DIY Speakers and Subs > 4xSI18 + CV5000 + 11cu ft = a good option?

Hey all,

This is a cross-post with the UK AV forum I'm on but I thought I would post it here, the land of the multi-driver DIY beasts to get your views as well!

I have been playing with WinISD now I have (finally!!) got it to work on this knackered old laptop at home, and think I've worked out a box that might be ok for my needs... (wants )

So, I would welcome your feedback on the following!!

I'm thinking 4xSI18s, a CV5000, 11 cu ft sealed...

It 'feels' like it might be a bit small, with each driver only seeing 2.75 cu ft rather than the 6 cu ft needed for Qtc of 0.707, but the power would be there from the CV5000 and the graphs appear to show that XMax of c.22.5mm would only be reached at just about 4Hz, with XMech of 35mm / 43mm (Databass estimated measurements / SI ratings) reached at between 1.5 and 2Hz, which shouldn't be a problem because the amp starts roll off at 5Hz, I believe, and room gain should provide a half-decent boost (6-10dB on a good day??).

I think the 8033 can boost down to 5Hz as well, so that could be an option? (I can't work out how to apply a boost in WinISD to bump up the 5-20Hz range!)

Does the above sound like a decent idea? I have to thank those that have already offered assistance on this it's just that this is my first time actually playing with WinISD now I've got it to work, so I thought I'd post up my noob thoughts lol

Thank you in advance for any input you can give!

This is a cross-post with the UK AV forum I'm on but I thought I would post it here, the land of the multi-driver DIY beasts to get your views as well!

I have been playing with WinISD now I have (finally!!) got it to work on this knackered old laptop at home, and think I've worked out a box that might be ok for my needs... (wants )

So, I would welcome your feedback on the following!!

I'm thinking 4xSI18s, a CV5000, 11 cu ft sealed...

It 'feels' like it might be a bit small, with each driver only seeing 2.75 cu ft rather than the 6 cu ft needed for Qtc of 0.707, but the power would be there from the CV5000 and the graphs appear to show that XMax of c.22.5mm would only be reached at just about 4Hz, with XMech of 35mm / 43mm (Databass estimated measurements / SI ratings) reached at between 1.5 and 2Hz, which shouldn't be a problem because the amp starts roll off at 5Hz, I believe, and room gain should provide a half-decent boost (6-10dB on a good day??).

I think the 8033 can boost down to 5Hz as well, so that could be an option? (I can't work out how to apply a boost in WinISD to bump up the 5-20Hz range!)

Does the above sound like a decent idea? I have to thank those that have already offered assistance on this it's just that this is my first time actually playing with WinISD now I've got it to work, so I thought I'd post up my noob thoughts lol

Thank you in advance for any input you can give!

I would highly recommend that you go with two duel opposed enclosures instead of all four in the same enclosure. That way you will have a better frequency coverage over having just one enclosure, and if it were me, I would do 4 separate enclosures spread around the room!

Are you settled on 11 cuft total due to space constraints? I currently give each SI about 4 cuft, so 4 DO boxes at 8.0 cuft after bracing and driver displacement. Giving them 1100 watts per driver, and a 3-6dB boost down low depending on what I am watching, is near perfect for these drivers. Going as small as you are thinking on the box for 4 would require more power and then you are getting into the thermal limits of the driver to get it to operate properly. Any chance you can bump the box to 16cuft or possibly even more?

Thanks for the replies, it is very much appreciated

I've only been playing with WinISD for a few hours so I'm sure I am not fully aware of everything I should be, but I was basing 11 cu ft on the graphs:

SPL chart

Cone Excursion chart

Group Delay chart

Transfer Function Phase chart

I can't get the max power handling graph to work, though, so I'm clearly doing something wrong lol

I'm also not entirely sure about the Transfer Phase chart and what it will mean in real life! I presume the transition is quite smooth over the graph, therefore it will be ok?

From the graphs that I have got working, I was assuming I should aim to keep excursion under XMax for as much of the chart as possible - so I was thinking that having XMax at 4Hz would mean plenty of headroom throughout the rest of the range, and amplifier roll-off would reduce power below that and protect the driver, which would still be within XMech (either 35m or 43mm, going from the two different figures I've seen) all the way down to 1.5-2Hz.

If I bump the box size up to 16cuft I think I end up over XMax at 10Hz and below with the CV5000 amp (can't play on this computer, WinISD is on the home laptop!) so I guess dual alternative, lower power amps with a ~5Hz roll-off (EP4000?) could be an option instead? (More expensive, though! lol)

But I guess that boosting when it's already near XMax at ~10Hz would be dangerous, as I think it's double the cone excursion for +6dB, if I recall correctly from a Bill Fitzmaurice post??

I do hear what you're saying re: making them DO and spreading them round the room for a more consistent response but I only really have one set of three seating positions in a row about 8-10 feet from the TV (about 12 feet from where the sub would be as the sub would be behind the screen) so I am not to worried about having an inconsistent response across the room, just as long as it's loud enough at the seating position.

