Hi All,
Not sure which section to post this in. Moderators - apologies if this is the wrong forum.
Please could someone help me with determining the correct crossover set up for my minor speaker upgrade?
Have recently changed tweeters and now trying to determine the crossover point.
The setup was an Alcone AC8HE (in sealed enclosure of volume stated on the datasheet) along with the Alcone AC1 HAT tweeter (these are protected by a 22uF cap). The crossover for these was an active LR4 @ 2kHz. These speakers were built about 8 years ago when I had very little knowledge of DIY Hi-fi. Although, knew enough to brace the cabinet, use a double thickness (1.5") front baffle and for some reason lined the insides with 1/4" thick lead 🙂.
Was looking at the datasheet for the AC1 HAT and decided that it was the weak link in my set up. Replaced it with an MB Acoustics SB29RDC. The sound changed beyond recognition. It such a big change that it somehow doesn't sound right. This was using the same crossover (LR4 @ 2kHz).
I realise that the crossover probably needs changing but to what frequency? Using my minimal knowledge of loudspeakers I reckon about 1.5kHz. This is just over 2x the fs of the tweeter and well below the frequency where the cone break up of the AC8HE occurs.
Here are links to the datasheets of the drivers mentioned above.
http://www.lautsprechershop.de/pdf/alcone/ac8he.pdf
www.lautsprechershop.de/pdf/alcone/ac1hat.pdf
http://www.falconacoustics.co.uk/downloads/SBA/SB29RDC-C000-4.pdf
The end goal is to have an active LR4 with (optional) baffle step correction and time alignment.
Realise that, as usual, I'm probably trying to do too much too quickly but that's how I approach my DIY stuff. The only problem is that now I have got myself all confused, even with a lot of research. Been messing with the relative "gain" levels between the tweeter and woofer and it just doesn't sound right whatever I do. Probably something obvious but it escapes me.
So here I am humbly asking for help from those that know so much more.
Thank you for taking the time to read this.
Paul
Not sure which section to post this in. Moderators - apologies if this is the wrong forum.
Please could someone help me with determining the correct crossover set up for my minor speaker upgrade?
Have recently changed tweeters and now trying to determine the crossover point.
The setup was an Alcone AC8HE (in sealed enclosure of volume stated on the datasheet) along with the Alcone AC1 HAT tweeter (these are protected by a 22uF cap). The crossover for these was an active LR4 @ 2kHz. These speakers were built about 8 years ago when I had very little knowledge of DIY Hi-fi. Although, knew enough to brace the cabinet, use a double thickness (1.5") front baffle and for some reason lined the insides with 1/4" thick lead 🙂.
Was looking at the datasheet for the AC1 HAT and decided that it was the weak link in my set up. Replaced it with an MB Acoustics SB29RDC. The sound changed beyond recognition. It such a big change that it somehow doesn't sound right. This was using the same crossover (LR4 @ 2kHz).
I realise that the crossover probably needs changing but to what frequency? Using my minimal knowledge of loudspeakers I reckon about 1.5kHz. This is just over 2x the fs of the tweeter and well below the frequency where the cone break up of the AC8HE occurs.
Here are links to the datasheets of the drivers mentioned above.
http://www.lautsprechershop.de/pdf/alcone/ac8he.pdf
www.lautsprechershop.de/pdf/alcone/ac1hat.pdf
http://www.falconacoustics.co.uk/downloads/SBA/SB29RDC-C000-4.pdf
The end goal is to have an active LR4 with (optional) baffle step correction and time alignment.
Realise that, as usual, I'm probably trying to do too much too quickly but that's how I approach my DIY stuff. The only problem is that now I have got myself all confused, even with a lot of research. Been messing with the relative "gain" levels between the tweeter and woofer and it just doesn't sound right whatever I do. Probably something obvious but it escapes me.
So here I am humbly asking for help from those that know so much more.
Thank you for taking the time to read this.
