The off-the-shelf L-pad is bulky, has to be securely mounted in some way, and costs more than the 2 resistors. If it isn't calibrated in dB, then you can't know whether or not the subjectively correct amount of attenuation is less or more than what would be correct according to making the sensitivities of woofer and tweeter the same.
MCPete speaks the truth. However, I still chose the off-the-shelf route as I change my setup around periodically and wanted the flexibility to be able to use it with different drivers. Plus I wanted to keep it neat and tidy, so I used a small enclosure and some decent binding posts.
Attachments
Or alternatively, you can use just a single resistor in series with the tweeter (before the HP xover section so it won't effect the tweeter xover pionts) and that makes it very easy and simple to try out different values.
I'd say the tweeter would need to have a fairly linear impedance curve though for this to work properly. Most decent tweeters should though.
I've found this to be adequate in all my applications.
My 2c
I'd say the tweeter would need to have a fairly linear impedance curve though for this to work properly. Most decent tweeters should though.
I've found this to be adequate in all my applications.
My 2c
Hi,
The L-pad is best placed across the driver so as not to affect the x/o.
A single resistor before the x/o changes the source impedance and
does affect both the x/o point and the Q of the tweeter high pass.
Obviously the higher value the resistor the more effect it has.
Tweaking values 1R to 2R are usually OK.
rgds, sreten.
Are you sure about that? Yes, I agree it increases the source impedence for the amplifier, but an L/C circuit resonant freq and Q factor should only be effected by the load resistance/impedence it works into.
Therefore a resistor in series before the L/C circuit shouldn't effect anything.
Or am I missing something?
Hi,
Yes you are missing something, an assumed zero source impedance for
filters. Proper filter design requires known source* and load impedance.
You can use use series resistors, but you need to change x/o values.
Dead simple example : A 1st order high pass on a 8 ohm tweeter
via a series capacitor. You knobble the tweeter by 6dB by adding
a 8 ohm series resistor, you've then also halved the x/o point.
If you L-pad the driver by 6dB the x/o point remains the same.
rgds, sreten.
Try it using Tina-Ti, an excellent free circuit emulator.
*Yes high impedance amplifiers do screw up x/o filters.
Last edited:
All methods of resistive attenuation will alter the action of the high pass filter. Whether the crossover point, the damping or just by the shape of the impedance curve.
The L-pad would be the first choice as it does more to flatten the impedance. Of course if you design so from the start, the different resistor locations/types of attenuation, can each be made to do the job in the end.
Tweaking the amount of attenuation in an L-pad will affect the filter and tweeter response slightly, but the L-pad is good for larger amounts of attenuation as it becomes more consistent the more it attenuates.
The L-pad would be the first choice as it does more to flatten the impedance. Of course if you design so from the start, the different resistor locations/types of attenuation, can each be made to do the job in the end.
Tweaking the amount of attenuation in an L-pad will affect the filter and tweeter response slightly, but the L-pad is good for larger amounts of attenuation as it becomes more consistent the more it attenuates.
Interesting, you may be correct here, I did this test today with my own home made speakers, I removed the series resistor before the xover and fitted the 2 correct value resistors after the xover and I find the speakers much smoother sounding with much better stereo imaging.
Its a 12dboct slope on the tweeter, so the xover freq can't change with impedence, however (using your logic) the Q of the filter may change? Would the Q factor increase?
If that is indeed the case, that explains why the sound is smoother now, ie: lower Q = smoother/soggier rolloff.
Looks like it will need more tweaking and listening now.
Didn't mean to hijack this thread, sorry.
I'm not sure I agree that a lower Q will sound much better in this case, I know it can in some circumstances. Perhaps a more smooth response could be helping.
I've prepared a small test simulation. I'm not calling it conclusive, but hope it helps. There are four plots, each is a tweeter (same tweeter) whose impedance is shown in grey. All four have an identical second order linkwitz riley filter at 2000Hz based on 6 ohms.
1. The green plot is just the tweeter and filter.
2. Light blue uses one resistor before the filter.
3. Red uses an L-pad.
4. Dark blue uses one resistor after the filter.
Draw your own conclusions.
I've prepared a small test simulation. I'm not calling it conclusive, but hope it helps. There are four plots, each is a tweeter (same tweeter) whose impedance is shown in grey. All four have an identical second order linkwitz riley filter at 2000Hz based on 6 ohms.
1. The green plot is just the tweeter and filter.
2. Light blue uses one resistor before the filter.
3. Red uses an L-pad.
4. Dark blue uses one resistor after the filter.
Draw your own conclusions.
Attachments
Hi,
What you think doesn't affect the actual reality.
All your doing is showing what you don't understand.
Its very lazy miss-applied thinking with the wrong conclusion.
rgds, sreten.
