Deep End DIY - My first speaker project

[acg]

It has been a while...damn work keeping me too busy...but I'va managed to achieve quite a bit in small fits and starts.  Got up some 'acoustic' wall panels to cover the steel VPR's littering the room. 

 

This was one corner before...

 

 

 

Now...

 

 

 

HUUUGGGEEE difference to the sound, and so much for the better.  Apparently the panelling absorbs from about 1.5kHz up but the way I've used them with an inch or so airgap behind them to the VPR's perhaps this reduces to 1kHz.  Twas a lot of work cutting and drilling and screwing and all that but it got done and I am very happy.  My pre-teen daughter thinks it is an assault to her eyes and my teenage son really liked the steel aesthetic, my wife refuses to give an opinion but I am satisfied.

 

No room measurements since panelling but I have this weekend off (finally!) so expect to get to it then.  Bits have been ordered to build some ceiling VPR's to go above the listening position to see if that 60Hz-90Hz suckout can be tamed and maybe that will all happen in the next week of three...fingers crossed.

 

Other exciting news is that I have been talking to an acoustic consultant that uses BEM modelling of the room and its contents to recommend/design diffusers to achieve a diffuse sound field at the listening position.  So excited about that, and so is he apparently, so big hopes that something useful comes from it.  The more I learn about QRD diffusers the less I want to use them so very happy to bring a more modern perspective to stereo room acoustics.

 

 

 

Edited June 9, 2023 by acg

[acg]

In analog news, something arrived in the mail from Lithuania...a new cartridge...

 

 

 

What's that...six pins instead of four???

 

 

 

That's right, a field coil cartridge...here is the power supply...

 

 

Of course the first thing I did was pull the power supply apart because I've never played with field coils before and trace out a schematic to get a good understanding of what I had purchased.  It is a voltage adjustable, centre-tapped full wave rectifier using EY500A rectifier diode tubes.  Very simple but seemingly well executed.  Two switchable outputs so can be used to power two field coil carts should the need arise.  I'd have wired it much more untidily to reduce inductance but all in I like it.  Those rectifier tubes are monsters capable of 440mA.

 

All a bit moot at the moment though at the TT is out of the room during renovations and I need a new tonearm capable of handling this hefty cart.  The plan is to finish the room, bring the Helix Two back in and then the maker of my TT will come up with a bunch of tonearms and we will make a weekend out of finding the optimum sound.  Really looking forward to that. 

 

DaVa FC Ref.

Edited June 9, 2023 by acg

[acg]

So here we are at the end of the bass treatment program for the walls.  Green is the empty room where I started and Orange is where I have ended up using all DIY room treatment:  14 VPR's on three walls and straddling the front corners; two mega superchunk traps that basically fill in between the triangulated VPR's in the corners with light polyester batts; the porous traps underneath the vinyl racks on the back; covering of the VPR's with off-the-shelf slatted timber (acoustic?) panels.

 

No consideration yet for ceiling bass treatments, multiple subs or midrange absorption or diffusion within the room.

 

 

 

Those are some E.N.O.R.M.O.U.S frequency response changes in the room due to the bass panels.  The VPR's have effectively killed the room modes from 40Hz and up and have even been effective at 30Hz.  The 30Hz is interesting because at the side walls (it is a lateral mode between the side walls) the SPL is now over 20dB lower than pre-treatment.

 

So, what are my current frequency response issues? 

 

First is that enormous suckout from 60Hz to 110Hz, somewhere near an octave wide, that already has been attenuated 10dB-25dB by the wall treatments but has not flattened out as much as I would have liked.  Earlier thinking is that this is at least partially a product of the sloped ceiling above the listening position, but playing bass tones and walking around with the microphone it is plainly also a sub placement issue.  That is, the subs in their current position are not able to appropriately excite that part of the room within that octave.  I would like to think that putting some VPR's on the ceiling will fix the issue but I don't think it will get me all the way there.  From my experiments, placing a second pair of subs on the wall behind the listening position is a good chance of filling that octave wide suckout.

