WinISD and sensitivity ratings?

The 3015 sensitivity I believe is measured out of cab. Check the data sheet foot notes to be sure. Yhe box controls the cone, so less sensitivity might be a compromise for better mechanical power handling.

 

Data sheet charts show response with the driver literally mounted in a wall, so they're only accurate above 150 Hz or so. Data sheet SPL figures are usually an average across a wide portion of the cab bandwidth, where as a bass player you're main concern is in the lows. WinISD is what you should use to compare drivers below 250 Hz, data sheets above that.

 

15PR400 manufacturer rating sensitivity 1W/1m 99dB
predicted sensitivity (BassBox pro)

response chart modelling 1Watt/1Meter


Kappalite 3015 manufacturer rating sensitivity 1W/m 100.8dB
predicted sensitivity (BassBox pro)

response chart modelling 1Watt/1Meter




15MI100 manufacturer rating sensitivity 1W/1m 100dB
predicted sensitivity (BassBox pro)

response chart modelling 1Watt/1Meter


Not sure the MI100 was really suitable for (modern) bass guitar application, except for coffee house gigs where there is no need for deep lows or if amp power is a limiting factor on a larger stage.

 

I wrote my own speaker modeling program, starting from what I thought was the underlying physics principles. (You can find it at my "non commercial" web page, but there are much better programs for actual useful speaker design). So far so good. My program consistently predicts sensitivity higher than WinISD. I believe this is due to a difference in how my program and WinISD convert "1 Watt" into a corresponding input voltage. Computing sensitivity the way I think WinISD does it, I get the same results. But I like my method better.

The difference hasn't been big enough for me to worry about in my designs, and is a detail where our respective approaches are both documented, so an expert could weigh in on whose method is preferable. But it points to possible modeling differences between programs, and meanwhile, there's also the measured curve.

 

As you have seen, there are different ways to conpute sensitivity, both power and voltage will give the same value if nominal impedance is used but different values if actual impedance versus frequency is used.

These days, most manufacturers use the nominal impedance method because the difference is usually pretty small.

 

That's what I do in my program. My main rationale is that I don't really care about power, but about the voltage driving the speaker terminals, because voltage gain is what's constant with frequency when an amp is flat.

At least, that's a good enough model for me to think my way through my own DIY speaker designs.

Output current just goes along for the ride. Until it doesn't, but that's your job to deal with.

 

WinISD electro-mechanical parameters has no entry for Z, only for Re and Le. It uses Re to calculate sensitivity, so the voltage drive figure is going to be lower than expected. For the 3015, for instance, it calculates one watt as a 2.3v signal. It's a program quirk that you get used to. I always change the entry in the signal tab to the appropriate voltage, 2.8 for 8 ohms, 2 for 4 ohms, etc.