I need to set up my active crossover to match the passive this speaker was designed for. Can anyone tell me how I could start ???
What are my crossover points and such ?
drivers are the
Peerless 830874
SEAS Prestige 27TBFC/G
woofer is 8 ohms, tweeter is 6 ohms
CARDINAL-TBFC-XOVER — ImgBB
What are my crossover points and such ?
drivers are the
Peerless 830874
SEAS Prestige 27TBFC/G
woofer is 8 ohms, tweeter is 6 ohms
CARDINAL-TBFC-XOVER — ImgBB
Analog or DSP?
Three options:
1. Measure the frequency response at the drivers terminals. Then simulate the active crossover and tweak the parameters to match these frequency responses.
2. Simulate the crossover in software in order to determine the frequency response at the drivers terminals. Continue as in the first option.
3. Design a crossover from scratch.
The last option requires a measurement microphone, but takes advantage of the possibilities of an active crossover and gives the best results.
Three options:
1. Measure the frequency response at the drivers terminals. Then simulate the active crossover and tweak the parameters to match these frequency responses.
2. Simulate the crossover in software in order to determine the frequency response at the drivers terminals. Continue as in the first option.
3. Design a crossover from scratch.
The last option requires a measurement microphone, but takes advantage of the possibilities of an active crossover and gives the best results.
I already have the passive crossover. I just need to turn it digital so I can set it up !
Last edited by a moderator:
Interesting, so if you can trace those before and after plots and take the difference, you'll have a guide for the new filters.
There are other ways to get the same information.
There are other ways to get the same information.
When I "read" the crossover, I see a 4th order highpass, and a 2nd order lowpass (with a few extra bits).
"What are my crossover points and such ?"
If you read the designer notes, they tell you exactly this, and more:
Cardinal-TBFC – AmpsLab
Figure 4 shows the crossover point "They are crossing at about 1.8kHz"
...so that + the crossover slopes I mentioned are a good starting point for an electronic version.
---
I'm gonna 100% support what TBTL said. Get a measurement mic.
---
The Dayton mic + an app for your phone + a cable is about $33.
It beats guesswork.
Audiotool is really quick, really good for taking averages (e.g. it is a fast way to measure the overall FR of your speaker, across the listening area rather than at a single point).
Where Audiotool fails is if you want to get in depth, e.g. take 10 sets of measurements and compare them all (you can do it, by exporting the data to PC and plotting it - but it is clunky).
For about $100 you can get a USB mic + REW which is good for easily comparing a lot of single point measurements.
Or you can use both. REW for detailed and accurate on-axis tests, and run Audiotool as an idiot check, to make sure you aren't ruining the power response.
On-axis vs. Power Response
These are not your only mic + software options, of course; they are just what I'm familiar with.
"What are my crossover points and such ?"
If you read the designer notes, they tell you exactly this, and more:
Cardinal-TBFC – AmpsLab
Figure 4 shows the crossover point "They are crossing at about 1.8kHz"
...so that + the crossover slopes I mentioned are a good starting point for an electronic version.
---
I'm gonna 100% support what TBTL said. Get a measurement mic.
---
The Dayton mic + an app for your phone + a cable is about $33.
It beats guesswork.
Audiotool is really quick, really good for taking averages (e.g. it is a fast way to measure the overall FR of your speaker, across the listening area rather than at a single point).
Where Audiotool fails is if you want to get in depth, e.g. take 10 sets of measurements and compare them all (you can do it, by exporting the data to PC and plotting it - but it is clunky).
For about $100 you can get a USB mic + REW which is good for easily comparing a lot of single point measurements.
Or you can use both. REW for detailed and accurate on-axis tests, and run Audiotool as an idiot check, to make sure you aren't ruining the power response.
On-axis vs. Power Response
These are not your only mic + software options, of course; they are just what I'm familiar with.
Neither TBTL nor myself are suggesting a microphone as the first option. Doing a crossover from scratch would need one, but transcribing an existing crossover potentially more accurate if not taken into the acoustic domain.
I was hoping because He already did all the work to make this design I could just set it up actively 🙂
I saw that "crossing at at 1.8khz" and was thinking that is the effect of a higher and lower crossover point with steep enough slopes that they rolled off and met around 1.8k. But I guess I am over thinking it!
I have REW and scarlet solo with measurement mic. I am not so great at adjusting my EQ setting with it. are there any good youtube videos that I could watch about flattening out my responce curve and such??
I kinda got lost because I read somewhere that when I had a 24db crossover I have to reverse polarity of one of the drivers. But I'm not really sure how to even read the phase chart on REW.
I saw that "crossing at at 1.8khz" and was thinking that is the effect of a higher and lower crossover point with steep enough slopes that they rolled off and met around 1.8k. But I guess I am over thinking it!
I have REW and scarlet solo with measurement mic. I am not so great at adjusting my EQ setting with it. are there any good youtube videos that I could watch about flattening out my responce curve and such??
I kinda got lost because I read somewhere that when I had a 24db crossover I have to reverse polarity of one of the drivers. But I'm not really sure how to even read the phase chart on REW.
If you want to do 24dB or different to the passive then you have much more work to do, and that is fine if you want.