 

Second problem is the 12dB/octave SPL drop between 20Hz and 40Hz.  This is potentially something that can also be fixed with the aforementioned second pair of subs...not sure...or perhaps with a filter on the amplifier channel for that sub although passive filters also take a little bit of gain which I seem to be short on for that channel.

 

Third problem is that the bass sub 140Hz is generally 6dB below the remainder of the speaker output.  There are a number of ways to fix this within the amplifiers so hopefully easy to do, and the second pair of subs may also alleviate the problem.

 

What is the next step?  Not sure whether to put the VPR's on the ceiling and re-measure the room or build the second pair of subs to go on the back wall.  VPR's are pretty much already made, they just have to be put up but it would also be nice if they were not required at all.  Little planning has been put into further subs other than the maximum permissible size.  Looks like I will have to parallel multiple 10"/12"/15"/18"/21" drivers in an effort to maximise sensitivity.

[andyr]

Great room treatment result @acg.  

 

If you have access to EQ (like Roon Convolution - or even a miniDSP unit), I suggest you would end up with a pretty damn good FR if you implemented:

1. an 8dB shallow cut, centred on 40Hz
2. plus another cut, centred on a higher frequency - which I can't define, as your graph only goes up to 144Hz.

 

[playdough]

Nice work.

Any waterfall graph ? I've been through similar experience last month, DIY treating a lounge.

Reading about suckouts and so on,  sounded quite familiar and had luck with timing the subs to the listening position, more than EQ, which does not do all that much at below 100Hz for those holes in FR.  

[acg]

@andyrNo DSP in my system.  I wish to keep all EQ analogue.  I could possibly design a passive filter for the "8dB shallow cut" but fear I will run out of system gain from my lower output sources.

 

Having thought about the steps forward some more, the ceiling bass trapping with VPR's should help to smooth out some modal behaviour from 120Hz - 250Hz that porous absorption will be unable to alter, so they are the obvious next step.  Perhaps they will also do enough in the 60Hz-110Hz range but my thinking now is that is another problem to be alleviates with more subs.

[acg]

8 minutes ago, playdough said:

Nice work.

Any waterfall graph ? I've been through similar experience last month, DIY treating a lounge.

Reading about suckouts and so on,  sounded quite familiar and had luck with timing the subs to the listening position, more than EQ, which does not do all that much at below 100Hz for those holes in FR.  

 

@playdoughI've glanced at your thread but not had time to follow it through.  Looks like you have properly got stuck in!

 

Here is the current room waterfall post wall treatments.  That 200Hz ringing is a computer fan.  Listening room noise floor right now is about 40dB-45dB which is where I have ended the graph.  Pretty sure I can get that suckout centred at 200Hz with ceiling bass treatment (VPR's). 

 

[playdough]

The Waterfall looks decent. 

Little bit of EQ here and there may improve things,  FR is not too bad, maybe try for another +3dB in the lowest 2 octaves crank the subs up . The narrow Q hole at 200Hz, is there a crossover there ? Try moving the mic 200mm left or right, you might find that will give a slightly  better picture of it overall. 

 

Wouldn't sweat on it too much. All looks quite well considering the trouble you have gone to to get this "reference horn system" actually working.

Respect,

 

 

[playdough]

1 hour ago, acg said:

/21" drivers in an effort to maximise sensitivity.

 Normal sized/sensitivity to match, preference as pairs, for the free +3dB driven with a single ended Triode

 

Edited September 9, 2023 by playdough

[frednork]

@acg , thanks for continuing to document this massive endeavour. Thinking about some big picture separate issues you may have with "back of envelope" type accuracy.

 

To understand the drop in gain for the same power output pre and post treatment perhaps you could adjust your graph comparing the 2 above to show that, as it shows a higher output after treatment. but using your waterfall as a basis instead

 

1).Taking your waterfall above at face value the drop below 40Hz wont change for the better due to any treatment (as I understand that is the natural slope) so not using extra subs or eq is not an option.  The other thing is that you need around a 10db per octave slope between 20 and 40Hz if you want it flat to say 20hz or belowish. 

