Have you looked at Zaph's ZD5 project? It uses LR2 xover and an all-pass (aka "ladder network") for delay. Measurements galore and listening results.
Zaph|Audio - ZD5 - Scan Speak 15W8530K00 and Vifa XT25
Zaph|Audio - ZD5 - Scan Speak 15W8530K00 and Vifa XT25
I guess it's worth a try. Interesting experiment would be leaving passive xovers on the drivers but adding delay by DSP to one loudspeaker, and by all-pass network to the other and then compare the sound.
I've rarely seen benefit from any passive network as complicated as that that doesn't have any drawbacks - which doesn't mean anything for this particular circuit untill it is tested.
I know Dynaudio Contour 1.3mk2 has it and i liked the sound. But would i like it better with digital delay with the same passive xover - that i don't know.
I've rarely seen benefit from any passive network as complicated as that that doesn't have any drawbacks - which doesn't mean anything for this particular circuit untill it is tested.
I know Dynaudio Contour 1.3mk2 has it and i liked the sound. But would i like it better with digital delay with the same passive xover - that i don't know.
Last edited:
Hi Line,
I've never felt any need to implement one in reality, but I have played with
it simulating. It's not the same what kind of values you choose. Affects the
high pass of a tweeter as well, so it shapes the overall response with addition
of a phase shift, naturally.
I know Dynaudio has used these networks before and nowdays too. Focus 110
has one.
Noticed you added delay in your sim. Did you define it as 0,54 or -0,54?
I know that convention of XSim is different than the one in PCD to achieve
the same result.
Xsim states: "Positive delay moves the driver away from the mic"
For a flat baffle, I used -0.54" to position the tweeter wave launch closer to the mic. The delay crossover circuit sims with a +/- 30_degree phase shift up to 10kHz.
If we machined a 0.54" back-step baffle for the tweeter I would use 0.00" in Xsim.
--------
SO, do the extra ladder delay L-C components "crap-up" the sound more than:
-poor time alignment on a flat baffle?
-extra tweeter baffle edge diffraction from a stepped-baffle?
-360_degree phase shift and extra poles from LR4/LR4?
-off-axis listening from a 13-15 degree slanted baffle?
-unequal slopes and unequal phase shifts from a BW3 tweet with LR2 woof?
When it comes to two way passive xovers, i haven't seen loudspeaker that has simple xover and sounds and mesaures good. It can be simple (bsc incorporated in series woofer inductor etc.) but then mids sound pretty awful. Making two way xover for flat frequency response and good reverse null inevitably leads to complicated (large) xovers. Not a problem with floorstanding loudspeakers but with bookshelf it can be. Orientation of coils is important and if you have 18/12 od 18/24 electrical and only one notch filter, you get 4-5 coils and it becomes pain in the neck to orientate them properly. Passive xover with added time delay complicates the xover further more adding 2 more coils to the equation. I divided tweeter and woofer xovers on separate boards to avoid coil to coil and coil to magnet interaction.
So, i don't think that delay network will crap up the sound too much (if done and implemented properly) but i do think that aditional complication of doing it properly can be problematic. You'll need separate box just for xover or use of smaller in size (cored) inductors - both of which i'd never do unless instructed by the one i am making the speakers for.
So, i don't think that delay network will crap up the sound too much (if done and implemented properly) but i do think that aditional complication of doing it properly can be problematic. You'll need separate box just for xover or use of smaller in size (cored) inductors - both of which i'd never do unless instructed by the one i am making the speakers for.
Last edited:
Line,
thanks for clarifying. So I guess your point of reference is the
woofer, because you could have positively delayed the woofer
with the exact SPL summation, only different phase plot.
BTW, Regarding phase calculation, I have extracted from Bagby's paper
"Finding the Relative Acoustic Offset in PCD" and the "Instructions
in PCD" on how to define the Z offset, and now it looks as if there is
a discrepancy between the two. Do I conclude correctly that the
PCD Instructions are misleading or am I misinterpreting things?
I know for sure that Z offset in PCD has to be entered with a negative
sign if you want to present the woofer voicer coil being behind
the tweeter's voice coil.
p.s. Zvu, I'm sorry if this appears as an OT. Interesting project and I don't
doubt it will sound satisfactory to you and the client of yours.
thanks for clarifying. So I guess your point of reference is the
woofer, because you could have positively delayed the woofer
with the exact SPL summation, only different phase plot.
BTW, Regarding phase calculation, I have extracted from Bagby's paper
"Finding the Relative Acoustic Offset in PCD" and the "Instructions
in PCD" on how to define the Z offset, and now it looks as if there is
a discrepancy between the two. Do I conclude correctly that the
PCD Instructions are misleading or am I misinterpreting things?
I know for sure that Z offset in PCD has to be entered with a negative
sign if you want to present the woofer voicer coil being behind
the tweeter's voice coil.
p.s. Zvu, I'm sorry if this appears as an OT. Interesting project and I don't
doubt it will sound satisfactory to you and the client of yours.
Attachments
Few final touches and xover point ended up at~1650Hz. I wanted to see if i stressed them too much with low xover point so i measured distortion. Here are few distortion measurements - HD2 and HD3.
These are very good drivers as expected. It is very hard to find drivers of this quality in Europe for much more money, but at these prices, SB is a bargain. I just hope they don't change their company policy.
These are very good drivers as expected. It is very hard to find drivers of this quality in Europe for much more money, but at these prices, SB is a bargain. I just hope they don't change their company policy.
Attachments
Earlier versions of PCD did use a positive offset. I think the instructions didn't get updated when PCD was changed to use negative offsets.
- What if the tweeter is replaced with Sb26stc-c000-4 ?
- Can anyone help me design make crossover ?
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.