Hi all...I'm new here but not entirely new to DIY or car audio. Basically, I got a new head unit which has a nice clean signal but almost no bass. I had 4 6.5" coaxials, and I replaced the rear ones with midbass drivers. Now the bass is quite good but the high mids are very harsh and fatiguing. The freq. response on the rears is 60-5000Hz, and my thought is I should start experimenting with a low-pass filter. I think a first-order filter with a cutoff of 1250Hz or so should provide a smooth enough transition so I don't lose too much fill. I'm wondering if I'm right, and if a Zobel circuit would help too?
TIA!
TIA!
GhettoSQ said:
This is a problem with adding rear-fills. What is happening is the frequencies between the front and rear speakers are overlapping and creating a 'resonant hump' through the overlapping freqs. This is also caused by time-delay or phase differences between the speakers. Really, the best way to negate this is with an amp. either get some nice component speakers, with a good low freq. response, and forgo the rears, or get a sub and allow the front speakers to provide the mids and highs. Rear-fills are always a subjective subject, but most car audio people don't like rear speakers because of what I stated above. Besides, who cares if the people riding in back think the system sounds good. 😉
interesting...I figured it was the overlap, but right now the overlap is providing a nice fill and some amount of thump on the low end. I'm not doing a sub anytime soon, though I might end up amping the rear speakers at some point. My moniker reflects my budget 🙂
So, I'm assuming that while a simple crossover fix wouldn't make it a glorious system, it might be a good place to start? I'll try reversing the rear speaker polarity too, it might not affect the low end as much. Do you think 6db/oct. at 1250Hz would be about right? by 5000Hz it would be down 12db, audible but (hopefully) not unnatural. I'm also interested in doing a Zobel circuit, although I don't have the Le and Re specs for these cheapo speakers.
thanks!
So, I'm assuming that while a simple crossover fix wouldn't make it a glorious system, it might be a good place to start? I'll try reversing the rear speaker polarity too, it might not affect the low end as much. Do you think 6db/oct. at 1250Hz would be about right? by 5000Hz it would be down 12db, audible but (hopefully) not unnatural. I'm also interested in doing a Zobel circuit, although I don't have the Le and Re specs for these cheapo speakers.
thanks!
I think that as a starting point your idea is OK. Have you any test equipment such as signal generator and voltmeter? You can then measure Le quite easily. I think you would need to use a zobel on a first order network, firstly to get the response right and secondly because at your chosen frequency we can only guess what the impedance is.
I have a multimeter but no signal generator -- could I use a test tone CD if I knew the frequency?
I don't fully understand how a Zobel affects sound...it would seem to me that if I use a crossover it should increase the speaker's impedance at high frequencies faster than the Le does, but a zobel would lower the impedance?
I don't fully understand how a Zobel affects sound...it would seem to me that if I use a crossover it should increase the speaker's impedance at high frequencies faster than the Le does, but a zobel would lower the impedance?
You could burn a tone of the desired frequency to a CD and use that, yes.
A zobel can actually mean different things depending on the application it's used for in a circuit. In this particular instance it serves to compensate for the rise in voicecoil impedance due to its natural inductance. This then gives your crossover filter a fairly even and constant impedance to work with, meaning that it will produce a response much more inline with what theory predicts.
A zobel can actually mean different things depending on the application it's used for in a circuit. In this particular instance it serves to compensate for the rise in voicecoil impedance due to its natural inductance. This then gives your crossover filter a fairly even and constant impedance to work with, meaning that it will produce a response much more inline with what theory predicts.
Also, if the zobel lowers the impedance at resonance (67Hz), will that increase the volume of that frequency in a good way? Dare I ask what would happen to the cutoff frequency without a zobel?
I didn't mean to ignore TO-3's post above...the next stage would be to get separates up front, I think. I'm also toying with the idea of doing a 2-way xover in back and running extra tweeters up front, but I think that gets into too many speakers and trying to overcompensate...
thanks again btw, this is very helpful stuff!