If you imagine a 20'x14'x7'8" room, the TV is tucked into one corner against the short wall, but behind that corner is actually an alcove of c. 38"w x 25"d x 7'8" tall. I can therefore build (a) box(es) within that 38"x25" footprint and extend up as required!

I'm guessing that couple of DO boxes could either be placed across the alcove, firing left and right and then radiating out into the room, or stood up? If I had them firing left and right, would the fact the walls of the alcove were so close add to the deep bass or load the cone too much? (I may be getting confused with the effect of extending the port in a vented box that is placed in a corner...)

I was thinking that two FF and two DF (but with a 'surround' behind and round the sides of the DFs, so the airflow is contained and then 'fires' into the room in, I presume, the same phase as the FF?) would perhaps tidy up the phasing because they would both be firing into the room at the same time and not bouncing round the alcove before eminating out?

I thought this box design lark would be easy, but alas... lol

I've only been playing with WinISD for a few hours so I'm sure I am not fully aware of everything I should be, but I was basing 11 cu ft on the graphs:

SPL chart

Cone Excursion chart

Group Delay chart

Transfer Function Phase chart

I can't get the max power handling graph to work, though, so I'm clearly doing something wrong lol

I'm also not entirely sure about the Transfer Phase chart and what it will mean in real life! I presume the transition is quite smooth over the graph, therefore it will be ok?

From the graphs that I have got working, I was assuming I should aim to keep excursion under XMax for as much of the chart as possible - so I was thinking that having XMax at 4Hz would mean plenty of headroom throughout the rest of the range, and amplifier roll-off would reduce power below that and protect the driver, which would still be within XMech (either 35m or 43mm, going from the two different figures I've seen) all the way down to 1.5-2Hz.

If I bump the box size up to 16cuft I think I end up over XMax at 10Hz and below with the CV5000 amp (can't play on this computer, WinISD is on the home laptop!) so I guess dual alternative, lower power amps with a ~5Hz roll-off (EP4000?) could be an option instead? (More expensive, though! lol)

But I guess that boosting when it's already near XMax at ~10Hz would be dangerous, as I think it's double the cone excursion for +6dB, if I recall correctly from a Bill Fitzmaurice post??

I do hear what you're saying re: making them DO and spreading them round the room for a more consistent response but I only really have one set of three seating positions in a row about 8-10 feet from the TV (about 12 feet from where the sub would be as the sub would be behind the screen) so I am not to worried about having an inconsistent response across the room, just as long as it's loud enough at the seating position.

If you imagine a 20'x14'x7'8" room, the TV is tucked into one corner against the short wall, but behind that corner is actually an alcove of c. 38"w x 25"d x 7'8" tall. I can therefore build (a) box(es) within that 38"x25" footprint and extend up as required!

I'm guessing that couple of DO boxes could either be placed across the alcove, firing left and right and then radiating out into the room, or stood up? If I had them firing left and right, would the fact the walls of the alcove were so close add to the deep bass or load the cone too much? (I may be getting confused with the effect of extending the port in a vented box that is placed in a corner...)

I was thinking that two FF and two DF (but with a 'surround' behind and round the sides of the DFs, so the airflow is contained and then 'fires' into the room in, I presume, the same phase as the FF?) would perhaps tidy up the phasing because they would both be firing into the room at the same time and not bouncing round the alcove before eminating out?

I thought this box design lark would be easy, but alas... lol

It really is pretty easy, you just have to get your other design aspects down first like space constraints, amount of power per driver and VoiceCoil configuration for how you plan on wiring. The problem is on the model it doesn't take any signal chain rolloff into effect and odds are between your BD player, AVR, MiniDSP or other EQ device, and amp, your chain will roll off significantly enough starting at worst around 20hz, and at best 7-10hz. putting these drivers in that small a box and giving them a CV (assuming 2500x2 a 2ohm stereo) is the right amount of power, but you'll need to go a little bigger as I suggested to compensate better for the signal rolloff.

Well, I've had a play and I think I've settled on 15 cu ft and 5000w of power, thank you for the recommendation!

Excursion rises as you go down the scale, with a plateau around 3mm below XMax between about 17Hz and 9Hz, at which point it starts to rise, reaching XMax around 7Hz. If XMech is 35mm then it hits that at 2Hz... so with that, the Xmax at 7Hz, and the signal chain roll off, it will hopefully limit excursion and keep the drivers safe, and room gain will hopefully bump in-room response back up again!

I*think* I've got the configuration correct in WinISD - if I've got it right I've got them as Znom(?) of 4Ohm (for the D4 version) with VCs wired in parallel...

Although as I write this I realise they would have to be the D2 version instead, so I'd have two drivers to each channel, each driver wired in parallel (to give 1 Ohm) then wired together in series (to give 2 Ohm)... :facepalm:

I don't think there's a way to wire 4 x D4s to show a 2Ohm load other than if all 4 are wired parallel and each to individual 2Ohm capable channels of a 4 channel amp (which the CV5000 isn't)...

