Crossover Inductor Determination

[omni]

I am using Vance Dickasons formulas for a 3 way APC crossover. After plugging in the numbers into the formula using my measured impedances, I get an 8mH coil for the lowpass section. My question is this : The voice coil inductance for my woofer is 3.1mH. Should I subtract the 3.1 from the 8mH to get the value coil I use or should I simply use the 8mH ? This same question applies to my midbass filter as well. Any assistance will be appreciated........Respectfully........ Omni

[owdi]

I believe you simply use the 8mH coil. If one of your lowpass crossover points falls in an area where the impedance of the driver is already rising quickly, consider implementing an impedance equalization circuit.

Link to some useful calculators - http://www.mhsoft.nl/spk_calc.asp

Dan

[grantnsw]

Hi Omni,
I'm a beginner, but I found a very useful page here:

http://www.lalena.com/audio/calculator/apcxover/

(the audio section from the home page has lots of useful
stuff too)
I'm trying to do a 3way APC from the 'cookbook' as well,
and received some very helpful replies to my post
"bandpass formulas dilemma - a newbie Q"

hope this helps, Grant

[omni]

Grant,thanks for the link, I checked it out and its pretty cool.What I am trying to determine was not on the site but it still is helpful. Lets say you have a woofer with a voice coil inductance of 3 mH........ now, the calculation based on impedance renders an inductor of 8mH for the lowpass filter. Now that the woofer already has 3mH inductance, would this make the inductor for the crossover to be 5mH thereby adding up the Voice coil 3 mH inductance to 5 mH to equal 8 ? I perhaps hope some of the more learned audiophiles on this site might be able to share their experience. Also, I am hoping to find out what method to calculate a zobel impedance compensation network works the best, as there are a variety of theories, any input out there, fellow speakerbuilders ? Respectfully..........Omni

[Svante]

I am not familiar with the formulas that you use, but assuming that the filter is first order you should in principle subtract the inductance of the woofer. There is a problem with this, however, due to the nature of the voice coil inductance: it is frequency dependant, larger at low frequencies and smaller at high frequencies. The number stated in the datasheet is only valid at a single frequency (and it is often not stated which frequency this is).

I suggest that you try to simulate the entire system instead of blindly relying on formulas for the crossover. Crossover design is tricky and any attempt to reduce it to a couple of formulas is bound to produced results that were not intended.

[omni]

Svante :Thank you for your post. I always appreciate feedback, as I am certainly NOT extremely experienced. At any rate, the slope will be 12bB per octave. Here are my thoughts on creating my crossover; and maybe you can help ? First, I will be doing impedance measurements with an old tube frequency generator. I have done some preliminary measurements with voltage divider method. I am thinking I may have to calibrate after each measurement to maintain proper voltage readings at each frequency. Does this seem reasonable? I also thought of experimenting with Speaker Workshop......Is this an accurate measurement system ? I have limited finances.........Then put Zobels on the midbass, and tweeter, and re measure impedances, mounted in cabinets. Crossover frequencies of 300, and 3000Hz. 12 dB per octave using APC formulas in Vance Dickasons book. Then tweak to get the sound I am looking for. As I stated, I do not have a lot of money to afford other measurement systems, would rather spend it on decent crossover components. Peerless 830669 SLS 12 inch woofer, SEAS CA15RLY midbass, SEAS 27TDFC tweeter, All good performers with reasonably good characteristics. Cabinet is a pyramid shaped 4 cubic feet sealed for woofer, and about 15 liters, sealed for midbass. 7 degree slanted baffle, all made of 1 inch particle board. I made an earlier post pertaining to this project and TINITUS and a few others convinced me my earlier bass cabinet size was too small, so I decided to start from scratch, and build a larger cabinet. Thank You TINITUS, Johnathon Brite, 5th Element, and others for your insights. At any rate my father passed away, and the project was shelved for a while, hence me not pursuing more info on this forum, but I have returned to gain more insights if possible. Clarity Caps bypassed with Mundorf Supreme silver/oil, 14 guage inductors throughout all of crossover except in low end of midbass in the bandpass, and woofer section. Mills resistors for Zobels and tweeter attenuation, when that time comes. This project has been very time consuming,with several cost overruns, but fun, to say the least, and very enjoyable communicating with you all,on this forum.......At any rate ANYONE else who cares to chime in on this thread, please feel free, I can use the help. This is like Dorothys' house, back in Kansas, in the Wizard of Oz: everything starting out in Black and White, and then Dorothy opens the door of the house after the whirlwind tornado,to experience a whole new world of COLOR. I look forward to that experience when I plug these speakers in and turn up the volume..............But it's still got a little bit of work to be done..........Omni