I didn't mean to ignore TO-3's post above...the next stage would be to get separates up front, I think. I'm also toying with the idea of doing a 2-way xover in back and running extra tweeters up front, but I think that gets into too many speakers and trying to overcompensate...
thanks again btw, this is very helpful stuff!
Properly implemented (not hard) a zobel will not make a blind bit of difference to the impedance at resonance as that part of the spectrum is well below the inductive segment.
And yes, extra tweeters in the front is a bad idea.
And yes, extra tweeters in the front is a bad idea.
Hehehe, thought so...
so, back to the original question -- do you think 1250Hz is a good point for a 6db xover? partsexpress has those nifty 800Hz 12db low-pass ones but I think that's a bit too low and at $20 each I'd want to be sure. I just don't know at what frequency the harshness is occurring, but it's mostly higher vocals and snare drums.
so, back to the original question -- do you think 1250Hz is a good point for a 6db xover? partsexpress has those nifty 800Hz 12db low-pass ones but I think that's a bit too low and at $20 each I'd want to be sure. I just don't know at what frequency the harshness is occurring, but it's mostly higher vocals and snare drums.
Update: I switched the polarity in back, the harshness is ever so slightly less, the bass seems like it might even be a tiny bit tighter, but now I can hear it coming from the rear whereas before it was pulled up front.
I'm also noticing that my front speakers do in fact sound harsh in that range...would components help, or just a better set of coaxials?
I'm also noticing that my front speakers do in fact sound harsh in that range...would components help, or just a better set of coaxials?
Personally, I like coaxials because they are more like a point source. Unfortunately car audio people don't seem to have grasped this fact so they are marketed more as budget stuff and the better stuff is component only. Maybe some better coaxials will help. What is the tweeter material on your present ones?
As for your original idea, you would just have to suck it and see what crossover frequency to use as it would depend on music preferences and vehicle acoustics, but I would think that 1st-order would be better as that introduces less phase change and is better damped at the cutoff frequency.
As for your original idea, you would just have to suck it and see what crossover frequency to use as it would depend on music preferences and vehicle acoustics, but I would think that 1st-order would be better as that introduces less phase change and is better damped at the cutoff frequency.
I'm not sure about the tweeter material but they're kenwood so I'm guessing some kind of mylar. Would silk be better?
I agree about the point-sourcing...it's very hard to find an honest opinion about components. Right now I kind of like the audiobahn ACS2063P which can function as both component or coaxial but have outboard crossovers. If I could afford CDT Braxials I would 🙂
That was what I was hoping about 1st-order crossovers; it seems like they're not very popular. I was thinking it might be ideal to have 1st or 3rd-order crossovers in front to match phasing, but maybe it's better to have the mids a bit out of phase anyway.
So, I get some .51mH air-core inductors, glue them to a board and solder them to disconnects and plug them in-line on the positive side, right? 🙂
I agree about the point-sourcing...it's very hard to find an honest opinion about components. Right now I kind of like the audiobahn ACS2063P which can function as both component or coaxial but have outboard crossovers. If I could afford CDT Braxials I would 🙂
That was what I was hoping about 1st-order crossovers; it seems like they're not very popular. I was thinking it might be ideal to have 1st or 3rd-order crossovers in front to match phasing, but maybe it's better to have the mids a bit out of phase anyway.
So, I get some .51mH air-core inductors, glue them to a board and solder them to disconnects and plug them in-line on the positive side, right? 🙂
Those Audiobahn's sound interesting. Outboard crossovers and coaxial... hmmm might have to see about those as one of my best mates is friendly with an Audiobahn dealer.
1st-order crossovers are often bad mouthed because as a *crossover* they aren't very good. The don't have enough stopband attenuation. However, in your application as a *filter* just to roll off gently and provide fill, they are ideal.
Yes, all you need to do is put an inductor inline with one of the speaker leads. Just make sure it's DC resistance is less than 0.2 ohms.
1st-order crossovers are often bad mouthed because as a *crossover* they aren't very good. The don't have enough stopband attenuation. However, in your application as a *filter* just to roll off gently and provide fill, they are ideal.
