Scan-Speak Revelator

[blue934]

i persume the excursion graph in winisd or bassbox that i use are similar to what you are saying about unibox. my problem is knowing how much power to apply to the graph. does my 240w plate amp need 240w applied to the graph ,or, since i will never have my amp at max, should i use 50% or 75% or?

i guess i am asking do i model for max power or normal listening levels?

blue934

 

[GK]

Thanks for the replies.
So 200W is OK, good. I'll go play a bit more with unibox now

Cheers,
Glen

 

[sreten]

I model max power.
A lot of what you are saying was mentioned above in sreten's post.
I don't do it that way.
For me, there is no "normal" listening level. I like to crank it more often than most and I don't want my drivers struggling with the load.
You need to decide what you want out of the design - what will be the spl at the power level you think you won't exceed? To get an acceptable level at low frequency requires power, therefore power handling ability.

In my post #33, the projected frequency response has the speaker (driver and port) producing 104db at 20Hz at 130 watts. This is within the drivers linear excursion limit.
Is this enough or is more output needed? If your amp is only able to produce 100 watts, you don't need to worry that your driver is distorting from over excursion.

Hi,

Well that might the way you do it but it makes assumptions that
to me do not stand up to sensible analysis. They might if you are
modelling a powered high pass filtered subwoofer but they do
not stack up for wideband music signals.

You are suggesting that the bass excursion limited power handling
should exceed the amplifiers maximum rms power, which will of
course mean superb bass power handling, but there is of course
a consequence :

the 12" Peerless would work in a ~ 86dB/2.83V/1m 3 way speaker
and you are then limiting peak midband levels to 100W = 106dB.

The corollary is that speaker should / can easily handle far more
than 100W music programme (200W peak with mild clipping) if
it is designed correctly.

The related issue is the crest factor of the music programme, peak
levels to average levels, generally the higher the quality of the
music recording the higher this ratio is, though modern CD's and
radio stations are doing their best to compress every ounce of
life out of the reproduced signal, they are not my bag at all.

~ 6dB crest factor is lift music and the bass channel in a PA.
~ 10dB to 20dB crest factor is good quality non classical.
~ 20dB to 30dB crest factor is audiophile quality classical.

(Crappy CD mastering and FM radio compression ignored)

Another ~ for decent music is the signal spends 80% of the time
below 20% of the peak level, average is much lower than peak.
Unfortunately the source of the above figure did not seperate
the above into frequency bands, just analysed typical CD's.

Not taking these factors into account leads to erroneous assumptions,
especially if one likes cranking it up, what usually is the limiting factor ?

/sreten.

 

[MJL21193]

Well that might the way you do it but it makes assumptions that
to me do not stand up to sensible analysis. They might if you are
modelling a powered high pass filtered subwoofer but they do
not stack up for wideband music signals.

You are suggesting that the bass excursion limited power handling
should exceed the amplifiers maximum rms power, which will of
course mean superb bass power handling, but there is of course
a consequence :

the 12" Peerless would work in a ~ 86dB/2.83V/1m 3 way speaker
and you are then limiting peak midband levels to 100W = 106dB.

The corollary is that speaker should / can easily handle far more
than 100W music programme (200W peak with mild clipping) if
it is designed correctly.

Hi,
This is how I do it, others are free do as they like. Everything that I have built lately has each driver separately amplified, so the other drivers have no bearing on the model. There are no passive crossover components to consider either.
There are necessary compromises, and there are needless ones. I want performance. I want spl without distortion at maximum power. I do not want to settle for "mild clipping". If there is another driver (lots to choose from) that will fulfill the requirements, then I'll use that one. Modeling for max power gives you all of the story - how you use that is entirely up to what you consider an acceptable amount of distortion.

So much talk about cables and interconnects and even components such as capacitors and inductors for passive crossovers and the need to get the best or pay a high price for these things. This is for not if the weakest link in the chain (speaker) distorts during operation due to a preventable flaw. How much clipping do you consider acceptable in a power amp? How about a preamp?
Back to the speaker again, how much should you expect the tweeter to clip during operation?

The related issue is the crest factor of the music programme, peak
levels to average levels, generally the higher the quality of the
music recording the higher this ratio is, though modern CD's and
radio stations are doing their best to compress every ounce of
life out of the reproduced signal, they are not my bag at all.

Here's my perspective: The amp will not exceed its maximum linear output without distortion. If the signal drives the amp to 1% below clipping, then the speaker that is handling that signal needs to be below it's clipping threshold, otherwise it will distort.
If these incidents of high power output are brief and do not add up to a large percentage of the total musical program is immaterial to me.
If I'm going to take the time, money and effort to build it, it's going to be worth it.

 

[sreten]

Hi,
This is how I do it, others are free do as they like.
Everything that I have built lately has each driver separately amplified.

Hi,

So why say you do it differently to what I stated when
in reality what you do agrees far more with what I said
about high pass filtered amplifiers than disagrees ?

I only took your statement literally because you said you do it differently ....

/sreten.

 

[jupiterjune]

Here is perhaps another contender for the bass driver (assuming a sealed box):

Glen-

Did you run the specs for this driver in unibox? I am curious as to what the SPL predicted by Unibox for this driver, as Unibox predicted an SPL about 3dB lower for the Dayton reference sub than that predicted by the software I use.

Thanks,
JWJ

 

[GK]

Hi JJ

Yes I did. I just returned today from a work trip and I'm currently upgrading my computer. I'll redo and post up the unibox results maybe tomorrow. Can't remember the exact numbers ATM.

Anyway, I've decided to go with the 10" Scan-speaks as they work better with a smaller box volume. I might buy one of those 12" XXLS units to have a play at a subwoofer.

I'm going to make a Sonus Faber Stradivari clone(ish), but sealed. I plan to make the front baffle removable, with a pair made to experiment with (as per the original) the WWMT layout and a WMTMW layout.

I'll power them actively with seperate test amps through a digital XO to subjectively/objectively determine the optimal XO points / filter slopes / BSC / etc and then replicate passively.

I'm also interested in making distortion measurements on the midrange (two in series for less coil heating / cone excursion Vs just a single unit).

Anyway, attached below is a pic of my mobile office. Maybe some "anechoic" measurements to be made with the crane fully extended and the test speaker dangling from the pulley?


Cheers,
Glen