Deep End DIY - My first speaker project

[almikel]

37 minutes ago, acg said:

then perhaps another sub or two with Class D amps and DSP.  I've made accommodations for that outcome but if I don't have to do it then I would prefer not to...

as much as I'm a fan of EQ/DSP generally, in your setup where you've avoided it so far, it would be a shame to have to resort to it now...but not that surprising you're considering it for the bottom octave...

 

...in a well treated room with the bass under control, a sub or 2 with a bit of DSP will be "icing on the cake", to create the ultimate "no compromise" system!

 

cheers,

Mike

[acg]

1 hour ago, davewantsmoore said:

So? (I'm not understanding the problem here)

 

You either add the filter yourself.... or you use a driver which already has the filter built into it.

The "filter built into the driver" is not a good thing.  You would very much want to avoid that, if possible, if you want high performance.

 

EDIT:  is it an amplifier gain issue?

 

 

It is a filter issue.  At such low frequencies when adding orders to a passive filter the filter transition gets wider (in Hz) and thus eats away at the gain at say 20Hz.  It may only be a dB or two but I've only got 9w to play with for my Bass Cannons so it also becomes an amplifier issue.

 

Saying this, I've figured out a way to add about 6dB of extra gain to the Bass Cannons.  Or more appropriately 6dB more difference in gain between the amplifier channel for the Bass Cannons and all the remaining amplifier channels.  Should be very useful.

 

 

1 hour ago, davewantsmoore said:

 

 

2 hours ago, acg said:

Previously I have experimented with a midbass channel in the parts of the room

Just moving one source around the room, isn't going to do anything very much.

It is about multiple sources.

 

Was more than one bass source as the midbass box was in conjunction with the Bass Cannons.  Still made two-fifths of stuff-all change to SPL in that 60Hz-110Hz range because the room was being driven from similar positions.  This new area is quite different and the sims I've used here indicate it could work.

 

[acg]

1 hour ago, almikel said:

as much as I'm a fan of EQ/DSP generally, in your setup where you've avoided it so far, it would be a shame to have to resort to it now...but not that surprising you're considering it for the bottom octave...

 

...in a well treated room with the bass under control, a sub or 2 with a bit of DSP will be "icing on the cake", to create the ultimate "no compromise" system!

 

cheers,

Mike

 

Mike, I'm certainly hoping not to have to resort to DSP and Class D, but you never know.

[davewantsmoore]

48 minutes ago, almikel said:

where you've avoided it so far

Yeah.  I think I've been pretty clear through the thread the the goal of "keeping it analogue" is unnecessary from a "it will sound better" perspective ... BUT, you can still do pretty much everything you need with analogue filters, as this is a minimum phase system.   The point being, don't let anyone tell you that "DSP is needed", or whatever.  Analogue will be "harder" .... but what is "hard" is knowing what filters to make, etc. (speaker design) ..... they can be implemented "however".

 

Digital / DSP won't necessarily sound better, or worse ... and there is nothing "magic" about "DSP filters" (like "FIR" filters), and to be blunt those who go around saying so (or things which sound like it) are full of it... but I digress.   The point being that it isn't the "how" you do the filters ...... it's the "what" (ie. what filters.... what shape are they, and is that the right one).    Filters which correct and crossover the drivers, are what is needed .... what those filters are made from, is just about irrelevant, in comparison.

 

55 minutes ago, almikel said:

a sub or 2 with a bit of DSP will be "icing on the cake"

Does not even really need DSP.    All you need to do, is design the woofer first ... (place the woofer, and) design the filter second... and the amp 3rd (so it can have the required filter/gain embedded in it)

 

No EQ (beyond what you need to correct the physics of the drivers)  is actually somewhat noble/lofty goal (as EQing against the room is somewhat folly) ..... but small rooms are very difficult.  Multiple sources very very very helpful.  Treatment, no so much.

 

1 hour ago, almikel said:

I'd appreciate your insights into what these graphs show

Hard to assign causes to things, without more information or a little speculation.

 

I think it is very helpful for people to see what "the chart is supposed to look like"... as it helps you enter the thinking mode of "what is wrong with this chart, and why could that happen?!"..

 

 

 

 

[davewantsmoore]

3 minutes ago, acg said:

It is a filter issue.  At such low frequencies when adding orders to a passive filter the filter transition gets wider (in Hz) and thus eats away at the gain at say 20Hz.  It may only be a dB or two but I've only got 9w to play with for my Bass Cannons so it also becomes an amplifier issue.