Yes, all you need to do is put an inductor inline with one of the speaker leads. Just make sure it's DC resistance is less than 0.2 ohms.
I was wondering about the DC resistance -- I was looking at the Jantzen 18ga. with a Re of .34, but maybe I should spend a few bucks more on better ones.
Here's my take Ghetto -
You want the cheapest and simplest solution to the harshness problem. As has been mentioned in previous posts...this problem is due to either poor front co-ax speakers ( I'm not a fan of Kenwood speakers) OR by the problem mentioned by TO-3 i.e. "What is happening is the frequencies between the front and rear speakers are overlapping and creating a 'resonant hump' through the overlapping freqs."
Both valid problems and in fact both may be contributing to the harshness. To find out what is going on...simply listen to the fronts speakers alone...then slowly fade in the rears. Decide which situation produced the most harshness. If it is the front co-ax's then it's a no-brainer...buy better ones. (in my opinion Polk and JBL are decent low-cost co-ax's)
On the other hand, if it seems that the harshness is being caused by the overlapping as articulated by TO-3 then you already had, in your first post, the simplest way to cure that: a low pass filter in line with the positive lead to each rear speaker.
In this context a "low pass filter" is the same as the low-end portion of a "first order (6db) crossover" which is, simply, an inductor coil.
If you decide to cure the frequency overlapping...buy two inductor coils from parts express. Do NOT buy ready made crossovers, especially anything "12db per octave" because that is a more complex circuit that would be very difficult if not impossible to tune by ear and is not necessary anyway.
Do NOT worry about Zoebels...that is a circuit sometimes used to assist in matching the low-end portion of a crossover to the high-end portion and is desigened to work with specific drivers...again too complex to do without the proper equipment and in your case is not even called for.
Get EITHER the 18 gauge Jantzen coils you already are thinking about OR the "laminated I core" inductors. The laminated inductors have less resistance (also mentioned in a previous post) and would be my choice. But either one will work fine for a system with low-to-medium power...
http://www.partsexpress.com/webpage.cfm?&DID=7&WebPage_ID=313
http://www.partsexpress.com/webpage.cfm?&DID=7&WebPage_ID=197#18gauge
As for what size coils (inductors) to get...well the size (measured in mH) is what determines the crossover point. You mentioned already 1250Hz I believe. That may be adequate to reduce high-mid harshness. However, you could also go much lower...say to around 150 - 200Hz because in your current set-up your real speakers are essentially acting as subs to your front speakers...in other words, your front speakers are most likely playing down to at least 150Hz.
The trade-off is that to cross your rear speakers at 150Hz requires a bigger and more expensive coil than to cross it at 1250Hz.
Another issue is that you will be doing trial-and-error tuning and it's not practical to buy several pairs of coils.
Assuming your rear speakers are 4 ohms: a 2.5mH coil will cross them at 250Hz; a 1.25mH will cross at 500Hz; a .64 mH will cross them at 1000Hz ...
It becomes guess work from here...my FIRST try would be to buy a pair of either 2.0mH or 1.5mH coils (again this assumes the rear speakers are 4 ohms...if the are 8 ohms you need to double the value of the inductor coil to cross at the same frequency)
Finally, I would wire all four speakers with the same polarity...undo the change you made.
good luck, hope this all makes sense.... 🙂
You want the cheapest and simplest solution to the harshness problem. As has been mentioned in previous posts...this problem is due to either poor front co-ax speakers ( I'm not a fan of Kenwood speakers) OR by the problem mentioned by TO-3 i.e. "What is happening is the frequencies between the front and rear speakers are overlapping and creating a 'resonant hump' through the overlapping freqs."
Both valid problems and in fact both may be contributing to the harshness. To find out what is going on...simply listen to the fronts speakers alone...then slowly fade in the rears. Decide which situation produced the most harshness. If it is the front co-ax's then it's a no-brainer...buy better ones. (in my opinion Polk and JBL are decent low-cost co-ax's)
On the other hand, if it seems that the harshness is being caused by the overlapping as articulated by TO-3 then you already had, in your first post, the simplest way to cure that: a low pass filter in line with the positive lead to each rear speaker.
