calculating power of each Driver.

Hello,

I was wondering how calculations are made for getting power rating of each driver used in multi-way system. Like let me say. I have 2 way system rating 10w with Fq response of 120Hz to 20 Khz . Well the design is completed to give a perfect flat response in the mentioned frequency range. To achieve the response I use crossovers , attenuator, Impedance equalizer, this all stuff absorb some energy. so what is the calculation for this. Also after using the above attenuator and crossovers the power rating of drivers required would not be same say 10w for each. I guess this depends on where the crossover point is set. so finally if I say my system can handle 10w continuous over the given range. then possibly the mid-woofer I would need to use would only be 7.5W where the tweeter can just be 3W. I mean that was just for example.

So is there any software available for the calculations, or can anybody actually help me with actual calculation method and equations.
 
In a typical situation, you would get a driver that's rated for continuous power handling that's greater than what you need. Then you would design your enclosure and crossover to make sure that with the max voltage you intend to apply to the speaker, the excursion will not exceed the limits of the driver. If you'd like to get more detailed than that, you'd calculate the voltage versus frequency at the driver for a given input voltage to the crossover network. Then since you'd know the impedance versus frequency of the driver, you could calculate the real part of the impedance and then calculate the power dissipated in the real impedance. Then you could calculate an RMS dissipated power by taking the square root of the quantity sum of the squared power divided by number of power data points. A pseudo excel formula would be RMS=sqrt(sumsq(P1:px)/count(P1:px))

But no, I don't imagine most people calculate things this way. If you use the simple method I outlined, you'd be fine in most cases.
 
John, Richie, Casull & Thosuk,

Thanks for your reply,

I am went through the link that you provided. Also thinking deep on suggestion given by John.

I am thinking If its possible to Simulate this using Orcad- Pspice.

A detail of each component would have to be entered for more precise calculation. say, Exact equivalent circuit of speaker.Then the crossover network. That should include dc resistance of coil used.

cause i think there is no software readily available that can do the job for us. Although I am not getting stuck on the calculations. but its always good to know. how much power is wasted? And actual power that each driver needs to handle.

Can I get some help or direction in getting the simulation done on orcad - pspice.
 
Hello,

I was wondering how calculations are made for getting power rating of each driver used in multi-way system.

I think you are asking how to measure or calculate how much of the input gets to each driver in a system.

I've found the easiest way is to feed pink noise into the system input (even frequency shaped noise such as the EIA shape used for power testing) and then measure, with a true RMS meter, the voltage across each driver.

So 10VRMS input might give 8VRMS on the woofer and 3VRMS on the tweeter. V sqrd over R will give the approximate power. You can also use this method to see if your power resistors are good for some rated system input.

You can do the same thing with voltage curves across each unit, but you must consider not only the network drop but also the bandwidth across that unit.

David S.
 
AES E-Library: Power Transmission Through Crossover Networks
This has some charts you can calculate from.

But not free sorry to say. If you had a particular frequency/slope in mind I could dig out my hardcopy of his paper.

The Achilles heels hindering your entire question are:
- Good research on the spectrum of music simply has not been done, so everything is rather a guess. (But the spectrum approximation in that paper seems at least vaguely reasonable).
- Manufacturers simply don't quote thermal-only power handling features.
:(
 
I used to have a program that ran in the DOS shell that may have helped here. You would hook up your driver to the sound card, enter it's specs and then engage the program.

Now I don't know how accurate it was but, it gave you the speakers true RMS handling at specific frequencies which is why some very good speakers die from being run by lower than their rated power. If you ran it through the speaker while it was hooked up to a crossover, you of course got different results.

A 1000 watt RMS sub would sometimes only be able to actually handle 150 watts RMS at 40Hz. The loss of power handling was always astonishing and some speakers amazed me with what they would handle. I'm not sure if the testing standard has changed but as far as I know, the accepted method of rating a speaker is by using a 1kHz sine wave to determine it's RMS wattage tolerance.
I would have to think that by now though, companies are using a 100Hz sine to generate effective ratings for subwoofers as they will be used in real life situations rather than deceptive lab conditions. It would make sense to use a frequency in relation to the speakers intended purpose.

