Vacuphile,
Did you normalize the on axis response, it looks like a flat line with no ripples or anything going on? I see the 1/3 octave smoothing but that is just so flat even for a smoothed curve.
Did you normalize the on axis response, it looks like a flat line with no ripples or anything going on? I see the 1/3 octave smoothing but that is just so flat even for a smoothed curve.
Kindhornman, it is indeed normalized. However, this speaker is almost flat (+/-1.5 dB) without any kind of DSP-flattery, so the non-normalized curve looks very similar.
I don't think you could do much to improve that. I wonder if you change the vertical if you will make the horizontal dispersion worse?
What are you doing for a crossover network and is this also using dsp correction?
What are you doing for a crossover network and is this also using dsp correction?
I have developed on a dsp platform, and the slopes were subsequently translated into an analog filter (which beats the MiniDSP with about 15dB on THD+N). It is a three way 4th order network acoustic, with baffle step compensation on the mids and a delay line on the tweeter.
Coming back to a suggestion which was made on this thread to have a passive xover between mid and high, I can tell you why I did not opt for that.
Firstly, if you already have an active xover, addition of another channel is relatively cheap. Quality amplification is almost free nowadays. Total additional cost of a fully active additional channel including amplification is less than 10 USD. For that price, it is impossible to make an equivalent passive filter.
Secondly, it would have been impossible to achieve the same precision with a passive xover, certainly if VC heating is taken into account.
Thirdly, it would have been neigh well impossible to realize passively the delay required to align mid and high. Working with third order filters to align these drivers as Richard suggested not a solution I like, as you typically loose a degree of freedom by doing so - ime the xover frequency cannot be choosen freely, but is prescribed by the difference in acoustic centre of the drivers.
Coming back to a suggestion which was made on this thread to have a passive xover between mid and high, I can tell you why I did not opt for that.
Firstly, if you already have an active xover, addition of another channel is relatively cheap. Quality amplification is almost free nowadays. Total additional cost of a fully active additional channel including amplification is less than 10 USD. For that price, it is impossible to make an equivalent passive filter.
Secondly, it would have been impossible to achieve the same precision with a passive xover, certainly if VC heating is taken into account.
Thirdly, it would have been neigh well impossible to realize passively the delay required to align mid and high. Working with third order filters to align these drivers as Richard suggested not a solution I like, as you typically loose a degree of freedom by doing so - ime the xover frequency cannot be choosen freely, but is prescribed by the difference in acoustic centre of the drivers.
Last edited:
Vacuphile,
Thank you for that explanation of what you have done. I actually looked at changing the shape of my enclosure to physically align the devices for the acoustic center, this was a mind exercise that in the end just changed one aspect of the overall system while bringing in other acoustic factors such as sloping baffle and angular section between devices which was going to just create new problems to deal with, there really is no easy solution for having two or more device that are separated by a distance between centers. I understand what you are doing by going to very small drivers and having the center to center distances as small as possible. I imagine that your final concern about the suck out in the vertical axis is caused by the MTM arrangement you have chosen, there is no perfect solution to this but you surely have come close to that ideal.
Now building an amp channel for only $10.00 is another thing. I would think you would have to be using monolithic amplifier chips to do that?
Thank you for that explanation of what you have done. I actually looked at changing the shape of my enclosure to physically align the devices for the acoustic center, this was a mind exercise that in the end just changed one aspect of the overall system while bringing in other acoustic factors such as sloping baffle and angular section between devices which was going to just create new problems to deal with, there really is no easy solution for having two or more device that are separated by a distance between centers. I understand what you are doing by going to very small drivers and having the center to center distances as small as possible. I imagine that your final concern about the suck out in the vertical axis is caused by the MTM arrangement you have chosen, there is no perfect solution to this but you surely have come close to that ideal.
Now building an amp channel for only $10.00 is another thing. I would think you would have to be using monolithic amplifier chips to do that?
Weyasayso, I am traveling until the 7th so impulse response will have to come later.
Kindhornman, LM3886 is a great building block. The bare chip can be had for around 2 USD, sometimes a bit more, sometimes a bit less. Main costs are power supply and casing, but since you need these anyhow, each additional channel costs peabutts. It is difficult to imagine a two way system that can't run on two paralleled LM3886 for midbass and one for the highs, add one for a three-way to drive the mids.
The suckout is there partly because even 4th order keeps the mids too long busy after the xover point. Therefore I am going to buy a MiniSharc to see if steeper filters would improve this.
If you want to go all the way, put it in a loop with a great opamp and you can get silly low distortion figures.
Kindhornman, LM3886 is a great building block. The bare chip can be had for around 2 USD, sometimes a bit more, sometimes a bit less. Main costs are power supply and casing, but since you need these anyhow, each additional channel costs peabutts. It is difficult to imagine a two way system that can't run on two paralleled LM3886 for midbass and one for the highs, add one for a three-way to drive the mids.
The suckout is there partly because even 4th order keeps the mids too long busy after the xover point. Therefore I am going to buy a MiniSharc to see if steeper filters would improve this.
If you want to go all the way, put it in a loop with a great opamp and you can get silly low distortion figures.
Thanks again Vacuphile. I can see using a 3886 for my tweeter but not for the bass/mid as the base efficiency is only about 86db@ 1 watt. That will be something where discrete parts will most likely be the answer.
Behringer B2031 Active runs 3886 for highs and parallel 3886 for lows.
I have no idea of bass driver efficiency but plenty enough level to give the neighbours cause to throw bricks, and the missus to yell from the kitchen.
Dan.
I have no idea of bass driver efficiency but plenty enough level to give the neighbours cause to throw bricks, and the missus to yell from the kitchen.
Dan.
Max,
The Behringer is huge in the sense that I have to put this inside the enclosure. I originally looked at using 3886 chip amps to drive my speakers. I read many of the threads on this site and over and over once you left a single unit and started to parallel or bridge those chips many problems were noted. Many have moved to other higher powered chip amps or even using an opamp to drive these high powered chips, it gets very confusing to go that direction. I understand how cheap the 3886 is to use and that is why we see it in so many places but it does seem to have some real limitation once you are looking for some real output with higher current capabilities. I am looking for 120 watts @ 8 ohms to give me 3db of overhead at max output of my speakers as determined by the suspension limitations. If I could overcome those limits of the suspension I would be looking for 300 watts, that would just be to much heat to dissipate in the heat sink that I can fit to the back of the speaker and I don't want to even consider class D at this point in time. I am not trying to develop a $10K self powered speaker system, I am not going there.
The Behringer is huge in the sense that I have to put this inside the enclosure. I originally looked at using 3886 chip amps to drive my speakers. I read many of the threads on this site and over and over once you left a single unit and started to parallel or bridge those chips many problems were noted. Many have moved to other higher powered chip amps or even using an opamp to drive these high powered chips, it gets very confusing to go that direction. I understand how cheap the 3886 is to use and that is why we see it in so many places but it does seem to have some real limitation once you are looking for some real output with higher current capabilities. I am looking for 120 watts @ 8 ohms to give me 3db of overhead at max output of my speakers as determined by the suspension limitations. If I could overcome those limits of the suspension I would be looking for 300 watts, that would just be to much heat to dissipate in the heat sink that I can fit to the back of the speaker and I don't want to even consider class D at this point in time. I am not trying to develop a $10K self powered speaker system, I am not going there.
Kindhornman, having followed this thread from the beginning, my best advise for you would be to hire a guy who can do a design for you. Your job should focus on drafting detailed specs for the electronics package, based on your loudspeaker design. Don't specify technologies but set performance standards.
Vacuphile,
I am in agreement on that, I do have a couple people in mind if they are willing to do it. I would need to spend years learning all the ins and outs of this and all the finer details. I am at least learning enough to understand what I required and the different way to go about it. When you said earlier that you worked out the network for your speakers using dsp and then made an analog equivalent what exactly did you do. Can I assume it is still an active analog circuit and that you didn't get those waterfalls using a traditional passive network?
I am in agreement on that, I do have a couple people in mind if they are willing to do it. I would need to spend years learning all the ins and outs of this and all the finer details. I am at least learning enough to understand what I required and the different way to go about it. When you said earlier that you worked out the network for your speakers using dsp and then made an analog equivalent what exactly did you do. Can I assume it is still an active analog circuit and that you didn't get those waterfalls using a traditional passive network?
Kindhornman, it is convenient to work out the optimum slopes and compensations using a dsp because you can make iterations much quicker. However, as I posted earlier, dsp for production has a number of disadvantages. So, after you have established the optimum slopes, you whip out your slide rule or other preferred calculation device, and translate all those curves into an analog equivalent. Basically a bunch of opamps, caps and resistors. Lots of people will be able to do that for you. In order to save space and cost, it would be nifty to design in the power stage at the same time so that you can put all on one pcb.
Edit: the waterfalls were actually a directiviy diagram the same way Stereophile presents these data, and they were done with an analog xover as described above connected to class d amps.
Edit: the waterfalls were actually a directiviy diagram the same way Stereophile presents these data, and they were done with an analog xover as described above connected to class d amps.
Last edited:
Thanks Vacuphile, I was assuming but wanted to make sure you were indeed using an active network and not telling me you were doing this with passive components. On the integration of this with the power amps that is also one of my goals. I don't want to have multiple boards all strung together with connectors and jumpers or ribbon cables if at all possible. That I why as you say I need someone who can integrate all of this in a nice small package. I won't say it is going to be easy but that is one of the goals as you say. I have drawn myself a block diagram and now I need someone who can create that in silicone and get it off a piece of paper. I will talk to one of the people I know can do this and see if they are interested.
I suggest hunting down a pair of the Behringers and taking a closer look/listen or buy a pair used...they can be had pretty cheaply.
I have altered the internal damping arrangements but deliberately not altered the amplifier/crossover module....for now.
I am running balanced feed direct from usb soundcard (plus isolation transformer 240V power) and the sonic result is quite spectacularly good, with delightful clarity and depth imaging going for miles (signal polarity is perfectly revealed).
The only criticism perhaps is very low end extension and low end power capability, but of course this is to be expected for an 8" 2 way.
As an example of what does work quite decently these are worth a study, and the schematics are on the net if you search.
Dan.
Thanks Dan. I wish kgrlee wasn't on your side of the globe, I think he and I could work real well together. Perhaps he may have to brush off the sand and play a bit of catch up to what is currently going on but he has so much test and development knowledge. If I still didn't have two more kids to take care of, one is about to leave for college in Hawaii, I would love to move down under. Just getting away from Southern California would be a nice thing.
Thanks for this Vac.
The display is of course what I mentioned in #306. It's not in AES so Julian probably published in IoA, Windermere and Graham Bank was probably co-author.
What size coil on the 2 1/2" mids?
Is the 2k5 kHz xover treble/mid?
What is the mid/bass xover?
Can you tell us who you are doing this work for?
I'm tempted to suggest how you can improve these .. SHUT UP LEE! JUST SHUT UP!!
Hello
maybe you interested in a free GPL software solution for a 4 way - crossover.
You can see my setup here:
diyAudio server HTTPS page
HIFI-FORUM » Do it yourself » Lautsprecher » Aktive Frequenzweiche per Standard Software mit GUI
Sorry it is in German
You have
4 x 2 crossover
many parametic equalizer
limiter
....
much more
HDMI 8 Channel output
8 Channel analog output
For activating for example 2, 3 or 4 way loudspeaker
PC based so bluetooth shal be possible
Regards
Loafmeat
maybe you interested in a free GPL software solution for a 4 way - crossover.
You can see my setup here:
diyAudio server HTTPS page
HIFI-FORUM » Do it yourself » Lautsprecher » Aktive Frequenzweiche per Standard Software mit GUI
Sorry it is in German
You have
4 x 2 crossover
many parametic equalizer
limiter
....
much more
HDMI 8 Channel output
8 Channel analog output
For activating for example 2, 3 or 4 way loudspeaker
PC based so bluetooth shal be possible
Regards
Loafmeat
- Status
- Not open for further replies.