Cones vs Compression

cilla.scope wrote:   out they pop with a 4 x 6K in a 3U box ...

tallmike wrote: cilla.scope wrote:   out they pop with a 4 x 6K in a 3U box ...   oh jeez

Hehehe :) .. well you know, just saying .. trouble with class D is there are so many design options ...

Lets take a *completely*  hypothetical design scenario :)

 their are 4 real areas to consider in a class D amp,
1) PSU
2) switching speed
3) output topology
4) output filter

the rest is failry standard control and protection that anyone who has designed amps or servo drives will be familiar with.

big switchmode PSU is a failry mature technology theses days, it shouldnt cause too many issues

switching speed is coming up, new devices around that are faster and more efficient .. this means things are gettign smaller!

output topology is a big area ... so many to choose from .. my fave though is grounded bridge ... you take a standar "full bridge" driver , and ground one side, then let the psu float up and down. The downside is you need 1 psu per channel, but it has other benefits ... its still bridgeable, you don't get rail pumping like you do in half-bridge designs.

Having 1 psu per channel means it is sensible to build the channels as "blocks" ... drop 2 into a case and voila .. a 12K ...  drop 4 into a case and .. there you go .. a 4x6K .. bridgeable as 2 x 12K

now ... heres the thing .. if you put a pair of 6K modules in a 2U case, they dont quite fill it vertically .. but they pretty much fill the floor area ... basically, each module ends up being half a rack wide ...   you cant just stand them edge on and put 4 in a box, as the box ends up too tall, and you dont cover all the floor ... so what to do .. one idea was to effectively folding them in half ... PSU on one side, amp on the other in a box 3U high, and 1/4 of a rack wide ... when in 4 channel mode, and 2U high and 1/2 a rack wide with the PSU and amp side-by-side in 2 channel mode .. but that would have meant different modules for 2 and 4 channel amps ..  so on to plan B

Plan B is to make the module 1/2 a rack wide and sort of L shaped in section .. half od i 2 U high (with the caps and Big Stuff in it ... half of it 1U high with the "other bits" ... in 2 channel mode, you get 2 modules on the bottom of the amp, in 4 channel mode you get 2 on the bottom, 2 on the top ... with the 2 U bots of the upper amps sitting over the 1U bits of the lower amps giving a total of 3U ..  all the connectors  could be mounted on the modules themselves .. picking up their audio inputs from the front panel, which has the DSP and volume controls.

The big advantage of this is that you can make 4 or 2 channel units up from the same modules, plus, by building 2K, 4K and 6K modules in the same footprint, give people a "mix and match" solution .. and if you just fill your racks with 6K units, and have a failure, it allows you to swap out modules rather than whole amps ..

Of course this is totally hypothetical ... I've not seen any solidworks modles for this, and the metalworker up the road hasnt just delivered half a dozen module boxes and the 2 channel case. That would be just crazy talk

What  do you guys think of this:

cilla.scope wrote: tallmike wrote: cilla.scope wrote:   out they pop with a 4 x 6K in a 3U box ...   oh jeez   Hehehe :) .. well you know, just saying .. trouble with class D is there are so many design options ... Lets take a *completely*  hypothetical design scenario :)  their are 4 real areas to consider in a class D amp, 1) PSU 2) switching speed 3) output topology 4) output filter the rest is failry standard control and protection that anyone who has designed amps or servo drives will be familiar with. big switchmode PSU is a failry mature technology theses days, it shouldnt cause too many issues switching speed is coming up, new devices around that are faster and more efficient .. this means things are gettign smaller! output topology is a big area ... so many to choose from .. my fave though is grounded bridge ... you take a standar "full bridge" driver , and ground one side, then let the psu float up and down. The downside is you need 1 psu per channel, but it has other benefits ... its still bridgeable, you don't get rail pumping like you do in half-bridge designs. Having 1 psu per channel means it is sensible to build the channels as "blocks" ... drop 2 into a case and voila .. a 12K ...  drop 4 into a case and .. there you go .. a 4x6K .. bridgeable as 2 x 12K now ... heres the thing .. if you put a pair of 6K modules in a 2U case, they dont quite fill it vertically .. but they pretty much fill the floor area ... basically, each module ends up being half a rack wide ...   you cant just stand them edge on and put 4 in a box, as the box ends up too tall, and you dont cover all the floor ... so what to do .. one idea was to effectively folding them in half ... PSU on one side, amp on the other in a box 3U high, and 1/4 of a rack wide ... when in 4 channel mode, and 2U high and 1/2 a rack wide with the PSU and amp side-by-side in 2 channel mode .. but that would have meant different modules for 2 and 4 channel amps ..  so on to plan B Plan B is to make the module 1/2 a rack wide and sort of L shaped in section .. half od i 2 U high (with the caps and Big Stuff in it ... half of it 1U high with the "other bits" ... in 2 channel mode, you get 2 modules on the bottom of the amp, in 4 channel mode you get 2 on the bottom, 2 on the top ... with the 2 U bots of the upper amps sitting over the 1U bits of the lower amps giving a total of 3U ..  all the connectors  could be mounted on the modules themselves .. picking up their audio inputs from the front panel, which has the DSP and volume controls. The big advantage of this is that you can make 4 or 2 channel units up from the same modules, plus, by building 2K, 4K and 6K modules in the same footprint, give people a "mix and match" solution .. and if you just fill your racks with 6K units, and have a failure, it allows you to swap out modules rather than whole amps .. Of course this is totally hypothetical ... I've not seen any solidworks modles for this, and the metalworker up the road hasnt just delivered half a dozen module boxes and the 2 channel case. That would be just crazy talk

pfly wrote: I think that the modern 6" or 8" mid horns that are often preferred against large compression drivers, are suprisingly close to compression drivers, tigh compression, high tech phase rings & plugs, really low xmax, high fs and some manufacturers use drivers that have sealed chassis. So it is mainly paper vs. metal thing?

So why then, if they are so similar? Distortion? What's the REAL difference?

Anyone...

pfly wrote: I think that the modern 6" or 8" mid horns that are often preferred against large compression drivers, are suprisingly close to compression drivers, tigh compression, high tech phase rings & plugs, really low xmax, high fs and some manufacturers use drivers that have sealed chassis. So it is mainly paper vs. metal thing?

it would certainly seem so.
if you take a look at some of the hifi bods they would probably argue that the aim of using a metallic or more importantly rigid diaphragm is to shift break up resonances out of the human hearing range, and this is true when loaded into free air.
 
however when the stress is applied in a high compression ratio horn where the diaphragm is meant to be more or less constant excursion the diaghrpam can be stressed by uneven (non conical) horn loading of the diaphragm and caused to aplly force unevenly on the diaphragm this causes break up. with a paper cone the malleability allows any break up to be somewhat tempered the stiff compression driver doesn't.

in reference to line array horns they are the worst offenders for compromising on an ideal horn in order to achieve a constant coverage, as such they will apply the greatest uneven loading and detriment the compression drivers sound the most.

on the plus side new developments in materials tech (such as bms) means the break up in comp drivers might be less harsh in the future but it still comes down to the compromise between a perfect horn and audience coverage.