I will have to do some more playing tonight but with the VCs as 2Ohm (I think?) although how does WinISD know if they are all in series or if each driver is being run separately off one channel of an amp?? Won't it just assume they are being run at 1 Ohm and therefore increase (double?) their output?

Man, this is confusing... lol

Popalock's build with each driver seeing about 3.87 cu ft (IIRC?) does at least give me confidence that they can take the power in a box of similar size, as I'm sure the Clone is giving them about the same as the CV5000 would, and he has said that he appears to be amp-limited(!!) in some scenes, with the subs lapping it up!

WRT amplification, it appears the American Audio V-6001 is identical / virtually identical to the CV5000, but perhaps sligtly more powerful?

http://www.directproaudio.com/product.cfm?directid=57377

I am a little worried regarding power draw (!) but I did find an archived webpage on a CV forum that stated the following WRT power consumption:

so hopefully things should be alright on my breaker board as long as I'm not attempting to blow my windows out with bass?? lol

Thanks for all your help with this, it's really useful!

Excursion rises as you go down the scale, with a plateau around 3mm below XMax between about 17Hz and 9Hz, at which point it starts to rise, reaching XMax around 7Hz. If XMech is 35mm then it hits that at 2Hz... so with that, the Xmax at 7Hz, and the signal chain roll off, it will hopefully limit excursion and keep the drivers safe, and room gain will hopefully bump in-room response back up again!

I

Although as I write this I realise they would have to be the D2 version instead, so I'd have two drivers to each channel, each driver wired in parallel (to give 1 Ohm) then wired together in series (to give 2 Ohm)... :facepalm:

I don't think there's a way to wire 4 x D4s to show a 2Ohm load other than if all 4 are wired parallel and each to individual 2Ohm capable channels of a 4 channel amp (which the CV5000 isn't)...

I will have to do some more playing tonight but with the VCs as 2Ohm (I think?) although how does WinISD know if they are all in series or if each driver is being run separately off one channel of an amp?? Won't it just assume they are being run at 1 Ohm and therefore increase (double?) their output?

Man, this is confusing... lol

Popalock's build with each driver seeing about 3.87 cu ft (IIRC?) does at least give me confidence that they can take the power in a box of similar size, as I'm sure the Clone is giving them about the same as the CV5000 would, and he has said that he appears to be amp-limited(!!) in some scenes, with the subs lapping it up!

WRT amplification, it appears the American Audio V-6001 is identical / virtually identical to the CV5000, but perhaps sligtly more powerful?

http://www.directproaudio.com/product.cfm?directid=57377

I am a little worried regarding power draw (!) but I did find an archived webpage on a CV forum that stated the following WRT power consumption:

Quote:

About the current, we will have to do some extrapolation. if I remember correctly, the CV-2800 is a 2-step class H, while the CV-5000 is a 3-step class H. This means the CV-5000 will be slightly more efficient, so the calculations here should have a little safety factor since I will not adjust for the extra efficiency.

The peak current draw of the CV-5000 is 27A @ 230V, while the CV-2800 is 16A @ 230V. Realistically, the highest average current draw you should see is the 1/3 power rating, which is 8.3A for the CV-2800, yielding an efficiency of about 49%. This extrapolates to a draw of 14.8A on the CV-5000 for a 4 Ohm load bridged or a 2 Ohm load stereo.

...

Be certain to properly configure your subsonic filter on your Driverack PA, around 35 to 40 Hz at 18 dB / octave or so. Setting the subsonic filter (high-pass filter) too low will significantly increases power draw and leads to blown woofers, but you probably already knew that. Keep in mind however with the dynamics of music, the peaks may pose the threat of tripping a breaker; running a double-kick too hot or some low-end feedback might still be enough to trip the breaker though, it's hard to say.

Anyhow, a properly run system should have an average draw at the 1/8 power rating, which I have calculated to be 7.3A @ 230V on a 2 Ohm stereo load for the CV-5000 but using the lower efficiency of the CV-2800. I will send a message to our member 1Rolli to see if he can verify these numbers with some measurements.

The peak current draw of the CV-5000 is 27A @ 230V, while the CV-2800 is 16A @ 230V. Realistically, the highest average current draw you should see is the 1/3 power rating, which is 8.3A for the CV-2800, yielding an efficiency of about 49%. This extrapolates to a draw of 14.8A on the CV-5000 for a 4 Ohm load bridged or a 2 Ohm load stereo.

...

Be certain to properly configure your subsonic filter on your Driverack PA, around 35 to 40 Hz at 18 dB / octave or so. Setting the subsonic filter (high-pass filter) too low will significantly increases power draw and leads to blown woofers, but you probably already knew that. Keep in mind however with the dynamics of music, the peaks may pose the threat of tripping a breaker; running a double-kick too hot or some low-end feedback might still be enough to trip the breaker though, it's hard to say.

Anyhow, a properly run system should have an average draw at the 1/8 power rating, which I have calculated to be 7.3A @ 230V on a 2 Ohm stereo load for the CV-5000 but using the lower efficiency of the CV-2800. I will send a message to our member 1Rolli to see if he can verify these numbers with some measurements.

so hopefully things should be alright on my breaker board as long as I'm not attempting to blow my windows out with bass?? lol

Thanks for all your help with this, it's really useful!

Ok, quick update on my snail-like progress lol

I am currently torn between building the 7.5cuft (internal) x 2 to get 15 cu ft in total, or drop down to perhaps 13 cu ft to keep size slightly less imposing in my tiny UK room lol

Modelling suggests the smaller size should only make a small different in output, should hopefully protect the driver from bottoming out a bit better, and once EQd with XT32, I doubt I will be running the CV5000 anywhere near max on its gain dial, therefore I think I should be pretty safe! I am still confused as to how to get the most low end if Audyssey is going to EQ to 75dB and probably only the 3rd notch on the gain dial, though...

The CV5000 has arrived at least and it is a beast, it is so heavy!! Fans are 24V, stupid noisy (45dB each) and seemingly flow 100.5CFM, according to the specs I've found, so hopefully I can replace them with 12V quiet computer fans in series (although loss of warranty is a concern, so I may see if I can build a small box that extends from the back into a padded but vented area in an attempt to quieten them down and not replace them). I've confirmed that my ring main is rated to 32Amps so the CV5000 should be fine with its 27amps or so on our 240v power at 5000w.

Drivers have been ordered from SI so should hopefully not take too long, and once I've finally sorted out the box design I can get the guy to construct it for me

Just to check, in a dual opposed setup, do the drivers actually have to be directly back to back or facing each other? or as long as one is on the front wall and one on the back wall, is that enough, perhaps one at the top of the front wall and one at the bottom of the back wall??

Thanks for everyone's help on this forum, not just in this thread but in all the posts, it's really useful!

I am currently torn between building the 7.5cuft (internal) x 2 to get 15 cu ft in total, or drop down to perhaps 13 cu ft to keep size slightly less imposing in my tiny UK room lol

Modelling suggests the smaller size should only make a small different in output, should hopefully protect the driver from bottoming out a bit better, and once EQd with XT32, I doubt I will be running the CV5000 anywhere near max on its gain dial, therefore I think I should be pretty safe! I am still confused as to how to get the most low end if Audyssey is going to EQ to 75dB and probably only the 3rd notch on the gain dial, though...

The CV5000 has arrived at least and it is a beast, it is so heavy!! Fans are 24V, stupid noisy (45dB each) and seemingly flow 100.5CFM, according to the specs I've found, so hopefully I can replace them with 12V quiet computer fans in series (although loss of warranty is a concern, so I may see if I can build a small box that extends from the back into a padded but vented area in an attempt to quieten them down and not replace them). I've confirmed that my ring main is rated to 32Amps so the CV5000 should be fine with its 27amps or so on our 240v power at 5000w.

Drivers have been ordered from SI so should hopefully not take too long, and once I've finally sorted out the box design I can get the guy to construct it for me

Just to check, in a dual opposed setup, do the drivers actually have to be directly back to back or facing each other? or as long as one is on the front wall and one on the back wall, is that enough, perhaps one at the top of the front wall and one at the bottom of the back wall??

Thanks for everyone's help on this forum, not just in this thread but in all the posts, it's really useful!

"Just to check, in a dual opposed setup, do the drivers actually have to be directly back to back or facing each other? or as long as one is on the front wall and one on the back wall, is that enough, perhaps one at the top of the front wall and one at the bottom of the back wall??"

to cancel vibrations, the drivers would be back-to-back and wired so that they both move out and both move in at the same time (drivers wired in phase).

BUT, there are plenty of successful builds with both drivers pointing forward on one side of the enclosure (such as both drivers facing forward or both facing down). if the cabinet is heavy enough, there shouldn't be much vibration anyways.

having a dual opposed setup where one driver has its magnet sticking out of the enclosure and drivers wired in reverse phase will cancel a lot of even order distortions. an "inny" version of this is known as the clamshell baffle, while i'm not sure what the "outty" version would be called.

mounting them dual opposed with one driver at the top of a tall cab pointing forward and one driver at the bottom of the cab pointing backward would be most likely to create losses due to rocking the cabinet, so you may wish to avoid that kind of setup.

any in the black are natural cancellation, so fine. the blue require heavier cabs, but are fine. the only one to avoid unless you make a really heavy cab is red circled out.

to cancel vibrations, the drivers would be back-to-back and wired so that they both move out and both move in at the same time (drivers wired in phase).

