OK I think we are getting somewhere. I was using a crossover of 300Hz but it is probably a bit low for the Fostex. Lets take it up to about 400Hz. Try 66uF for the main cap and 2.1mH for the inductor.
Also try 88dB for the FF85wk, from the data sheets I know it says 86.5 dB but looking at the 1 to 2kHz range it wiggles around and touches 90dB in places so 88 dB may be a good average in that region.
Also try 88dB for the FF85wk, from the data sheets I know it says 86.5 dB but looking at the 1 to 2kHz range it wiggles around and touches 90dB in places so 88 dB may be a good average in that region.
OK don't worry about changing the driver parameters we'll go with what's there.
So far we've added a parallel resistor to the Fostex and you can see that we have reduced the peak at about 115Hz, (blue trace) the peak is due to the rising impedance at resonance.
We've added in a Zobel to the woofer which makes sure the low pass roll off continues up into higher frequencies, (red trace). It counteracts the rising impedance of the woofer.
The blue trace still has a shelf to it, moving to 400Hz hasn't reduced it much. We can reduce the capacitor's parallel resistor from 27R down to 8R and increase the capacitor from 68uF to 100uF.
The mid range has a mild hump from 1kHz to 2kHz this might be due to not a strong enough Zobel. Try reducing the 8R resistor to 7.5R.
So far we've added a parallel resistor to the Fostex and you can see that we have reduced the peak at about 115Hz, (blue trace) the peak is due to the rising impedance at resonance.
We've added in a Zobel to the woofer which makes sure the low pass roll off continues up into higher frequencies, (red trace). It counteracts the rising impedance of the woofer.
The blue trace still has a shelf to it, moving to 400Hz hasn't reduced it much. We can reduce the capacitor's parallel resistor from 27R down to 8R and increase the capacitor from 68uF to 100uF.
The mid range has a mild hump from 1kHz to 2kHz this might be due to not a strong enough Zobel. Try reducing the 8R resistor to 7.5R.
So what do we have so far. We have a bass sensitivity near 90dB (100Hz) and it gradually falls to 86dB at higher frequencies. This gives us some baffle step compensation (BSC). At higher frequencies the sound is projected forward from the speaker front baffle. At lower frequencies where the wavelength becomes larger than the width of the baffle the sound starts to become omni-directional. This is why some sub-woofers are down firing, they don't need to face the listner.
The maximum bass loss at lower frequencies is -6dB. However at frequencies below about 100Hz room gain starts to take effect. Also placement of the speakers will also have an effect. Small stand mount speakers away from walls will need more BSC than floor standing speakers that may be nearer walls.
The actual amount of BSC needed will vary and therefore we may need only about 3-4dB. We can add more BSC by attenuating the mid/tweeter (FF85WK) with a series resistor. You would adjust the amount of attenuation once you have built the speaker and listened to it.
The maximum bass loss at lower frequencies is -6dB. However at frequencies below about 100Hz room gain starts to take effect. Also placement of the speakers will also have an effect. Small stand mount speakers away from walls will need more BSC than floor standing speakers that may be nearer walls.
The actual amount of BSC needed will vary and therefore we may need only about 3-4dB. We can add more BSC by attenuating the mid/tweeter (FF85WK) with a series resistor. You would adjust the amount of attenuation once you have built the speaker and listened to it.
Thank you very much for the XO and BSC design, adjustments, and explanations/teaching.
I started reading about BSC and I think I understand the basics of it, but it would have taken me ages to get to the point where I could design and tune it, and same goes with the XO.
I did the last adjustments you suggested (attached).
I started reading about BSC and I think I understand the basics of it, but it would have taken me ages to get to the point where I could design and tune it, and same goes with the XO.
I did the last adjustments you suggested (attached).
Attachments
Cool that's looking better. One thing to keep an eye on is the overall impedance. So far we are keeping at or above 4 ohms. Anything lower and amps could start to struggle.
I'm not sure what's causing the slight hump at 1kHz. Normally we wouldn't worry about it as its only about 1dB. However looking at the FF85WK graphs there is already a raised sensitivity between about 700Hz and 2Khz of about 3dB and we don't want to add to it.
It may be that you just need to tweak various values until it looks nicer.
We are using the crossover frequency to obtain our baffle step.
My speakers are about 32cm wide and the baffle step would start at about 350Hz. However my crossover is set at about 270Hz. This because the baffle step response is still falling above this. For example if you look at the box in your sofware marked filter you can see that the two lines cross at about 400Hz.
In the box marked SPL (sound pressure level), the red and blue line cross higher at about 700Hz (acoustic crossover), this is because the woofer is more sensitive. So normally we would want a lower crossover, however in this case looking at the FF85WK again we can see that sensitivity drops below 700Hz. So our higher baffle step frequency might balance that out.
