I've been wanting to have a go at OB but I don't think my room/WAF will allow decent bass, so I built some bass modules. Initially I was going to build a compact 3 way into these cabinets. They are old Wharfedale Laser 100 boxes of about 28litres. Turned 90 degrees to how they were originally oriented and MDF applied.
My other speakers are XRK Ref Monitors so I have experience of RS225 and I bought another pair. I had swapped out the 3"FR in that build for a Morel CAT378 with good results.
So after much simming and reading I thought I had enough knowledge to 'have a go' and so it began. Pretty much cobbled together and used various values of XO components that were on the old Laser crossover board. Far from optimal but a start .
First iteration above with SB15NRXC30-8 and SB29RDNC.
This with 3.7mh and 25uf on the woofer as per the Laser Xo which is about where I wanted to be. Mids were 1st order electrical HP of series 16uf and LP being 0.8mh, 8uf. HP on tweeter was 6.8uf 0.3mH. Padded down about 6db.
Initial impression were that it 'worked' but was a bit bright. I tried padding the mid and also upped the HP on it to 50uf to push them lower. Sounding decent now but I think the break up of the SB15 is causing the brightness.
I remembered I had some RST28F in WG300 so I tried that as a 2 way.
It's good to have an external crossover to be able to swap bits in and out easily! Not sure the wife is so thrilled at this point!
So maybe I should be deciding exactly what I want to build now but the 3 way was still an itch so I went back and tried me a bit of a dipole naked setup .
Having read the dipole tweeter thread and seeing the use of small and cheap Vifas, I had already bought some and employed them here now. BC25SC15-04 on the front for improved WAFnand the OC25SC15 on the back. Doesn't give as small a dipole distance as 2 OC back to back but looks better. Wired them series + to - but I'm not sure if that's correct. I think the rear membrane should move in the same direction as the front firing one. Here it will fire backwards as the front fires forwards.
HF circuit now modified for 8 ohm as the 2 4ohms in series, so now on 3.3uf and I think 0.56mH. I am aware these tweeters have quite high FS so I'm probably pushing them a bit with this approx 3k HZ filter. Can't hear anything untoward from them mind . The main issue as I look at the data sheets is the the SB15 won't cross nicely to the tweeters up at 3k. They start to break up fast after 3k. I considered a notch filter but remembered I had some Seas MCA15RCY. These look much flatter and so they went in next. Now we're getting somewhere. Sounding nice!
I keep playing with the series cap on the mid, I've pushed it a bit lower to meet the RS225 with a 50uf. Having read a bit more on dipole I understand with a calculation that the mid should be rolling off 1st order at about 1100Hz. (172/ 0.15m). The 50uf should be rolling it off again at about 450hz. I removed the padding on the mid. All this being done by ear and looking at the data sheets and playing with XSim. I might try 1st order LP on the mid with 0.3mH and simming a lower value HP cap on it gives a much sharper reverse null.
I can hear the groans of dissaproval if no measurements but I find this iterative approach very interesting and informative. I've learned a lot. But I have bought a Dayton Audio microphone and do intend to see what sort of mess I've created! They do sound very nice. I'd like a bit more bass and know it's there so probably crossing nearer to the baffle step frequency and attenuating the upper 2 drivers is in order. But I dont know if the naked MCA15 will do that if it's rolling off at the dipole frequency of 1146Hz.
Latest XO below.
I have some newer bigger guage Monacor 3.9mH inductors to try on the woofer. Much lower DCR. And I'll try a film cap on the mid. It's been handy this old Wharfedale crossover! That's the story so far. Thanks for tuning in and I'll update accordingly.
My other speakers are XRK Ref Monitors so I have experience of RS225 and I bought another pair. I had swapped out the 3"FR in that build for a Morel CAT378 with good results.
So after much simming and reading I thought I had enough knowledge to 'have a go' and so it began. Pretty much cobbled together and used various values of XO components that were on the old Laser crossover board. Far from optimal but a start .
First iteration above with SB15NRXC30-8 and SB29RDNC.
This with 3.7mh and 25uf on the woofer as per the Laser Xo which is about where I wanted to be. Mids were 1st order electrical HP of series 16uf and LP being 0.8mh, 8uf. HP on tweeter was 6.8uf 0.3mH. Padded down about 6db.
Initial impression were that it 'worked' but was a bit bright. I tried padding the mid and also upped the HP on it to 50uf to push them lower. Sounding decent now but I think the break up of the SB15 is causing the brightness.