In this context a "low pass filter" is the same as the low-end portion of a "first order (6db) crossover" which is, simply, an inductor coil.
If you decide to cure the frequency overlapping...buy two inductor coils from parts express. Do NOT buy ready made crossovers, especially anything "12db per octave" because that is a more complex circuit that would be very difficult if not impossible to tune by ear and is not necessary anyway.
Do NOT worry about Zoebels...that is a circuit sometimes used to assist in matching the low-end portion of a crossover to the high-end portion and is desigened to work with specific drivers...again too complex to do without the proper equipment and in your case is not even called for.
Get EITHER the 18 gauge Jantzen coils you already are thinking about OR the "laminated I core" inductors. The laminated inductors have less resistance (also mentioned in a previous post) and would be my choice. But either one will work fine for a system with low-to-medium power...
http://www.partsexpress.com/webpage.cfm?&DID=7&WebPage_ID=313
http://www.partsexpress.com/webpage.cfm?&DID=7&WebPage_ID=197#18gauge
As for what size coils (inductors) to get...well the size (measured in mH) is what determines the crossover point. You mentioned already 1250Hz I believe. That may be adequate to reduce high-mid harshness. However, you could also go much lower...say to around 150 - 200Hz because in your current set-up your real speakers are essentially acting as subs to your front speakers...in other words, your front speakers are most likely playing down to at least 150Hz.
The trade-off is that to cross your rear speakers at 150Hz requires a bigger and more expensive coil than to cross it at 1250Hz.
Another issue is that you will be doing trial-and-error tuning and it's not practical to buy several pairs of coils.
Assuming your rear speakers are 4 ohms: a 2.5mH coil will cross them at 250Hz; a 1.25mH will cross at 500Hz; a .64 mH will cross them at 1000Hz ...
It becomes guess work from here...my FIRST try would be to buy a pair of either 2.0mH or 1.5mH coils (again this assumes the rear speakers are 4 ohms...if the are 8 ohms you need to double the value of the inductor coil to cross at the same frequency)
Finally, I would wire all four speakers with the same polarity...undo the change you made.
good luck, hope this all makes sense.... 🙂
Whilst all that advice is good, I disagree with the zobel part. At the frequency he is looking to filter at, the impedance will NOT be 4 ohms due to voicecoil inductance, and it will also be rising throughout the stopband, possibly subduing any attenuation that the filter is trying to achieve.
All he needs to do is burn a couple of tones to a CD and get his meter out, a 1k and a 4.7 ohm resistor.
All he needs to do is burn a couple of tones to a CD and get his meter out, a 1k and a 4.7 ohm resistor.
Awesome guys, thanks! I think it is a little bit of both, the front speakers have a harsh upper-mid tone all by themselves, but it's much more pronounced when I fade in the rears. I did undo the polarity switch, it ended up sounding much worse out of phase.
My local electronics superstore for some reason doesn't carry inductors bigger than 1uH, which makes the trial-and-error more difficult. I think I will start at about 1000-1250Hz since I like the depth all the way through the range. Eventually I will also upgrade the front speakers but that's a bit more research-intensive, and when I do I might find I want a lower filter point in back.
I think since I can get all manner of caps and resistors locally, I will try the inductors solo first and see if they have any effect. This is also because it's a learning project and I want to isolate each step. Richie, how do I go about measuring voice-coil inductance?
thanks again, this is great help!
My local electronics superstore for some reason doesn't carry inductors bigger than 1uH, which makes the trial-and-error more difficult. I think I will start at about 1000-1250Hz since I like the depth all the way through the range. Eventually I will also upgrade the front speakers but that's a bit more research-intensive, and when I do I might find I want a lower filter point in back.
I think since I can get all manner of caps and resistors locally, I will try the inductors solo first and see if they have any effect. This is also because it's a learning project and I want to isolate each step. Richie, how do I go about measuring voice-coil inductance?
thanks again, this is great help!
- Status
- Not open for further replies.