Otherwise, keep it the same. We are ready to know whether you are able to do as TBTL has suggested and record a response from your terminals, then copy it on your signal processor?
Otherwise, keep it the same. We are ready to know whether you are able to do as TBTL has suggested and record a response from your terminals, then copy it on your signal processor?
"If you want to do 24dB or different to the passive then you have much more work to do, and that is fine if you want.
Otherwise, keep it the same."
The passive is 24dB (for the tweeter) - or have I misread this?
Otherwise, keep it the same."
The passive is 24dB (for the tweeter) - or have I misread this?
"I have REW and scarlet solo with measurement mic. I am not so great at adjusting my EQ setting with it."
There are heaps of tips and tutorials, e.g. Community Tutorials
...but they depend on what device / interface you have.
Me, I do it in multiple steps.
Measure. See 4 problems.
Fix one problem.
Measure again. See 3 problems.
Fix one problem...
There are heaps of tips and tutorials, e.g. Community Tutorials
...but they depend on what device / interface you have.
Me, I do it in multiple steps.
Measure. See 4 problems.
Fix one problem.
Measure again. See 3 problems.
Fix one problem...
So when you "see a problem", are you basically just trying to make the frequency response line as flat as possible?
I have a zoudio 4 channel Amp with built in DSP. I basically just add notches, peaks, bass and treble shelves at points where it spikes, to make it flatter. Is this the right idea?
I have a zoudio 4 channel Amp with built in DSP. I basically just add notches, peaks, bass and treble shelves at points where it spikes, to make it flatter. Is this the right idea?
Last edited:
"So when you "see a problem", are you basically just trying to make the frequency response line as flat as possible?"
Yes, but don't focus overmuch on a single line / measuring position - e.g. if a tweak (like setting the X-over point to 5kHz) gets the FR marginally better when perfectly on axis, but makes the overall power response much worse, then it is probably a bad tweak.
"I have a zoudio 4 channel Amp with built in DSP. I basically just add notches, peaks, bass and treble shelves at points where it spikes, to make it flatter. Is this the right idea?"
Yeah.
I'd start with the X-over point & slopes, then do overall level matching of the drivers, and only then go into the finer eq.
The main thing to avoid: don't try to boost bass nulls - you'll waste power / lose headroom without getting much result.
I've also read that very sharp peaks and notches, even with DSP, degrade the sound - but I honestly don't understand how this happens.
There are probably tutorials for that device that are way more articulate / helpful 🙂
Yes, but don't focus overmuch on a single line / measuring position - e.g. if a tweak (like setting the X-over point to 5kHz) gets the FR marginally better when perfectly on axis, but makes the overall power response much worse, then it is probably a bad tweak.
"I have a zoudio 4 channel Amp with built in DSP. I basically just add notches, peaks, bass and treble shelves at points where it spikes, to make it flatter. Is this the right idea?"
Yeah.
I'd start with the X-over point & slopes, then do overall level matching of the drivers, and only then go into the finer eq.
The main thing to avoid: don't try to boost bass nulls - you'll waste power / lose headroom without getting much result.
I've also read that very sharp peaks and notches, even with DSP, degrade the sound - but I honestly don't understand how this happens.
There are probably tutorials for that device that are way more articulate / helpful 🙂
Awesome, sounds like I'm on the right track. Is there anything besides the level matching and frequency response i should be worried about? Like my drivers being effected by their crossover slopes and out of phase?? How would I even know if they are off?
If you trust your current crossover then you shouldn't have a problem to begin with.
If you want to check phase, or even to EQ, that's fine but these three things are complex enough that you'd find it easier doing them one at a time.
You then impose this curve onto your signal processor. Do both drivers and set levels.
If you want to check phase, or even to EQ, that's fine but these three things are complex enough that you'd find it easier doing them one at a time.
In basic terms.. Drop any EQ, set the levels so you don't damage your sound card inputs and feed the speaker positive terminal back to the input and measure it the same as you would when taking a microphone measurement.
You then impose this curve onto your signal processor. Do both drivers and set levels.
I fought with it mostly lol. I have played around enough to figure out how to get a reading from it. But I'm apparently missing some steps...
Yeah, but many potential problems won't come up if you stay fairly close to the original crossover - assuming the original was good 🙂
e.g. if you halve the X-over frequency, you will increase the load on the tweeter (and drastically reduce the max SPL of the system).
So don't do that.
If you triple the X-over frequency, your speaker will beam more in the top end of the woofer's range (it might measure fine on-axis, but the overall power response will be worse).
So don't do that (unless you actually prefer it).
You could experiment with shifting the X-over a little, but I'd advise against any big changes.
If this is a concern, you could simply keep the same slopes as the original.
Looking at the passive X-over, I see 24dB high pass, 12dB low pass (+ some bits for eq and level matching).
I think it'd be fine to keep those slopes, and just add an extra notch on the LF part to fix the issue the designer noted.
"Note the slight cancellation at 4.5kHz. That’s the effect of the cone breakup I mentioned earlier. "
...or you could experiment with other slopes, while keeping the original as a basis for comparison.
---
I don't know enough to comment on the phase thing, I'll leave that to an official clever person.