Crossover Characteristics

jakelm

Im trying to figure some out since my TrueRTA does NOT work!! (darnnett)

In my crossover, from the positive intput, I have a 12uf cap and a 4.5ohms resistor in series with each other, going to the tweeter.

I recently replaced the 2.7ohm resistor that was originally there. I noticed when I changed the 2.7 out with the 4.5, the tweeter toned down alot (to my liking).

Did the tweeter's intire dB drop or was only certain frequencies cut, because of the higher resistor value? The cap was left the same.

If I were to change the value of the cap and leave the resistor the same, what is going on?

I'm asking if anyone knows what the effects would look like on paper when you change a cap or resistor? I know what my ears hear, but not quite sure what "electronicaly" is going on.

Jake

jakelm

ahhaa....Resistors take care of the sensitivity/load. And caps do the filtering.

Now I understand.


Now to figure out , when using a 4ohm driver (weather a woofer or a tweeter), where the slope changes with different caps and inductors and how many db's drop with reistors changing in 1ohm intervals. And how the overall ohm changes for the entire speaker with a given resistor change, cap or inductor.

jcaut

I think that's a bit of a tricky question to answer, without having the impedance data from the tweeter on hand. In general, however, changing the value of the series resistor should affect the output above the crossover frequency in a relatively uniform manner. Going from a 2.7 Ohm resistor to the 4.5 should result in about a 1.5db attenuation of the tweeter's output. You have to be careful, though because going too high on the series resistor could begin to create a peak toward the low end of the tweeter's range. Changing the capacitor would change the point below which the tweeter's output would be attenuated (crossover frequency). Lower values on the cap would move the crossover frequency up, and higher values would move it down. There are also issues with phase relationships that are even more difficult to deal with, without acutal measurement data.

What speaker is this crossover in? Is it actually just a cap and a resistor, or are there other components?

Jason

jakelm

Thanks Jason. Its accually a M10 crossover. I built a center channel, using 2 Polk's mw6503's and a peerless tweeter. The peerless to my knowledge is 4ohms, the mw6503 are 3.7ohms. By taking drivers that are supposed to on in a larger cabinet and putting them in a 1cuft cabinet, the characteristics of the drivers changed.

As you can see in the diagram, the original resistor for the tweeter was 2.5ohms, which I changed to 4.5ohms and removed the 12uf cap that was in parallel with the resistor, due to the fact that, being in the new cabinet, the db's of the tweeter were much higher. I was also thinking of changing the crossover points to accomedate the new box and the fact that I am now using these drivers to make up a center channel. I also increased the reistor value to come closer "tember wise" to my 7b's (which in thier crossover they have 4.5ohm resistors). So by changing the reisitor, I was wondering what else was affected and what would be affected by changing the caps and inductors. I'm really just starting to learn about crossovers. I found some great literature at this website here....

http://www.lalena.com/Audio/FAQ/XOver/

Here is the schematic of the crossover I am using as a learning project.

jcaut

Here's a visual representation, using an 8 Ohm Audax tweeter.



The top line is the response of the tweeter with a 12uF cap and no resistor. The lines below represent what happens when a series resistor is added and increased in value ( I think I went up to about 8 Ohms).

The next picture represents the response of the same tweeter as the value of the capacitor is changed. I used a series 4.5 Ohm resistor for this calculation. The top line is the response of the tweeter with no capacitor (just the series resistor) and then the subsequent lines represent values of capacitors from about 15uF (top) down to 4uF.

Obviously different tweeters will produce different results, but I think you'll get the general idea of what's going on.


Jason

jakelm

Yes Jason I do get the idea. So by changing a single resistor, everything is changed..db and crossover point. And depending on the inductors, if you change the resistor too much everything can be scewed.

I will go back, I have a 3.5ohm Mills (I have replaced all my resistors with Mills) resistor, I wil replace the 4.5 with the 3.5 and see what happens.

And with caps. The lower the value of the cap the higher the crossover slope is.

Is inductors the same way? If I found that I had too much midrange, would I raise or lower the inductor? In my crossover diagram you can see that I do have a small inductor (.4mH) that can be replaced to tweek the woofer crossover.

jcaut

Okay. What you're working on is not exactly what I am describing. The 2.5 Ohm resistor and the 12uF capacitor are in parallel with each other, and will alter *mostly* the low end response of the tweeter, but I can't say exactly what changing the values will do. It's kind of like a padding resistor that changes value as the frequency changes. At lower frequencies, the impedance of the capacitor approaches infinity, so the resistor value is dominant. As the frequency goes up, the impedance of the capacitor approaches zero, so the resistor has less and less of an effect. I'll look into it a bit more a little later when I have more time.

Jason

jakelm

Ok Jason I do appreciate your input.

