In this very technical video we cover user submitted crossover designs discussed in Part 3. Thanks to everyone who submitted. I'll be responding to everyone by email in the next few days.
The idea is just brilliant! Really building, measuring, comparing to simulations and sharing your thoughts is just awesome! I have learned so much from your videos. Can't wait for next one!
Thank you, exactly what I was looking for, xsim crossovers with design considerations. I've learned a lot! I'm building a crossover for a 3 way speaker, glassfiber woofer, 50mm softdome mid and softdome tweeter. I use a modified off the shelf crossover, better than nothing but far from perfect. I have 4 in quad stereo setup. First attempt parts for just one already runs into 100 euros, it certainly adds up quickly!
this series gonna be useful for me, today i just received my woofer and tweeter from Sb Acoustics, they're SB15NRXC30-8 and SB19ST-C000-4, and hopefully i can make a good sounding speaker with good box design too
I know you probably don't take requests but I was thinking as its nowhere to be found and no one has any such videos, maybe you should make a video teaching how to take measurements, properly, like mic placement etc, and when designing a crossover most people only talk about 1 woofer and 1 tweeter, someone like me wants to do 3 woofers and 1 tweeter per speaker, like the PSB t65, how will the crossover be different? How will it be measured differently? When we use the mic to measure response, do we play just 1 woofer or all 3 together to get a frequency response for crossover design? Where should the mic be placed if we do multiple drivers? These are things I've not found an answer anywhere to! Love your videos. You're helpful and have always replied to help me with all you can, much appreciated.
I do listen to requests! Whether or not I can find the time to do them is another story. Right now being so dark and cold it's hard to get videos done. A few people have requested this kind of video and I'd like to do one. But it would be a lot of work. For now I have a few things to get done. Please spread the word. The more people watching the more I can justify spending the time. I need to buy a better camera for this dark winter weather!
Impulse Audio fair enough. I will do so man! Thank you. I understand. Hopefully you will cone through with such a video as its nowhere to be found. And there is no answers to all this no matter where I look, not just RU-vid. I will share your channel as much as I can. Thanks again.
First of all, I have to thank you for working so hard in putting together those crossovers, and test them... I know how hard this think is, and time consuming! I only know the basic crossover tipes, but how to shape an SPL or an Impedance response, that I do not know! In a previows video of yours, there is a simulation of a crossover of the second order, if I recall correctly, and besides of a coil and a capacitor, there was a resistor in series with this capacitor... I tryied it myself, and I've noticed that the woofer SPL was shaped perfectly! I want to ask you what is this, actualy? And, most of all, where can I find any theoretical formulas regarding the ... contour networks ... I presume this is the name... Thank you again!
I’m not sure exactly what it was you’re referring to, but most likely a zobel filter. The best is to simulate the response using actual measured responses and software. The old calculations are a waste of time.
@@ImpulseAudioSpeakers no, it wasnt a zobel filter... on the woofer, because this is the type of driver that I was reffering to, I have used a second order filter. In series with the capacitor I have used a resistor in such a manner as one pole of the resistor is connected to the capacitor, and the other one to the ground! I have seen this in one of your videos... I dont know what this is... I am using Xsim and VituixCad2. Vituix helped me to create the *.frd and *.zma files for the SPL and impedance graph, because I allredy had them as images. I did not measure the driver myself... I do not have the means to do this... besides, I do not have the experience necessary for doing such a job... Thanks for the reply!
Could you do something similar but using an active set up? Ideally using a PC for source and DSP (e.g. Equaliser Apo can do IIR and FIR) and comparing filter performance with graphs AND music samples. This would be BRILLIANT!! Or what about an active vs passive vs hybrid and compare performance (though you might have to wait for the alligator clip nostalgia to set in). Thanks for all the work, and excellent technical break-down!!
Hi Mark, that's not impossible to do, although I don't have any projects planned that could work that way at the moment. I will keep it in mind though.
