcome any issues you will have with impedance.

[mgrabow]

That tweeter will easily handle 2000Hz and will do 2000Hz far better than the 7" Usher is going to do 3200Hz.

Crossing low eliminates a lot of issues especially when using a 7" driver for midrange duty. I think if you look most of the projects done with the Usher midwoofer cross at 1500Hz to 2000Hz. You are going to lose the ability to integrate the tweeter and midwoofer well that high along with poor off axis response crossing it that high. I only suggested this so you could get the most out of the drivers you are using. Plus if you are concerned with the tweeter there are several moderately priced tweeters that will handle crossover points as low as 1500Hz with out any concerns.

I agree that you should have no need for impedance equalization and the the filter itself can be optimized to overcome any issues you will have with impedance.

Thanks. On the impedance circuit x-over pro shows a funky peak on the 7 inch woofers FR without it...

[r-carpenter]

are several moderately priced tweeters that will handle crossover points as low as 1500Hz with out any concerns.



ehhhh I have to respectfully disagree. Perhaps there are (and I just haven't gotten to a good one) but I've tested a few budget (around and under $50) tweeters and 1500hz crossover point appears to be a bit too low. Do you mean to cross the tweeter around 1500hz and have 20 db down at ... say a 1000hz or do you mean to cross it around 2k and have it 20db down at 1500hz? I think that the real trade off here, given selected drivers is off axis response vs getting in to the distortion area of the tweeter. Are we talking 4th order electrical here?
I'd go just a bit higher, say to 2200-2600hz on the midwoofer/tweeter. It would definitely simplify a tweeter selection. 1inch silk dome would do great. Also, 8945A (not exactly a perfect choice for a mid tho) is not horrible in off axis region so I think between 1" silky they would have an ok power response.
A simple test for Mark, would be to build crossover for a chosen tweeter and run FR sweep with just a tweeter connected.
But yes, definitely, go as low as the tweeter will allow!
Just an opinion.:)

[dthomas]

ehhhh I have to respectfully disagree. Perhaps there are (and I just haven't gotten to a good one) but I've tested a few budget (around and under $50) tweeters and 1500hz crossover point appears to be a bit too low. Do you mean to cross the tweeter around 1500hz and have 20 db down at ... say a 1000hz or do you mean to cross it around 2k and have it 20db down at 1500hz? I think that the real trade off here, given selected drivers is off axis response vs getting in to the distortion area of the tweeter. Are we talking 4th order electrical here?
I'd go just a bit higher, say to 2200-2600hz on the midwoofer/tweeter. It would definitely simplify a tweeter selection. 1inch silk dome would do great. Also, 8945A (not exactly a perfect choice for a mid tho) is not horrible in off axis region so I think between 1" silky they would have an ok power response.
A simple test for Mark, would be to build crossover for a chosen tweeter and run FR sweep with just a tweeter connected.
But yes, definitely, go as low as the tweeter will allow!
Just an opinion.:)


You should try the Seas 27TBFC or 27TDFC or RS28, the North Creek D2605 all handle 1500Hz well. I guess you did not test these tweeters but I think 4 qualify as several. Of course 1500 Hz is not best for all tweeters but there are several that are reasonably prices and based on the midwoofer cost I think all of the ones I mentioned ought to be in the right price range. The Peerless HDS works great at 1500Hz or even slightly sub 1500Hz if you needed it to it might be a bit pricey but it is not anymore expensive than the Usher midwoofer. I assumed based on the fact he was using the Usher midwoofer his tweeter budget was more than a Dayton silkie.


I think the Usher is an excellent choice for a mid if you cross it low enough. That was my point. As I stated earlier crossing at 1500 to 2000Hz makes the most sense. I assumed that he would understand not all tweeters can be crossed down as low as 1500Hz to 2000Hz If you move up to 2000Hz there are several more tweeters choices. Not sure I would mate up the the Dayton Silkie and an Usher 8945A. While I am sure you can get the Usher to play higher I only recommended he consider a crossover point of 2000Hz to 1500Hz. If you go with 2000Hz there are several more tweeters that would fit that bill.