400Hz is about a 28cm wide baffle.
I'm not sure what's causing the slight hump at 1kHz. Normally we wouldn't worry about it as its only about 1dB. However looking at the FF85WK graphs there is already a raised sensitivity between about 700Hz and 2Khz of about 3dB and we don't want to add to it.
It may be that you just need to tweak various values until it looks nicer.
We are using the crossover frequency to obtain our baffle step.
My speakers are about 32cm wide and the baffle step would start at about 350Hz. However my crossover is set at about 270Hz. This because the baffle step response is still falling above this. For example if you look at the box in your sofware marked filter you can see that the two lines cross at about 400Hz.
In the box marked SPL (sound pressure level), the red and blue line cross higher at about 700Hz (acoustic crossover), this is because the woofer is more sensitive. So normally we would want a lower crossover, however in this case looking at the FF85WK again we can see that sensitivity drops below 700Hz. So our higher baffle step frequency might balance that out.
400Hz is about a 28cm wide baffle.
Last edited:
All right. About amp, all I have is an old SONY from a semi integrated hifi, rated at 45W 6ohms @1kHz (according to the specs attached, mine is ex50), I have no idea how it will behave with 4 ohms.
The 1kHz hump I think is because of the slope of the low pass. I tried reducing it by increasing the zobel capacitor from 15µF to 20, and reducing the woofer inductor from 2.1mH to 1.8 (attached). It does not disappear though (the curve only raises on both sides), and I don't know whether these values are realistic.
May I ask what the FR is on your speakers ? These are crazy good looking speakers btw, I wish I could do something that good one day...
More about your system, are both drivers in separate sealed enclosures ? I expect the SPH-8M yield about the same LF response as the SPH-220HQ should, a ~55Hz f3 and 35Hz f10 ?
If that is a sub to the left of the picture, does that mean you felt the LF output of the SPH-8M was not enough ?
The 1kHz hump I think is because of the slope of the low pass. I tried reducing it by increasing the zobel capacitor from 15µF to 20, and reducing the woofer inductor from 2.1mH to 1.8 (attached). It does not disappear though (the curve only raises on both sides), and I don't know whether these values are realistic.
May I ask what the FR is on your speakers ? These are crazy good looking speakers btw, I wish I could do something that good one day...
More about your system, are both drivers in separate sealed enclosures ? I expect the SPH-8M yield about the same LF response as the SPH-220HQ should, a ~55Hz f3 and 35Hz f10 ?
If that is a sub to the left of the picture, does that mean you felt the LF output of the SPH-8M was not enough ?
Attachments
Should be fine with 4 ohms. Most transistor amps will cope no problem.
My woofers are in 46 litre bass reflex enclosures. Although at the moment I have stuffed the ports so its more like a sealed enclosure with a Q of around 0.5 to 0.6. I rarely use the sub.
The top box is sealed 'mid transmission line' designed to reduce the impedance peak at resonance a bit. Its not a permanent fixture and can be changed. It contains a Mark Audio Pluvia 7 which has an impedance peak at around 68Hz and X max of 4mm, so it can handle a bit of bass itself.
The sides of the speakers are Ikea kitchen wall unit end panels in gloss grey. They weren't cheap but worth it for a good finish. The black is a cheap leatherette material from ebay glued onto MDF. So no painting or varnishing.
My woofers are in 46 litre bass reflex enclosures. Although at the moment I have stuffed the ports so its more like a sealed enclosure with a Q of around 0.5 to 0.6. I rarely use the sub.
The top box is sealed 'mid transmission line' designed to reduce the impedance peak at resonance a bit. Its not a permanent fixture and can be changed. It contains a Mark Audio Pluvia 7 which has an impedance peak at around 68Hz and X max of 4mm, so it can handle a bit of bass itself.
The sides of the speakers are Ikea kitchen wall unit end panels in gloss grey. They weren't cheap but worth it for a good finish. The black is a cheap leatherette material from ebay glued onto MDF. So no painting or varnishing.
Attached
Thank you for the suggestion though I'm afraid I don't understand what it means. What are those figures 400 and 2000 and how do they relate to a kneejerk response ?GM said:
Attachments
Urban Dictionary: kneejerk
The sag is ~ in the 400-2000 Hz bandwidth [BW] and since sound progresses at 1/f [exponentially] the mean is the square root [sqrt] between these two points, so shifting the XO point higher should in theory flatten out the overall response.
GM
The sag is ~ in the 400-2000 Hz bandwidth [BW] and since sound progresses at 1/f [exponentially] the mean is the square root [sqrt] between these two points, so shifting the XO point higher should in theory flatten out the overall response.
GM
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.