My RS180 MTM Design

[billschulte]

Re: My RS180 MTM Design

as an uninformed observer of this site, I have been following this section closely. I would like to do a bi-amp of a single Dayton RS180(8 or 4) and found Jeff Bagby's Denhams response shaping network. Does anyone have a network like this worked up for a single RS 180?

[Jeff B.]

Re: My RS180 MTM Design

Hi Jeff (or someone else that understands ),

I'm struggling a little understanding all the relationships of the network... I think I'm lost in the simplicity.
Simple in number of parts, but very finely tuned to mate with the complex impedance and frequency response of the drivers.

The system sensitivity plot shows ~90dB @2.83V (correct?). Yes

Sensitivity is 94.1 dB @2.83V for the two RS180S-8's in parallel (per Jeff's post). Is this the low frequency sensitivity or the mid-band sensitivity? I'm assuming mid-band, based on what I see in the response curves.
Sensitivity is ~90.5-91 for the RS28AS-4 @2.83 V, pretty flat out to 11k. Sensitivity is usually a low frequency or average parameter.
When designing crossovers you have to deal with the actual responses which can vary quite a bit, sometimes several dB above the rated sensitivity at some frequencies.

Here's my understanding (confusion?):
1) With no pad on the tweeter, that means everything should be operating at it's nominal sensitivity. But, I'm only seeing a 3-3.5dB difference between the woofers' and tweeter's sensitivities (see above, ~90.5 vs 94.1) for baffle step. The sensitivity appears to drop off a few dB below 1kHz on the woofers. So, is this the ~6dB total difference that's being stated for baffle step? I'm not sure where the 6dB is coming from.
2) I see that the shelf is dropping ~6dB from 100Hz to 1000Hz, but what is creating it? If my calculations are correct, the 1.2mH coil would appear to begin its influence at ~460Hz (-3dB center) for a flat 3.5 Ohm impedance; This is close to the -3dB point on the shelving rolloff. Obviously, impedance isn't flat on the drivers; is this changing impedance what's causes the shallower (only 3dB/octave) shelving rolloff?
The baffle step compensation is coming from the 1.2mH inductor and its initial roll-off. Your thinking is correct.

3) Is the notch filter responsible for the sharper rolloff after the XO center?
The 20uF cap and 0.025mH coil are forming in the trap for the cone breakup (by my calculations, centered ~7.1kHz, which matches with the notch in the transfer filter). I'm think I'm square on this.
The notch filter has been tuned so that its corner behavior matches that of an LR4 crossover. Remember, the notch filter includes the 1.2mH inductor. Without it there is only a drop in the impedance due to the cap and small inductor, but no change in the response.

4) Why is the crossing point (1.7kHz) -9dB (in the filter transfer function graph)? Shouldn't it be -6dB for 4th order? Wouldn't this create a -3dB dip in this region? Or is this accounting for the peaking in the RS180's acoustic response in this region?
Yes. You are trying to understand things by looking at only the transfer function of the filters. Although the transfer function shows the effects of the complex impedance of the drivers on the circuit, it does not show how it mates up with the natural on-baffle frequency response of the drivers. The final results are shown in the first post in the thread. The filter accounts for baffle step, rising and peaking response in the woofers, and the baffle diffraction bump in the tweeter's response.


Thanks much in advance... I'm going in circles

[scottq]

Re: My RS180 MTM Design

I intended to post this chart too, but forgot. Here are the transfer functions for the crossover for those that are interested in this.


Jeff

I guess about 6dB based on the lowpass transfer function plot which shows a 6dB transition between 1khz and 100Hz. However, I don't tend to think in these terms when I design. My measurements are a composite of 1 meter on the stands and near field spliced, then I target a flat response. I also measure using an RTA with pink noise in the listening position to check the overall balance. I usually never think in terms of how much baffle step compensation I am applying while I am doing this. I am only interested in the final result.

Hi Jeff (or someone else that understands ),

I'm struggling a little understanding all the relationships of the network... I think I'm lost in the simplicity.

The system sensitivity plot shows ~90dB @2.83V (correct?).

