Speakerplans.com Homepage
Forum Home Forum Home > General > Advanced Discussion
  New Posts New Posts RSS Feed - Bandpass overlap with High and Low pass
  FAQ FAQ  Forum Search   Events   Register Register  Login Login

Bandpass overlap with High and Low pass
 Post Reply Post Reply Page  12>
Author
Message
  Topic Search Topic Search  Topic Options Topic Options
JoelDD View Drop Down
Registered User
Registered User


Joined: 20 April 2018
Status: Offline
Points: 12 		Post Options Post Options   Thanks (0) Thanks(0)   Quote JoelDD Quote  Post ReplyReply Direct Link To This Post Topic: Bandpass overlap with High and Low pass
    Posted: 20 April 2018 at 11:04am
Hello, new here so be gentle..

I have been building passive crossovers for personal use for a few years now and I have had experience building everything from first order two way all the way to fourth order three ways. I know enough about crossovers now that I can reverse engineer most crossovers just by looking at the components and values, but one thing still mystifies me. 

I have noticed in almost all cases where a bandpass is used, there is always a 160% of high pass crossover point overlap and around -36% low pass crossover point overlap. Even the auto calculators online do this and I am wondering, why? It would seem to me that if you want the crossover point to be as flat as possible, you would want the crossover to begin at or just before the actual crossover point and not beyond it. 

The only explanation that I can come up with in my head is that there is something electrical that is different about the circuitry of a bandpass filter that changes the required values to reach the same crossover point. Only problem is that I have never seen this explained anywhere. From a simple observers point of view it is just a high pass and low pass sandwiched together and should function basically the same as they run in parallel to the high and low pass filters and not in series.

I understand that some designers may overlap drivers to make up for deficiencies in the drivers themselves and sometime box reflections, but there has to be a reason why it seems to be the go to standard. It can't be just because someone successful did it and now everyone just does it and doesn't question it. At least I'd hope not.

The reason why I want to figure this out is because I want to start using my own custom calculations to build my crossovers. So far I have just been using online calculators and adapting the designs using a LCR meter as reference for the drivers, but this limits me to typical designs that only make use of a single bandpass. I want to figure this out so I can stack bandpasses and make four way passive crossovers. 

If this just comes down to allowance for margin of error, then I may just start using my own calculations regardless because I have learned to trust my own measurements over rough guesses.


Edited by JoelDD - 20 April 2018 at 11:08am
Back to Top
gen0me View Drop Down
Young Croc
Young Croc
Avatar

Joined: 20 February 2016
Location: UK
Status: Offline
Points: 999 		Post Options Post Options   Thanks (0) Thanks(0)   Quote gen0me Quote  Post ReplyReply Direct Link To This Post Posted: 20 April 2018 at 2:02pm
Bandpass impulse is already horrible crossover doesnt matter in terms of destroying the phase.

Edited by gen0me - 20 April 2018 at 11:10pm
I appreciate every like :)) https//www.facebook.com/genomesoundsystems
Mixes: https://www.mixcloud.com/gen-ome/
Back to Top
LeruSound View Drop Down
Registered User
Registered User
Avatar

Joined: 27 August 2015
Status: Offline
Points: 171 		Post Options Post Options   Thanks (0) Thanks(0)   Quote LeruSound Quote  Post ReplyReply Direct Link To This Post Posted: 20 April 2018 at 2:24pm
I am now setting a passive crossover for 15" BPH to an 8" horn.
I measured and i did not overlap the cutoffs. I measure and load the measurements files in a simualtor in where i can build the crossover and see the results with real response on and off axis. Due to pretty much different response, i decided to set a very flat L-R 12db/oct, phase is locked, i'll se how it permorms and let you know. Its first time i try with a BPH, and it turned out i did the simpler ever, but it's a first try, will see...
Back to Top
JoelDD View Drop Down
Registered User
Registered User


Joined: 20 April 2018
Status: Offline
Points: 12 		Post Options Post Options   Thanks (0) Thanks(0)   Quote JoelDD Quote  Post ReplyReply Direct Link To This Post Posted: 20 April 2018 at 10:47pm
Yeah, the more I look into this the more I think it is just to "cover all the bases". I did some manual calculation last night and there was so little difference between the capacitor and inductance values that your driver selection and zobel network implementation would have a much greater effect on where your crossover is actually happening. Something to note is that I see this the most in cheaper speaker crossovers. For example, some cheap Chinese in-wall units that I picked up have the crossovers set to 4500hz high pass, 5800hz-270hz bandpass and 300hz low pass. Mind you they omitted a phase inductor on the bandpass, didn't zobel network anything and have no driver attenuation on the bandpass. I plan to rebuild these networks entirely as they sound god awful.