 

2)So to get a flat response at 20Hz to match levels at >100hz there must be (either through subs or eq) around 20db gain at 20hz and more below that. so  for a target of 15Hz say 25db gain through either subs or eq needed

 

if we take the 20hz issue as an example. If youR subs have infinite headroom then simplistically it is just a matter of more power with some eq, so how much more power needed. We know that to get a 3db increase we need 2 x the power right,  but, you would need 100x the power to achieve a 20db increase.

 

hope I got this wrong.

[acg]

@playdough there is no crossover near the 200Hz suckouts.  The upperbass horn goes down to 140Hz.  At the horn mouth the frequency response is super flat through that range but once out in the room things change (of course).

 

I've made enquiries regarding a 24" driver from the US but so far the company has been reluctant to reply.  Low sensitivity, huge cone area, strong motor transducers are what I need.

 

 

[playdough]

Might be helpful

Here is a plot of my lounge, 21's only on a very tiny AB power amp 25w

Note the big suckout measured flat outside. EQ doesn't do a lot for this type of room related problem.

Oh, and you are lucky, nice reverb response at very low Fs, better than this one, must be a light building maybe an upstairs room on it's own.

26 minutes ago, acg said:

enquiries regarding a 24" driver

Personally, would be looking for something else from what experience  have with these larger Transducers, operating at less than 1mm cone excursion.

[acg]

37 minutes ago, frednork said:

@acg , thanks for continuing to document this massive endeavour. Thinking about some big picture separate issues you may have with "back of envelope" type accuracy.

 

To understand the drop in gain for the same power output pre and post treatment perhaps you could adjust your graph comparing the 2 above to show that, as it shows a higher output after treatment. but using your waterfall as a basis instead

 

 

All good @frednork, thanks for taking an interest.

 

The "pre" graph was taken just using the Bass Cannons whereas the "post" graph also includes the hornstack, so the upperbass horn will have influence on the SPL down to say 100Hz.  During all of the room treatment iterations the humps at 40Hz and 20Hz barely moved so I used these to match the SPL graphs.  The match should be accurate to within a few dB I would say.  I've also measured huge spl reductions at the sidewall boundaries of more than 20dB at frequencies like 30Hz to feel comfortable enough that the modal influence in that range has been lowered enough mid-room to perhaps increase SPL output at the listening position.

 

 

41 minutes ago, frednork said:

1).Taking your waterfall above at face value the drop below 40Hz wont change for the better due to any treatment (as I understand that is the natural slope) so not using extra subs or eq is not an option.  The other thing is that you need around a 10db per octave slope between 20 and 40Hz if you want it flat to say 20hz or belowish. 

 

 

Agreed.  Would have been nice to have some more room gain but the lowest mode is circa 30Hz so there is not so much to be had.

 

 

42 minutes ago, frednork said:

 

2)So to get a flat response at 20Hz to match levels at >100hz there must be (either through subs or eq) around 20db gain at 20hz and more below that. so  for a target of 15Hz say 25db gain through either subs or eq needed

 

 

Agreed.

 

50 minutes ago, frednork said:

if we take the 20hz issue as an example. If youR subs have infinite headroom then simplistically it is just a matter of more power with some eq, so how much more power needed. We know that to get a 3db increase we need 2 x the power right,  but, you would need 100x the power to achieve a 20db increase.

 

hope I got this wrong.

 

Agreed.  But there are other options than 1000w on the Bass Cannons.

 

The horn stack currently runs un-attenuated and rampant at 109dB/w/m and the Bass Cannons un-attenuated and less rampant at 98dB/w/m.  The SET amp for the Bass Cannons is capable of driving them to 105dB at 40Hz at the listening chair and probably 95dB at 20Hz based on in-room response.  This should be loud enough.  There are options to adjust gains within the amplifiers, particularly within the passive line level filters, and I can also attenuate the horn-stack channels individually so rather than boosting the Bass Cannons to match the higher sensitivity horn stack, I can add gain to the Bass Cannon amplifier.  Probably not another 20dB gain as that is a lot but I can simultaneously cut signal to the horn stack to make up any shortfall.  That is the beauty of passive line-level filters as opposed to speaker level filters...gain structure is always optimised such that the amplifiers are never doing more work than needed (i.e. amplifying to a large voltage and then attenuating before the speaker).

 

The second pair of subs for the back wall may possibly make things a little easier by increasing output in that 20Hz-40Hz range.

[acg]

16 minutes ago, playdough said:

Might be helpful

Here is a plot of my lounge, 21's only on a very tiny AB power amp 25w

Note the big suckout measured flat outside. EQ doesn't do a lot for this type of room related problem.

Oh, and you are lucky, nice reverb response at very low Fs, better than this one, must be a light building maybe an upstairs room on it's own.

Personally, would be looking for something else from what experience  have with these larger Transducers, operating at less than 1mm cone excursion.

 

Which 21's are you using @playdough?  I'm looking to low sensitivity non-pro subs that have just as much output in that lowest octave as high-sensitivity subs but don't need as steep a low pass filter.

 

Would you believe that my room is sound-proofed?  A very heavy construction, sealed watertight...those VPR's are so effective!

[davewantsmoore]

3 hours ago, acg said:

multiple subs

I suspect we have discussed this before.

 

... but if you drive the room from one location (or the L+R, if we call that "one) ....  vs if you drive the room from multiple locations .... things look very differently, and so may need different "treatment".   Driving the room from one spot, and then trying to "fix" it by absorbing all the sound, is kinda a bit (not sure of the right word), backwards?

 

3 hours ago, acg said:

Second problem is the 12dB/octave SPL drop between 20Hz and 40Hz. 

Yeah, my immediate thought here is just that you haven't corrected the sealed box rolloff of the woofers properly (12dB/octave).

 

1 hour ago, frednork said:

thanks for continuing to document this massive endeavour.

Very much.

 

1 hour ago, acg said:

Here is the current room waterfall post wall treatments.

It is a really bad chart for showing anything.  I recommend not to use it.

 

Use spectrogram, with normalise to peak frequency ticked.

Otherwise, all that is really seen on waterfall is the uneven frequency response.

[davewantsmoore]

13 minutes ago, acg said:

but don't need as steep a low pass filter.

Why would you want to avoid that, though?

 

The reason they don't need such a steep low pass filter, is that the low pass filter is built into the driver (eg. inductance, mass, rms).   Your 'external' low pass filer will be much more linear (read: high performance).

[playdough]

Hi 

I use a PRO 21" sub, high sensitivity, sadly no longer made and cannot find the specs any more. The enclosure in all practicality ignored the QTC and Fs of the actual driver  to the horror of some ppl, I've done this with WinIsd in the past successfully). Main reason I can bend these rules is knowing they are to be used mainly at 1w or less, not 2000 W and 140dB, They do what I want at 1w, make prestigious 20Hz wide Q, first crossover is 130Hz. 

Yes the crossover, if passive would suggest probably a big inductor/cap 2nd order roll off at say 40Hz, as the output rises to 100Hz. But that's just me and my high sensitivity theory.

18 minutes ago, acg said:

A very heavy construction, sealed watertight

Looks weather boarded by the WFGraph not double brick, happy to be incorrect.

20 minutes ago, acg said:

VPR's are so effective

Yes congratulations. Amazingly effective and if double brick,,,,,,,,,, ,,astonishing truly.

Large MLV  traps here a similar principal although your membrane is a piston, mine is a limp weighted 6Kgm2 loaded  vinyl, air sealed enclosure/casement, some 5.5m2 installed area can't go wrong really, phenomenal SQ improvement in room.

Cheers, won't cut in anymore. 

[acg]

3 minutes ago, davewantsmoore said:

Why would you want to avoid that, though?

 

The reason they don't need such a steep low pass filter, is that the low pass filter is built into the driver (eg. inductance, mass, rms).   Your 'external' low pass filer will be much more linear (read: high performance).

 

Dave, I've modelled some of the "high-sensitivity" more pro type drivers in Hornresp and they have frequency responses out into the several kilohertz.  The passive filter required to cut that extra output needs to start so low that it also cuts away much of the gain in the passband I am interested in sub 100Hz, plus it needs to be at least one extra order (more likely two) to have a similar frequency response.  See below...

 

High-sensitivity subwoofer (2 x 21" in a sealed box)... 

  

 

Low sensitivity subwoofer (1 x 24" in the same sized sealed box)...

 

 

10dB extra to filter out from 100Hz through 1000Hz with the high-sensitivity woofer and no real advantage at 20Hz.

 

 

 

[acg]

24 minutes ago, davewantsmoore said:

I suspect we have discussed this before.

 

... but if you drive the room from one location (or the L+R, if we call that "one) ....  vs if you drive the room from multiple locations .... things look very differently, and so may need different "treatment".   Driving the room from one spot, and then trying to "fix" it by absorbing all the sound, is kinda a bit (not sure of the right word), backwards?

 

Yes, I suppose backwards, in some ways.  But the bass treatment thus far has served multiple purposes not just improvements in frequency response sub 100Hz at the listening chair but also up to 600Hz or more and it has massively altered reverberation times which has improved clarity.  At my desk in the corner of the room (this is my office) once unbearable modal resonances (that sometimes approached brown-note territory if the wick was fully lit) are all but eliminated which makes my days much more pleasant.  That alone is worth the effort.  Across the entire room bass is so much more even now.

 

Previously I have experimented with a midbass channel in the parts of the room in which it could be accommodated but unfortunately those positions had almost zero effect on the 60Hz-110Hz suckout.  Although now, I suppose, I could bring those boxes back in and try them in the new back-wall locations to see what they do.  

 

 

 

 

39 minutes ago, davewantsmoore said:

Yeah, my immediate thought here is just that you haven't corrected the sealed box rolloff of the woofers properly (12dB/octave).

 

No correction applied at all.  Not sure it is feasible with a passive line-level filter and the amount of gain I have on hand.

 

 

41 minutes ago, davewantsmoore said:

It is a really bad chart for showing anything.  I recommend not to use it.

 

Use spectrogram, with normalise to peak frequency ticked.

Otherwise, all that is really seen on waterfall is the uneven frequency response.

 

Not sure which presentation is more useful, so both are included below...

 

 

 

[frednork]

43 minutes ago, acg said:

 

All good @frednork, thanks for taking an interest.

 

The "pre" graph was taken just using the Bass Cannons whereas the "post" graph also includes the hornstack, so the upperbass horn will have influence on the SPL down to say 100Hz.  During all of the room treatment iterations the humps at 40Hz and 20Hz barely moved so I used these to match the SPL graphs.  The match should be accurate to within a few dB I would say.  I've also measured huge spl reductions at the sidewall boundaries of more than 20dB at frequencies like 30Hz to feel comfortable enough that the modal influence in that range has been lowered enough mid-room to perhaps increase SPL output at the listening position.

 

 

 

Agreed.  Would have been nice to have some more room gain but the lowest mode is circa 30Hz so there is not so much to be had.

 

 

 

Agreed.

 

 

Agreed.  But there are other options than 1000w on the Bass Cannons.

 

The horn stack currently runs un-attenuated and rampant at 109dB/w/m and the Bass Cannons un-attenuated and less rampant at 98dB/w/m.  The SET amp for the Bass Cannons is capable of driving them to 105dB at 40Hz at the listening chair and probably 95dB at 20Hz based on in-room response.  This should be loud enough.  There are options to adjust gains within the amplifiers, particularly within the passive line level filters, and I can also attenuate the horn-stack channels individually so rather than boosting the Bass Cannons to match the higher sensitivity horn stack, I can add gain to the Bass Cannon amplifier.  Probably not another 20dB gain as that is a lot but I can simultaneously cut signal to the horn stack to make up any shortfall.  That is the beauty of passive line-level filters as opposed to speaker level filters...gain structure is always optimised such that the amplifiers are never doing more work than needed (i.e. amplifying to a large voltage and then attenuating before the speaker).

 

The second pair of subs for the back wall may possibly make things a little easier by increasing output in that 20Hz-40Hz range.

 

Should have said first (prior to all the doom and gloom which was responding to the later parts of your post which were front of mind at the time ),

 

with the VPR's, the smoothing you have achieved for passive treatment is enormous, Great job, massive effort and  gutsy exploration which also shows what is required to really make a significant difference at these frequencies with passive treatments.

