So I would like to replicate the crossover shown below, but having a hard time understanding it.
Partly because I'm a newbie at this and partly because of all the glue!
I think left inputs are negative. Looks to me like a second order filter, and I know there is supposed to be a notch filter for the HF.
And possibly an RC filter for the LF? Speaker has biwiring, so two sets of inputs. Lower half is LF circuit, and upper half is HF circuit.
If someone is able to make a schematic from this, that would be fantastic 🙂 Not looking for component values.
EDIT; here's another pic I found of the same crossover that shows the component in glue better.
Conserning the component with a questionmark (hidden in glue) in the first pic, it looks like this (and the one between the two bottom inductors) is an Ohmite wirewound resistor;
Partly because I'm a newbie at this and partly because of all the glue!
I think left inputs are negative. Looks to me like a second order filter, and I know there is supposed to be a notch filter for the HF.
And possibly an RC filter for the LF? Speaker has biwiring, so two sets of inputs. Lower half is LF circuit, and upper half is HF circuit.
If someone is able to make a schematic from this, that would be fantastic 🙂 Not looking for component values.
EDIT; here's another pic I found of the same crossover that shows the component in glue better.
Conserning the component with a questionmark (hidden in glue) in the first pic, it looks like this (and the one between the two bottom inductors) is an Ohmite wirewound resistor;
Last edited:
Yes sorry I didn't mean component values of course, forgot to mention that 🙂
But a schematic with components w/o values may be possible?
I found another pic where I think the mysterious component (?) from the crossover above is shown.
The brown components shown below. I wonder if they are capacitors or some exotic resistors?
It's from Living Voice Auditorium R25A. A two-way MTM with offset tweeter. 6.5" Vifa C17 woofers and Scanspeak D2608/93000 tweeter.
Found the photo in an online review of the speaker.
I have a roughed in XO for the second, there are 2 wires heading off to the right (actually the top?). Do these have terminals for bi-wiring?
The first xo is a mess. I know it is painful to do but can you pull off a bunch of the hot glue?
dave
The first xo is a mess. I know it is painful to do but can you pull off a bunch of the hot glue?
dave
Hi Dave,
I realise now that I should not have posted the xo in second picture, as it will probably just be confusing.
I'm only interested in the xo in my fist post. The second pic was only added to show the brown components that I think may be whats inside the glue, where there is a questionmark on the first picture.
I don't have access to this crossover, it's just a photo published in an online review. So there will be some guesswork involved, I know. Especially the aforementioned brown components completely hidden by the glue..
And yes, they do have terminals for bi-wiring. The top two are for the HF section and bottom two for the LF section.
If you expand your Google searching to include a Living Voice OBX crossover you'll see that all variants seem to have a similar topology for their networks.
which is ;
2-pole x 2-pole with a series LCR strapped across the woofers outputs ( Yellow//Green terminals are for the mid//bass woofers ).
I'm not really sure how you knowing the schematics layout helps move along your clone project ( stalled ?? from last year ).
🙂
![img_5140[1].jpg](https://www.diyaudio.com/community/attachments/img_5140-1-jpg.1139878/ "img_5140[1].jpg")
which is ;
2-pole x 2-pole with a series LCR strapped across the woofers outputs ( Yellow//Green terminals are for the mid//bass woofers ).
I'm not really sure how you knowing the schematics layout helps move along your clone project ( stalled ?? from last year ).
🙂
Thank you Earl 🙂
I think the R25A network is very similar, in that it incorporates a series notch filter for the HF and is probably 2nd order for both ways. I say probably because I'm not sure what's going on in the HF section;
I'm pretty sure the notch filter is composed of the components shown. The parallell inductor and series resistor are also fine I think.
It gets literally blurry when it comes to the capacitors. I can't see if the green cap is parallelled to the big yellow cap, but I assume it is.
However the connections in the resistor area is a mess, but it looks to me like the caps are in parallell instead of series? That would be unusual, and the reason why I'm not sure its a regular 2nd order filter.
On to the LF network;
The series inductor is fine, and also the parallell resistor I think. But again it gets blurry with the 4 parallelled caps and the white encapsulated cap next to them. I assume they are in parallell with the network but looks like they are in series to eachother, which would be strange..
I managed to make a filter that sounds reasonably good, I'm just very curious about the original filter and how it sounds 🙂
Did you get anywhere from sleuthing the Dali 104?
Randomly found someone's photo of those, brochure says XO is 3khz, owner said Dali told him they used LR2
Topology seems tweeter 1R resistor + 5.6uf cap + ~0.39mh coil, woofer ~ 1.8mh coil + 2.2uf cap + 3.9R zobel
https://www.skelbiu.lt/skelbimai/krosoveris-2-taku-dali-104-koloneliu-66225656.html
https://www.avforums.com/threads/dali-104-crossover-help.2349503/
https://www.dali-speakers.com/media/3268/dali-104-brochure.pdf
My guess is LV add a notch for the D2608 (or whatever) FS resonance, and the bass zobel becomes an LCR notch of some breakup peak in their custom Vifas. I think they like to tweak and tweak so maybe that explains the fancypants paralleled small caps and bypasses.
The key seems to be knowing what XO point they currently use.
Depending on the drivers used in your clone, you may or may not need to implement their notches at all, less match their values for the notches if you have measurements of your own to address bumps and peaks at the ends of the driver passbands.
Randomly found someone's photo of those, brochure says XO is 3khz, owner said Dali told him they used LR2
Topology seems tweeter 1R resistor + 5.6uf cap + ~0.39mh coil, woofer ~ 1.8mh coil + 2.2uf cap + 3.9R zobel
https://www.skelbiu.lt/skelbimai/krosoveris-2-taku-dali-104-koloneliu-66225656.html
https://www.avforums.com/threads/dali-104-crossover-help.2349503/
https://www.dali-speakers.com/media/3268/dali-104-brochure.pdf
My guess is LV add a notch for the D2608 (or whatever) FS resonance, and the bass zobel becomes an LCR notch of some breakup peak in their custom Vifas. I think they like to tweak and tweak so maybe that explains the fancypants paralleled small caps and bypasses.
