Parallel driver interaction in a 4-way system.

Status
Not open for further replies.
Jump to Latest
#122
SimontY said:
You seem to be getting very defensive now! These speaker guys are a tough bunch to please, and they revel in this kind of debate / conflict. Just step back and think about how to use the negative advice positively - if someone says there's a problem consider how to overcome it.

I think your idea of using one coil per driver is going to be fine. The only drawback I can see is if you need to fine-tune the value, then you need to do that FOUR times, which will be annoying. Perhaps settling for two larger 8ohm drivers really will be better, I just don't know how much you want to be pushing the amp, as it might end up sacrificing sound quality if you can't find passive components worthy of the end goal. Another idea here would be to mess with one large coil, then when you're set on the value after listening trials and measurement you can replace it with 4 coils - one for each driver, to minimise the DCR and distortion. You may have to allow for the fact that altered resistance will shift the crossover point and efficiency of the bass part.

The caps will be inexpensive as you say, it requires no debate, especially considering your budget.

The hardest part of this project might end up being the cabinet, which will want to be solid, but also inevitably large, and you need to have a good height for your mid(s) and tweeter. If chosen carefully there's no reason these can't match your woofers for sensitivity. Usually we need to pad the tweeter and mid down somewhat.

You certainly don't want to pad the bass drivers down, as damping factor is considered very important for excellent bass reproduction. And throwing power away more than you're going to have to do already is not on.

Another point to consider is baffle step correction - compensating for baffle step losses - which requires your bass section on a 3-way to be at least 3dB louder than the midrange and treble. This implies crossing over at a very considered frequency (ie. where baffle losses begin). That extra sensitivity is very handy indeed.

Simon
Click to expand...


The possible need to fine fune multiple inductors really doesn't phase me (that's only a little labour in the whole scheme of things) and I agree that the cost (and physical size - what the?) of the capacitors isn't something that should be debated at all. It's a truly trivial expense in a project of this scale.

It has been declared here that such a project is unfeasible due to the demands that a low impedance / high current speaker design puts on the crossover components. I think that this is a complete red herring.

With seperate passive crossovers for each woofer, the demands are not any different at all from a much higher impedance, conventional design that only uses one woofer. The only real demand is that placed on the speaker leads!

If I come across as defensive it is because these silly debates don't really contribute much in the end. Also, this is also quite obviously a passive speaker project. Those who want to argue about the merits of active speaker design start their own thread or do that elsewhere.

I totall agree with you comments on the cabinet design. This is something which will require a lot more study on my behalf before the project even begins.
However, a line array with 4 woofers still seems the way to go.
My amp has enough capacity to drive two such speakers connected in parallel per channel! There will be plenty of reserve with just one pair.

The issue of loudness has been brought up again. I think the HiFi merits of a competent line array reach well beyond the sole objective of obscene SPL, which is not a priority objective of mine (I have stated this already, but someone is blissfully ignoring it). I am not building a PA system. I am building a deliberately OTT HiFi system to cope with high dynamic range material.

Anyway, with a nominal impedance of a little under 2 ohms and over 400W RMS per channel to drive it, I don't think I'm going to have much of a issue WRT high SPL. 🙂

Cheers,
Glen
 
#123
SimontY said:


Ahh ok, and now I realise what you were getting at - a lower impedance speaker actually requires a smaller coil - great! I just assumed it was the other way around, as with caps.

That solves the inductor problem then. Something good could be used e.g. Mundorf "feron-core" 0.82mH has a DCR of 0.06R - a splendidly low figure. And not a problem on the budget if you only need two: maybe around $130AUD.
Click to expand...


That is still a little steep!

Seriously, if I was after a really low DCR inductor, I would just wind my own.

Cheers,
Glen
 
#124
Actually 0.06 ohms DCR is an extremely good figure for a crossover coil.

Remember that the DCR of speaker drivers is actually a parasitic parameter. It does not contribute in any way to sound output. Voice coil DCR is high because the coil must be small and lightweight and the current must travel around the field gap many many times in order to get a decent BL factor and some efficiency, thus imposing a limit on magnet wire thickness. Yes, voice coil RDC in speakers is absolutely parasitic, it works exactly like an external dummy resistor connected in series with a speaker having an ideal 0 ohm voice coil.

Sorry again if you expended too much time improving damping factor on your amplifier. Real damping factor is almost always in the 1 to 2 range regardless of amplifier performance. Low amplifier output impedance is only required because speaker impedance may be very uneven across the audio band.
 
#126
Eva said:
Actually 0.06 ohms DCR is an extremely good figure for a crossover coil.
Click to expand...


I was refering to the price.


Eva said:
Remember that the DCR of speaker drivers is actually a parasitic parameter. It does not contribute in any way to sound output. Voice coil DCR is high because the coil must be small and lightweight and the current must travel around the field gap many many times in order to get a decent BL factor and some efficiency, thus imposing a limit on magnet wire thickness. Yes, voice coil RDC in speakers is absolutely parasitic, it works exactly like an external dummy resistor connected in series with a speaker having an ideal 0 ohm voice coil.

Sorry again if you expended too much time improving damping factor on your amplifier. Real damping factor is almost always in the 1 to 2 range regardless of amplifier performance. Low amplifier output impedance is only required because speaker impedance may be very uneven across the audio band. [/B]
Click to expand...


No, you should be sorry for wasting my time by deliberately derailing the topic of this thread by giving me a lecture about something that was never under contention.

And you still have yet to explain why this project is rendered unfeasible due to the limitations of the passive crossover components (which are no less an issue in any other passive speaker design) as per your previously unthought through comments.



tinitus said:
it sound too much like a "punkrk" dejavu to me

Its a complete waste of money and has nothing to do with reaching fore upmost quality, I wont participate any further

wish you luck
Click to expand...


Thank you!
 
#127
Back on topic

I'm going to attempt to steer this thread back in the desired, productive direction – that being WWWWMTM design – not active speaker or amplifier design PLEASE!

A line array speaker of this type, as far as I can tell, appears to be a perfectly feasible solution for a low impedance, high power, high-fidelity system.

This seems to be evidenced to some degree by the fact that there are some very talented and capable speaker builders out there putting a great deal of resources into such speaker systems.

I’ve just been politely accused of being some kind of “punkrk” (whatever that is) idiot for even considering such a thing, but I’m not sure how that criticism holds up for the obviously talented guy building one such system starting at post 35 in this thread (link previously provided by Steve/BlueWizzard)

http://htguide.com/forum/showthread.php4?t=30645


Now there is still so much ground to cover, I’m really don’t know which questions are the most pertinent to ask in order to further the WWWWMTM design discussion, but if anyone with more knowledge on the topic can oblige I’d really appreciate it!

Cheers,
Glen
 
#128
Hmmm........

Certainly some pukrokr deja-vu, he did not know anything about
loudspeakers and the trouble with him AFAICT he still doesn't.

If 2 ohm nominal speakers are wanted, that is easy but suitable
driver choices will be restricted compared to 4ohm / 8 ohm.

Controlling senstivity is also easy, especially with isobaric bass loading.

What is very difficult is making a high quality design to the correct
requirements without getting it horribly wrong, even though to
the technically challenged there seems to be nothing wrong with it.

2 ohms nominal means ~ 400W program (not RMS) power handling.
It is quite a lot but relatively do-able, 200W program (i.e. typical
music with a small % of clipping on peaks) can be handled by e.g. :

http://www.rjbaudio.com/AlpheusMkII/alpheusmkii.html

Obviously the programs crest ratio, peaks to average, will determine
whether the 200W / channel blows a driver or not, FWIW the LS3/5A
was shown years ago to be able to handle 1kW transient peaks.
(obviously not in the bass)

So .....

Budget ? Size ? Quality Expectations ? Purpose ? (not what is
it for, what is it supposed to do ?), Cabinet building skill ? etc .....

Nothing particularly unusual is needed for a 2 ohm speaker ....
(Except big inductors ....)

🙂/sreten.
 
#129
just calculated the DCR and inductance of a big coil.
for 0.75mH and DCR~=0r065 requires 1.91kgs of 3.1mm diam copper wire wound onto a 66.8mm diam 31.8mm wide former resulting in a coil with an overall diameter of 124.1mm using 94.8Turns.

2kilos of enameled copper wire will cost about £20.
 
#130
AndrewT said:
just calculated the DCR and inductance of a big coil.
for 0.75mH and DCR~=0r065 requires 1.91kgs of 3.1mm diam copper wire wound onto a 66.8mm diam 31.8mm wide former resulting in a coil with an overall diameter of 124.1mm using 94.8Turns.

2kilos of enameled copper wire will cost about £20.
Click to expand...

Hi,

Or alternatively slap a proper core in the inductor.

It is simple, a 2 ohm speaker will need ~ twice as big inductors as 4 ohm.
So what ? the same is true for 4 ohm versus 8 ohm, no big deal.

🙂/sreten.
 
#131
sreten said:
Hmmm........

Certainly some pukrokr deja-vu, he did not know anything about
loudspeakers and the trouble with him AFAICT he still doesn't.

If 2 ohm nominal speakers are wanted, that is easy but suitable
driver choices will be restricted compared to 4ohm / 8 ohm.

Controlling senstivity is also easy, especially with isobaric bass loading.

What is very difficult is making a high quality design to the correct
requirements without getting it horribly wrong, even though to
the technically challenged there seems to be nothing wrong with it.

2 ohms nominal means ~ 400W program (not RMS) power handling.
It is quite a lot but relatively do-able, 200W program (i.e. typical
music with a small % of clipping on peaks) can be handled by e.g. :

http://www.rjbaudio.com/AlpheusMkII/alpheusmkii.html

Obviously the programs crest ratio, peaks to average, will determine
whether the 200W / channel blows a driver or not, FWIW the LS3/5A
was shown years ago to be able to handle 1kW transient peaks.
(obviously not in the bass)

So .....

Budget ? Size ? Quality Expectations ? Purpose ? (not what is
it for, what is it supposed to do ?), Cabinet building skill ? etc .....

Nothing particularly unusual is needed for a 2 ohm speaker ....
(Except big inductors ....)

🙂/sreten.
Click to expand...



As stated previously, budget is ~4k AUD.
Purpose, OTT HiFi
Cabinet building skills - not an issue.

That Alpheus still isn't a WWWWMTM.

Is the concept of a line array intrinsically flawed? And are there any established WWWWMTM or WWWWMT HiFi designs of particular merit?

I'll be able to continue this discussion tomorrow.

Cheers,
Glen
 
#132
Have you looked at Pro Sound drivers to handle the power?? You will need some hefty drivers to absorb that kind of power and survive anything beyond short peaks. You could easilly parallel up two pairs of 8 ohm drivers for a 2 ohm load that can handle over 1000 watts longer than you could. You would have your low end limited to say 40hz depending on the drivers chosen. If you can get drivers with 4 ohms nominal you could get it closer to 1 ohm.

Rob🙂
 
#133
G.Kleinschmidt said:




As stated previously, budget is ~4k AUD.
Purpose, OTT HiFi
Cabinet building skills - not an issue.

That Alpheus still isn't a WWWWMTM.

Is the concept of a line array intrinsically flawed? And are there any established WWWWMTM or WWWWMT HiFi designs of particular merit?

I'll be able to continue this discussion tomorrow.

Cheers,
Glen
Click to expand...

Hi,

I presume you mean OTT hifi quality.

4 x 8" 8 ohm woofers or 2x10" 4 ohm woofers - there not a lot in it.

I've always liked the basics of this design :

virgooic1.jpg


http://www.stereophile.com/floorloudspeakers/147/index.html

Force cancelling of the bass drivers is an appealling idea.

MTMing the front is easy enough. So is doubling up the bass drivers.

some numbers :

Say the TB titanium drivers and a ribbon for the MTM, as used here :

http://www.htguide.com/forum/showthread.php4?t=28728

The midrange will be nominally 85-87 dB/2.83V (4 ohm)
(See Zaphs specs for the TB at Zaphaudio .com)
This is because the MTM must include the baffle step : 4dB to 6dB.

http://www.madisound.com/catalog/PDF/neocd3.pdf

The treble will be padded back to the mid and high impedance,
consequently actual "power handling" will be very much higher.

This leaves the bass section which depending on the amount of
BSC needs to appear to be around 91dB/2.83V, that is into
half space whilst in fact it is full space and around 85dB.
Lets say the bass is 90dB/2.83V to allow mid/treble adjustment.

Enter the laws of physics regarding box size, bass extension
and efficiency, noting you need decent cone area and excursion
for bass. 4 8ohm drivers would need to be 78dB/2.83V each.
Hmmm .....
Then again 4 isobaric'd 8ohm drivers would need to be 84dB/2.83V
each, or in other words effectively be 2 x 84dB/2.83V 4 ohm drivers.

We are talking drivers for subwoofers here, not bass/mids - they
do not have these sort of parameters optimised for low bass in
relatively compact cabinets.

