| qguy |
cross posted in loudspeakers ...
WARNING ....not an electronics expert.. not even an electronics amateur...
but I just built an electronic subwoofer crossover from a kit, I simply soldered and prayed a lot It has a volume gain pot but it has no cut-off frequency pot, it has a DPDT that switches from 50 or 70 hz. Is the default 50 or 70 hz setting based on the values of the resistors and capacitors ? Does the two IC thats built into the kit play a role in the cut-off frequency. My goal is to change these to 30 and 40 hz, as I can still localize the position of the sub on the 50 hz setting.
 |
|
|
| pinkmouse |
The actual filter is the section in the middle, it's a new topology on me, but at a quick glance it looks like it is only a first order crossover as I can only see one pole. Of course, I reserve the right to be completely wrong. :)
This means you only have a 6dB slope, so no wonder you can still localise the sub. Changing caps or resistors won't help much I'm afraid, you need a much steeper slope. Where did that kit come from? Do you have any references? |
|
|
| qguy |
| its from a local electronics store.. unfortunately, its the only one available...how about daisy chaining two of these ? |
|
|
| Drew Eckhardt |
| quote: | Originally posted by pinkmouse
The actual filter is the section in the middle, it's a new topology on me, but at a quick glance it looks like it is only a first order crossover as I can only see one pole. Of course, I reserve the right to be completely wrong. :)
This means you only have a 6dB slope, so no wonder you can still localise the sub. Changing caps or resistors won't help much I'm afraid, you need a much steeper slope. Where did that kit come from? Do you have any references? |
IC1B is an inverting buffer, gain = 16.8. You'll need to turn down your sub-woofer volume 24.5dB compared to what you had before you tried to use the low-pass filter. Replacing the 500 and 510K resistors with a single 15K part will get you back to unity gain.
IC2D is a second order Sallen Key low-pass filter with unity gain.
With SW open, F0 = 50Hz and Q0 = .73
With SW closed, F0 = 70Hz and Q0 = .69
A third pole is added at 153 Hz before IC2C.
IC2C is a unity gain non-inverting buffer. You should have a resistor (try 196 Ohms) in series with its output to limit short-circuit current and avoid oscillation problems driving capacitive loads.
With calibrated output level (Measure with an SPL metter. Try pink noise a couple octaves wide in the sub's pass band and main speakers') no port noise, no rattles, and low distortion the sub should not be localizable.
Integration with the main speakers may be a problem.
http://en.wikipedia.org/wiki/Sallen_Key_filter
http://www.linkwitzlab.com/filters.htm#2 |
|
|
| qguy |
Hi Drew, thanks for the inputs.. to summarize
1. Replace the 500 and 510 K resistors with a 15K part
2. add a 196K ohm resistor in series with the ic2s output' , that would be before the 2.2K ohm resistor right ?
the new circuit would look like this right ?
After doing all these what would the effect on the sub be ? did I just lower the crossover point ?
thanks] |
|
|
| F1 FAN |
| The changes Drew suggested only change the gain to unity ,the crossover points won't change.The resistor he mentioned is actually 196ohms not 196K ohms and it should be at the output of IC2C pin14 not before the 2k2 resistor.The crossover actually has an 18db per octave slope.If you want to lower the crossover frequency you can do this by changing 3 resistors with larger values ,or by increasing the value of the .47uf and .22uf caps either by replacing them or paralleling new ones with them. |
|
|
| qguy |
| quote: | Originally posted by F1 FAN
The changes Drew suggested only change the gain to unity ,
Gain to unity ...hmmm what could this mean ? help !!!
you can do this by changing 3 resistors with larger values ,or by increasing the value of the .47uf and .22uf caps either by replacing them or paralleling new ones with them.
lets say target is 30 hz...should I change them to 20 ohms instead ? is there a simple formula that an idiot could use - thanks
|
|
|
|
| F1 FAN |
Unity gain means that the signal into the unit will not be boosted by the circuit at all.With the original resistors(500k,510k) it would boost the signal 16.8 db which is way too much for a sub crossover.
If you were to double the value of either the resistors or the caps in the filter section you would lower the crossover to exactly half the frequency.For instance doubling the two .47uf's and the .22uf you would then have a crossover point at 25hz (instead of 50hz) and 35hz( instead of 70hz.) |
|
|
| qguy |
| quote: | Originally posted by F1 FAN
Unity gain means that the signal into the unit will not be boosted by the circuit at all.With the original resistors(500k,510k) it would boost the signal 16.8 db which is way too much for a sub crossover.
If you were to double the value of either the resistors or the caps in the filter section you would lower the crossover to exactly half the frequency.For instance doubling the two .47uf's and the .22uf you would then have a crossover point at 25hz (instead of 50hz) and 35hz( instead of 70hz.) |
THANKS !!! |
|
|
| qguy |
| one more question... what would be a better way of lowering the crossover point.. would be via the resistor or via the capacitor ? |
|
|
| F1 FAN |
| Either way is fine as the results will be the same.It is a matter of what parts are easier to obtain and change. |
|
|
| pinkmouse |
| quote: | Originally posted by Drew Eckhardt
IC2D is a second order Sallen Key low-pass filter with unity gain. |
Of course it is. Doh!
Note to self: don't post late at night when you should be in bed.
qguy, this might help:
http://www.daycounter.com/Filters/S...alculator.phtml |
|
|
| AndrewT |
Hi Pink,
you're forgiven,
I too burnt the wick at too many ends over Christmas.
The last C could be increased from 0.47uF to 1uF and lower that final filter to about 72Hz & Q=0.71
The inverting input may be useful but be prepared to swap your sub input terminals to correct for the filter inversion.
If you want to preserve the Q of the middle filter then both 10k resistors need to be reduced by the switch (DPDT) or by adding extra parallel resistors. The range 5k to 10k is just fine.
If you want Butterworth for the middle filter then the top C =2 times lower C. Use a pair on top and single at lower location.
The final Q = Qmiddle * Qthird.
ie. Q= 0.69 * 0.71 = 0.9 or 0.73 * 0.71 = 0.52 (both close to Bessel or half of Linkwitz Reilly). |
|
|
| qguy |
| quote: | Originally posted by AndrewT
If you want to preserve the Q of the middle filter then both 10k resistors need to be reduced by the switch (DPDT) or by adding extra parallel resistors. The range 5k to 10k is just fine.
Hi Andrew,
When FI Fan said double the reisistance to lower the crossover point this means 10K to 5K and not 10K to 20k right
The two resistor that you mentioned are the one in red circle right ?
If you want Butterworth for the middle filter then the top C =2 times lower C. Use a pair on top and single at lower location.
The final Q = Qmiddle * Qthird.
ie. Q= 0.69 * 0.71 = 0.9 or 0.73 * 0.71 = 0.52 (both close to Bessel or half of Linkwitz Reilly).
What advantage would this changes make ?
|
|
|
|
| AndrewT |
Hi,
to set the frequency one needs to look at R*C. If one increases the combination one ends up with a lower frequency. So, you can increase R and/or increase C.
20k will increase the noise the circuit produces and at this level in a sub-base it will probably be tolerable, but it depends on the following gain and sensitivity.
I recommend you stay within the ranges below 12k. That's why I said 5k/10k before.
The red encircled resistors are the original (top) setting 50Hz and the extra (bottom) sets to the lower resistance and raises the frequency to 70Hz.
The other half of the pair is the next one to the left. You should change both of these to preserve the roll-off characteristic that you intended. The pair need to be adjusted to change frequency properly and the method being adopted is a cheap way of doing it.
I have now listened to Linkwitz Reilly and to me it sounds like there is a hole in the crossover where something is lacking (I had 150Hz 4pole). Many recommend L-R (Q=0.5) over Butterworth (Q=1/sqrt(2)=0.7071). This choice or anywhere in between will affect the sound. Once you have found the one that sounds right you want to keep that and fiddling with just one half of the pair of frequency setting resistors will make assessing/choosing more difficult. You should also be adjusting the passive filter to match but that is becoming even more complicated.
It just ocurred to me that Q=0.5 is not half of L-R, since L-R can only be achieved with an even number of poles i.e. 2pole or 4pole etc. 3pole(18db/octave) is just a Bessel characteristic. |
|
|
| qguy |
| tried the 20K ohm resistor prior to reading your post, the sound was distorted, will try 5-9K ohm resistor and post the result - thanks |
|
|
|
