Gestating large sensitive passive 15" augmenting subs

[Upfront]

Russell, it was very refreshing to have you and Linc down yesterday. Thank you (both) for all your advice and discussion. Nice pic from your phone, too.

Thanks for having us Brian. Always a pleasure listening to your system, although briefly! Really enjoyed just sitting and having a chat. My place next time. Will PM you when I work out a time and date.
Thanks again mate.

[davewantsmoore]

4 hours ago, BioBrian said:

Also, it kept changing the 2.83 V back to 2.8, so I've let it.

I don't use this version of winisd (I'm stuck on the older one).... but I suspect it is just rounding what it is displaying to x.x

 

 

[davewantsmoore]

On 28/04/2019 at 6:54 PM, BioBrian said:

Another issue that's come up since ordering is from reading a comment by @almikel, also I think from 2015, that has pointed to the SB "L" having an Le of 2.1 mH, and the AE 5 times lower, this having a significant effect on lower frequency response. My error here is that I had thought 2.1 was pretty low.  Not sure of the implications here.

It's complicated.

 

The inductance slows the change in current in the voice coil.    This means that the current can't change quickly .... this means the driver can't play high frequencies....  This is why an inductor, rolls off higher frequencies.

 

The other, and more problematic issues .... is that as the voice coil moves in and out ..... the amount of inductance (the Le parameter) in the driver varies.    If you looked at the impedance of the driver vs frequency .... while the driver was moving, the impedance curve would wobble around.

 

This created non-linear (harmonic and intermodulation) distortion .... which increases with increasing excursion.

 

It's one of the reasons why I said it's expected the AE driver would work better at high power (ie. high excursion)

[davewantsmoore]

5 hours ago, BioBrian said:

My design criterion is to get max volume (efficiency) at 20 Hz, so the boxes/vents have been tuned to achieve this

"Poor" is perhaps being a bit of a doomsayer.

 

This issue you'll face is that:

 

The small vent (making it bigger will make it too long) causes turbulence and compression...  that winisd doesn't model.

 

You will need multiple filters to make the driver behave.

• Perhaps - a high pass (otherwise there's a theoretical risk of damage at very LF, even at home levels)
• A filter in band to cure the rolloff stating below ~200Hz.    In my sim I just did, I was able to get away with a simple 6dB/octave filter.
• A low pass filter

All of them will require enormous values.... not to mention the bass is still uncorrected for any issues in room.

 

There are a lot of reasons to consider if a sealed box will work.   Typically you would only avoid a sealed box if you wanted to lower power and/or excursion for your desired max SPL....  but with drivers like these, they are within their comfort zone.

 

With the AE driver you only need 120w and half of xmax to get to 105dB@ ~20Hz

[BioBrian]

OK what about this.

 

SB42FHCL driver. 320 l box. 150 x 348mm vent (dare that to chuff!). 36W input. 60Hz LP filter (imagine around 18mH C-coil (rated to 2000W), and 100-300uF parallel cap as 2nd Order:

 

[davewantsmoore]

Personally, I'd want a steeper cutoff.    (It's too loud at 200Hz)

 

... but more so, the idea of putting 36w (perhaps even double that) through a huge inductor and a capacitor... on the way to my speaker driver - doesn't seem like a very good plan.

 

People argue about the distortion these components create even when they are much smaller, and the energy transfer is much lower.    If we accept the views (I'm not 100% sure) that there is something to these people claims  (for example, different components sound different) .... then using them at all - let along in this situation - seems like a bad strategy.

 

Quote

dare that to chuff!

It's only 6 and a bit litres.   

[almikel]

On 28/04/2019 at 6:54 PM, BioBrian said:

Another issue that's come up since ordering is from reading a comment by @almikel, also I think from 2015, that has pointed to the SB "L" having an Le of 2.1 mH, and the AE 5 times lower, this having a significant effect on lower frequency response. My error here is that I had thought 2.1 was pretty low.  Not sure of the implications here.

Quite a while since I've been back here...other priorities...but I think @davewantsmoore has covered it - lower inductance affects the upper end, not the lower end...

 

...but on my previous investigations of the Acoustic Elegance range of drivers, if looking to produce down to 20Hz without DSP, and you can manage a large box, the IB15 was the only driver I ever found that could get to 20Hz in a sealed cabinet without EQ - it was a big box though (around 240litres) - not surprising for a driver designed for Infinite Baffle alignments.

 

Around the same time I went to the darkside and started using EQ with sealed cabinets, so I never tried the IB15s - I ended up with TD18s instead - tiny boxes and loads of EQ - but requiring a sub underneath.

The IB15s are proper sub territory, but need a big box or Infinite Baffle...

...in your scenario have you considered Infinite Baffle?

Just a thought...

 

cheers

Mike

 

[BioBrian]

6 minutes ago, almikel said:

have you considered Infinite Baffle?

Thanks for your post, and welcome back! Hope those pesky distractions aren't too disruptive for too long.

 

Interesting about the IB15. I didn't follow through with the idea of this in a vented box. A little swayed by the presence of a pair of used TD15Hs for sale locally.

 

I was going to respond to one of Dave's posts about baffle shape, which would elicit much the same words as this:

 

My intended wall-ceiling box placement, with a 45 degree angled baffle, would effectively be an infinite baffle - ie no baffle step loss, in fact I'd expect a net gain over IB (on a flat wall) - approaching "corner placement" - huge room gain.

 

But IB implies a virtually infinite air space behind the baffle, doesn't it? It's all looking fairly similar though - my vented box size is 320 litres, so I'm guessing they won't get many Watts thrown at them.

[BioBrian]

13 hours ago, davewantsmoore said:

Personally, I'd want a steeper cutoff.    (It's too loud at 200Hz)

 

... but more so, the idea of putting 36w (perhaps even double that) through a huge inductor and a capacitor... on the way to my speaker driver - doesn't seem like a very good plan.

 

People argue about the distortion these components create even when they are much smaller, and the energy transfer is much lower.    If we accept the views (I'm not 100% sure) that there is something to these people claims  (for example, different components sound different) .... then using them at all - let along in this situation - seems like a bad strategy.

 

It's only 6 and a bit litres.   

The 105 dB is also a little loud!!   But yes, I could do the same as my other SB42 woofers (vented 240 l) - use a 4th order passive. Easy. I'd call that fine-tuning, and I suspect room-tuning will predominate those decisions.

 

Not sure that the parallel capacitor sound 'thing' is an issue with large bass drivers - much more a tweeter issue. Happy to do some tests. There is a new generation of electrolytic caps, but Jantzen Cross-caps aren't that much dearer (I have a couple of 150 uF here).

 

Huge inductors?  No idea how they affect sound.  Not at all, from my listening. From all the emotive stuff on "damping factor" threads, I have sort of gleaned that R is the only relevant issue there, and these C-coils have almost none. Thick, short wire. Phase and group delay may be important, but that's unfortunately something I'll have to swallow with these design parameters.

 

I had modelled the vents using 2 lengths of 102 mm ID sewer pipe, as in my previous build. At the GTG on Saturday, we played a 24/192 version of "Hotel California" through them, appreciably louder than my normal tolerance, and I held a tissue over one of the vents. It only flapped about a centimetre. I have zilch worries about a 150 mm vent chuffing, as its area is larger than the 2 x 4" pipes (177 cm^2 vs 163 cm^2).

 

Further on "passive" inductors - did you see what I think I saw, of the mH values on @acg's line-level filters?? I seriously want to know more about this approach.

 

Edited April 30, 2019 by BioBrian
GTG on Sat, not Sunday. Wish sun'd been there.

[almikel]

3 minutes ago, BioBrian said:

Thanks for your post, and welcome back! Hope those pesky distractions aren't too disruptive for too long.

I've been learning how to play harmonica - a fantastic distraction!

6 minutes ago, BioBrian said:

But IB implies a virtually infinite air space behind the baffle, doesn't it?

correct - but drivers designed for IB alignments can work in sealed boxes if the box is big enough - when you sim a driver meant for IB in a sealed box you'll see it peak before roll off - the bigger the box the lesser the peak...

...if you're considering a 320 litre vented box and haven't purchased a driver yet, then consider the AE IB range, and run them in a large sealed box...or proper IB - they don't have to be on a flat wall to be IB...most IBs have a manifold (that the drivers mount on), with the manifold exiting outside.

 

IMHO I would avoid ported alignments if sound quality is the goal - too many things go awry as SPL changes - much less so with sealed or IB subs.

 

cheers

Mike

[davewantsmoore]

9 hours ago, BioBrian said:

The 105 dB is also a little loud!!

Not in a max SPL sense.

 

If you think about what you measure on a SPL meter with music .... You need to add the crest factor of your content on top.    Is that 4dB?  10dB?  30dB????   

 

Otherwise what happens, is that on the "short, sharp, loud" bass tones ....  You get some sort of linear or non-linear distortion....  which typically results in either a raised noise floor (from non-linear), or a reduced level (from linear) .... both results in "indistinct" bass.

[davewantsmoore]

9 hours ago, BioBrian said:

Huge inductors?  No idea how they affect sound.  Not at all, from my listening. From all the emotive stuff on "damping factor" threads, I have sort of gleaned that R is the only relevant issue there, and these C-coils have almost none.

No, it isn't a "damping factor" thing.

9 hours ago, BioBrian said:

Phase and group delay may be important

They're directly related to the frequency response in this instance ..... so, just the FR is important.

 

9 hours ago, BioBrian said:

but that's unfortunately something I'll have to swallow with these design parameters.

Yes, I guess that's my point.   

 

9 hours ago, BioBrian said:

I have zilch worries about a 150 mm vent chuffing

I didn't mean to imply that was the primary problem.    It's more a "compression" issue, when you look at how the driver responds at LF vs time.

 

9 hours ago, BioBrian said:

did you see what I think I saw, of the mH values on @acg's line-level filters??

Yes.    You'll see some comments from me in his thread   

[davewantsmoore]

10 hours ago, BioBrian said:

My intended wall-ceiling box placement, with a 45 degree angled baffle, would effectively be an infinite baffle - ie no baffle step loss, in fact I'd expect a net gain over IB (on a flat wall) - approaching "corner placement" - huge room gain.

At the frequencies you are talking about, the concept "of baffle step loss" is kinda nonsensical.    At LF essentially anywhere in the room you place the subwoofer it is approaching being "in the corner".

 

10 hours ago, BioBrian said:

But IB implies a virtually infinite air space behind the baffle, doesn't it?

Yes.

The practical effect of doing that is reduced power requirements vs SPL at LF.

[davewantsmoore]

Looking at the drivers frequency response in a box, before you have shaped the response to your needs - is not the right way to evaluate a driver.

 

Need to mould their response to be the same (whatever your target is), then look.    A lot of the differences will go away, others will become more obvious.

[acg]

13 hours ago, BioBrian said:

Further on "passive" inductors - did you see what I think I saw, of the mH values on @acg's line-level filters?? I seriously want to know more about this approach.

It's pretty simple Brian, you jam the filter at the input of the amplifier where the voltages are much smaller, let's say 2V max (4V in my case) and use the input impedance of the amplifier as the load for the filter rather than the reactive impedance of a speaker level crossover network and transducers.  Then you direct connect your transducer to the amplifier.  It keeps the filter component values much smaller, so tiny caps and inductors.

 

In some circumstances you may even be able to build part of the filter into the amplifier itself.  For example an interstage coupling capacitor can be sized such that it creates a first order high pass filter at the desired frequency.  Couple that with a low pass filter  before the amplifier and you have a bandpass filter.

 

I've not yet read your thread Brian but the speaker photos on the front page look excellent.

[A9X]

2 hours ago, acg said:

For example an interstage coupling capacitor can be sized such that it creates a first order high pass filter at the desired frequency. 

Yes, that works, but a 1st order filter is woefully inadequate in almost all situations. You can do a 2nd order PLLXO easily enough, but 3 order and above are much harder, pretty much impossible without a lot of parts and adequate measurement gear.

[acg]

@A9X, yeah the bits add up.  I have a 4th order highpass set at 10kHz using two air-caps and two inductors.  It takes lots of space. ...like twice as much space as all the filters for the other five channels.   3rd or 4th order RC filters take next to no room.  The caps are tiny and the resistors only need to be quarter-watters.

[BioBrian]

On 01/05/2019 at 12:30 PM, acg said:

It's pretty simple Brian, you jam the filter at the input of the amplifier where the voltages are much smaller, let's say 2V max (4V in my case) and use the input impedance of the amplifier as the load for the filter rather than the reactive impedance of a speaker level crossover network and transducers.  Then you direct connect your transducer to the amplifier.  It keeps the filter component values much smaller, so tiny caps and inductors.

Yes, I've a handle on the concept, but how to calculate the values, and how to find the values seems less easy.

 

By "smaller, so tiny caps and inductors", I guess you mean like this (small by acg standards )

 

 

With the cable ties, i can only guess, but the values look humongous. (I'd expect only 0.1 uF and less than 0.1 mH if it were passive).

 

Thanks for your kindness re my speakers.

Edited May 3, 2019 by BioBrian
spelling

[acg]

Yeah @BioBrian, that's the 4th order high pass filter set at 10khz.  How to calculate them....that's easy.  I have the "Electronic Filter Design Handbook" and I took the equations for calculating that particular "CLCL T High Pass Filter with unequal terminations - Fourth Order Butterworth - Infinite Termination"  (simple when you say it like that haha) from within its pages and put it in an Excel spreadsheet, which also includes all the other types of much simpler filters that I have employed:  low pass RC; high pass RL; bandpass RCL; low pass RRC; and the high pass coupling capacitor filter.  Subtractive filters only.

 

With those larger shielded inductors in the photo 1.241H and 0.298H are also two air capacitors set at 346pF and 490pf (yes picofarad).  The filter is loaded by a 30k ladder type LPad.  Do not fear the inductor...the inductor is good... capacitor is bad...reduce or eliminate the capacitor. 

 

EDIT:  I should add that the caps used in my other passive line level filters range from  0.0047uF to 0.068uF with the large low passing at 78Hz.  Tiny little suckers.

Edited May 2, 2019 by acg

[davewantsmoore]

https://www.t-linespeakers.org/tech/filters/passiveHLxo.html

[A9X]

11 hours ago, acg said:

capacitor is bad...reduce or eliminate the capacitor. 

Oh good grief.

[afa]

capacitor not bad

room is bad

room can make mega buck speakers into pretty average speakers

room is bad, very bad...must fix room

Edited May 3, 2019 by afa

[BioBrian]

23 hours ago, afa said:

room is bad, very bad...must fix room

Hi Arthur, not sure if you mean generally, or my room in particular.

 

It's had some treatment, and I anticipate more in conjunction with this present project.

 

Of course, you are still most welcome to visit. I'd appreciate any help with this complex situation.

[BioBrian]

On 02/05/2019 at 5:22 PM, acg said:

I have the "Electronic Filter Design Handbook" and I took the equations for calculating that particular "CLCL T High Pass Filter with unequal terminations - Fourth Order Butterworth - Infinite Termination"  (simple when you say it like that haha) from within its pages and put it in an Excel spreadsheet, which also includes all the other types of much simpler filters that I have employed:  low pass RC; high pass RL; bandpass RCL; low pass RRC; and the high pass coupling capacitor filter.  Subtractive filters only.

 

With those larger shielded inductors in the photo 1.241H and 0.298H are also two air capacitors set at 346pF and 490pf (yes picofarad).  The filter is loaded by a 30k ladder type LPad.  Do not fear the inductor...the inductor is good... capacitor is bad...reduce or eliminate the capacitor. 

 

EDIT:  I should add that the caps used in my other passive line level filters range from  0.0047uF to 0.068uF with the large low passing at 78Hz.  Tiny little suckers.

I'd need to take a very deep breath to take this on. So many unknowns for me.

 

How to source such tiny, odd-valued Resistors and Caps?

How to test these components, and test if the filters work as predicted?

What if a bespoke inductor turns out to need adjustment?

 

I have to say, I've always loved those air-caps. Such precision engineering - used to play with them out of old radiograms in the '60s. Didn't know they were state-of-the-art for speakers!

 

The formula from Dave's post says a 2nd Order HP is 2 caps and 2 coils. Is it possible to do this without caps?

Can quality caps of these values be found at all?

 

Easy to see why people go to Phil Marchant or Accuphase, etc, to have ready-made filters. But then adjustability is expensive.

[davewantsmoore]

36 minutes ago, BioBrian said:

I'd need to take a very deep breath to take this on. So many unknowns for me.

When you take your deep breath.... there is one thing you should meditate on.

 

 

It is not the parts or method that you use to implement the filter the determines its ultimate performance .....  but it is the curve of the filter.

 

If you implement a filter with PLLXO, speaker level XO, or DSP, or whatever..... and they have the same curve - they will sound identical.

 

If you build filters with these different methods, and they don't have the same curves - they will sound different..... but if you are UNAWARE that they have different filter shapes, then you might be tempted to blame the different parts/methods used to implement the filter.

 

 

36 minutes ago, BioBrian said:

and test if the filters work as predicted?

The same way as you do for your current filters.