BUT, there are plenty of successful builds with both drivers pointing forward on one side of the enclosure (such as both drivers facing forward or both facing down). if the cabinet is heavy enough, there shouldn't be much vibration anyways.

having a dual opposed setup where one driver has its magnet sticking out of the enclosure and drivers wired in reverse phase will cancel a lot of even order distortions. an "inny" version of this is known as the clamshell baffle, while i'm not sure what the "outty" version would be called.

mounting them dual opposed with one driver at the top of a tall cab pointing forward and one driver at the bottom of the cab pointing backward would be most likely to create losses due to rocking the cabinet, so you may wish to avoid that kind of setup.

any in the black are natural cancellation, so fine. the blue require heavier cabs, but are fine. the only one to avoid unless you make a really heavy cab is red circled out.

LTD, as usual, is always spot on in his explanations.

Speaking from personal experience, I am running dual 5' tall cabs with 3 forward facing drivers in each one. Even at full tilt, I have very, very little vibration. I could place a lamp on top without worry of it ever falling off. I used 3/4" mdf for the entire build so YMMV.

Speaking from personal experience, I am running dual 5' tall cabs with 3 forward facing drivers in each one. Even at full tilt, I have very, very little vibration. I could place a lamp on top without worry of it ever falling off. I used 3/4" mdf for the entire build so YMMV.

That is an amazingly helpful diagram, you are a star!

My design was looking a lot like the one in red lol

I think I can shoot for one of the other ones easily, though, so back to the drawing board for me!

I will update again as soon as I can

Thanks for your input too, bass addict, I am surprised there's so little movement with 3 18s moving that much!

My design was looking a lot like the one in red lol

I think I can shoot for one of the other ones easily, though, so back to the drawing board for me!

I will update again as soon as I can

Thanks for your input too, bass addict, I am surprised there's so little movement with 3 18s moving that much!

Hey all

Just another update! I don't know how it's now December... this year is disappearing quickly!

I have re-modelled in WinISD following learning about how WinISD sets your box 'leakiness' (the QI value) to 10 ('very leaky', the Red line in the links below) - thank you to FoLLgoTT and LTD02 (IIRC??) for pointing this out, it is a big help

I have remodelled with a QI value of 50 ('no leaks'), therefore I have amended this value and re-graphed, as per the links below.

I've also upped the size to 7.5cuft or so per dual cabinet, as previously mentioned.

Having done so I have seen the cone excursion drop drastically from the original modelled response, and the drivers should hopefully now remain within XMax all the way down to 3Hz, so they should be safe from over-excursion and always within XMech, even with 5kw of power! (Assuming system roll-off, which I fully expect by 3Hz! lol)

I guess my main concern now will be cooking voicecoils from extended ultra-low bass scenes at high volumes in movies... (OHF, anyone? *concerned look* lol)

I also played with the QA (interior stuffing) values as well - going from minimal fill (QA value of 80, the Green line) to lined walls (QA of 20, the Blue line) to totally stuffed (QA of 5, the Yellow line). This has reduced the box Q from 0.841, to 0.815, and then down to 0.726 repectively (so much nearer to 0.707) but I am a little worried that too much stuffing will make it sound 'muffled'?

I may work to the Blue line and line the walls fairly well / stick a few pillows in there, while making sure the box is as airtight as possible!

Any comments or thoughts are gratefully received I am pretty happy with the results so just want some reassurance really! lol

Links to graphs are as follows - I still have no options to upload pictures to my posts here, I don't understand why

Output SPL

Excursion

Amplifier apparent load

Impedance

Group Delay

Just another update! I don't know how it's now December... this year is disappearing quickly!

I have re-modelled in WinISD following learning about how WinISD sets your box 'leakiness' (the QI value) to 10 ('very leaky', the Red line in the links below) - thank you to FoLLgoTT and LTD02 (IIRC??) for pointing this out, it is a big help

I have remodelled with a QI value of 50 ('no leaks'), therefore I have amended this value and re-graphed, as per the links below.

I've also upped the size to 7.5cuft or so per dual cabinet, as previously mentioned.

Having done so I have seen the cone excursion drop drastically from the original modelled response, and the drivers should hopefully now remain within XMax all the way down to 3Hz, so they should be safe from over-excursion and always within XMech, even with 5kw of power! (Assuming system roll-off, which I fully expect by 3Hz! lol)

I guess my main concern now will be cooking voicecoils from extended ultra-low bass scenes at high volumes in movies... (OHF, anyone? *concerned look* lol)

I also played with the QA (interior stuffing) values as well - going from minimal fill (QA value of 80, the Green line) to lined walls (QA of 20, the Blue line) to totally stuffed (QA of 5, the Yellow line). This has reduced the box Q from 0.841, to 0.815, and then down to 0.726 repectively (so much nearer to 0.707) but I am a little worried that too much stuffing will make it sound 'muffled'?

I may work to the Blue line and line the walls fairly well / stick a few pillows in there, while making sure the box is as airtight as possible!

Any comments or thoughts are gratefully received I am pretty happy with the results so just want some reassurance really! lol

Links to graphs are as follows - I still have no options to upload pictures to my posts here, I don't understand why

Output SPL

Excursion

Amplifier apparent load

Impedance

Group Delay

I don't have a diyaudio account

great post memx!

