I have built a triamp and currently use two pairs of Kef Coda 7 speakers wired for biamp so I only use the base speakers in one of the pairs.
This is temporary as I do not need any crossovers.
Although the sound is very good I would like to build a pair of cabinets/speakers without any crossovers.
I have never built any speakers and have only ever used the Kef's from new (my son had one set and I had the others).
I don't wish to spend large amounts of money as this would be the first attempt and I know from experience with the triamp that a few failures might be needed before success.
I am after some advice on what to use.
I have been recommended to try Scanspeak speakers but I have non experience of them.
I also am thinking whether it might be easier to build separate cabinets for the base speakers only with a combined cabinet for the tweeter and midrange.
I would be happy to achieve at least the same quality I get at the moment but if there could be an improvement that would be better.
I listen to my system most days; it does sound great and I keep thinking how on earth did I manage to build the triamp (took me about 5 years)
This is temporary as I do not need any crossovers.
Although the sound is very good I would like to build a pair of cabinets/speakers without any crossovers.
I have never built any speakers and have only ever used the Kef's from new (my son had one set and I had the others).
I don't wish to spend large amounts of money as this would be the first attempt and I know from experience with the triamp that a few failures might be needed before success.
I am after some advice on what to use.
I have been recommended to try Scanspeak speakers but I have non experience of them.
I also am thinking whether it might be easier to build separate cabinets for the base speakers only with a combined cabinet for the tweeter and midrange.
I would be happy to achieve at least the same quality I get at the moment but if there could be an improvement that would be better.
I listen to my system most days; it does sound great and I keep thinking how on earth did I manage to build the triamp (took me about 5 years)
I think you'll find that you have the speakers Biwired, not biamped. This is possible because the crossover is already inside the speaker.
The only way to build a speaker without crossovers is to use a 'full range' driver, such as a Visaton B200 or BG20, and live with their inherent limitations. Some love them, others are not so impressed.
Scanspeak drivers are excellent, but very expensive. You can get much better value with SB Acoustics or Peerless drivers.
Perhaps you could consider building a kit which would save you going through the crossover design process.
The only way to build a speaker without crossovers is to use a 'full range' driver, such as a Visaton B200 or BG20, and live with their inherent limitations. Some love them, others are not so impressed.
Scanspeak drivers are excellent, but very expensive. You can get much better value with SB Acoustics or Peerless drivers.
Perhaps you could consider building a kit which would save you going through the crossover design process.
Last edited:
I think you are setting up yourself for failure if you insist on using no crossover with a 3-way speaker. It won't be a good speaker, the tweeter for example will probably not survive that long if you feed it a full bass signal.
I'd recommend either building a speaker from an existing schematic or going active with DSP. A decent USB measurement microphone is cheap nowadays and it makes the driver crossovers and integration easy as you can build whatever box is convenient (within reason) and then make the response fit afterwards.
I'd recommend either building a speaker from an existing schematic or going active with DSP. A decent USB measurement microphone is cheap nowadays and it makes the driver crossovers and integration easy as you can build whatever box is convenient (within reason) and then make the response fit afterwards.
I definitely do not require crossovers.
My triamp has a front end linkwitz Riley filter which splits the audio signal into 3 frequency ranges. Each range is then fed into dedicated stereo amplifiers.
So there are 3 pairs of amplifiers, each pair only handling a narrow frequency band.
The high frequency amps cover approx 2500hz and above, mid range 275hz to 2500hz and low frequency below 275hz.
So there are no tone controls, each pair of amps is attenuated by an optical encoder.
So perhaps I should look at other speakers rather than scanspeak from what you say.
Not having any experience of speaker design makes it difficult to select them.
I wondered that not having to fit any crossovers perhaps would make it better to have separate cabinets for each speaker.
My triamp has a front end linkwitz Riley filter which splits the audio signal into 3 frequency ranges. Each range is then fed into dedicated stereo amplifiers.