It is a 2nd order crossover, right? 2 caps and 2 inductors.

I did remove the 12uf cap that was in paralell with the resistor. Any idea what that changed slope or frequency wise?

Im trying to get a better balance between the 2 mw6503's and the peerless tweeter. But I'm having problems getting it just right. I havent messed with the caps (other than removing that one). The only reason I did remove that one, was because the crossovers in my 7b's did not have one for the tweeter and I thought if I made the 2 crossovers for the tweeter identical I would get a better tember match, but like I said, I just cant get right. My 7b's have the exact same drivers in them
(peerless tweeter and Polk mw6500 [which is almost identical to the mw6503]), except that the 7b's have only 1 mw6503 instead of 2.

jakelm

This is a quote from that website...

A series-notch filter will be used to remove this spike.

Could this be the 12uf cap, I removed that was in paralell with the resistor?

Edit..never mind...a series notch filter is used for the midbass in my crossover no the tweeter.

Systems

The 12uf cap 2.5 ohm resistor combo will boost output at the top end of the tweeters response.

jakelm

GV#27 wrote: »

The 12uf cap 2.5 ohm resistor combo will boost output at the top end of the tweeters response.

So if the tweeter is -12dbs @ >16khz. This configuration will accually boost the 16khz and higher?

So by taking out the 12uf cap , I actually lowered the Fr responce of the speaker?

Will I still get a boost if I leave the 12uf cap and raise the 2.5ohm to 4.5ohm?

Systems

In the tweeters lower range the resistor will attenuate its output as normally adding resistance would.At some frequency the 12uf cap starts to short or (bypass)the resistor so that it is no longer reducing its level.So the crossover designer was trying to compensate for a tweeter with a rolled off top end.
Does that help.

Systems

jakelm wrote: »

So if the tweeter is -12dbs @ >16khz. This configuration will accually boost the 16khz and higher?

So by taking out the 12uf cap , I actually lowered the Fr responce of the speaker?

Will I still get a boost if I leave the 12uf cap and raise the 2.5ohm to 4.5ohm?

1. It can be used to add 3 -4 DB more and frequency range is determined by the relationship between the values of res and cap.

2.Yes the very high treble will be reduced a few DB.

3.Changing the resistor to a 4.5 ohm will mean greater reduction in output in the tweeters lower range and the frequency the cap begins to take effect will be changed as well.

jakelm

Thaks GV...This is more complex than I thought. I understand the effects a cap, inductor and a resistor has on db's and fr. But when they are combined (paralleled) it gets very confusing.

I dont want to do anything to the crossover that will prohibit the drivers in any way, but I want the best "blend" I can get between the 2 woofers and the tweeter. In the since of the tweeter, I dont want to limit its Fr's in the upper octives, or have it roll off in the upper khz prematurely. Thats why I was curious in finding out , when changing values in the caps and resistors what effects they have on the tweeter.

I have searched the web looking for more indepth information on crossover design. Seems every turn just says something about "what" a crossover does and not in depth of "how" it does this, or in what way different componets effect the db's and Fr roll off or correcting spikes.

I was reading on notch filters........that will take a while.

jakelm

GV#27 wrote: »

In the tweeters lower range the resistor will attenuate its output as normally adding resistance would.At some frequency the 12uf cap starts to short or (bypass)the resistor so that it is no longer reducing its level.So the crossover designer was trying to compensate for a tweeter with a rolled off top end.
Does that help.

So if a tweeter design was week and it rolled off prematurely, the designer would want to increase the top end by amplifying the Fr's that were rolling off?

If a tweeter had a roll off of 6db's starting at 15khz. The designer would want to increase the top end range of the tweeter by amplifying 16+khz. To get a flatter responce from the tweeter 15khz -20khz. So they (the designers) in a since, cheat.

jcaut

Okay. I had a little more time to look at this, so I'll weigh in again. First though let me say that what you're doing is very difficult. Phase relationships are extremely important in trying to acheive a good transition from one driver to the other. What happens to the frequency response when you change component values is fairly easy to predict, in general terms, if we assume a "typical" tweeter impedance curve and a generally flat response. Since you've already got the basic crossover for the tweeter, your values should still be close, as far as acheiving a frequency response that works for that tweeter. However, driver interaction- those phase issues--especially when you're trying to make a center channel (I presume with the MW on each side and tweeter in the middle), are troublesome. Without measurements, it's going to be tough.

I did some quick and dirty modeling, again using that Audax tweeter. Here's what happens when you apply the M10 tweeter crossover to this tweeter and change values on some components. Keep in mind that this is not necessarily what's going on with YOUR tweeter, as the impedance curve and the FR are going to be a little different. But perhaps it can help you get an idea of what happens when you make changes.

Here goes with the pictures again. This first one is the response of this tweeter (Audax TM025F1) with no crossover (the red line) and with the M10 crossover applied, but WITHOUT the capacitor-resistor combo (blue line).


Next, I added the 12uF-2.7Ohm combo, and the result was the pink line.


Next, I removed the 12uF cap, and got the teal-colored line.


Finally, I changed the resistor from 2.7Ohms to 4.5Ohms. The result is the bright green line.


I'm not sure how much that will help. The program I'm using to get the plots above is SpeakerWorkshop. It's free and you can get it at http://www.speakerworkshop.com/. It's not particularly easy to use, and it requires some additional parts to actually be able to take measurements, but it's almost an all-in-one speaker building program once you get it working and get the hang of it. There are better ones out there, but they're not free. You also might be interested to look at http://www.pvconsultants.com/audio/frdgroup.htm For some other tools.

Jason

jcaut

I should have changed the scale on those charts so that the Frequency on the bottom would be more readable. On those I posted, along the bottom the lines should be 20Hz,50,100,500, 1K, 2K, 5K, 10K, 20KHz. Below about 500 is meaningless.

jakelm

Thanks Jason. That did help out alot. And it is excactly what my ears are hearing. I went back and added the 12uf cap and changed the resistor to 3.5ohms. I think I found a happy medium.

Using the 4.5ohm resistor with no cap, causes a gap between the tweeter and the woofers. The tweeter was too dull, I thought it was more natural at first but then realized I wasnt taking advantage of the tweeters full potential in the upper bands.

I do really need to put on paper, what the intire specrum looks like with the woofers added. So I can see if my intire Fr responce for the intire speaker is flat or where the woofers roll off so I can see where to roll off the tweeters.

I thank you and GV millions for taking the time out to help me. It is very difficult, nearly imposible, to calculate where to roll off the drivers in their new cabinet, with just my ears. I do rely about 80% on my ears, but to see it on paper help tremendously.

Thanks for taking the time to give me a hand.

One more question. Does you program check for phase allignment when changing the values of the caps and resistors? Did it say weather my tweeter and woofers were still in phase with my original design (4.5ohm with no cap)?

Systems

jakelm wrote: »

If a tweeter had a roll off of 6db's starting at 15khz. The designer would want to increase the top end range of the tweeter by amplifying 16+khz. To get a flatter responce from the tweeter 15khz -20khz. So they (the designers) in a since, cheat.

Well not really cheating just doing some simple passive EQ to try and flatten the frequency response.No driver is perfectly linear and many designers of hi performance speakers will use passive EQ of some form (like contour networks or notch filters etc.) in their designs to compensate for an irregularity in a drivers response.

Btw Jcaut those are excellent simulations that cleary show what effect each part has on the tweeters response.

jcaut

Thanks. Personally, I'm able to understand things like that better if I can see a representation of what's happening. If I had some measurements from the drivers he's actually using we could probably whip something up. BTW, that's the only way to determine exactly what's going on as far as phase- It's affected by driver placement, as well as the crossover components. As you tinker with the crossover, you should try reversing the polarity of the tweeter.

On another note, I thought the Peerless tweeter was an 8 Ohm tweeter- I know the recently discontinued one that looked a lot like it was 8 Ohm. Here's the spec sheet http://www.partsexpress.com/pdf/277-400.pdf. Its FR and impedance are actually pretty comparable to that Audax that I modeled.

Jason

jakelm

jcaut wrote: »

Thanks. Personally, I'm able to understand things like that better if I can see a representation of what's happening. If I had some measurements from the drivers he's actually using we could probably whip something up. BTW, that's the only way to determine exactly what's going on as far as phase- It's affected by driver placement, as well as the crossover components. As you tinker with the crossover, you should try reversing the polarity of the tweeter.

On another note, I thought the Peerless tweeter was an 8 Ohm tweeter- I know the recently discontinued one that looked a lot like it was 8 Ohm. Here's the spec sheet http://www.partsexpress.com/pdf/277-400.pdf. Its FR and impedance are actually pretty comparable to that Audax that I modeled.

Jason

Your right Jason, the peerless is a 8ohm tweeter, my bad. I have specs of a similar tweeter that I just purchased to replace the peerless incase it ever failed. Here is the other tweeter,,,,

http://www.mcminone.com/product.asp?catalog%5Fname=MCMProducts&category%5Fname=3829626&product%5Fid=LS00058

I dropped this tweeter in to see if there was a difference between it and the peerless. No difference. It might be 1db more efficient but as for range and sq, my ears could not tell a different.

Yes, I much prefer to see the results on paper. I know what my ears hear but to actually see it on paper , is much more helpful.

Your diagrams show a great deal of info. I wish someone here had your software and a peerless or a MCM tweeter. That would be cool to see what the results are.

I might just invest in the equipment to measure myself, and learn along the way.