Thanks LTD, and thanks again for your useful posts - they are much appreciated!

I do feel a bit like I'm 'taking a risk' with this as not (m)any people have done the PPSL thing, and I'm still worried I will kill the drivers with the CV5000 following discussion in another thread about the low volts (57v?) Josh needed to max out the drivers at certain frequencies (30Hz or so?) even with the relatively small boxes these will be, so I will just have to attempt to set the gain structure up correctly and then turn it up slowly on a few scenes that dig deep, and cross my fingers I don't bottom them out or cook them...

splotten - sorry about that! I can't access a photo-hosting site from work thanks to 'security restrictions' so I've had to rely on attachments and I have no option here to attach anything for some reason!

I do feel a bit like I'm 'taking a risk' with this as not (m)any people have done the PPSL thing, and I'm still worried I will kill the drivers with the CV5000 following discussion in another thread about the low volts (57v?) Josh needed to max out the drivers at certain frequencies (30Hz or so?) even with the relatively small boxes these will be, so I will just have to attempt to set the gain structure up correctly and then turn it up slowly on a few scenes that dig deep, and cross my fingers I don't bottom them out or cook them...

splotten - sorry about that! I can't access a photo-hosting site from work thanks to 'security restrictions' so I've had to rely on attachments and I have no option here to attach anything for some reason!

xmax is only the linear limit of the bl field or approximately the 10% distortion point, not the end of the road.

xmech, the point where excursion will cause some sort of damage potentially, is 43mm. not sure what you are worried about?

http://stereointegrity.com/index.php?id=60

xmech, the point where excursion will cause some sort of damage potentially, is 43mm. not sure what you are worried about?

http://stereointegrity.com/index.php?id=60

I am completely confused on this XMax / XMech thing on the SI subs - the website and Josh seem to have different opinions!

The website says 22.5mm Xmax (one-way linear) and 43mm Xmech (one-way)

Josh's review (http://www.data-bass.com/data?page=driver&id=46) says "The xmax is rated at 20.5mm with an xmech of 36.5mm (Confirmed by Klippel testing)." but then says "Manufacture's Xmax 23.5 mm" in the table, and "At what looked like maybe 60 or 65mm peak to peak the driver abruptly tapped the back plate. Xmech effectively seems to be closer to 30 or 35mm inward".

I have therefore been a bit paranoid about it lol

Hopefully the WinISD graphs are accurate and all should be safe!

And regarding another thing I'm paranoid about...

Looking at the table on Josh's site again (http://www.data-bass.com/data?page=system&id=90&mset=97) (under the Static Graphs tab):

the 31.5Hz measurement seems to need 95v to reach excursion limits (XMech, going from Josh's text above the tables on the DB webpage).

If Watts = Volts x Amps, a 5000w CV5000 on a 240v 32amp Ring-main power circuit (breaker rating, as I understand it) would give 156.25v at full power, I think?? (5000w/32amps=156.25v?) Would this therefore mean that the following are max safe amplifier power outputs?

10Hz -- 126v x 32 Amps = 4032watts

12.5Hz -- 126v x 32 Amps = 4032watts

16Hz -- 119v x 32 Amps = 3808 watts

20Hz -- 113v x 32 Amps = 3616 watts

25Hz -- 109v x 32 Amps = 3488 watts

31.5Hz -- 95v x 32 Amps = 3040 watts

40Hz -- 126v x 32 Amps = 4032 watts

With the following being power supply / amplifier limited?

50Hz -- 189v x 32 Amps = 6048 watts

63Hz -- 283v x 32 Amps = 9056 watts

80Hz -- 283v x 32 Amps = 9056 watts

100Hz -- 283v x 32 Amps = 9056 watts

125Hz -- 283v x 32 Amps = 9056 watts

**Or...**

Do I need to take account of the resistance (Impedance) graph I posted a link to?

i.e. if Ohms = Volts / Amps, that means that Ohm / Volts = Amps, therefore the Impedance graph would indicate the following Amps required at each frequency to deliver 5000w (as modelled in WinISD to get the Impedance graph in Ohms) assuming a 240v power supply?

10Hz -- 4 Ohms / 240v = 60 Amps

12.5Hz -- 4.1 Ohms / 240v = 58.5 Amps

16Hz -- 4.3 Ohms / 240v = 55.8 Amps

20Hz -- 5 Ohms / 240v = 48 Amps

25Hz -- 6 Ohms / 240v = 40 Amps

31.5Hz -- 10 Ohms / 240v = 24 Amps

40Hz -- 15.2 Ohms / 240v = 15.8 Amps (stuffed)

40Hz -- 24 Ohms / 240v = 10 Amps (lined)

50Hz -- 8 Ohms / 240v = 30 Amps

63Hz -- 5.7 Ohms / 240v = 42 Amps

80Hz -- 4.5 Ohms / 240v = 53.3 Amps

100Hz -- 4 Ohms / 240v = 60 Amps

125Hz -- 3.9 Ohms / 240v = 61.5 Amps

which would then be fed back into the first calculations to give the following max amplifier outputs before the single driver tested by Josh was maxed out, using the Max SPL No THD Limit voltages from his table??