So there are 3 pairs of amplifiers, each pair only handling a narrow frequency band.
The high frequency amps cover approx 2500hz and above, mid range 275hz to 2500hz and low frequency below 275hz.
So there are no tone controls, each pair of amps is attenuated by an optical encoder.
So perhaps I should look at other speakers rather than scanspeak from what you say.
Not having any experience of speaker design makes it difficult to select them.
I wondered that not having to fit any crossovers perhaps would make it better to have separate cabinets for each speaker.
Have a look at Troel's site to get some ideas for speaker combinations that work well together:
DIY-Loudspeakers
DIY-Loudspeakers
What we have here is a failure to communicate.
You do not have 'no crossovers'.
You have active(line level) crossovers, instead of passive crossovers.
As far as separate enclosures, yes the drivers should be separate so that they do no push on each others' diaphragms.
Tweeters are usually sealed, so you're good there, just keep the mid and woofer from beating up on each other(usually a one-sided fight, but the TS parameters of the woofer are screwed up by the mid acting as a poor passive radiator.
Since you are new yet to filters and drivers, you might look for projects with active filters. Michael Chua has several active systems on his Ampslab site, and is a very helpful guy.
After you have played with a few systems, you will find that you can make a good active version of almost any passive setup, and then it's time to experiment with random drivers.
For now, use use someone else's design as a starting point to play with filters without buying lots of fiddly bits.
You do not have 'no crossovers'.
You have active(line level) crossovers, instead of passive crossovers.
As far as separate enclosures, yes the drivers should be separate so that they do no push on each others' diaphragms.
Tweeters are usually sealed, so you're good there, just keep the mid and woofer from beating up on each other(usually a one-sided fight, but the TS parameters of the woofer are screwed up by the mid acting as a poor passive radiator.
Since you are new yet to filters and drivers, you might look for projects with active filters. Michael Chua has several active systems on his Ampslab site, and is a very helpful guy.
After you have played with a few systems, you will find that you can make a good active version of almost any passive setup, and then it's time to experiment with random drivers.
For now, use use someone else's design as a starting point to play with filters without buying lots of fiddly bits.
Thanks,
I think I will seriously consider separate cabinets.
I have experimented with some filters, I tried building a Butterworth filter 8th order but it was not as good as the linkwitz Riley filter even though there is some similarity on the design.
As I design the PCB it does take a lot of time to do the layout, etching and building but hopefully I will get back to doing some more.
It is the attenuator and pic which has taken the most effort, but now I will press on with the speakers and look at cabinet design.
I think I will seriously consider separate cabinets.
I have experimented with some filters, I tried building a Butterworth filter 8th order but it was not as good as the linkwitz Riley filter even though there is some similarity on the design.
As I design the PCB it does take a lot of time to do the layout, etching and building but hopefully I will get back to doing some more.
It is the attenuator and pic which has taken the most effort, but now I will press on with the speakers and look at cabinet design.
Read Linkwitz thoroughly. And understand the importance of especially low-pass shelving and delay parts. In other words: you not only need the filter, you need the enclosure design and driver specific corrections for a more or less linear transfer. Not just filters. If it only were that simple.
I was aware of the additional circuitry for phase correction and driver alignment but after building the basic filter using two 12db/oct filters cascaded with a total of eight filters (4 high pass and 4 low pass in each channel) I found the sound quality was excellent so I haven't gone any further yet.
I am not sure if my ears are good enough to pick up much improvements but I hope to put some more effort into the filters after I have built the speaker enclosures.
It just takes so much time; I am currently trying to fit a redesigned amp circuit into a limited space pcb, the amp is based on a BC design with some modifications to prevent oscillation.
I am not sure if my ears are good enough to pick up much improvements but I hope to put some more effort into the filters after I have built the speaker enclosures.
It just takes so much time; I am currently trying to fit a redesigned amp circuit into a limited space pcb, the amp is based on a BC design with some modifications to prevent oscillation.
