Hi everyone,
First of all a happy newyear!
I'm working on my first speaker design. It's basically an MTM with the woofers in a serparated vented box.
Woofers: HDS-P830883, connected in parallel
Crossover: 3th order (see below)
I'm having some baffle step issues and implemented a bafflestep correction of 3db, after seeing the not-pleasing response I added them together to use as a 6db baffle step correction to see what this 'll do.
Paramters used to calculate the bafflestep:
ReVC= 5,85ohm/2= 2,925ohm (because they are connected parallel)
baffle width= 30cm (baffle measured left-to-right)
This gave as bafflestep values (for 3db attenuation)
R= 1,2ohm
L= 0,5mH
As you can see in the measurements of the woofers it seems like the bafflestep is just giving an offset mostly in the range of 250Hz-5kHz
Does someone see what's going on here?
Thanks!
Regards
First of all a happy newyear!
I'm working on my first speaker design. It's basically an MTM with the woofers in a serparated vented box.
Woofers: HDS-P830883, connected in parallel
Crossover: 3th order (see below)
I'm having some baffle step issues and implemented a bafflestep correction of 3db, after seeing the not-pleasing response I added them together to use as a 6db baffle step correction to see what this 'll do.
Paramters used to calculate the bafflestep:
ReVC= 5,85ohm/2= 2,925ohm (because they are connected parallel)
baffle width= 30cm (baffle measured left-to-right)
This gave as bafflestep values (for 3db attenuation)
R= 1,2ohm
L= 0,5mH
As you can see in the measurements of the woofers it seems like the bafflestep is just giving an offset mostly in the range of 250Hz-5kHz
Does someone see what's going on here?
Thanks!
Regards
Can you tell us how you arrived at your current crossover design? Since you are able to measure, which XO simulation program did you use?
Normally - BSC is not done via a separate circuit. You usually use a lead series inductor around 1.8mH - 3.0mH.
Your impedance estimates for your BSC calculation... are they driver Re values? or actual values over the frequency range in question? When combined with other components in the woofer circuit - the impedance will vary quite a bit - meaning a textbook BSC calculation will never hit the mark.
Normally - BSC is not done via a separate circuit. You usually use a lead series inductor around 1.8mH - 3.0mH.
Your impedance estimates for your BSC calculation... are they driver Re values? or actual values over the frequency range in question? When combined with other components in the woofer circuit - the impedance will vary quite a bit - meaning a textbook BSC calculation will never hit the mark.
Just a thought.... looks like the natural rising reposnse of the dual woofers is swamping your baffle step correction. You need to adjust your low pass filter to flaten the rise out. How are you fixed for xover parts? A second order - zobel combo set down around 400 - 500 Hz may get you bettor over all results.
The crossover was simulated in Xsim, and the measurements were done in REW.
How do you mean "seperate circuit"? It's there... betwheen the Crossover and the woofers (and zobel).
Well, the Re was just taken from the driver specifications 🙂...
For a baffle correction it needs the ReDC, being the DC resistance of the driver, not the impedance...?
Last edited:
What I mean is you normally don't have to add a separate BSC circuit as you have. by replacing the 0.4mH inductor with a larger one - you get BSC with less components.
Doesn't really matter what the response is due to the crossover, as I understand the baffle step compensation should do what it's ment to do... no?
Hi,
Ok, I'm a bit confused by this comment...
Replacing the 0.4mH Inductor does not simply compensate for baffle step losses... It's simply changing the crossover, including the crossover frequention.
Why is it even called a "baffle step compensation circuit" then?
Call it a trick of the trade. A well designed crossover isn't simply taking separate circuits and bolting them together. This is for 2 reasons:
1. Separate circuits do not account for the driver specifics and impedance changes introduced by other circuits
2. You can save money by having one component do 2 jobs.
For example - impedance compensation (zobel) and resonant peak networks are not required if you are not crossing over near those points or your woofer impedance rises gradually not need compensation. Even so, a zobel may not be required if a shaping resistor is added to align phase in the parallel leg of a 2nd order network. This is an example of combing 2 treatments into one set of components.
I am suggesting a rethink of your woofer circuit which you must design in conjunction with your tweeter circuit (which we have not seen). Can you show us the tweeter circuit and response including overall response as well as what target acoustic slopes you are aiming for?
In a large MTM like this one it is advisable to have different filtersections for both midbass drivers. So, the best solution in this situation would be to filter one of the midbass drivers at a much lower frequency than the other one. This would improver the fr immediately and lead to a much cleaner vertical dispersion and better power response.
Quiz question: which of the midbass drivers should have the lower cutoff frequency?
Edit: clue Phase Response in Active Filters Part 2, the Low-Pass and High-Pass Response | Analog Devices
Quiz question: which of the midbass drivers should have the lower cutoff frequency?
Edit: clue Phase Response in Active Filters Part 2, the Low-Pass and High-Pass Response | Analog Devices
Last edited:
Hi Dave,
Sorry but your explanation doesn't make a lot of sense to me.
Conserning the Zobels, I'm not sure how you will compensate for the rising impedance of a woofer just by changing the values of your crossover...
Anyway, the zobels don't have anything to do with this issue.
I'm also not sure how you can compensate for baffle step losses just by "playing" with the values of the crossover itself.
So Zobel networks and bsc circuits are al unnessasery in a good crossover design? What about notch filters? What about the LS3/5A crossovers? It's almost a computer based on the amount of components? Or is it just a bad design?
Not ment to be rude or whatever 😉
Sorry but your explanation doesn't make a lot of sense to me.
Conserning the Zobels, I'm not sure how you will compensate for the rising impedance of a woofer just by changing the values of your crossover...
Anyway, the zobels don't have anything to do with this issue.
I'm also not sure how you can compensate for baffle step losses just by "playing" with the values of the crossover itself.
So Zobel networks and bsc circuits are al unnessasery in a good crossover design? What about notch filters? What about the LS3/5A crossovers? It's almost a computer based on the amount of components? Or is it just a bad design?
Not ment to be rude or whatever 😉
This seems to me like running away of present problems by just changing the concept to a 3-way?
You posted something on the same subject in another thread. The forum etiquette is not to post multiple threads about the same subject. Anyway I will say here what I already said: you have to completely rework your crossover, for the reasons Dave explained here. A Zobel has almost no place in a properly designed crossover, its use was something mandatory when no crossover simulators were available and the only way to calculate the components was in maintaining fixed the impedance. The baffle step circuit has a sense only with a full range driver used alone. Examples of baffle step compensation done right with the use of an appropriately sized first coil in series with the woofer abound. For an example read here and have a look at the image labeled "Transfer functions": Zaph|Audio - ZA-SR71
In general Zaph designs are well described and are a mine of knowledge, if you are willing to understand.
The key in crossover design using a simulator like XSim is in using valid data, in other words you need to correctly measure your drivers in the intended cabinet. A sweep from REW imported into XSim is not a valid measure for a crossover design. Period. What has to be done is an appropriate gated far field measure with phase for both woofers and tweeter. A near field measure of the woofer will show the baffle step you need.
Ralf
You don't need to, at least for the vast majority of cases, surely not with you woofer.
Have a look at this thread where I posted my journey on a speaker with your same woofer but with a different tweeter. The crossover details is in the post #29 here: Another 6.5”+1” BR design (Peerless HDS + Seas Prestige)
I used an evolution of this crossover as a base to obtain what I presented in your other thread (post #19): LS2019, time to make these better
Clearly this crossover won't exactly work this way because my baffle is different than yours, and the tweeter is different. But it is clear that with only one element more I avoided all your problems (baffle step non addressed correctly, probable poor phase alignment in the crossover point, the 5dB drop between 3KHz and 4KHz, and very poor sum of FR between 4KHz and 5KHz).
Ralf
I applaud your effort.
I like the driver, and thinking for me, a 6db time/phase aligned mtm to morel cat378, 6.5" mtm requiring 10' listening distance (like dunlavy).
But baffle step, hmm........
For you, increase the resistor some more maybe ?
Or go for more than 3db, I'd go for 6db, i think you change the inductor also though.
Easy try.
I like the driver, and thinking for me, a 6db time/phase aligned mtm to morel cat378, 6.5" mtm requiring 10' listening distance (like dunlavy).
But baffle step, hmm........
For you, increase the resistor some more maybe ?
Or go for more than 3db, I'd go for 6db, i think you change the inductor also though.
Easy try.
What is happening is entirely as expected. You're thinking about this the wrong way. Passive baffle step compensation doesn't boost the frequencies below the baffle step frequency, it cuts everything above it. Effectively this gives a boost to the bass relative to mid/treble but also an overall reduction in sensitivity. You can't increase the amount of bass with passive filters because the woofers are already run flat out (unattenuated) over those frequencies even before you add BSC.
If you increase the amount of BSC, you need to lower the level of the tweeter by the same amount, otherwise you end up making a 'V' shaped frequency response.
Looks like you need to tweak the Q of your woofer lowpass filter as well because when you increase the BSC, the level at 1.5kHz isn't changing appropriately. This is because the load impedance your lowpass filter 'sees', which is the series combination of the BSC network and the woofers, is increased when you add more BSC.
Last edited:
LSspeaker,
the purpose of a crossover and additional networks/filters is frequency response and impedance shaping. If one can achieve satisfactory results with fewer parts, fine. Otherwise, additional ones are needed. The main issue with loudspeaker design is to determine what kind of a frequency response, where in space and how many of these should we record to draw the right kind of a conclusion.
the purpose of a crossover and additional networks/filters is frequency response and impedance shaping. If one can achieve satisfactory results with fewer parts, fine. Otherwise, additional ones are needed. The main issue with loudspeaker design is to determine what kind of a frequency response, where in space and how many of these should we record to draw the right kind of a conclusion.
The only reason to employ 2 midbass drivers in a loudspeaker is because 1 midbass driver would hit excursion limits at the desired max SPL.
When you employ 2 midbass drivers in a 2-way, you need to make it a 2.5. Running both midbass drivers to the full xover frequency to the tweeter leads to unacceptable lobing and power response issues.
Putting the tweeter in between the two midbass drivers only excacerbates these lobing and power response issues, because of the increased distance between the two midbass drivers.
There are no two ways about this.
When you employ 2 midbass drivers in a 2-way, you need to make it a 2.5. Running both midbass drivers to the full xover frequency to the tweeter leads to unacceptable lobing and power response issues.
Putting the tweeter in between the two midbass drivers only excacerbates these lobing and power response issues, because of the increased distance between the two midbass drivers.
There are no two ways about this.