You're not going to use any more watts (as it is doing nothing to the speaker efficiency).    It just needs more gain on the input signal.

Right?! (what have I missed?)

[davewantsmoore]

1 hour ago, almikel said:

I'd appreciate your insights into what these graphs show? 

In short you can see, any least what I'd suspect it is:

 

Some crossovers where the drivers aren't in phase (at least not over a wide range)

Diffraction (eg. sources reflecting off each other)

An extremely well damped room (quite impressive)

 

 

[acg]

5 minutes ago, davewantsmoore said:

You're not going to use any more watts (as it is doing nothing to the speaker efficiency).    It just needs more gain on the input signal.

Right?! (what have I missed?)

 

That is corrrect.  No problems with my high gain sources but the low gain sources such as a couple of phono stages and the AVR are such that there is very little volume attenuation happening right now when I want those sources LOUD.  The amp for the Bass Cannons can drive them to about 105dB at 40Hz (enough) or 95dB at 20Hz (not enough).  Was hoping for more room gain but it ain't there.

[almikel]

35 minutes ago, acg said:

The amp for the Bass Cannons can drive them to about 105dB at 40Hz (enough) or 95dB at 20Hz (not enough).  Was hoping for more room gain but it ain't there.

gotta hate evolution that determined Fletcher Munson curves were best for our survival, but not so great for fantastic "in room" audio ...

 

[davewantsmoore]

2 hours ago, acg said:

That is corrrect.

Ok, so I'm confused then.  Make the filters you need, and then build your amps to have the right gain (?!)

[acg]

@davewantsmooreI can't find my amplifier measurements for output voltage at onset of clipping for any of the channels let alone the one for the Bass Cannons so will have to do them again in order to precisely determine how loud the Cannons can be made at the listening chair.  My memory of 95dB at 20Hz may be wrong... best to know for sure. 

 

 

[davewantsmoore]

9 hours ago, acg said:

I can't find my amplifier measurements for output voltage at onset of clipping

How many stages is your amp?

If it is a case of not enough gain rather than not enough possible output power ... then you will need to boost the input signal somehow.

I imagined it has more than one stage, and so you could adjust the gain in the first stage.... but I'm sort of still not completely clear on what your problem is (!?)

 

This is why the age old rule about first picking speakers (including whatever filters they need), and then getting an amp to drive them.   The shape of the signal, is more important than how it is shaped (eg. "all analogue", some specific type of amp, some type of EQ/crossover, etc. etc.).  12dB/octave is no small thing.

 

13 hours ago, acg said:

Was hoping for more room gain but it ain't there.

Being a fairly small room, your room gain being up quite high, and so you might not see it where you're looking for it.

 

Where you will see it is at the lowest room mode.   Your speaker is rolling off at 12dB per octave ... and at the lowest mode, it will start to increase at 12dB per octave.... the result being it is flat.   So it will decrease by 12dB / octave, and stop at the lowest mode.    That is of course, if your source actually continues 12dB/octave rolloff as you get very low in Hz, as opposed to rolloff faster (which they probably do) .... so what you will actually see is 12dB/octave, until that rolloff stops or slows down.   Hope that makes sense.

 

The effect is very very difficult to see in your chart posted earlier, but if you really really quint through the lumps and dips due to the modes ... you can see that the 12dB/octave slope starts to plateau at 30Hz.

[acg]

@davewantsmoorethat amp channel is two stages and needs circa 4vrms input to make full power output.  The phono and AVR are about 1vrms output and my preamp is x8 gain which means I have about 6dB system gain headroom for those sources which is only just enough in practice...I've not run out of volume dial yet but do get close.  All other sources are high output which provides loads of system gain headroom.

 

My major problem is attainable volume at 20Hz given the low amplifier output power and the present gain mismatch with the remaining channels.  I can win back 6dB of the gain mismatch by changing a couple of resistors and I can win 6dB of extra low frequency output with a second sub channel which on paper could get me there or at least close enough at 20Hz.  And even then, if I want a bit more 20Hz output I can add another sub just for that task.

 

First though, I need to be sure how loud I am able to push 20Hz at the listening chair with things as they are now are...that means getting out the oscilloscope, audio analyser and microphone.