OK, I think I know what is going on. AllenB, could you please do the same simulation with just one capacitor as the filter? (Could you choose a resistor of 6 ohm to make the plots vary more from the unattenuated one, no need for an L-pad simulation before/after unless you can find the values that build a 6 ohm L-pad?) The resistor before or after the capacitor shouldn't matter then. But in case of a second order filter or higher, it does matter, becuase the load, or load plus resistor, throws off the balance of impedance between the legs of the filter. A resistor before the crossover messes with that leg of a second order crossover. Thus, there is no way to avoid a small push of the crossover point if one wants to attenuate a driver when using a second order filter or higher. In the case of an L-pad it can be designed to hold the total load, as seen by the crossover, closer to the load of the driver, and less pushing of the crossover point is the result.
What I will take away from this lesson is that the performance of a tweeter is closer to the raw tweeter at the crossover point with a second order crossover when the resistor is placed before the crossover. That's good news, because tweeters fall too quickly around their desired crossover point in a 2-way system.
AllenB, for education and fun, could you also run a simulation with a 6.5" woofer, 6 ohm, 0.5 mH, otherwise standard parameters, through a second order at 2000 Hz with a 6 ohm resistor before/after, no need for a L-pad simulation unless you can easily find a 6 ohm.
What I will take away from this lesson is that the performance of a tweeter is closer to the raw tweeter at the crossover point with a second order crossover when the resistor is placed before the crossover. That's good news, because tweeters fall too quickly around their desired crossover point in a 2-way system.
AllenB, for education and fun, could you also run a simulation with a 6.5" woofer, 6 ohm, 0.5 mH, otherwise standard parameters, through a second order at 2000 Hz with a 6 ohm resistor before/after, no need for a L-pad simulation unless you can easily find a 6 ohm.
It'll always matter.
You should probably have a crossover simulator like the one included in Speaker Workshop so you can experiment yourself.
I can tell you now, that when you have a capacitor and a resistor in series with each other, and nothing in between, their values can be combined into a single complex value for network analysis. Indeed it does not matter which way they are ordered.
In fact the case of the resistor before the capacitor in series, followed by the inductor and tweeter in parallel, would be identical if the capacitor was first, then the resistor followed by the inductor and tweeter.
Now you're on the right track.
Not quite. It will appear this way though if you focus too closely on the knee area of the response.
Anything after the filter that is seen by the whole filter affects damping. Damping is like disposal of the energy that the two components are managing. This energy management occurs primarily around their resonance. As you move far above or below the resonance, damping has little effect on the filter. The slope at low frequencies should remain the same regardless.
This is entirely a design choice.
You may use a second order filter that combines with the natural tweeter rolloff to give a virtual 4th order slope, but it may be down too far at the crossover point because the tweeter is naturally already down a little at that point. By using a high Q electrical filter you can alleviate some of the droop.
So when you see what you need acoustically from a filter, then apply it so that it functions correctly with the available impedance, than you can have what you want, within reason.
AllenB, for education and fun, could you also run a simulation with a 6.5" woofer, 6 ohm, 0.5 mH, otherwise standard parameters, through a second order at 2000 Hz with a 6 ohm resistor before/after, no need for a L-pad simulation unless you can easily find a 6 ohm.[/QUOTE]
OK. Of course using a resistor in series with a woofer will expose the impedance peak. It will bring the whole response down, and I assume you were expecting 6dB with a total impedance based at 12 ohms.
Starting with a flat response to expose the filter action, using the impedance curve in grey. Green shows the resistor before, blue is after. Red is using a 6 ohm, 6dB L-pad like you asked, and yellow uses a 12 ohm, 6dB L-pad for comparison with the first two plots.
Attachments
Thanks, AllenB. Could you do a redo and include the phase shift resulting from a 6 ohm resistor before/after a second order crossover, and include the phase shift for a crossover with no resistor? Is that a freeware program, or where can I get a similar freeware?
I moved the resistor in my 2-way speakers to in front of the second order crossover and the sound got more level and clean. Nice to know I can lift a woofer too around the crossover point. I often use a 0.33-0.47 ohm resistor on a woofer to lift the low end which is sagging otherwise, but have put it after the crossover.
I moved the resistor in my 2-way speakers to in front of the second order crossover and the sound got more level and clean. Nice to know I can lift a woofer too around the crossover point. I often use a 0.33-0.47 ohm resistor on a woofer to lift the low end which is sagging otherwise, but have put it after the crossover.
You're lucky, I was about to close that one off.
Here they are, using flat responses and impedances to indicate the filter action. With the resistor after in yellow, the damping is affected but the crossover point (looking at the downward slope) is at the same frequency. The red curve shows the crossover point has been shifted even though the damping is fine.
Here they are, using flat responses and impedances to indicate the filter action. With the resistor after in yellow, the damping is affected but the crossover point (looking at the downward slope) is at the same frequency. The red curve shows the crossover point has been shifted even though the damping is fine.
Attachments
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.