I remembered I had some RST28F in WG300 so I tried that as a 2 way.
It's good to have an external crossover to be able to swap bits in and out easily! Not sure the wife is so thrilled at this point!
So maybe I should be deciding exactly what I want to build now but the 3 way was still an itch so I went back and tried me a bit of a dipole naked setup .
Having read the dipole tweeter thread and seeing the use of small and cheap Vifas, I had already bought some and employed them here now. BC25SC15-04 on the front for improved WAFnand the OC25SC15 on the back. Doesn't give as small a dipole distance as 2 OC back to back but looks better. Wired them series + to - but I'm not sure if that's correct. I think the rear membrane should move in the same direction as the front firing one. Here it will fire backwards as the front fires forwards.
HF circuit now modified for 8 ohm as the 2 4ohms in series, so now on 3.3uf and I think 0.56mH. I am aware these tweeters have quite high FS so I'm probably pushing them a bit with this approx 3k HZ filter. Can't hear anything untoward from them mind . The main issue as I look at the data sheets is the the SB15 won't cross nicely to the tweeters up at 3k. They start to break up fast after 3k. I considered a notch filter but remembered I had some Seas MCA15RCY. These look much flatter and so they went in next. Now we're getting somewhere. Sounding nice!
I keep playing with the series cap on the mid, I've pushed it a bit lower to meet the RS225 with a 50uf. Having read a bit more on dipole I understand with a calculation that the mid should be rolling off 1st order at about 1100Hz. (172/ 0.15m). The 50uf should be rolling it off again at about 450hz. I removed the padding on the mid. All this being done by ear and looking at the data sheets and playing with XSim. I might try 1st order LP on the mid with 0.3mH and simming a lower value HP cap on it gives a much sharper reverse null.
I can hear the groans of dissaproval if no measurements but I find this iterative approach very interesting and informative. I've learned a lot. But I have bought a Dayton Audio microphone and do intend to see what sort of mess I've created! They do sound very nice. I'd like a bit more bass and know it's there so probably crossing nearer to the baffle step frequency and attenuating the upper 2 drivers is in order. But I dont know if the naked MCA15 will do that if it's rolling off at the dipole frequency of 1146Hz.
Latest XO below.
I have some newer bigger guage Monacor 3.9mH inductors to try on the woofer. Much lower DCR. And I'll try a film cap on the mid. It's been handy this old Wharfedale crossover! That's the story so far. Thanks for tuning in and I'll update accordingly.
Thanks guys.
A big bonus with them is pretty much even sound from most angles. You can walk around the room/sit off axis and they still sound right.
@abstract I'm just listening to music and thinking...that's better....or worse. Generally music is sounding good but I must have a hole somewhere as with speech when watching television through them there are times when certain voices are very recessed.
I'm just tweaking the XO now.
A big bonus with them is pretty much even sound from most angles. You can walk around the room/sit off axis and they still sound right.
@abstract I'm just listening to music and thinking...that's better....or worse. Generally music is sounding good but I must have a hole somewhere as with speech when watching television through them there are times when certain voices are very recessed.
I'm just tweaking the XO now.
Looks like fun. There's nothing like a bit of experimentation.
But you will find measurements will help hugely. It's ridiculously hard to do it all by ear.
I recently put together a passive crossover for a two way wide-range plus woofer. I measured the drivers and designed the crossover, then wired it all up and sat down for an initial listening session. I was very pleased with how they sounded - not perfect but lots of positives. I thought I should measure them as well as listen, though. To my horror, the response was way off, with a massive hump in the mids.
It turned out I'd made a stupid schoolboy error - my crossover needed a 3mH inductor on the woofer, but I didn't have any to hand, so I wired up three smaller inductors, totalling the correct value. Except, it turned out I had connected the three in parallel rather than in series, giving about 0.3 mH instead of 3 mH! Hence the massive error in the mids!
It was very sobering that I thought they sounded really good on that initial listening session. They do sound better with the error corrected, though! But it really showed me how hard it is to tune by ear.
But you will find measurements will help hugely. It's ridiculously hard to do it all by ear.
I recently put together a passive crossover for a two way wide-range plus woofer. I measured the drivers and designed the crossover, then wired it all up and sat down for an initial listening session. I was very pleased with how they sounded - not perfect but lots of positives. I thought I should measure them as well as listen, though. To my horror, the response was way off, with a massive hump in the mids.