Paul
We'd be better able to help you if we knew what the difference in sound was like. Clearly, you've not measured the results. It could be as simple as reverse phase on the tweeters or a level difference....
Thank you for the reply.
Trouble is, my thinking tends to head straight to the datasheets. Try to get the design right from a technical perspective first... That's how I confused myself.
No, I admit, I haven't measured the results at all. Normally an amplifier / DAC person. The speakers got left behind as a result...
"Time to address that" is the thinking.
Double checked phase a few hours after the first listening and it is continuous (no reversals) from amp to driver.
Level difference, maybe. Currently, have it set to cut the tweeter by approx 1.5dB.
Now to try and describe the sound...
The difference between the tweeters is night and day. The AC1 was almost harsh looking back now. With this new tweeter the sound of cymbals / percussion just stand right out. Almost too much. So much more detail in the top end. So I started slowly turning down the tweeters, in steps over a period of hours, until the percussion sounded ok (to me). The overall sound was much softer and had lost definition in midrange. Imaging is there but not very good. Is it all purely psychological? Am I just conditioned to the sound of the old set up. I listen to it most evenings. (Got rid of the TV).
So I ended up in a cycle of setting the levels too high and too low. Could not find a sweet spot. That's when I had to question what I thought I knew.
Well, at this point, you don't really have much of a reference... I've not measured any of your components, and manufacturers datasheets may not be enough to design speakers with. LR4 crossovers will give you a very nice crossover null if the phase is reversed on the tweeter side. You might try that and see if your design simply needs the reversal. Short of that, measure......
Paul,
You mentioned the tweeters are protected by a 22uF cap. The cap will still affect the phase through the crossover point, as a result either polarity of the tweeters may result in a dip or a peak in the crossover region, as well as a different acoustic crossover (the additional phase shift from the caps), which affects vertical dispersion.
This makes setting relative levels (especially by ear) difficult, since the acoustic crossover will change more than the relatively small level changes would indicate.
Unless your HF amp is prone to DC, I'd suggest getting rid of the caps.
Art
I suspect that your next step will be to collect together a modest measurement setup.
A very brief perusal of the datasheets you posted a link to suggests that you might need more than 1.5db attenuation on the tweeter, but you really (really!) need to do some measurements at this point.
Some hardcore people use very expensive gear for measurement. I border on this myself.
But... To dip your toes in, and particularly with the setup of an active crossover fit 2khz xo, I would strongly recommend that you start with a very simple mic and borrow a channel of your power amp to make the following measurements. Look up "wallin jig" for how to do this.
I would suggest that you set up the speaker and mic, with the mic in a makeshift mic stand (imagine the mic taped to a broom handle if necessary).
You need to measure the woofer and tweeter without moving either the mic or speaker.
Measure the amplitude response (will come to phase later).
Get both plots on the same screen.
Things to look for are:
- that the actual response of the drivers in your box are behaved around your target xo point
- what the difference in sensitivity is, you should expect the tweeter to be several db more sensitive than the woofer
- make these measurements at a few cm, about a metre and maybe a bit more.
- watch out for floor reflections that will generate ripple and crud in the measurements.
- you are really after the specific data you need at the xo point. Try not to get distracted by the room artefacts. (More advanced measurements will require you to use gated measurements etc, but I am trying to get you up and running here)
When measuring you are particularly concerned "around" your xo point
You will see all sorts of ripple and crud in the dc- few hundred hz region. Use the close field and increasing distance measurements to validate that these measurements are valid.
Try the do set up with the parameters from above.
Now for delay/phase.
I did not see if you have a dsp or analogue xo. If analogue then:
- guess the tweeter phase, Make a frequency response measurement
- check it out
- flip the tweeter phase and repeat measurement
- compare to previously
One of these will likely have a dip. This is out of phase. Use the other connection.
- work from here and fiddle to your hearts content.