Sensitivity is 94.1 dB @2.83V for the two RS180S-8's in parallel (per Jeff's post). Is this the low frequency sensitivity or the mid-band sensitivity? I'm assuming mid-band, based on what I see in the response curves.
Sensitivity is ~90.5-91 for the RS28AS-4 @2.83 V, pretty flat out to 11k.

Here's my understanding (confusion?):
1) With no pad on the tweeter, that means everything should be operating at it's nominal sensitivity. But, I'm only seeing a 3-3.5dB difference between the woofers' and tweeter's sensitivities (see above, ~90.5 vs 94.1) for baffle step. The sensitivity appears to drop off a few dB below 1kHz on the woofers. So, is this the ~6dB total difference that's being stated for baffle step? I'm not sure where the 6dB is coming from.
2) I see that the shelf is dropping ~6dB from 100Hz to 1000Hz, but what is creating it? If my calculations are correct, the 1.2mH coil would appear to begin its influence at ~460Hz (-3dB center) for a flat 3.5 Ohm impedance; This is close to the -3dB point on the shelving rolloff. Obviously, impedance isn't flat on the drivers; is this changing impedance what's causes the shallower (only 3dB/octave) shelving rolloff?
3) Is the notch filter responsible for the sharper rolloff after the XO center?
The 20uF cap and 0.025mH coil are forming in the trap for the cone breakup (by my calculations, centered ~7.1kHz, which matches with the notch in the transfer filter). I'm think I'm square on this.
4) Why is the crossing point (1.7kHz) -9dB (in the filter transfer function graph)? Shouldn't it be -6dB for 4th order? Wouldn't this create a -3dB dip in this region? Or is this accounting for the peaking in the RS180's acoustic response in this region?

Thanks much in advance... I'm going in circles.

[Jeff B.]

Re: My RS180 MTM Design

Hmmmm..... missed that. Started thinking that the sensitivity would drop 6dB like it was a single 8 Ohm driver. However, looking at the specs, the 8 Ohm version calculates to 88.1/dB/2.83V. Put two in parallel and we are at 94.1dB/2.83V. The 4 Ohm version calculates at 90.3dB/W/2.83V, put them in series and you're still at 90.3dB/2.83V. Insertion losses due to series resistance will impact the parallel set a little more, so the difference will be pretty close to 3dB lower for the series 4 ohm pair. So, you're right, the tweeter shouldn't drop the full 6dB that I accounted for. Later when I have a chance I will reevaluate the crossover and post an update. (I don't have those files handy right now, they're on my flash drive, but I'll take a look later on.)

Jeff

OK. Revising the tweeter circuit for this new woofer level I come up with a third order circuit of: 8 uF, .35mH, and 40 uF with a series resistor of 1.5 ohms (you can tweak this up or down for small shifts in the tweeter balance). This yields the identical response curve, crossover point, and phase tracking as the original design posted at the top. Overall sensitivity is 3dB lower. It is mostly an 8 ohm speaker with a minimum impedance of 5 Ohms at 7kHz, but the phase angle is very small and inductively approaching 0 degrees, so the load should be pretty easy to handle.

Sorry for the confusion, good catch.

Jeff

[Jeff B.]

Re: My RS180 MTM Design

Thanks Jeff, I've ordered all the parts and will keep you posted. Likely start another thread. The only questionable aspect is the 4 ohm R in front of the RS28A. That's going to offer about 6db attenuation compared to your original schemo. I am under the impression that a pair of 8 ohm woofers wired in parallel will only be 3db louder than a pair of the 4 ohm version of those woofers in series.

There is a 6db difference for the same driver in series vs parallel configuration. But a 4 ohm woofer WILL be 3 db louder.

Hmmmm..... missed that. Started thinking that the sensitivity would drop 6dB like it was a single 8 Ohm driver. However, looking at the specs, the 8 Ohm version calculates to 88.1/dB/2.83V. Put two in parallel and we are at 94.1dB/2.83V. The 4 Ohm version calculates at 90.3dB/W/2.83V, put them in series and you're still at 90.3dB/2.83V. Insertion losses due to series resistance will impact the parallel set a little more, so the difference will be pretty close to 3dB lower for the series 4 ohm pair. So, you're right, the tweeter shouldn't drop the full 6dB that I accounted for. Later when I have a chance I will reevaluate the crossover and post an update. (I don't have those files handy right now, they're on my flash drive, but I'll take a look later on.)

Jeff

[gowa]

Re: My RS180 MTM Design

This actually looks like it should be a PE kit. Nice Jeff.

[Armstrong]

Re: My RS180 MTM Design

I REALLY want to hear these babies!!!

[ontariomaximus]

Re: My RS180 MTM Design

Thanks Jeff, I've ordered all the parts and will keep you posted. Likely start another thread. The only questionable aspect is the 4 ohm R in front of the RS28A. That's going to offer about 6db attenuation compared to your original schemo. I am under the impression that a pair of 8 ohm woofers wired in parallel will only be 3db louder than a pair of the 4 ohm version of those woofers in series.

There is a 6db difference for the same driver in series vs parallel configuration. But a 4 ohm woofer WILL be 3 db louder.

[Jeff B.]

Re: My RS180 MTM Design

No, the cap would scale to 10uf. However, I would need to see if the Q of notch would be right with those values. Let me see if I can take a closer look at my simulation tomorrow.

Jeff

OK. I went back to my original files and remodeled it as an two 4 ohm drivers in series. If you replace my woofer circuit above with these values: 2.5mH, 10uf, and .05mH, then the lowpass looks perfect. (I'd use a fairly low dcr inductor for that 2.5mH one).

For the tweeter you can use the same crossover as above and add an L-Pad network of Rs= 2 Ohm and Rp= 4 Ohm.

However, I have a better option. This one raises the tweeter load to 8 Ohms, making the speaker an 8 Ohm speaker over the entire audio band, and allows you to use more conventional component values:

Set up a 3rd order circuit on the tweeter comprised of 6 uf, .5 mH, and 25uf, and then use a 4 Ohm resistor in series between the filter and the tweeter. If you want to raise the top octave of the RS28A you can bypass the resistor with a 2.2uf cap, but it may not be necessary, and if you try the RS28F, then don't use the cap.

Give it a try, this should be very close based on the design I built.

Jeff B.

[Jeff B.]

Re: My RS180 MTM Design

I don't know exactly what an OPC is, but simpler! You can say that again.

Other People's Crossovers :D

I just can't do that! Be honest, you'd be disappointed in me if I did. ;)

[PunkSweeper]

Re: My RS180 MTM Design

Only my crossover is simpler, has excellent phase trackking, and doesn't have that notch at 3kHz. Plus I'm too neurotic to use OPC's. :D

I don't know exactly what an OPC is, but simpler! You can say that again.

[Jeff B.]

Re: My RS180 MTM Design

Did I miss something or are these strikingly similar to Dr. K't MTMs?
http://www.parts-express.com/project...project=DrKMTM

Yes, pretty similar. I started with the basics of that kit - cabinets and drivers, only I started with an all active system. The two amps and active filters later we went passive. I never even looked at this design till now. (I notice that our notch is pretty much identical, how about that). Only my crossover is simpler, has excellent phase trackking, and doesn't have that notch at 3kHz. Plus I'm too neurotic to use OPC's. :D

[PunkSweeper]

Re: My RS180 MTM Design

Did I miss something or are these strikingly similar to Dr. K't MTMs?
http://www.parts-express.com/project...project=DrKMTM

[Jeff B.]

Re: My RS180 MTM Design

So the 1.2mh, 20uf, and .025mh components in the woofer section would be 2.4mh, 40uf and .05mh, right?

And can I just add 3 ohms in front of the tweeter crossover to pad the level a bit, without changing the other values?

Perhaps I may even use the Seas TBFC instead as it is SA (still available). Maybe the Usher 9950, any other candidates spring to mind?

No, the cap would scale to 10uf. However, I would need to see if the Q of notch would be right with those values. Let me see if I can take a closer look at my simulation tomorrow.

Jeff

[ontariomaximus]

Re: My RS180 MTM Design

So the 1.2mh, 20uf, and .025mh components in the woofer section would be 2.4mh, 40uf and .05mh, right?

And can I just add 3 ohms in front of the tweeter crossover to pad the level a bit, without changing the other values?

Perhaps I may even use the Seas TBFC instead as it is SA (still available). Maybe the Usher 9950, any other candidates spring to mind?