No I mean cross it at 1500Hz with a 4th order electrical network this with the tweeters natural acoustic roll off should get you better than 4th order acoustic roll off.

[r-carpenter]

TBFC and TDFC I've tested and used a number of times. I personally didn't think they were very good below 2k. Third order electrical was used in particular systems.
I hear you, I just think it's pushing the unnecessary limit.:)
In any event, let's see what tweeter Mark will get.

[dthomas]

TBFC and TDFC I've tested and used a number of times. I personally didn't think they were very good below 2k. Third order electrical was used in particular systems.
I hear you, I just think it's pushing the unnecessary limit.:)
In any event, let's see what tweeter Mark will get.

Not to sure how you tested them but both of the Seas tweeters have very low distortion at 1500Hz and less than most midwoofers at this frequency. Also as I stated I used a 4th order electrical which obviously would better limit the applied voltage below 1500Hz. But I disagree that you are pushing their limits they perform flawlessly with very low distortion and not a hint of over excursion. I think these tweeters set new limits on how low a typical dome can go. There are numerous of measurements that verify the distortion and low frequency operation of these tweeters.

Many would consider pushing a 7" midwoofer to 3200 Hz is pushing the limits of wide dispersion and entering a region where beaming will become audible. By the way in my original post i suggested 2000Hz as a better crossover point for the Usher midwoofer and that plenty of tweeters were capable of reaching that low. But there have been many systems implemented with good success and reviews with both tweeters being used as low as 1500Hz. So for the record I disagree that you are pushing the limits with these tweeters at 1500Hz but are well within their range of operation.

[mgrabow]

I know what that feels like since I’m fairly new to all of this but I do run Xover Pro and PCD side by side and have found that the SPL/Frequency response graphs are pretty close to each other given the same input. This has held true for the half dozen designs I have done. And both are pretty close to the finished product frequency response graphs. I say all of this assuming I am doing it right and it appears that I am but I am pretty low on the learning curve.

Sometimes it is easier for me to use Xover Pro and sometimes PCD. I use each to check the other. I found Xover Pro was a good jumping off point. I don’t know if I could have tackled PCD without the Xover experience first (and Brian Powers) to get my feet wet. Besides, I didn’t even know about other software until I came here and found out many use PCD. I purchased BassBox Pro and Xover Pro together. They come with lots of documentation and good reading.

If you use the auto calculation features of Xover Pro for impedance compensation it gets the job done but as someone already said, you can accomplish much of that in the crossover design and some of the impedance peaks are not a problem. An example is the woofer peaks at low frequencies. Xover Pro calculates what it will take to flatten the impedance out but it often is not necessary. A lot of the design work comes from knowing what to do and then using the programs to get there. Your costly parts are likely due to the “include resonance peaks” features in Xover Pro. Just make sure you need them. I have found that haven’t needed them.

PCD will do more than Xover Pro and in the hands of an expert, PCD is the correct tool. But it is not true that a good design can’t be made with Xover Pro. It is true that you could come up with a poor design in Xover Pro but you could do that in PCD also. Another thing about Xover Pro is that it will not plot notch filters as of this time.

I don’t know all the workings of Xover Pro but it can utilize real in-box frequency response but not actual measured acoustical phase. It does plot phase but I think it uses electrical phase from the T/S parameters. I know that phase is plotted with or without measured FRD data. Xover Pro will plot system response using only T/S parameters but it is more accurate when you import measured FR data.


I downloaded PCD... Its gonna take some time to figure out how to use it. I have had very good luck with x-over pro.

[davepellegrene]

Not to sure how you tested them but both of the Seas tweeters have very low distortion at 1500Hz and less than most midwoofers at this frequency.

Would it be as simple as setting up your crossover to 1500 Hz and hook it up by itself and listen?
I assume that would go for any driver that you want to hear how it plays at a curtain frequency.
Dave