10Hz -- 126v x 60 Amps = 7560 watts

12.5Hz -- 126v x 58.5 Amps = 7375.6 watts

16Hz -- 119v x 55.8 Amps = 6641.9 watts

20Hz -- 113v x 48 Amps = 5424 watts

25Hz -- 109v x 40 Amps = 4360 watts

31.5Hz -- 95v x 24 Amps = 2280 watts

40Hz -- 126v x 15.8 Amps = 1989.5 watts (stuffed)

40Hz -- 126v x 10 Amps = 1260 watts (lined)

50Hz -- 189v x 30 Amps = 5670 watts

63Hz -- 283v x 42 Amps = 11915.8 watts

80Hz -- 283v x 53.3 Amps = 15093.3 watts

100Hz -- 283v x 60 Amps = 16980 watts

125Hz -- 283v x 61.5 Amps = 17415.4 watts

In which case my 4 drivers should be perfectly safe because they will each be getting 1250w at most, which is less than any of the power outputs above that were max safe limits of a single driver as tested by Josh??

And if I alter the last calculations and input the max voltages with clean THD rather than the Max SPL No THD Limit voltages, the following single driver figures show that with 1250 per driver I should be well under THD limits the whole time?

10Hz -- 57v x 60 Amps = 3420 watts

12.5Hz -- 53v x 58.5 Amps = 3102.5 watts

16Hz -- 58v x 55.8 Amps = 3237 watts

20Hz -- 80v x 48 Amps = 3840 watts

25Hz -- 109v x 40 Amps = 4360 watts

31.5Hz -- 95v x 24 Amps = 2280 watts

40Hz -- 126v x 15.8 Amps = 1989.5 watts (stuffed)

40Hz -- 126v x 10 Amps = 1260 watts (lined)

50Hz -- 189v x 30 Amps = 5670 watts

63Hz -- 283v x 42 Amps = 11915.8 watts

80Hz -- 283v x 53.3 Amps = 15093.3 watts

100Hz -- 283v x 60 Amps = 16980 watts

125Hz -- 283v x 61.5 Amps = 17415.4 watts

Man, my head hurts lol

The website says 22.5mm Xmax (one-way linear) and 43mm Xmech (one-way)

Josh's review (http://www.data-bass.com/data?page=driver&id=46) says "The xmax is rated at 20.5mm with an xmech of 36.5mm (Confirmed by Klippel testing)." but then says "Manufacture's Xmax 23.5 mm" in the table, and "At what looked like maybe 60 or 65mm peak to peak the driver abruptly tapped the back plate. Xmech effectively seems to be closer to 30 or 35mm inward".

I have therefore been a bit paranoid about it lol

Hopefully the WinISD graphs are accurate and all should be safe!

And regarding another thing I'm paranoid about...

Looking at the table on Josh's site again (http://www.data-bass.com/data?page=system&id=90&mset=97) (under the Static Graphs tab):

the 31.5Hz measurement seems to need 95v to reach excursion limits (XMech, going from Josh's text above the tables on the DB webpage).

If Watts = Volts x Amps, a 5000w CV5000 on a 240v 32amp Ring-main power circuit (breaker rating, as I understand it) would give 156.25v at full power, I think?? (5000w/32amps=156.25v?) Would this therefore mean that the following are max safe amplifier power outputs?

10Hz -- 126v x 32 Amps = 4032watts

12.5Hz -- 126v x 32 Amps = 4032watts

16Hz -- 119v x 32 Amps = 3808 watts

20Hz -- 113v x 32 Amps = 3616 watts

25Hz -- 109v x 32 Amps = 3488 watts

31.5Hz -- 95v x 32 Amps = 3040 watts

40Hz -- 126v x 32 Amps = 4032 watts

With the following being power supply / amplifier limited?

50Hz -- 189v x 32 Amps = 6048 watts

63Hz -- 283v x 32 Amps = 9056 watts

80Hz -- 283v x 32 Amps = 9056 watts

100Hz -- 283v x 32 Amps = 9056 watts

125Hz -- 283v x 32 Amps = 9056 watts

Do I need to take account of the resistance (Impedance) graph I posted a link to?

i.e. if Ohms = Volts / Amps, that means that Ohm / Volts = Amps, therefore the Impedance graph would indicate the following Amps required at each frequency to deliver 5000w (as modelled in WinISD to get the Impedance graph in Ohms) assuming a 240v power supply?