 

Now, back to the doom and gloom,

Yes, hopefully other options available rather than just power amp gain. Just remember whatever your db requirement at the lowest frequency you want to reproduce accurately, sets the overall level for max spl as that is your biggest hole to fill.

 

95dB or lower may not  quite be enough for the whole system. Suspect you may want to go louder than that at times.

 

With the roof VPR is it targeting the 200 valley specifically, or will it affect lower frequencies also?

 

it will be interesting to see what some subs in other locations might do. Am sure you are aware but just a reminder going from one sub to 2 subs outputting the same spl ends up in a 6db increase whereas doubling the power to one sub increases by 3db.

 

 

[playdough]

Had a dig about found the reflex enclosure.  measurement.

Testing.
The 2 lines at the top (highest dB) are the top and bottom 21" driver close mic, top most line the bottom cone. it is a little louder because of the close to ground boundary.
Second pair of lines that reach furthest down the frequency is close mic on the vent, 
The bottom 2 lines are of the whole speaker at 1m away.

 

Sorry thought this actual measurement might be helpful. Sort of supports my theory.

Can delete on request no problem.

 

[acg]

21 minutes ago, frednork said:

 

with the VPR's, the smoothing you have achieved for passive treatment is enormous, Great job, massive effort and  gutsy exploration which also shows what is required to really make a significant difference at these frequencies with passive treatments.

 

 

Thanks.  I am super happy with the results thus far.

 

22 minutes ago, frednork said:

ow, back to the doom and gloom,

Yes, hopefully other options available rather than just power amp gain. Just remember whatever your db requirement at the lowest frequency you want to reproduce accurately, sets the overall level for max spl as that is your biggest hole to fill.

 

95dB or lower may not  quite be enough for the whole system. Suspect you may want to go louder than that at times.

 

For sure I will want to play music louder than 95dB.  105dB peaks are loud in the midrange...really loud...but the lowest note of a double bass is circa 40Hz at 95dB max, and a kick drum, although not as low will max out at say 115dB at 50Hz.  The way I see things right now is that down to 40Hz my system can do those volumes, below 40Hz not so much.  Not a lot of musical content at 20Hz so if it needs filling up at the end of the process 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...it is the last step once all other solutions have been exhausted.

 

40 minutes ago, frednork said:

 

With the roof VPR is it targeting the 200 valley specifically, or will it affect lower frequencies also?

 

 

The ceiling VPR's are a slightly different design (different foam) for safety reasons but should be effective to 60Hz and below and as far up as 20kHz if needed.  There are still some HF reflections to sort out in the room and I suspect the ceiling is a primary culprit.

 

43 minutes ago, frednork said:

it will be interesting to see what some subs in other locations might do. Am sure you are aware but just a reminder going from one sub to 2 subs outputting the same spl ends up in a 6db increase whereas doubling the power to one sub increases by 3db.

 

Yes, in my situation that second pair of subs covering a similar frequency range is a large bonus.  If it manages to boost 20Hz by 6dB then that is some of my problem gone in that area. 

[acg]

48 minutes ago, playdough said:

Can delete on request no problem.

 

All good.  Please keep posting.

[davewantsmoore]

1 hour ago, acg said:

that it also cuts away much of the gain in the passband

 

So increase the gain.

.... or use a steeper filter slope.

<shrug>

 

1 hour ago, acg said:

10dB extra to filter out from 100Hz through 1000Hz

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?

 

1 hour 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.

 

1 hour ago, acg said:

No correction applied at all.  Not sure it is feasible with a passive line-level filter and the amount of gain I have on hand.

I remember some in depth discussion about this years (?!) back now.   These sorts of things needed to be in the design from the get go.... so the drivers have a flat response in their intended band.

 

1 hour ago, acg said:

Not sure which presentation is more useful, so both are included below...

Both are ok. 

 

[almikel]

12 minutes ago, davewantsmoore said:

Both are ok. 

Hi Dave,

1 hour ago, acg said:

Not sure which presentation is more useful, so both are included below...

 

 

 

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

 

cheers

Mike