Something interesting though is that this calculator, https://www.diyaudioandvideo.com/Calculator/ApcSpeakerCrossover/ , sets the the first half of the high pass section of the bandpass of a fourth order three way much wider than the second half of the same high pass filter section of that very same bandpass. At 3500/437.5 8Ohm, it sets C3 at 38.12uF, L3 at 1.44mH, C4 at 53.86uF and L4 8.23mH. While if I calculate it manually using fourth order Linkwitz Riley formula, I get C3 at 24.11uF, L3 at 1.83mH, C4 at 48.23uf and L4 at 8.23mH. As you can see the calculator is making the first half of the high pass section much wider, but falls closer in line in the second half of the high pass section. The High pass and low pass filters are always the same as my calculations.


Edited by JoelDD - 20 April 2018 at 11:10pm
Back to Top
gen0me View Drop Down
Young Croc
Young Croc
Avatar

Joined: 20 February 2016
Location: UK
Status: Offline
Points: 999 		Post Options Post Options   Thanks (0) Thanks(0)   Quote gen0me Quote  Post ReplyReply Direct Link To This Post Posted: 20 April 2018 at 11:12pm
Look at the group delay mountain on the higher spike. Trying to cross it you will get cancelations anyway around the crossover point. Notice on ear it sounds like just changing spl  is controlling crossover. I think any random boxes crossed together without setting delay times sound similar dont they?
I appreciate every like :)) https//www.facebook.com/genomesoundsystems
Mixes: https://www.mixcloud.com/gen-ome/
Back to Top
JoelDD View Drop Down
Registered User
Registered User


Joined: 20 April 2018
Status: Offline
Points: 12 		Post Options Post Options   Thanks (0) Thanks(0)   Quote JoelDD Quote  Post ReplyReply Direct Link To This Post Posted: 21 April 2018 at 12:17am
Originally posted by gen0me gen0me wrote:

Look at the group delay mountain on the higher spike. Trying to cross it you will get cancelations anyway around the crossover point. Notice on ear it sounds like just changing spl  is controlling crossover. I think any random boxes crossed together without setting delay times sound similar dont they?

Yes, driver characteristics will play into where you cross a driver. You should have at least an octave of relative calm near the crossover point. Is this a single driver measure or a cabinet?
Back to Top
DMorison View Drop Down
Old Croc
Old Croc


Joined: 14 March 2007
Location: Aberdeen
Status: Offline
Points: 1649 		Post Options Post Options   Thanks (0) Thanks(0)   Quote DMorison Quote  Post ReplyReply Direct Link To This Post Posted: 21 April 2018 at 1:28pm
Originally posted by gen0me gen0me wrote:


gen0me, he's talking about something completely different - that's a model of a Bandpass type bass bin, the OP is talking about the electrical filters for the mid band of a three way crossover: https://en.wikipedia.org/wiki/Band-pass_filter.


Edited by DMorison - 21 April 2018 at 1:39pm
Back to Top
DMorison View Drop Down
Old Croc
Old Croc


Joined: 14 March 2007
Location: Aberdeen
Status: Offline
Points: 1649 		Post Options Post Options   Thanks (0) Thanks(0)   Quote DMorison Quote  Post ReplyReply Direct Link To This Post Posted: 21 April 2018 at 1:33pm
Originally posted by JoelDD JoelDD wrote:

Yeah, the more I look into this the more I think it is just to "cover all the bases". I did some manual calculation last night and there was so little difference between the capacitor and inductance values that your driver selection and zobel network implementation would have a much greater effect on where your crossover is actually happening. Something to note is that I see this the most in cheaper speaker crossovers. For example, some cheap Chinese in-wall units that I picked up have the crossovers set to 4500hz high pass, 5800hz-270hz bandpass and 300hz low pass. Mind you they omitted a phase inductor on the bandpass, didn't zobel network anything and have no driver attenuation on the bandpass. I plan to rebuild these networks entirely as they sound god awful.