[sreten]

Originally posted by Svante
I am not familiar with the formulas that you use, but assuming that the filter
is first order you should in principle subtract the inductance of the woofer........

Hi,

In principle for such simple "cookbook" crossovers this is not true.
In principle you should zobel the driver to be resistive and use
the "cookbook" value for the inductor.

All too trivial though, basically wrong, it may work, it may not,
actually understanding what is going on and the resultant
acoustic slopes including fundamental resonance, driver response,
baffle step and diffraction effects is far more important.

/sreten.

[tinitus]

Its better to start with a too big inductor - you can always take some off - when will you show what you have done so far

8mH seems fine

[omni]

Sreten, I agree with your post, and hopefully I have addressed the issues you have mentioned. I am figuring on a 4th order acoustic slope with the combination 2nd order crossover in my sealed cabinet. Maybe maybe not. At any rate I have treated the baffle edges with a 3/4 radius as per Zaphs website. On the baffle step issue, I am not too concerned. I may have the best of both worlds here... I have made the width of the cabinet a bit wider at the bottom of the pyramid to provide more reinforcement to the bass. It appears these days the designs of narrower cabinets, especially in the woofer section, has brought about the baffle step issue. I have decided to make mine a bit wider. Also, any baffle step I do get may be somewhat of a blessing, as my midbass sensitivity is a bit lower than the woofer. After Tinitus voiced his concern about my cabinet size, I decided to shift gears into a larger design: many hours of drawing prototypes, calculating volumes, to come up with my final design, which I think will be worth it, both in the bass as well as the midbass cabinet. See the specs in my last post. Thank you Tinitus. By the way Tinitus, its funny you mentioned about starting off with a larger coil with the option of removing some wire, because just before I logged on here, today, I was contemplating my future tweaking routine,and that thought came to mind. My cabinets are built, however I have not put the laminate on them yet, also, I will need to put one more brace in them. Tinitus, I will take some photos in the next day or so, and post them here. I know someone mentioned, I think it was you:that a pyramid could get ugly. You are right. You should see some of my preliminary drawings. Alas........the final design worked out beautifully, albeit a bit taller than my original concept, but very pleasing to the eye, as well as acoustically sound. Look for the pictures in the next few days..............Respectfully....Omni

[Svante]

Originally posted by sreten


Hi,

In principle for such simple "cookbook" crossovers this is not true.
In principle you should zobel the driver to be resistive and use
the "cookbook" value for the inductor.

All too trivial though, basically wrong, it may work, it may not,
actually understanding what is going on and the resultant
acoustic slopes including fundamental resonance, driver response,
baffle step and diffraction effects is far more important.

/sreten.

Yes, I agree, that's why the "in principle" was there in my post, and the recommendation at the end of the post. It is hard to know how to handle questions like these, it appears as if many people look for equations for how to design crossover filters, and indeed there are quite a few of them around on the web. My opinion is that simple equations never results in the response that you really want, and that you need to either simulate the whole loudspeaker system or do measurements on an actual system and modify the component values until the response is as desired. The impedance and the response of the drivers and also the baffle step all put non-cookbook requirements on the filter.

However, such a response to a person who just has found a bunch of equations on the web probably leads just nowhere. So I tried to answer the question with an "in principle" as a reservation that the line of thinking is at best as good as using a cookbook filter design in the first place. I am not sure whether this is the right thing to do, but that was my thinking behind it, anyway.

Now, as it turns out, the filter was second order, and that makes my line of reasoning not even in principle right, and the conjugate link combined with the full inductance seems more appropriate, but I would still like to recommend the full simulation of the system to include the response of the drivers and the baffle step etc.