Digital filters do not perform anything like passive, so I don't see how it would be possible! I have done it before in the past for fun and I was surprised how poorly the digital filter "modeled" the passive version with phase and knee characteristics (for starters).
Ryan, have you tried the Digital Filter in SoundEasy? I have used it myself, and find that it saves me plenty of prototyping time, not having to do a physical build until the final design iteration. In fact, I would recommend it as a possible future video topic: you could do an evaluation and report on its effectiveness.
Shaun Onverwacht I actually have not. I find SoundEasy so cumbersom that I havent used quite a few of the features. I dont find building test XOers that big of a deal normally, but for this video it would have probably saved me some time to use the simulator.
Hi, thank you for these videos. Could you possibly explain why a lower crossover point works well for these desk speakers and why a higher one would suit living room speakers.
SuperChilled I plan to cover this in much more detail in a video cause its a bit of a challenging topic to descibe in a comment. Basically it has to do with wavelengths and distance between the drivers can cause nulls in the response. The power handling required so close to the speakers means the tweeter can handle it. In a living room where you sit further away the speaker has to produce 10-15db more output. Thats more than 10x the power handling required. So a higher XO is important to protect the speaker and mitigate high levels of distortion.
SuperChilled well yes there is some acoustical reason. The lower XO mitigates nulls in the response above and below the listening axis. Watch for a future video about this.
Pentode3000 just googled that. I have a bunch of those and don’t really care for them to be honest. I find them a bit cumbersome. More of a direct connection though so that is nice.
i started on xsim a few days ago i want to make a 3 way crossover but because i have professional components, the frd and zma files weren't available i have the sica 10fe 3 cp 8Ω woofer the sica 8m 1,5cs mid and the visaton sc 10 n tweeter i'll send the pdfs too.Now i got the curves from graph tracer and edited them, with paint 3d i erased everything except the frequency and impedance charts everything works in xsim, i noticed tho that the curves of the drivers, are not aligned with the x and y axes of xsim based on the factory pdf. Is it possible to adjust the axes of xsim so that the curves appear as they do in the factory pdf?? and if so how? As you guys can see for yourself in the crossover photo i sent the frequency of each driver if you compare them to the factory pdfs does not align they start at more hz and more dbs i would really appreciate an answer cause i cannot align them.
Hi Ryan. I saw in your video always use a white noise or pink noise to plot the frequency response. How to do that ? I only knew how to do it in sweep way but wonder why you do with noise file? Hope you can tell me.
There’s basically no difference. Most programs use a sweep. Sound Easy uses white. Some have the option of either. I actually prefer a sweep cause I can hear any issues with the speaker during the sweep.
howard stever I’m not sure if it’s the same, but sure sounds like it. It basically takes the frequency response and derives the minimum phase. Perfectly acceptable way of doing it.
Marcel Ooms neither is necessarily better than the other. Just different. The Lpad advantage is that it can be used to smooth the tweeter impedance. This helps the filter do its job better. Hard to explain, I sense a new video idea. The resistor in front of the filter is basically a power sink which is advantageous in keeping the impedance load on the amp high (good). I have used both the resistor in front as well as the Lpad when a lot of tweeter padding was required.
Sorry it's been a while since I made this video and don't have the time to watch it at the moment, but from what I can remember, the woofer and tweeter were out of phase by 30 degrees. This means that the sound waves from the woofer and the tweeter are not in time by a small amount and the resulting output will not be as full as it could be. It's a complex subject, but maybe one day I'll put together a video about phase.
Impulse Audio thank you for replying. I wish to get a picture, what spec would I need as a minimum requirement for doing speaker calculation and measurement. What softwares shopping list can I get hold of. It's been a passion of mine to make a pair of loudspeaker and learn how and what's involved. That's for your videos as I love to see what projects your doing.
That sort of depends on how technical you want to get with it. As a first design project, I would follow exactly what I did in this series on the computer. Build your results if you wish but start by at least downloading the files from Part 2A and 2B and following along in Xsim etc. See if you can reach the same place as myself and some of the other submissions. Then if you feel confident I would download files from drivers that interest you and try to come up with a crossover that way. If you feel confident, build it and see how it sounds. Start cheap in case it gives you trouble. Too many people start with expensive drivers because they think it will help. It won't. From there you can consider getting measurement equipment and doing the whole thing start to finish, but get some experience and spend money on a few project before dropping $200 on measurement equipment. At the same time, $200 for these tools is very cheap. Hope that helps.
WV591 haha. Sorry but not really. Besides I always measure when I’m done to make sure nothing is hooked up wrong. The XO in the box has never made a difference.
W엉덩 when drawing the circuit I focus on the 0 degree measurement. I will use the other angles as a guide. When I measure I usually take them from 0-60 degrees in 15 degree increments at a minimum. Often more.
@@ImpulseAudioSpeakers Speaker position a. The position between the human ears is not 0 degrees frontal. 0 Is it right to design the circuit with a road measurement graph?
W엉덩 rarely do we listen at perfectly 0 degrees but usually close enough. Even 10 degrees is often quite close to the on axis results. And if they’re very different I wonder how good the speaker is. Designing with a 5 or 10 degree measurement isn’t a bad idea though.
After watching your video on trusting crossover calculation websites, I'm even further convinced I'm not going to be able to do this myself, as I don't have access to xcel in order to use Xsim. Would you be willing to engineer a 3-way 4th order crossover circuit for me and I can bill you for your time? Or do you perhaps have a link where I can get more help or read up on the matter
NatesiKness you don’t need excel. Xsim runs in windows. If that doesn’t help unfortunately I won’t be able to help with your request. I’m about to start an addition on my house and can’t work on any new projects. There are quite a few designs people have shared on the internet you could build. Find one that uses measurements and good detail. Not just a couple of lines on a graph. You want one where the person knows what they’re doing. Cause it’s a gamble buying all the parts. You don’t know what you’ll get.
qwerty asdf completely depends on everything. Ideal is no XO at all but that never ever works. Generally lower is better provided the tweeter can handle the lower frequencies.
Quake Audio you mean how did I determine that it was 0.65”? Or what did I do with it? Sorry I’m not quite sure what you’re asking. I found it using the 3 measurement technique. It can also be estimated. Typically just under 1” for a small woofer and dome tweeter like this. It’s basically how much further the woofer’ acoustic center sits behind the tweeter. Some tweeters are in horns and it gets trickier to estimate.
Hi Impulse, thanks for your response. I meant to say what did you end up doing with that offset. did you introduce a delay circuit into the crossover? cheers
Quake Audio oh I see. I did nothing. There’s no need to do anything with it. As you adjust the crossover the phase relationship will change. Phase is distance and/or time for the most part. So the crossover filter changes the the phase to compensate for the offset. No other special requirements.
Christian Murace it sorta depends. But usually you have to tell the software the acoustic centres of the driver. Xsim uses delay in the driver tune menus. Every software is a bit different.
does this not technically mean when we design all our speakers on a flat vertical baffle we should design the crossover with an offset for example tweeter 20mm in front? this is what confuses me. but because both drivers are measured at 1m, it works out so we don't have to. hence why if we align the drivers, then measure it at 1m should it work out? Im sure that simple experimenting with cheaper caps will answer that question for me. forgive my barrage of questions, your words are gold to me.
Huh, i have those drivers. Nothing to measure with so i am using your files and your xover and my graph looks nothing like that. almost flat across 85db then tweeter range is like 89db
Derek Jarman pretty soon actually. The past month or so I've been getting the XO right where I want it. Which I now have and the parts are here. Need to put on the finishing touches and I'll post Part 5. Probably early new year. Thanks for checking in!