84dB/2.83V 4 ohm is ball park for 8"/ 10" subwoofers for compact cabinets.
Here you have to be careful with specs as dB/W and dB/2.83V
get very mixed up, we want 84dB/2.83V 4ohm i.e. 81dB/W/1m.

http://partsexpress.com/pdf/264-855s.pdf seems to fit, cannot be sure.

The above describes a speaker of typical senstivity but 2 ohms
nominal where the 2 ohms is effectively used for really good
bass extension and level for the speakers cabinet volume.

The relative numbers do not change for a 3-way. For sure if you
use 4 8 ohm bass drivers in parallel you end up with + 12dB
sensitivity, a box 4 times as big and no more bass extension
than one driver in a quarter size box, obviously goes louder.

You also run into problems with quality mid drivers and tweeters
having sufficient senstivity and power handling to keep up with
the high senstivity high power relatively no bass bass end.

🙂/sreten.
 
#134
But then, with 2 ohms in the bass range (from four 8 ohm woofers), 4 ohms in the midrange (from two 8 ohm mids) and a 8 ohm tweeter, peak current demand on the amplifier will be barely 15A peak when playing music (may reach 18A or 20A briefly when playing bass heavy stuff). This is quite far from the 120A capability and goal.
 
#135
Eva said:
But then, with 2 ohms in the bass range (from four 8 ohm woofers), 4 ohms in the midrange (from two 8 ohm mids) and a 8 ohm tweeter, peak current demand on the amplifier will be barely 15A peak when playing music (may reach 18A or 20A briefly when playing bass heavy stuff). This is quite far from the 120A capability and goal.
Click to expand...


Hi,

Whatever. It is still a 2 ohm nominal speaker and the same arguments
apply to 4 ohm 3-ways. As 2 ohm nominal impedance worst case
loading on the amplifier if you do not mess up the crossover will be
equivalent to a 1R to 0.7R resistive load.
Which implies peak currents of the order of 40A to 60A.

🙂/sreten.

The midrange could be ~ 2 ohms nominal if desired, but what is the point ?
 
#137
A WWWWMTM configuration is definitely looking like a practical and feasible way of reaching my goal. I think with this configuration a spectacular pair of speakers that would really take advantage of my OTT amplifiers capabilities could be pulled off with enough education and attention to detail.
Click to expand...

Excuse me for being not negative, but this will be very easy to do.

The only drawback with 8" drivers is that it will be bit tough to get a sealed box with a closed box frequency and F3 down to 40Hz. But that is righteously deep bass for a closed box, you may be more than happy with that.

Scanspeak used to make some drivers capable of that, but they seem to be currently unavailable. I don't know if they are discontinued or not.

My subs use 8" drivers in a sealed box, with a box freqeuncy of around 38Hz. They shake the house. And only 2 drivers per sub!

I've spent the last few weeks trying to design a WWWWMTM that goes down to low 30's for Fcb and F3, but still has 90dB@2.83V/m sensitivity. Very hard to do without getting 2ohm impedence dips. But it is oh so easy if you allow low impedence.

JJ
 
#138
Cheers guys.

Lots of good information to slowly absord here 🙂


Eva said:
This is quite far from the 120A capability and goal.
Click to expand...

🙄
Not sure how many times I've indicated that I'm not interested in building a single pair of speakers to push the amplifier anywhere near its limit.


Cheers,
Glen
 
#139
Here is quickie of how it might work......


C1 = 7.543 µF, Polypropylene, 0.00588 ohms
C2 = 55 µF, Polypropylene, 0.00209 ohms
C3 = 6.295 µF, Polypropylene, 0.00642 ohms
C4 = 400 µF, Polypropylene, 0.00137 ohms
L1 = 0.274 mH, Air Core(#16), 0.282 ohms
L2 = 2 mH, Air Core(#16), 0.477 ohms
L3 = 0.5 mH, Air Core(#16), 0.28 ohms
L4 = 1.5 mH, Air Core(#16), 0.434 ohms
BPGain = 3.186 dB

Tweeter
L-Pad:
Rp1 = 1 ohms
Rp2 = 12 ohms

Midrange
L-Pad:
Rp1 = 2 ohms
Rp2 = 100 ohms

JJ
 

Attachments

  • the kleinschmidt part2.jpg
    the kleinschmidt part2.jpg
    56.6 KB · Views: 287
Status
Not open for further replies.
Top Bottom