It turned out I'd made a stupid schoolboy error - my crossover needed a 3mH inductor on the woofer, but I didn't have any to hand, so I wired up three smaller inductors, totalling the correct value. Except, it turned out I had connected the three in parallel rather than in series, giving about 0.3 mH instead of 3 mH! Hence the massive error in the mids!
It was very sobering that I thought they sounded really good on that initial listening session. They do sound better with the error corrected, though! But it really showed me how hard it is to tune by ear.
Indeed! And I'm not sure you can 'add' inductors together can you due to some interaction ....might be wrong though.
So my last series of XO mods seemed to make things worse. The mid got really bright. I've added 3r series and 5.8r shunt as an L Pad on the mid for now. Much better.
Using a series of test tones on Spotify and an android based spectrum analyser....and also by ear....seems things really roll off quite soon. 250Hz is appreciably quieter than 400. And the midrange is still putting out quite some level still at 4000hz. I expected that to be quite a bit lower up there. Obviously suspect microphone in the phone but my ears say the same.
Uploading the crossover schematic when I suss out how to!
Edit....the important midrange should be very low distortion as the driver is barely moving now
So my last series of XO mods seemed to make things worse. The mid got really bright. I've added 3r series and 5.8r shunt as an L Pad on the mid for now. Much better.
Using a series of test tones on Spotify and an android based spectrum analyser....and also by ear....seems things really roll off quite soon. 250Hz is appreciably quieter than 400. And the midrange is still putting out quite some level still at 4000hz. I expected that to be quite a bit lower up there. Obviously suspect microphone in the phone but my ears say the same.
Uploading the crossover schematic when I suss out how to!
Edit....the important midrange should be very low distortion as the driver is barely moving now
You can add inductors, but as always you have to be careful how you position them, because of the potential for interaction. It's only a temporary solution, though. Bigger inductors have been ordered, and are on the way.
I'm struggling to understand why my system impedance is so low. I'd overlooked it on XSim whilst trying to get a flat FR and nice deep nulls. I've got a nice flat impedance and phase sticks around 0degrees.... whatever that means. But impedance is hovering between 2 and 3.5ohms.
Did a static DCR measurement of the XO and drivers connected and it was indeed in the 3 ohm zone. Way too low methinks.
When I remove the parallel resistor on the tweeter and mid L Pads it raises back up but obviously now I don't have the attenuation I need.
Need to look into this further.
Both L Pads are approx 3r series and 5.8r shunt. Over 8ohm drivers with DCR of nearly 6ohms.
So I understand the parallel 5r8 over a 5r8 voice coil will give me a 2r9 load but then adding the series 3r should have me back up to the correct load value. So I don't know why it's so low and what I can do about it!
More simming and calculating to be done.
Old laptop gifted today from my brother so may get some measurements soon. Once I learn how to do that.!
Did a static DCR measurement of the XO and drivers connected and it was indeed in the 3 ohm zone. Way too low methinks.
When I remove the parallel resistor on the tweeter and mid L Pads it raises back up but obviously now I don't have the attenuation I need.
Need to look into this further.
Both L Pads are approx 3r series and 5.8r shunt. Over 8ohm drivers with DCR of nearly 6ohms.
So I understand the parallel 5r8 over a 5r8 voice coil will give me a 2r9 load but then adding the series 3r should have me back up to the correct load value. So I don't know why it's so low and what I can do about it!
More simming and calculating to be done.
Old laptop gifted today from my brother so may get some measurements soon. Once I learn how to do that.!
Ive swapped the series attenuator resistors to larger values and now I'm getting some balance to the levels I like. Maybe I've gone a bit far so will try some 4r4 on the tweeter.
I think I've convinced myself there is an audible shift to things that might necessitate a rehash and possible needs a zobel now. Or because I've just been reading about zobels I think I need to implement one!
This is an approximation of the XO at the moment. Apologies for poor screen photo.
The 2 X tweeters are there just for simming the impedance. I guess I could have just used 1 X 8ohm as I don't think XSim can simulate a dipole. It helps give me an idea of values of filters if not for level.
And for what it's worth the hypothetical FR.
When I had the Lpads in I had a beautiful flat impedance and phase nice and close to zero, but albeit a very low impedance.
I think I've convinced myself there is an audible shift to things that might necessitate a rehash and possible needs a zobel now. Or because I've just been reading about zobels I think I need to implement one!
This is an approximation of the XO at the moment. Apologies for poor screen photo.