If you have a dsp, you really ought to measure the delay of the tweeter and woofer at the xo frequency. This is a bit fiddly, and needs a cro. What I do is:
- generate a three cycle tone burst at the crossover frequency followed by 200ms of silence.
- I do this in audacity, and run it a a loop
- trigger channel 1 of the cro from the amplifier output
- measure the mic output using channel 2 (refer above, but do it at listening position)
- measure the woofer response. Remember there is a long delay to the mic signal. Measure this delay accurately, I set a marker at this time position on a digital cro
- now swap to the tweeter. Repeat above
You are looking to measure the difference in time arrival of the second zero crossing of the tone burst as measured from the woofer and tweeter.
Program this delay into your crossover to correct this difference.
Repeat the above to check you got it all right. You should be able to twiddle these parameters in real time on most crossovers, it is cool to watch. Look for this time alignment to be within a few tens of microseconds.
Oh, and also check the phaseing of your drivers while you are at it!
My experience is that once you have the sensitivity (first part above) and time alignment as just described right, active crossovers "just work".
User warning: the time delay needs to be measured at the actual crossover frequency, this changes as you change frequencies.
The above is also really useful for working out how much to offset drivers on a sloped or stepped baffle🙂
Phil
A very brief perusal of the datasheets you posted a link to suggests that you might need more than 1.5db attenuation on the tweeter, but you really (really!) need to do some measurements at this point.
Some hardcore people use very expensive gear for measurement. I border on this myself.
But... To dip your toes in, and particularly with the setup of an active crossover fit 2khz xo, I would strongly recommend that you start with a very simple mic and borrow a channel of your power amp to make the following measurements. Look up "wallin jig" for how to do this.
I would suggest that you set up the speaker and mic, with the mic in a makeshift mic stand (imagine the mic taped to a broom handle if necessary).
You need to measure the woofer and tweeter without moving either the mic or speaker.
Measure the amplitude response (will come to phase later).
Get both plots on the same screen.
Things to look for are:
- that the actual response of the drivers in your box are behaved around your target xo point
- what the difference in sensitivity is, you should expect the tweeter to be several db more sensitive than the woofer
- make these measurements at a few cm, about a metre and maybe a bit more.
- watch out for floor reflections that will generate ripple and crud in the measurements.
- you are really after the specific data you need at the xo point. Try not to get distracted by the room artefacts. (More advanced measurements will require you to use gated measurements etc, but I am trying to get you up and running here)
When measuring you are particularly concerned "around" your xo point
You will see all sorts of ripple and crud in the dc- few hundred hz region. Use the close field and increasing distance measurements to validate that these measurements are valid.
Try the do set up with the parameters from above.
Now for delay/phase.
I did not see if you have a dsp or analogue xo. If analogue then:
- guess the tweeter phase, Make a frequency response measurement
- check it out
- flip the tweeter phase and repeat measurement
- compare to previously
One of these will likely have a dip. This is out of phase. Use the other connection.
- work from here and fiddle to your hearts content.
If you have a dsp, you really ought to measure the delay of the tweeter and woofer at the xo frequency. This is a bit fiddly, and needs a cro. What I do is:
- generate a three cycle tone burst at the crossover frequency followed by 200ms of silence.
- I do this in audacity, and run it a a loop
- trigger channel 1 of the cro from the amplifier output
- measure the mic output using channel 2 (refer above, but do it at listening position)
- measure the woofer response. Remember there is a long delay to the mic signal. Measure this delay accurately, I set a marker at this time position on a digital cro
- now swap to the tweeter. Repeat above
You are looking to measure the difference in time arrival of the second zero crossing of the tone burst as measured from the woofer and tweeter.
Program this delay into your crossover to correct this difference.
Repeat the above to check you got it all right. You should be able to twiddle these parameters in real time on most crossovers, it is cool to watch. Look for this time alignment to be within a few tens of microseconds.
Oh, and also check the phaseing of your drivers while you are at it!
My experience is that once you have the sensitivity (first part above) and time alignment as just described right, active crossovers "just work".
User warning: the time delay needs to be measured at the actual crossover frequency, this changes as you change frequencies.
The above is also really useful for working out how much to offset drivers on a sloped or stepped baffle🙂
Phil
Last edited:
Oh and the above can be done with the cheapest mic that works. You are making comparative measurements at the crossover frequency.
I would counsel you to start there, and once you decide you want to make "system" measurements, consider buying a measurement mic.
I would counsel you to start there, and once you decide you want to make "system" measurements, consider buying a measurement mic.
My two cents...
the protection cap before the tweeter should only limit DC when switching on the amplifier or whatever it may happen; it should be of a value of capacitance that it doesn't affect the sound, so about an octave below the crossover point.
A value of 180 uF might be ok, so about an order of magnitude above the actual size.
No particular discussion about the amplifiers and their gain/sensitivity
which are part of the system and moreover are directly attached to the drivers
the protection cap before the tweeter should only limit DC when switching on the amplifier or whatever it may happen; it should be of a value of capacitance that it doesn't affect the sound, so about an octave below the crossover point.
A value of 180 uF might be ok, so about an order of magnitude above the actual size.
No particular discussion about the amplifiers and their gain/sensitivity
which are part of the system and moreover are directly attached to the drivers
Thank you all for the replies! Now have a direction to go in. 🙂
jplesset,
I will try reversing the tweeter phase. It's a 30 second job. And then set myself up for doing some simple measurements next week.
Art (Weltersys),
Would like to keep the protection cap. Only because I like to design amps and have yet to design a protection system that I fully trust (solid state relay at some point). However, you may have a point about the vertical dispersion though. Even a small change in listening position height has a significant affect.
Phil (googlyone),
Thank you for the comprehensive reply! Given me a lot to think about.
Running purely analogue at present.
Your testing routine makes sense. It is good to read that some simple but ultimately useful tests can be carried out with very little expense. Will be getting a mic of some description and performing these tests.
Like to take a more scientific approach... Especially if it can be done at a price I can afford.
Now have a new avenue to research. 🙂
picowallspeaker,
180uF.... I use paper caps in this position. That's going to have quite a big physical size. Not a problem though. I'll just move them to the amp. Have to redo my calculations.
As for the amps... Currently there are 4 channels of Leach amplification running the speakers. Will be using my own designed amps at some point but at all times the HF and LF amps will be identical.
Again, thank you, all, for the very helpful replies. Really appreciate you all taking the time to help. Had a horrible feeling that I had bought a tweeter that wouldn't match to the woofer.
Need to get a mic.
Paul
jplesset,
I will try reversing the tweeter phase. It's a 30 second job. And then set myself up for doing some simple measurements next week.
Art (Weltersys),
Would like to keep the protection cap. Only because I like to design amps and have yet to design a protection system that I fully trust (solid state relay at some point). However, you may have a point about the vertical dispersion though. Even a small change in listening position height has a significant affect.
Phil (googlyone),
Thank you for the comprehensive reply! Given me a lot to think about.
Running purely analogue at present.
Your testing routine makes sense. It is good to read that some simple but ultimately useful tests can be carried out with very little expense. Will be getting a mic of some description and performing these tests.
Like to take a more scientific approach... Especially if it can be done at a price I can afford.
Now have a new avenue to research. 🙂
picowallspeaker,
180uF.... I use paper caps in this position. That's going to have quite a big physical size. Not a problem though. I'll just move them to the amp. Have to redo my calculations.
As for the amps... Currently there are 4 channels of Leach amplification running the speakers. Will be using my own designed amps at some point but at all times the HF and LF amps will be identical.
Again, thank you, all, for the very helpful replies. Really appreciate you all taking the time to help. Had a horrible feeling that I had bought a tweeter that wouldn't match to the woofer.
Need to get a mic.
Paul
- Status
- Not open for further replies.