Since making the analogue circuits and making them work well is pretty hard I'd suggest taking a peek at using a line level DSP solution and hook up you own amps. You could even build the DSP boards into your amps if you want to save space and reduce the amount of cables.
DSP makes configuring all the crossovers a lot easier. Even Linkwitz himself was pleased with DSP as an alternative to building the analogue filters and included a MiniDSP 4x10 configuration file in the building plans for the LX521.
Nowadays the DSP kits have matured even more so we can get even better performance for the equal cost.
DSP makes configuring all the crossovers a lot easier. Even Linkwitz himself was pleased with DSP as an alternative to building the analogue filters and included a MiniDSP 4x10 configuration file in the building plans for the LX521.
Nowadays the DSP kits have matured even more so we can get even better performance for the equal cost.
Last edited:
This has plusses and pitfalls. What are you hoping to achieve with separate cabinets?
Not to mention the many years of acoustics calculations and experiments that are what makes this speaker successful.
I think with a separate base speaker cabinet it may be easier to design and build rather than combining with the midrange speaker.
However before I start anything I will do a lot more investigation.
I would be looking to build using 25thk MDF with laminate finish. I have done quite a lot of MDF/laminate construction, my last job being a 6 level audio unit with machined aluminium supports.
I will also look through some of the forums here.
As far as the DSP goes I assume this is a digital line filter, do you have any experience of this.
However before I start anything I will do a lot more investigation.
I would be looking to build using 25thk MDF with laminate finish. I have done quite a lot of MDF/laminate construction, my last job being a 6 level audio unit with machined aluminium supports.
I will also look through some of the forums here.
As far as the DSP goes I assume this is a digital line filter, do you have any experience of this.
Yes, and yes =)
The best result is when combined with a measurement microphone like UMIK-1 which includes a calibration file.
In practice it is mostly the same way of doing things as before, just because you can easily do crazy sharp filters with DSP doesn't mean it is a good idea. There is a quality in simplicity, sharp high Q filters usually cause more problems than they solve.
Just generally continue as before and build reasonably flat measuring boxes. When this is done then measure each individually and make the response flat-ish over the whole range + an octave above and below the crossover. If you use REW then you can even use the software to calculate the filters you need to get the desired response (I use this a lot). Then you add the desired crossovers and start time aligning the drivers.
Since DSP has flat time delay we don't have to fiddle with all-pass workarounds anymore. When time aligning I playfully call it "chase the null" where I invert the highpassed driver and then switch the delay back and forth while checking the impulse of how close I am until I get the greatest null at the crossover frequency. Then i continue to the next driver and do the same one step up.
Another cool thing is that we can even correct the phase distortion of the crossovers and make the speaker linear phase.
That's very interesting Ollboll but a bit above my current level.
However I will look into this as I will soon be designing and building another triamp.
Even on my current all analogue system I get good results but I would like to develop it further.
Have you any suggestions for learning about DSP.
I have some experience of assembly programming and maths.
However I will look into this as I will soon be designing and building another triamp.
Even on my current all analogue system I get good results but I would like to develop it further.
Have you any suggestions for learning about DSP.
I have some experience of assembly programming and maths.
MiniDSP is very popular one. Home | miniDSP
With that one (or other alternatives for audio use) it is not necessary to know about assembly or maths.
With that one (or other alternatives for audio use) it is not necessary to know about assembly or maths.
MiniDSP on top of having the easiest UI to work with of those I've tried also also have really good guides of how it works and how to use their stuff:
DSP basics
Stereo 2 Way Xover
Auto-EQ tuning with REW
Is the one I talked about where REW (a measurement software) can auto-tune the filters. But this one is a bit more advanced and so you can skip it until you feel more comfortable with the concept.
DSP basics
Stereo 2 Way Xover
Auto-EQ tuning with REW
Is the one I talked about where REW (a measurement software) can auto-tune the filters. But this one is a bit more advanced and so you can skip it until you feel more comfortable with the concept.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.