The 2 X tweeters are there just for simming the impedance. I guess I could have just used 1 X 8ohm as I don't think XSim can simulate a dipole. It helps give me an idea of values of filters if not for level.
And for what it's worth the hypothetical FR.
When I had the Lpads in I had a beautiful flat impedance and phase nice and close to zero, but albeit a very low impedance.
Have you tried auditioning the drivers individually and changing filter alignments to see what happens? I'm looking at the woofer - what if you tried it with just an inductor without the capacitor? How does it compare with 2nd order? And what about L-C-L for 3rd order? Good integration is of course important for the drivers to play together in concert with minimal phasing effects, but the individual voicing also adds up.
You could feel that one driver should take over a lot higher or lower than expected as well, based on listening tests.
You could feel that one driver should take over a lot higher or lower than expected as well, based on listening tests.
Yes I find that very informative. Unplugging the mid helped me decide that a harshness was coming from that and not the HF.
A quick go on XSim this morning and looks like a zobel will really help. Definitely on the mid. Maybe 10r and 10uf or so. An online calculator (MH Audio) gives R7.25 C 13.89uf. Shall try that and then look to tweak the filter after that.
A quick go on XSim this morning and looks like a zobel will really help. Definitely on the mid. Maybe 10r and 10uf or so. An online calculator (MH Audio) gives R7.25 C 13.89uf. Shall try that and then look to tweak the filter after that.
A few thoughts:
1. Regarding polarity, I'd definitely try having the tweeters in opposing polarity, so that the the rear tweeter moves in sync with the front one. If you have them wired pos to neg, as you do at the moment, you'll not get a dipole radiation pattern, but something closer to omni.
2. The polarities on your crossover schematic look a bit odd. You've got the woofer and mid shown as both being reversed, while the tweeters are shown in normal polarity. Your bass/mid crossover is 2nd order, so it would be more usual to have the woofer connected normal, and the mid reversed. You might also need to have the mid delayed a bit. If you did that, the front tweeter should be reversed, to match the mid, and the rear tweeter normal, so that it moves in sync with the front tweeter.
3. I've had a go at a sim of this in Basta. Note that it's not using any measured responses - just TS parameters - so it isn't going to be terribly accurate. It won't show the effect of driver resonances/break up etc. But I do find it quite helpful in this sort of situation, especially for getting an idea of relative sensitivities etc in a given box/baffle. I've set it up to use your baffle/box, with the woofer sealed and the mids and tweeters OB. And I've entered your crossover values, with a bit of guesswork for the inductor resistance.
The design looks decent, but it does suggest that you might need to pad the mid down a fair bit more, and the treble a bit less. The Seas mid is pretty sensitive, and crossing where you are you don't lose anything to baffle step. By the time you allow for baffle step, the Dayton woofer is only giving low 80s. It look to me as if you might need something like at least a 6 ohm series resistor on the mid, and just 2 or 3 ohms in series on the tweeters.
The other thing the sim suggests is that even with these changes, the bass might be a little lightweight. You might do better with a slightly bigger inductor on the woofer. But then you'd have to take a bit more off the mids. To beef up the bass a little, you could also try a vent in that box. A vent of 60mm diameter, 180mm long would give a tuning frequency of about 35Hz, and that looks better in my sim.
But really, all of this is guesswork, groping about in the dark, without measurements. Hope it might help a little, though!
1. Regarding polarity, I'd definitely try having the tweeters in opposing polarity, so that the the rear tweeter moves in sync with the front one. If you have them wired pos to neg, as you do at the moment, you'll not get a dipole radiation pattern, but something closer to omni.
2. The polarities on your crossover schematic look a bit odd. You've got the woofer and mid shown as both being reversed, while the tweeters are shown in normal polarity. Your bass/mid crossover is 2nd order, so it would be more usual to have the woofer connected normal, and the mid reversed. You might also need to have the mid delayed a bit. If you did that, the front tweeter should be reversed, to match the mid, and the rear tweeter normal, so that it moves in sync with the front tweeter.
3. I've had a go at a sim of this in Basta. Note that it's not using any measured responses - just TS parameters - so it isn't going to be terribly accurate. It won't show the effect of driver resonances/break up etc. But I do find it quite helpful in this sort of situation, especially for getting an idea of relative sensitivities etc in a given box/baffle. I've set it up to use your baffle/box, with the woofer sealed and the mids and tweeters OB. And I've entered your crossover values, with a bit of guesswork for the inductor resistance.