Anyone happen to know the program?
This was a good 10 years ago, I've no idea what happened to it.
 
AES E-Library: Power Transmission Through Crossover Networks
This has some charts you can calculate from.

The Achilles heels hindering your entire question are:
- Good research on the spectrum of music simply has not been done, so everything is rather a guess. (But the spectrum approximation in that paper seems at least vaguely reasonable).
- Manufacturers simply don't quote thermal-only power handling features.
:(

Charts and tables will never be of much help because every crossover will be different in terms of response shape, attenuation and bandwidth that it delivers to the driver. You either have to do a full network simulation or measure an actual system to see what power is delivered to a driver from a given power input.

I've seen a number of curves of musical spectra. You have to choose a music type and whether you want average spectrum or peak spectrum, but the data is available.

You're right though, in the end data from manufacturers on power handling is "highly suspect".

David S.
 
Thanks to all for your replies.




I'm not a big spice user so I can't help there, but I can tell you it isn't too tough to do in excel. Then there's no need for an equivalent circuit for the speaker - just import an impedance curve that you measured.

As far as the power dissipated by the crossover components, you mainly need to worry about the resistances of the components.


Dear John

I hope I have well understood you. Your idea of importing the measured impedance is good and that will help us in calculating power over the frequency range.But, we are looking for power absorbtion by each circuit module. Say, LPF Crossover,HPF Crossover,Equilization networks attenetuors and finally what each driver delivers.
This will help us to understand the effeciency of our design, final power that is delivered to drivers. We also need to get equivalent of speaker,because this will help us getting the load response vs frequency.
Also this tool if developed would give us the need of actual power rating of drivers required even before we procure them. also that would greatly improve our comfort level.And ofcourse even steps to calculate MFBF can be taken if required.


I think you are asking how to measure or calculate how much of the input gets to each driver in a system.

I've found the easiest way is to feed pink noise into the system input (even frequency shaped noise such as the EIA shape used for power testing) and then measure, with a true RMS meter, the voltage across each driver.

So 10VRMS input might give 8VRMS on the woofer and 3VRMS on the tweeter. V sqrd over R will give the approximate power. You can also use this method to see if your power resistors are good for some rated system input.

You can do the same thing with voltage curves across each unit, but you must consider not only the network drop but also the bandwidth across that unit.

David S.

Dear Dave

Idea of measuring is really good.This wont bother us about most of parameters. but when we are building multiple speakers we need to do this exercise. Also the idea of estimating power and behaviour even before we procure components/drivers wont be achived here.
what I feel is its much better to get the software/ tool made and put extra efforts in doing this. as this is one time. for rest of simulation, its just few click away and just no physical efforts and the overall project cost and time can be pullde down considerably.


This might be helpful.
Bullock and White's Home Page
Go go bottom of page and download passfil.exe (Passive Crossover Component Stress Analyzer).
There's other Bullock and White stuff as well.

Dear Thosuk,

That tool is really good. But I was looking for more.changing component values, aligning of different filters altogether. And some parameters regarding components which are important and needs to be considered are dc resistance of the components. Actual characterstic of drivers.

Hey and Voilance pls try to find the program.
 
Charts and tables will never be of much help...

I've seen a number of curves of musical spectra. You have to choose a music type and whether you want average spectrum or peak spectrum, but the data is available.

You're right though, in the end data from manufacturers on power handling is "highly suspect".

David S.

1) The paper showed that actually, the amount of transmitted power was pretty much the same whether you used 6, 12, 18 dB. The steeper filters cut more below the crossover point, but actually cut less above the crossover point because the "knee" of their response curve is steeper (if you plot this not in dB but linear, since power is linear not dB, you can see it kinda averages out the same)
--> Unless your crossovers have some huge odd boost or cut, the results will be "close enough"
because...

2) Actual power to each driver will vary wildly depending on each song. All this simulation is only an approximation, so it is kind of pointless to try and be TOO anal striving for "accuracy" which in the end is not real music anyway.

3) However, I laud the OP's goals of assessing the load on the passive components to be sure they are up to snuff. And if they want to make a whole simulation, it doesn't hurt.
BUT, I might suggest running such simulations with a variety of music spectra, and also looking at the peak vs. RMS issue.