Something interesting though is that this calculator, https://www.diyaudioandvideo.com/Calculator/ApcSpeakerCrossover/ , sets the the first half of the high pass section of the bandpass of a fourth order three way much wider than the second half of the same high pass filter section of that very same bandpass. At 3500/437.5 8Ohm, it sets C3 at 38.12uF, L3 at 1.44mH, C4 at 53.86uF and L4 8.23mH. While if I calculate it manually using fourth order Linkwitz Riley formula, I get C3 at 24.11uF, L3 at 1.83mH, C4 at 48.23uf and L4 at 8.23mH. As you can see the calculator is making the first half of the high pass section much wider, but falls closer in line in the second half of the high pass section. The High pass and low pass filters are always the same as my calculations.

Easiest thing to do is forget those king of online calculators exist in the first place.
They simplify too much to be truly useful.
XSim is free and easy to learn so you can model much more fully.
Of course to get any model properly right you need to measure the frequency response and Impedance of the drivers in cabinet, but even without that you could see visually what's going on with different crossover component values much more easily.


Edited by DMorison - 21 April 2018 at 2:30pm
Back to Top
JoelDD View Drop Down
Registered User
Registered User


Joined: 20 April 2018
Status: Offline
Points: 12 		Post Options Post Options   Thanks (0) Thanks(0)   Quote JoelDD Quote  Post ReplyReply Direct Link To This Post Posted: 21 April 2018 at 1:52pm
Originally posted by DMorison DMorison wrote:

Originally posted by JoelDD JoelDD wrote:

Yeah, the more I look into this the more I think it is just to "cover all the bases". I did some manual calculation last night and there was so little difference between the capacitor and inductance values that your driver selection and zobel network implementation would have a much greater effect on where your crossover is actually happening. Something to note is that I see this the most in cheaper speaker crossovers. For example, some cheap Chinese in-wall units that I picked up have the crossovers set to 4500hz high pass, 5800hz-270hz bandpass and 300hz low pass. Mind you they omitted a phase inductor on the bandpass, didn't zobel network anything and have no driver attenuation on the bandpass. I plan to rebuild these networks entirely as they sound god awful.

Something interesting though is that this calculator,  https://www.diyaudioand.com/Calculator/ApcSpeakerCrossover/ - https://www.diyaudioandvideo.com/Calculator/ApcSpeakerCrossover/ , sets the the first half of the high pass section of the bandpass of a fourth order three way much wider than the second half of the same high pass filter section of that very same bandpass. At 3500/437.5 8Ohm, it sets C3 at 38.12uF, L3 at 1.44mH, C4 at 53.86uF and L4 8.23mH. While if I calculate it manually using fourth order Linkwitz Riley formula, I get C3 at 24.11uF, L3 at 1.83mH, C4 at 48.23uf and L4 at 8.23mH. As you can see the calculator is making the first half of the high pass section much wider, but falls closer in line in the second half of the high pass section. The High pass and low pass filters are always the same as my calculations.

Easiest thing to do is forget those king of online calculators exist in the first place.
They simplify too much to be truly useful.
XSim is free and easy to learn so you can model much more fully.
Of course to get any model properly right you need to measure the frequency response and Impedance of the drivers in cabinet, but even without that you could see visually what's going on with different crossover component values much more easily.

Thats a really helpful program! Thanks! I do know that the online calculators like the one I linked are just strictly for the frequency slopes and I use them as such. I have just been using them as a base to build from, which is why I was curious about the bandpass overlapping because if your foundation is garbage then you are shooting yourself in the foot before breaching the wall.


Edited by JoelDD - 21 April 2018 at 2:46pm
Back to Top
DMorison View Drop Down
Old Croc
Old Croc


Joined: 14 March 2007
Location: Aberdeen
Status: Offline
Points: 1649 		Post Options Post Options   Thanks (0) Thanks(0)   Quote DMorison Quote  Post ReplyReply Direct Link To This Post Posted: 21 April 2018 at 2:30pm
Originally posted by JoelDD JoelDD wrote:

Originally posted by DMorison DMorison wrote:


Easiest thing to do is forget those king of online calculators exist in the first place.
They simplify too much to be truly useful.
XSim is free and easy to learn so you can model much more fully.
Of course to get any model properly right you need to measure the frequency response and Impedance of the drivers in cabinet, but even without that you could see visually what's going on with different crossover component values much more easily.

Thats a really helpful program! Thanks! I do know that the online calculators like the one I linked are just strictly for the frequency slopes and I use them as such. I have just been using them as a base to build from, which is why I was curious about the bandpass overlapping they do because if your foundation is garbage then you are shooting yourself in the foot before breaching the wall.

The thing is, you may still be doing exactly what you're trying to avoid by using these calculators - the simplifications they must make (as they don't account for the Freq Response and Impedance of the driver in cabinet) mean that their output is only valid for a driver that has an absolutely flat freq and impedance curve through the crossover region, and drivers like that basically don't exist for practical purposes.

Anyhoo, if you read the "get help for the 3 way crossover designer" on the link you posted, it explains a bit about why it's not just a stacked HPF & LPF based on the vales you'd use for each of those sections in isolation.
Back to Top
JoelDD View Drop Down
Registered User
Registered User


Joined: 20 April 2018
Status: Offline
Points: 12 		Post Options Post Options   Thanks (0) Thanks(0)   Quote JoelDD Quote  Post ReplyReply Direct Link To This Post Posted: 21 April 2018 at 2:58pm
Originally posted by DMorison DMorison wrote:

Originally posted by JoelDD JoelDD wrote:

Originally posted by DMorison DMorison wrote:


Easiest thing to do is forget those king of online calculators exist in the first place.
They simplify too much to be truly useful.
XSim is free and easy to learn so you can model much more fully.
Of course to get any model properly right you need to measure the frequency response and Impedance of the drivers in cabinet, but even without that you could see visually what's going on with different crossover component values much more easily.

Thats a really helpful program! Thanks! I do know that the online calculators like the one I linked are just strictly for the frequency slopes and I use them as such. I have just been using them as a base to build from, which is why I was curious about the bandpass overlapping they do because if your foundation is garbage then you are shooting yourself in the foot before breaching the wall.

The thing is, you may still be doing exactly what you're trying to avoid by using these calculators - the simplifications they must make (as they don't account for the Freq Response and Impedance of the driver in cabinet) mean that their output is only valid for a driver that has an absolutely flat freq and impedance curve through the crossover region, and drivers like that basically don't exist for practical purposes.

Anyhoo, if you read the "get help for the 3 way crossover designer" on the link you posted, it explains a bit about why it's not just a stacked HPF & LPF based on the vales you'd use for each of those sections in isolation.

Oh, thanks for pointing that out. Guess I should read a little closer. So basically, the the crossover shift is because of the bandpass gain and the bandpass gain is variable based on slope used and driver sensitivity. But if we factor in a flat dB reducing L-pad across the bandpass driver reducing both the bandpass gain and equalizing the sensitivity of the driver, could we not use the standard stacked high/low pass as the slope generator? It seems that we are trying to unscrew a bolt with a set of pliers here.
Back to Top
gen0me View Drop Down
Young Croc
Young Croc
Avatar

Joined: 20 February 2016
Location: UK
Status: Offline
Points: 999 		Post Options Post Options   Thanks (0) Thanks(0)   Quote gen0me Quote  Post ReplyReply Direct Link To This Post Posted: 21 April 2018 at 3:04pm
Originally posted by JoelDD JoelDD wrote:

Is this a single driver measure or a cabinet?
Thats 18" driver simulation in BP6. 1 cabinet.


Edited by gen0me - 21 April 2018 at 3:41pm
I appreciate every like :)) https//www.facebook.com/genomesoundsystems
Mixes: https://www.mixcloud.com/gen-ome/
Back to Top
 Post Reply Post Reply Page  12>

Forum Jump Forum Permissions View Drop Down

Forum Software by Web Wiz Forums® version 12.06
Copyright ©2001-2023 Web Wiz Ltd.

This page was generated in 9.313 seconds.