The design looks decent, but it does suggest that you might need to pad the mid down a fair bit more, and the treble a bit less. The Seas mid is pretty sensitive, and crossing where you are you don't lose anything to baffle step. By the time you allow for baffle step, the Dayton woofer is only giving low 80s. It look to me as if you might need something like at least a 6 ohm series resistor on the mid, and just 2 or 3 ohms in series on the tweeters.
The other thing the sim suggests is that even with these changes, the bass might be a little lightweight. You might do better with a slightly bigger inductor on the woofer. But then you'd have to take a bit more off the mids. To beef up the bass a little, you could also try a vent in that box. A vent of 60mm diameter, 180mm long would give a tuning frequency of about 35Hz, and that looks better in my sim.
But really, all of this is guesswork, groping about in the dark, without measurements. Hope it might help a little, though!
Last edited:
Here are some screenshots of my sim. The first one shows the predicted combined response of your speaker as is (but with the polarities as I suggested above). The second one shows the individual drivers as is. The third one shows the predicted response with some changes - vented box tuned to 35 Hz, woofer inductor upped to 5 mH, 8 ohm series resistor on the mid, 3 ohm series resistor on the tweeters, mid and tweeters set back 60mm. (Note, I've assumed both tweeters are the BC25. Too complicated otherwise!)
To be honest, though, I'd wouldn't put any time or money into trying these changes - I'd want to measure first. The sim is only a sim, and could be badly off.
To be honest, though, I'd wouldn't put any time or money into trying these changes - I'd want to measure first. The sim is only a sim, and could be badly off.
Attachments
Last edited:
If you did want to go OB the whole way, you could do worse than the SB Bianco 15OB350. Here's a sim of it with your mids and tweets. Woofer is on a 50 x 60cm rectangular baffle, low pass crossover is 2nd order (13mH low resistance inductor, Janzen C-core for example, and 100 uF). Midrange highpass is changed to 1st order (68 uF), and it has an L pad, 2 ohms series, 12 shunt. Lowpass midrange cap reduced to 6 uF. Tweeters have no resistive padding at all. Mids and tweeters are set back 80mm from the plane of the woofer.
Again, it's only a sim, not based on any driver measurements, and it's certainly not perfect. The actual crossover would undoubtedly need to be different. But food for thought maybe.
Again, it's only a sim, not based on any driver measurements, and it's certainly not perfect. The actual crossover would undoubtedly need to be different. But food for thought maybe.
Last edited:
Thanks very much for all that!
1. The relative polarity of the tweeters was something I wasn't sure about from the go so yes I will try swapping the rear so they both move in the same direction.
2. The reversed mid was to show the nulls. Normal practice is reverse woofer and normal on the mid and tweets. That what XSim seemed the think.
I had sort of planned to cross at the baffle step or there abouts....500 or so and then reduce the rest to suit. I have another pair of RS225 that could come into play....3.5 way those 2 at the baffle step.?! Bass is plenty adequate as is really. Probably a lot of room gain. Vented always an option but I seem to like to stick with sealed.
I think you prempted what I've done with the padding. I've got 5r6 on both mid and high now but think it needs less on the tweeter to my ears. It lost some sparkle. Plan on parallel 12r over it as a test.
Regarding delay, I do have the little baffle with the 2 upper drivers tilted back some random angle ! I guess when I measure I can look at this
A 15" is definitely out of the WAF for full OB..Maybe the 2 X RS225 with the back taken off the box!
Thanks again for your time
1. The relative polarity of the tweeters was something I wasn't sure about from the go so yes I will try swapping the rear so they both move in the same direction.
2. The reversed mid was to show the nulls. Normal practice is reverse woofer and normal on the mid and tweets. That what XSim seemed the think.
I had sort of planned to cross at the baffle step or there abouts....500 or so and then reduce the rest to suit. I have another pair of RS225 that could come into play....3.5 way those 2 at the baffle step.?! Bass is plenty adequate as is really. Probably a lot of room gain. Vented always an option but I seem to like to stick with sealed.
I think you prempted what I've done with the padding. I've got 5r6 on both mid and high now but think it needs less on the tweeter to my ears. It lost some sparkle. Plan on parallel 12r over it as a test.
Regarding delay, I do have the little baffle with the 2 upper drivers tilted back some random angle ! I guess when I measure I can look at this
A 15" is definitely out of the WAF for full OB..Maybe the 2 X RS225 with the back taken off the box!
Thanks again for your time
Last edited: