Need Help with speaker and crossover power handling calculations

[Nvr2manybikes]

Hi, I designed a 2 way series crossover speaker system a while back (Fig 1 based on System 7's concept giving very good phase alignment) and was just going back over the design trying to determine the power that the drivers and resistors will see. Original design was with Vision's BOXSIM program. I replicated the crossover in 5SPICE (Fig 2) and took in the BOXSIM equivalent circuit for the drivers. Using that equivalent circuit gave a pretty good match of the calculated impedance curves between BOXSIM and 5SPICE (Figs 3 and 4). Based on BOXSIM the woofer SPL looks to be more than 4dB attenuated in crossover/box (Fig3) based on raw output of driver being 87db but seeing ~83dB or less in the box. Using 5 spice I calculated PwrDiss (Speaker Driver V, Driver Current) divided by PwrDiss(System Input V, Input Current)(Fig 5). This shows much less attenuation - more like 1.5 to 2 dB for a good part of the range. (I also ran the calculation just multiplying V and Current rather than using the PwrDiss function in 5SPICE and got similar results)

Why the large discrepancy? For this 40W rated woofer can I put in 100W like BOXSIM implies or only 50W like 5SPICE implies?

Then for the resistors - I used 5 spice to calculate the power dissipated as for the driver. For the 6 Ohm resistor in the woofer circuit (R12) it sees a peak of -4.7 dB or about one third of the input power at around 2.3kHz (Fig 6). Do I have to size this resistor to handle 1/3rd of the input power even if it is not over a broad frequency range? Should I even believe this calculation based on the discrepancy between BOXSIM and 5SPICE on the Woofer driver above?

Thanks in advance for any help!

Fig1

Fig2:

Fig3:
:

Fig4:

Fig5

Fig6

 

[Nvr2manybikes]

** Visaton ** BOXSIM - not Vision

 

[wolf_teeth]

The attenuation you are seeing is not an attenuation at all, but a loss altogether. This is commonly known as baffle-step. The acoustic concept of this loss is a -6dB total, reducing in loss to the wavelength of the baffle width where there is a peak, and then levels off to nominal sensitivity. Since you are seeing a -4dB attenuation, the design was not fully compensated for a little closer to wall placement. The way it is typically compensated is by BSC, and involves attenuating the upper range on a woofer to lower the output to approximately that of the region with the loss to flatten it out. This is common in all free-standing designs for away from wall placement to yield flat response.

 

[Nvr2manybikes]

wolf_teeth - I just ran BOXSIM with no crossover (amp straight to woofer but woofer still in same enclosure) The result is shown in the plot below.
Relative to the result in Fig 3 above there is still significantly more loss than predicted by the 5spice simulation, and they should both have the same B-S loss right? So I'm still at the same place wondering which is correct and how much power I can apply from the amp. 🙁

Rough numbers:

F (Hz)	Raw Woofer in enclosure (SPL)	Woofer through Xover in encl (SPL)	Delta (dB)
100	79.5	79.5	0
300	86	83.5	2.5
500	85	82	3
700	86	82.5	3.5
1000	86	82	4
2000	83.5	79.5	4

 

[wolf_teeth]

That actually does not look like BS loss, unless it's a very wide baffle and large woofer.

I do not know that program well at all, so it's hard to wager what is going on. System7 really is the person I see use this program the most.

 

[Nvr2manybikes]

Yeah - it's a small speaker - 7L closed box, front face is about 30cmx20cm and speaker is a 5". Putting aside BOXSIM for a second - is the approach to evaluate powers with 5spice valid? And if I believe the simulated power levels - does one size resistors for worst case at a single frequency or assume some broader input spectrum from the amp?

 

[WhiteDragon]

For this 40W rated woofer can I put in 100W like BOXSIM implies or only 50W like 5SPICE implies?

The thermal rating for the woofer is 40 watts
So in real life you cant put more than 40 watts.

Far as linear distortion in a reflex or sealed.
Depends.
Its a 5" speaker with 4mm linear travel.
So you wont get much more than 15 to 20 watts
into it before distortion starts kicking in.

100 watts and 50 watts will never happen.
pretty straight forward everyday crossover parts
will work fine.
normal wire wound 10 watt resistors be fine.
Some of the 1% resistors on PE are 12 watt
rated. be way more than enough.

5" speaker on a highpass / bandpass for
a mid. Removing the bass you can get
far more travel before linear distortion.
So then possible to exceed the 40 watt
thermal rating and have no audible distortion.

As a woofer in a 2 way.
The bass frequencies will exceed the linear travel
quicker. usually hard to exceed thermal level.
Normal listening youll know when the speaker\
has reached its limit and be audible . Which will be way way under 40 watts.
Either way 40 watts is all it is rated for.
and 10 watt resistors fine. coils dont have to be massive
huge either. normal listening wont be over 5 to 15 watts
to the drivers. Box sim will show you excursion levels.
Be the usual expected for 3 to 4mm cone travel.

 

[Nvr2manybikes]

WhiteDragon
Thanks for the guidance and for pointing me back to BOXSIM for the Xmax, there is actually a max V/Pwr for linear excursion plot I never paid much attention to before. I've included a shot of it for this set up and like you said, the problem is below 100Hz so if this were used as a Midrange it could handle much more without distortion.
As you suggested I can see how that part of the curve is not impacted by the two way crossover design and only by the box.

Exactly how do I interpret this plot though? I presume it applies exactly to the case of a single frequency being sent through the amp and getting the full power of the amp? If you had your amp set at 40 Watts does all the power ever go to a single note instantaneously in practice? (Other than maybe dropping a phono needle for instance)

As for the thermal (RMS) power dissipation rules of thumb are nice but I am really trying to figure out how to get real calculations for a given design. This is the approach I came up with, let me know if it sounds reasonable to you.

I took the SPL for various music genres from a post by Pano (I actually used the curves by Msibilia pasted below). I broke the spectrums manually into 10 octaves and then determined roughly how much power would be in each octave for 100W total power. I then convolved those numbers across the woofer or each resistor 5 spice curves - posted in original post - of (Pwr dissipated(F)/(Pwr Input(F))

From this approach I could see that the 6ohm resistor (R12) would never see more than about 5 Watts but the 3 ohm resistor (R5) would see as much as 12Watts for some types of music (Classical and Classic Rock) where Hip Hop and EDM had more power in the bass where this resistor was not absorbing much. Based on this approach I also concluded that the driver itself would see anywhere from 36W to 60W RMS for a range of music genres. So based on this an input power of 66W would be at the thermal limit for these drivers and R5 in this ckt, though of course there would be much distortion.

Is this a fair approach? Not that I intend to operate these speakers anywhere near this level, I just want to know how to determine the limits.

Thanks for your help!

 

[shadowplay62]

At 36W this driver outputs 100dB SPL over 100Hz but it also reaches the maximum cone excursion limit (thin blue line).

 

[AllenB]

I presume it applies exactly to the case of a single frequency being sent through the amp and getting the full power of the amp?

That's right... You're not going to get a direct answer to the needed power ratings for resistors, you'll need to glean the information from what you have and then experiment to find out what you need.