The key seems to be knowing what XO point they currently use.
Depending on the drivers used in your clone, you may or may not need to implement their notches at all, less match their values for the notches if you have measurements of your own to address bumps and peaks at the ends of the driver passbands.
Last edited:
Reading this again, yes, for nominal speaker impedance 6ohms, LV's C17 must be a high ohm custom version, so any woofer midrange notch they use would be difficult to transpose precisely to an off the shelf driver.
Some might say running a 7" to 3k is a little on the high side, so that would change matters quite a lot, depending on what power response one wants. Lastly, symmetric, non-inverted LR2 on a flat baffle with no waveguide or all-pass circuit is another wrinkle to consider.
All said, the LV's sound great, and in my mind the goal should be not to religiously clone the crossover (dubious cause custom drivers and opaque tweaks) but use elements as a template (cabinet, baffle layout, tuning frequency, shallow slope on woofer).
Some might say running a 7" to 3k is a little on the high side, so that would change matters quite a lot, depending on what power response one wants. Lastly, symmetric, non-inverted LR2 on a flat baffle with no waveguide or all-pass circuit is another wrinkle to consider.
All said, the LV's sound great, and in my mind the goal should be not to religiously clone the crossover (dubious cause custom drivers and opaque tweaks) but use elements as a template (cabinet, baffle layout, tuning frequency, shallow slope on woofer).
Last edited:
Hi motokok, thank you for your interest in my thread 🙂
It all started from a pair of Dali 104 I purchased for $70. I modified the cabinets with bracing and bitumen glued to the inside panels.
Pulled out the original crossovers and tweeters, and replaced with Scanspeak D2608 as used in the LV R25A.
After alot of measurements and vituixcad simulation, I settled on a 2nd order slope for both ways, centered around 3khz. Sounds good to me.
There is alot of diffraction though from the edges of the woofers, doesn't seem to be a concern for Living Voice though.
2nd order is correct, but XO is definitely higher than 3khz. Here's my measurement of woofers + tweeters separately, more like 4khz I'd say;
Kevin Scott has openly discussed the notch filter at 700hz, so that is no secret. There is no impedance network besides the notch.
From LV website (Q&A about the R25A);
"We control the impedance peak at 700 Hz on the HF unit, with an LCR network but do not use any other impedance equalisation, as we have achieved subjectively perfect integration without having to resort to this."
There was mention of a 2850hz crossover point in an online review.
I'm using the exact same drivers. I think the LCR was used on the tweeter because the crossover point was lowered from the previous model.
I have obtained values for the notch in Vituixcad from impedance measurements of my D2608 tweeters.
Like LV I will not make zobel for the C17's.
Thats a good point motokok, however I have done my homework 🙂
There is no notch on the woofers. But I have also wondered how they get that 6 ohm nominal impedance, despite using 8 ohm woofers wired in parallell.
Are you referring to phase shift as a result of drivers not being physically time aligned? The phase doesn't look too bad actually with a 2nd order filter.
Why do you think it's preferable to use a shallow slope on the woofer? Do you mean a 1st order filter?
I didn't realize you have the exact drivers! I thought they had >=10ohm versions made?
Your on-axis measurements look smooth and suggest the drivers are easy to work with.
I think I've lost the plot - what are your goals?
A lot of people don't like the HDS tweeter (me included), btw. Don't LV use Revelators/Illuminators?
I don't have a preference for the shallow slope - just that from your photos they use 2nd order electrical on the woofer, and the C17's smooth response gives 2nd order acoustic. As that is the basis for the midrange, one would think to retain that to be close to their voicing.
Here's A Heissmann's take on MTM - lower XO, 4th order, and the kind of textbook measurements his designs share.
Your on-axis measurements look smooth and suggest the drivers are easy to work with.
I think I've lost the plot - what are your goals?
A lot of people don't like the HDS tweeter (me included), btw. Don't LV use Revelators/Illuminators?
I don't have a preference for the shallow slope - just that from your photos they use 2nd order electrical on the woofer, and the C17's smooth response gives 2nd order acoustic. As that is the basis for the midrange, one would think to retain that to be close to their voicing.
Here's A Heissmann's take on MTM - lower XO, 4th order, and the kind of textbook measurements his designs share.
Last edited:
Well I have the old Vifa drivers. But after alot of reading I've found out that the scanspeak drivers are exact copies of the old C17.
They may use 10 ohm drivers in some of their other models for all I know, but the R25A uses 8 ohm drivers.
This is for the original Dali 104 drivers. But yes the C17 are very well behaved!
My goal with this thread is to decode the Living Voice filter network shown above 🙂 In post #12 I have some questions about component locations and topology. So the goal is to understand and try this filter.
Their more expensive models use Revelator. I know the D2608 is related to the old HDS tweeter, but I think it has been revised.
I like the sound of this tweeter very much, smooth and detailed 🙂
Maybe this helps - this guy upgraded all caps and resistors on the OBX crossover and the parts are not smothered in glue. The bass LCR may be for system impedance correction / tube amps; otherwise optional. Basically a massive mish mash of 6-7 caps in the bass section! 🤣 . Interestingly, the tweeter in the OBX doesn't seem to need a padding resistor.
As the modder only replaced exact values, it can be seen where and with how many bypass caps LV uses.
As the modder only replaced exact values, it can be seen where and with how many bypass caps LV uses.
Last edited: