Hi Y'all,
I've been playing with Vituixcad for a bit and I think I'm almost where I understand the process. Still not sure about the merging of enclosure tool SPL/Z... but I not sure if it matters too much at this point as I'm more focused on the crossover and the enclosure for this woofer has already been designed, esentially I'm working on a version of P. Blix's MLTL using the Peerless 830656 and the Peerless DX20BF00-4 as the tweeter.
I've done the work with the difraction tool and I think I've merged the baffle step correctly and this is the result of quite some time of tinkering.
Directivity and Power Response are both quite lumpy though, and I don't know if there's anything that can be done. Optimiser does flatten both out a little, but it moves the x-over quite high and phase separates - so i'm not sure if thats any better or not.
Thanks in advance for any input.
P.S. I've googled quite a bit and I couldnt really find previous examples of crossover of these drivers to look at - so if you know of any kindly direct me.
I've been playing with Vituixcad for a bit and I think I'm almost where I understand the process. Still not sure about the merging of enclosure tool SPL/Z... but I not sure if it matters too much at this point as I'm more focused on the crossover and the enclosure for this woofer has already been designed, esentially I'm working on a version of P. Blix's MLTL using the Peerless 830656 and the Peerless DX20BF00-4 as the tweeter.
I've done the work with the difraction tool and I think I've merged the baffle step correctly and this is the result of quite some time of tinkering.
Directivity and Power Response are both quite lumpy though, and I don't know if there's anything that can be done. Optimiser does flatten both out a little, but it moves the x-over quite high and phase separates - so i'm not sure if thats any better or not.
Thanks in advance for any input.
P.S. I've googled quite a bit and I couldnt really find previous examples of crossover of these drivers to look at - so if you know of any kindly direct me.
Re lumpy DI. Two solutions:
1. Lower your XO point (so that your woofer is not rolling off) around 1500Hz should see some improvement. Caveat here is your tweeter may not like such a low XO point (I haven't looked at either driver specs)
2. Put the tweeter in a waveguide to allow a lower XO point. Building and measurement is your best option here unless others have experience in advanced 3d modeling and waveguide propagation analysis / prediction.
Z offset is more critical with higher XO points (as the wavelength is shorter). What Z value are you using? Did you remember to check the minimum phase checkboxes on the driver tab? (or did you extract minimum phase outside of VCad?)
Otherwise this looks good to me. I would accept / trade off DI for better power handling (i.e. not crossing the tweeter into the stress zone). I prefer my designs to be woofer limited.
1. Lower your XO point (so that your woofer is not rolling off) around 1500Hz should see some improvement. Caveat here is your tweeter may not like such a low XO point (I haven't looked at either driver specs)
2. Put the tweeter in a waveguide to allow a lower XO point. Building and measurement is your best option here unless others have experience in advanced 3d modeling and waveguide propagation analysis / prediction.
Z offset is more critical with higher XO points (as the wavelength is shorter). What Z value are you using? Did you remember to check the minimum phase checkboxes on the driver tab? (or did you extract minimum phase outside of VCad?)
Otherwise this looks good to me. I would accept / trade off DI for better power handling (i.e. not crossing the tweeter into the stress zone). I prefer my designs to be woofer limited.
Yes - min phase is checked. And here's what it looks like with x-over moved as low as I will dare (Tweeter FS is ~800)
I left the overlay of the previous config and I'm not sure if it isn't any less lumpy.
Z offset I'm not sure how to measure so it isn't included, but as the tweeter is in a shallow waveguide and will be countersunk - I suppose it might be 20-30mm. How to determine?
I left the overlay of the previous config and I'm not sure if it isn't any less lumpy.
Z offset I'm not sure how to measure so it isn't included, but as the tweeter is in a shallow waveguide and will be countersunk - I suppose it might be 20-30mm. How to determine?
The Hilbert transform does not “extract” minimum phase, it “converts” to it which creates errors out of non minimum phase information.
Well I'll eat my words. Nice work getting the DI more even at the expense of crossover passband phase tracking? (as you mentioned).
I'd be inclined to prototype and listen to both of these if you can. Quick A/B preferred (harder to do with passive, I admit). I'd put them to the test. It comes down to dare I say it the subjective aspects. Rank each crossover against imaging, soundstage, sweetspot as both look like they'll have equivalent timbre with both drivers playing roughly the same range.
Can you post inverted driver polarity for each crossover? (I find that easier to read than the phase graph).
I'd be inclined to prototype and listen to both of these if you can. Quick A/B preferred (harder to do with passive, I admit). I'd put them to the test. It comes down to dare I say it the subjective aspects. Rank each crossover against imaging, soundstage, sweetspot as both look like they'll have equivalent timbre with both drivers playing roughly the same range.
