Hi,
I have been trying to design a small 8-inch subwoofer in a bandpass case. The amplifier will be a single 3886, but I need help with the crossover. I am going to use the LM837 chip. I need to be able to set it up to use a 25K variable resistor to adjust the cutoff freq. from 50hz to 200hz.
How can I do this?
Thanks, Mike
I have been trying to design a small 8-inch subwoofer in a bandpass case. The amplifier will be a single 3886, but I need help with the crossover. I am going to use the LM837 chip. I need to be able to set it up to use a 25K variable resistor to adjust the cutoff freq. from 50hz to 200hz.
How can I do this?
Thanks, Mike
You can only do this if the pot is dual-ganged. That means two pots controlled by a single shaft.
To design your filter, try out Analog's Interactive Filter Synthesis tool. If the URL doesn't work, go to www.analog.com and navigate to Technical Support -> Interactive Design Tools -> Active Filter Synthesis. There are countless other filter design tools on the web. If you are googling for them, use the keywords "filter" and "low-pass", instead of "crossover".
Since you want a 4:1 ratio for the maximum to minimum cutoff frequencies, you need a 4:1 ratio for the filter's resistors. Use a 8.2 kOhm resistor in series with the pot, so that the combination of the two varies between 8.2 kOhm and 33.2 kOhm as you adjust the pot from one extreme to the other. Remember, both R's in the filter need to be the same value and need to change at the same time, so you need a dual-ganged pot. Now you just need to figure out what the C's are to get 50 to 200 Hz as R varies from 8.2 k to 33.2 k.
Mike,
In Analog's design tools, from top to bottom, choose the following:
Filter type: Lowpass, Butterworth, order=2.
Stage 1: Fo=200 Hz (don't change Q). Choose the Sallen-Key LP design.
Circuit: Gain must be 1 in order for R1=R2 (R1 and R2 is the ganged pot). Change C1 until R1 is close to 8.2 K. Now when you change F0 to 50 Hz, and change C1 to the same value, R1 should come out to be around 33 K, just like I mentioned in the above post.
I found that C1=133 nF is close. You may not be able to find a capacitor this exact size, but you could use a 100 nF and a 33 nF in parallel. C2 comes out to be 66 nF, which you can make from two 33 nF in parallel.
Note that R3 is arbitrary; the op-amp is configured as a voltage follower for gain =1 (R4 is infinite, i.e. non-existant) so R3 can be a piece of wire.
As always, use proper capacitor types. Mylar, Polyethlyne, Polyester, etc., but not ceramic.
In Analog's design tools, from top to bottom, choose the following:
Filter type: Lowpass, Butterworth, order=2.
Stage 1: Fo=200 Hz (don't change Q). Choose the Sallen-Key LP design.
Circuit: Gain must be 1 in order for R1=R2 (R1 and R2 is the ganged pot). Change C1 until R1 is close to 8.2 K. Now when you change F0 to 50 Hz, and change C1 to the same value, R1 should come out to be around 33 K, just like I mentioned in the above post.
I found that C1=133 nF is close. You may not be able to find a capacitor this exact size, but you could use a 100 nF and a 33 nF in parallel. C2 comes out to be 66 nF, which you can make from two 33 nF in parallel.
Note that R3 is arbitrary; the op-amp is configured as a voltage follower for gain =1 (R4 is infinite, i.e. non-existant) so R3 can be a piece of wire.
As always, use proper capacitor types. Mylar, Polyethlyne, Polyester, etc., but not ceramic.
Did you look at the filter design tool? If you did, you should notice that when you change the frequency, then two resistances change. Or two capacitors change. This is because the filter is a 2 pole (or "2nd order" or "12 dB/octave") filter. Each "pole" is made from one R and one C. You need to adjust both at the same time to adjust the overall filter. So you need the dual pot. Seriously. You just do.
Ceramic caps are very non-linear. Non-linear means distortion. Distortion sounds bad. Ceramics have their uses, though. They work great as power supply decoupling caps or RF snubbers, because of their very low parasitic inductance. But they should never be put in the signal path in an audio circuit.
Mike, you're a lucky guy, because that's what I've been doing the last few days.
It's done, and it works very well.
In my case, I have two OPA549s in parallel for the sub amp.
The filter is made of one OPA2132 as input buffer / bypass and a quad OPA4228 op-amp for summing channels, filter and phase inversion.
I post this document here, a bit crude, but these are my personal notes before I made it.
Please note that the numbers in bold are the order of the op-amps, from input to output.
It's done, and it works very well.
In my case, I have two OPA549s in parallel for the sub amp.
The filter is made of one OPA2132 as input buffer / bypass and a quad OPA4228 op-amp for summing channels, filter and phase inversion.
I post this document here, a bit crude, but these are my personal notes before I made it.
Please note that the numbers in bold are the order of the op-amps, from input to output.
soundNERD said:Thanks everybody!
carlosfm, a stupid question: where does the dual pot connect? I see all standard resistors in the schematics. Sorry for how stupid this question is
The dual pot is in place of the input resistors.
Don't you see two resistors bebore the NI input?
If you make that variable (with a pot), you have a variable filter.
soundNERD said:
I have no idea.
Maby someone can help with the maths.
BTW the OPA549s sounds so good (without the filter, that is) that it seams like a crime to use them on a sub.
When I connect the filter, I loose the midband and treble.
Well, since this isn't specifically an audio amplifier, I am not surprised that power output ratings for speaker loads (4 ohm, 8 ohm) are not given in the datasheet. However, if you look at the Total Harmonic Distortion graph, you will see several sets of lines, for different output power levels, and a note that the load is 4 ohms. One of the lines is for 75 W output. Note how the distortion rises sharply with frequency, even at low power output levels.
Actually, it is quite normal to see and increase in THD with frequency.
But check out the 3886. It's THD+N vs. Frequency curve is plotted with power out = 60 W into 4 ohms. At 20 kHz, the number is about 0.025%. For the 3875, with 40 W into 8 ohms, it is about 0.07%. And with the OPA549, at 75 W (or at 10 W) into 4 ohms it is about 0.2%. So what concerns me isn't that the THD increases with frequency, but how much it increases. The distortion in the 3886 increases by a factor of 10 from low frequencies to high frequencies. The distortion in the OPA549 increases by a factor of 100 from low frequencies to high frequencies.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.