[acg]

7 hours ago, davewantsmoore said:

Being a fairly small room, your room gain being up quite high, and so you might not see it where you're looking for it.

 

Where you will see it is at the lowest room mode.   Your speaker is rolling off at 12dB per octave ... and at the lowest mode, it will start to increase at 12dB per octave.... the result being it is flat.   So it will decrease by 12dB / octave, and stop at the lowest mode.    That is of course, if your source actually continues 12dB/octave rolloff as you get very low in Hz, as opposed to rolloff faster (which they probably do) .... so what you will actually see is 12dB/octave, until that rolloff stops or slows down.   Hope that makes sense.

 

The effect is very very difficult to see in your chart posted earlier, but if you really really quint through the lumps and dips due to the modes ... you can see that the 12dB/octave slope starts to plateau at 30Hz.

 

This is the aforementioned graph....

 

 

My lowest room mode is at 30Hz and before room treatment I did have a nice cancellation or flatline below to something like 18Hz.  But that orange line for the present situation is different.  I think that the VPR's have effectively "opened a window" by being a little bit effective circa 30Hz which has been enough to cancel some of the room gain.  Perhaps that is the price for such effective damping?  According to close gated measurements in-room, the Bass Cannons naturally roll off below about 40Hz at 12dB/octave all the way to below 10Hz.

 

 

On the topic of room reverberation and what to target, a while back I found a study where "real people" were asked to correlate sound quality in various sized rooms based on reverberation control, or reverberation time.  At the end of it was a formula to apply to our particular room size to determine the target RT range and using for my room 200ms was said to be "ideal".  Larger rooms tend to sound better with longer reverb and vice vera.  Interesting stuff, and if anybody is interested let me know and I will go search for the reference wherever it may be.  

[davewantsmoore]

36 minutes ago, acg said:

My major problem is attainable volume at 20Hz given the low amplifier output power

Ok, righto then.

So it IS a watts problem.... no matter what filters you put on to shape the response, if you can't make enough power to get 20Hz to the SPL you want, then....

[frednork]

1 hour ago, acg said:

Larger rooms tend to sound better with longer reverb and vice vera.  Interesting stuff, and if anybody is interested let me know and I will go search for the reference wherever it may be.  

Dont go to any trouble  to find it but if it pops up would be interested in having a look.

[acg]

1 hour ago, davewantsmoore said:

Ok, righto then.

So it IS a watts problem.... no matter what filters you put on to shape the response, if you can't make enough power to get 20Hz to the SPL you want, then....

 

Yes, sort of, partially.

 

Had a couple of hours to play around this afternoon and these are my findings (everybody loves dot points right!):

 

1. With 1.0Vrms input to the preamplifier (same output as my weakest output sources) I have enough SPL from 40Hz up apart from the octave wide suckout centred at 80Hz.
2. Sims show that the 80Hz octave wide suckout will be fixed with back wall subs.  Fingers crossed.
3. 20Hz - 40Hz is my biggest issue but I have found a slight complication (see below).  If the new back wall subs have a similar output to the Bass Cannons then I expect improvements of up to 6dB in the 20Hz-40Hz range.  If 20Hz is thus boosted by the new subs, and if I re-jig amplifier channel gains then I will have a little leeway to introduce a filter slope to flatten the 20Hz-40Hz response.  Problem solved.  Or close to it.

 

One interesting thing this afternoon is the low measured distortion at the listening chair when driving the Bass Cannon full throttle...THD was generally 0.5% - 2% down to 30Hz or so where it increased to about 6% at 20Hz.  From what I understand reading online subwoofer outdoor measurements these are excellent numbers.

 

 

Here is the new point of interest...

 

 

Sub-bass actually decreases about 3dB in room when the room is sealed i.e. the sound-proofing is complete with the trapdoor closed.  What the?  This was always the other way around.  The only thing that I can thing of is that the sealed room is helping the VPR's to be more effective perhaps in their pistonic mode which really only happens with big energy in the sub-bass (according to theory anyway).  If you look at the 30Hz mode see how much more effective it is dealt with when the trapdoor is closed.

 

 

[davewantsmoore]

24 minutes ago, acg said:

What the? 

Hmm... that is surprising.  3dB is a lot.

[acg]

4 minutes ago, davewantsmoore said:

 3dB is a lot.

 

It is a lot  (especially when I'm scraping for SPL in that range)

[frednork]

2 hours ago, acg said:

If you look at the 30Hz mode see how much more effective it is dealt with when the trapdoor is closed.

IS the loss of sound proofness going to be an issue?

[acg]

@frednorkonly for everybody else in the house.  The great bulk of listening is with the trapdoor open, but there are times when my family wish for sleep or silence or when I have the airconditioning on upstairs.  If I can get a 3dB sub bass boost most of the time I'm sure it would be appreciated, but I really should set the playback for the trapdoor being down. 

[acg]

On 10/09/2023 at 6:28 PM, frednork said:

Dont go to any trouble  to find it but if it pops up would be interested in having a look.

 

Can't find the actual reference for the work, but did find the formula in my notes.  From memory the study was done late '90's.

 

Tm=0.3(V/100)^1/3 +/- 15%

 

where:

     Tm = 60dB decay time 

     V = room volume in m3

 

So, figure out the room volume, plug that into the formula to get the time to decay 60dB in-room.  Other notes here suggest that you want the decay time to be within 15% of the nominal value across the frequency range although in the top octave this may drop to 50% of nominal without sounding lesser and that in the low frequencies it is ok for the decay time to double.

 

For example, my room is 5.9 x 4.25 x 2.4m (average ceiling height) = 60m3 so ideal Tm according to the formula is 0.25s which means across that I should try to keep the 60dB decay time between 20Hz and 20kHz within 0.22s < 0.29s (+/-15%).  It is ok if the top octave decays as low as 0.13s and sub 100Hz at up to 0.5s.  I remember watching a youtube video where an acoustic consultant went through this formula and talked about his experience chasing low end decay to match the formula.  His impressions were that such low decay times makes the room sound bass-shy and that the higher decay time recommendation (in my case 0.5s) would actually be prefered by many people.  He also said that if the room sounds bass-shy with such low decay that turning up the bass sounds fantastic.

 

Hope this is of interest.

[acg]

^^^^ further to this post.

 

The formula above is referenced to measuring room decay using a dodecahedron speaker that radiates sound it ALL directions thus lengthening the measured decay as opposed to using just your normal loudspeakers.  So if I want to measure this properly I need to get myself one of these dodecahedron speakers or shoot for a lower decay using my actual speakers.  For instance, a 0.2s measured decay time using my speakers is probably roughly equivalent to 0.25s measured decay time if using a loudspeaker that radiates in all directions.

Edited September 15, 2023 by acg
spelling

[playdough]

37 minutes ago, acg said:

He also said that if the room sounds bass-shy with such low decay that turning up the bass sounds fantastic

Agree on this. 

[frednork]

4 hours ago, acg said:

 

Can't find the actual reference for the work, but did find the formula in my notes.  From memory the study was done late '90's.

 

Tm=0.3(V/100)^1/3 +/- 15%

 

where:

     Tm = 60dB decay time 

     V = room volume in m3

 

So, figure out the room volume, plug that into the formula to get the time to decay 60dB in-room.  Other notes here suggest that you want the decay time to be within 15% of the nominal value across the frequency range although in the top octave this may drop to 50% of nominal without sounding lesser and that in the low frequencies it is ok for the decay time to double.

 

For example, my room is 5.9 x 4.25 x 2.4m (average ceiling height) = 60m3 so ideal Tm according to the formula is 0.25s which means across that I should try to keep the 60dB decay time between 20Hz and 20kHz within 0.22s < 0.29s (+/-15%).  It is ok if the top octave decays as low as 0.13s and sub 100Hz at up to 0.5s.  I remember watching a youtube video where an acoustic consultant went through this formula and talked about his experience chasing low end decay to match the formula.  His impressions were that such low decay times makes the room sound bass-shy and that the higher decay time recommendation (in my case 0.5s) would actually be prefered by many people.  He also said that if the room sounds bass-shy with such low decay that turning up the bass sounds fantastic.

 

Hope this is of interest.

Just to compare with a notional rt 60 for your room (blue line)

 

EBU standard - EBU Tech. 3276 - 2nd edition, May 1998, p.6

 

DIN 15996 - Studio

 

 

So, I guess EBU is your jam!

Worth having a look at if you are not familiar. lots of good pointers https://tech.ebu.ch/docs/tech/tech3276.pdf

 

Edited September 16, 2023 by frednork

[frednork]

this bit is interesting