Can you post inverted driver polarity for each crossover? (I find that easier to read than the phase graph).
Gross generalisation. If I have concern, I usually allow the woofer to play higher than the tweeter lower, especially with woofers without severe breakup like yours
Re Z - I do a parallel driver measurement then manually increase Z in the crossover tab on the driver I expect to be furthest away (i.e. woofer here) until the summed individual curves (without any crossover components) match the same peaks and dips. For a 6.5/1" 2 way - somewhere in the 30mm - 40mm should be about right. I don't use VCads tool, as I find it quick and easier enough to manually increase Z until they align.
eg Orange = summed measurement.
There is no appropriate term for our example. Yes, phase is derived from the response, but it only works when the response is the result of a minimum phase signal to begin with.
There is no way to preserve non minimum phase information in a response plot. You need an original impulse. Converting a response plot to an impulse is not enough.
Here is the tweeter inverted: First, with the flattend DI and high Xover point (doesn't look like it ought to)
And here is the Xover I derrived with the lumpy Power Response
This looks right to me - nice and deep. And again non-inverted (with a Z offeste of 15mm)
And here is the Xover I derrived with the lumpy Power Response
This looks right to me - nice and deep. And again non-inverted (with a Z offeste of 15mm)
Add a small value resistor in series to the parallel 15uF woofer cap on the "in phase" version. Start with 0.5ohm and increase by 0.5 ohm at a time. Use it to lessen the order on the woofer filter. It may fill in that dip and not wreck the phase alignment. Having said that, that almost BBC dip between 1KHz and 4KHz may be preferred. Also potentially increase the first series cap for the tweeter.
Edit: I'm referring to your original in phase crossover first posted with Xo point ~ 3KHz.
Edit: I'm referring to your original in phase crossover first posted with Xo point ~ 3KHz.
Tried that Dave... didn't have the desired effect. I think I'm going to go with the below. About as good as it gets - and then measure in box.
Parts are $105CAD. Drivers were $50CAD. Thats for 2 speakers everything but wood, wire, and fill (that I already have). These are supposed to be cheap n cheerful.
Parts are $105CAD. Drivers were $50CAD. Thats for 2 speakers everything but wood, wire, and fill (that I already have). These are supposed to be cheap n cheerful.
What happens if you use Peter Blix's crossover?
I know it's very simple, but you could try modelling it and see what that gives you.
He used a different tweeter, of course, a Vifa BC25SC55-04 I.
There's a German DIY project using your 830656, the 8 ohm version of the DX20 tweeter and a similar cabinet: it doesn't give you the schematic but it does provide the parts values, so it wouldn't be too hard to work out the crossover; no idea how it sounds, but.
Edit: I can't find the link but could PM you the details, if you're interested
Geoff
I know it's very simple, but you could try modelling it and see what that gives you.
He used a different tweeter, of course, a Vifa BC25SC55-04 I.
There's a German DIY project using your 830656, the 8 ohm version of the DX20 tweeter and a similar cabinet: it doesn't give you the schematic but it does provide the parts values, so it wouldn't be too hard to work out the crossover; no idea how it sounds, but.
Edit: I can't find the link but could PM you the details, if you're interested
Geoff
Last edited:
Well, I find it handy to post the reverse null. Especially for spotting any cancellation at the edges of the crossover passband which is harder to see in a pure phase on phase graph. One can also see the uniformity of the phase tracking across the power response, which is hard to judge from a single phase response graph.
I can't say that I share the same challenges.. but I could show a couple of examples where the null will send you down the garden path.
Firstly, an offset third order Butterworth. There is no null, and that is correct for this filter. (I could have used a regular Butterworth, both ways are identical.)
Secondly, the null is deep but the phase is not right.
Firstly, an offset third order Butterworth. There is no null, and that is correct for this filter. (I could have used a regular Butterworth, both ways are identical.)
Secondly, the null is deep but the phase is not right.
For the deep null, we see no destructive interference between drivers. i.e. when inverted, one driver is not cancelling the other out. Albeit this is only a single axis view. Although phase "is not right", we listen to the frequency response, not phase unless I am mistaken. Sure - someone could naively look at this inverted response and assume the on-axis inverted polarity would be linear, but we know that's not the case as we can see only a -3dB XO point, so combined with on-axis, I would not be worried about the relative phase between the drivers. they track well through the XO passband, would you not agree?
Anyway - I have no problem with people wanting different views of essentially the same information that leads to the same conclusions. Maybe I'm completely wrong, but I don't have a problem with phase only, or FR only views. Only if you mis-interpret the data to draw a wrong conclusion.
Anyway - I have no problem with people wanting different views of essentially the same information that leads to the same conclusions. Maybe I'm completely wrong, but I don't have a problem with phase only, or FR only views. Only if you mis-interpret the data to draw a wrong conclusion.