10Hz -- 4 Ohms / 240v = 60 Amps

12.5Hz -- 4.1 Ohms / 240v = 58.5 Amps

16Hz -- 4.3 Ohms / 240v = 55.8 Amps

20Hz -- 5 Ohms / 240v = 48 Amps

25Hz -- 6 Ohms / 240v = 40 Amps

31.5Hz -- 10 Ohms / 240v = 24 Amps

40Hz -- 15.2 Ohms / 240v = 15.8 Amps (stuffed)

40Hz -- 24 Ohms / 240v = 10 Amps (lined)

50Hz -- 8 Ohms / 240v = 30 Amps

63Hz -- 5.7 Ohms / 240v = 42 Amps

80Hz -- 4.5 Ohms / 240v = 53.3 Amps

100Hz -- 4 Ohms / 240v = 60 Amps

125Hz -- 3.9 Ohms / 240v = 61.5 Amps

which would then be fed back into the first calculations to give the following max amplifier outputs before the single driver tested by Josh was maxed out, using the Max SPL No THD Limit voltages from his table??

10Hz -- 126v x 60 Amps = 7560 watts

12.5Hz -- 126v x 58.5 Amps = 7375.6 watts

16Hz -- 119v x 55.8 Amps = 6641.9 watts

20Hz -- 113v x 48 Amps = 5424 watts

25Hz -- 109v x 40 Amps = 4360 watts

31.5Hz -- 95v x 24 Amps = 2280 watts

40Hz -- 126v x 15.8 Amps = 1989.5 watts (stuffed)

40Hz -- 126v x 10 Amps = 1260 watts (lined)

50Hz -- 189v x 30 Amps = 5670 watts

63Hz -- 283v x 42 Amps = 11915.8 watts

80Hz -- 283v x 53.3 Amps = 15093.3 watts

100Hz -- 283v x 60 Amps = 16980 watts

125Hz -- 283v x 61.5 Amps = 17415.4 watts

In which case my 4 drivers should be perfectly safe because they will each be getting 1250w at most, which is less than any of the power outputs above that were max safe limits of a single driver as tested by Josh??

And if I alter the last calculations and input the max voltages with clean THD rather than the Max SPL No THD Limit voltages, the following single driver figures show that with 1250 per driver I should be well under THD limits the whole time?

10Hz -- 57v x 60 Amps = 3420 watts

12.5Hz -- 53v x 58.5 Amps = 3102.5 watts

16Hz -- 58v x 55.8 Amps = 3237 watts

20Hz -- 80v x 48 Amps = 3840 watts

25Hz -- 109v x 40 Amps = 4360 watts

31.5Hz -- 95v x 24 Amps = 2280 watts

40Hz -- 126v x 15.8 Amps = 1989.5 watts (stuffed)

40Hz -- 126v x 10 Amps = 1260 watts (lined)

50Hz -- 189v x 30 Amps = 5670 watts

63Hz -- 283v x 42 Amps = 11915.8 watts

80Hz -- 283v x 53.3 Amps = 15093.3 watts

100Hz -- 283v x 60 Amps = 16980 watts

125Hz -- 283v x 61.5 Amps = 17415.4 watts

Man, my head hurts lol

"Man, my head hurts"

you complicated it too much. :-) before my head starts hurting, let's try to sort it out.

voltage squared/resistance = power; e^/r = p

95^2 / 3.5 = 2,578 watts

to calculate current...is impossible because subwoofers have an impedance that varies with frequency. you can see how this affects power and current on the "apparent amplifier tab" in winisd.

but if you want to i = v/r, so 95 / 3.5 = 27 amps where the impedance is at a minimum. obviously the driver isn't going to last too long at that rate as at least one member found out. even a fraction of that, sustained, will cook it.

as the driver gets closer to xmax, the bl field rolls off and even more power is required to move the driver by some amount. so its not a linear relationship as winisd makes it appear. winisd just takes the small signal performance and scales it up, which is fine for a first order approximation, but it won't be exact.

you complicated it too much. :-) before my head starts hurting, let's try to sort it out.

voltage squared/resistance = power; e^/r = p

95^2 / 3.5 = 2,578 watts

to calculate current...is impossible because subwoofers have an impedance that varies with frequency. you can see how this affects power and current on the "apparent amplifier tab" in winisd.

but if you want to i = v/r, so 95 / 3.5 = 27 amps where the impedance is at a minimum. obviously the driver isn't going to last too long at that rate as at least one member found out. even a fraction of that, sustained, will cook it.

as the driver gets closer to xmax, the bl field rolls off and even more power is required to move the driver by some amount. so its not a linear relationship as winisd makes it appear. winisd just takes the small signal performance and scales it up, which is fine for a first order approximation, but it won't be exact.

I think you will be much happier when you turn the things on for the first time and forget about all the charts and graphs. It's all in the sound and output.

Hopefully!

Apologies I didn't thank you for this immediately, LTD - I was attempting to digest it and make sense in my head (I still need to get all this stuff clear in my head lol) but was failing, and have been wiped out with illness over Christmas!

Going from those calculations, and the fact I have a CV5000 to power them at 1250wRMS each (2 drivers drawing 2Ohm per channel), working back I'd get:

Sq Rt of (1250 watts x 3.5) = 66.14378 volts

66.15 / 3.5 = 18.89 amps

which seems safe from a volts point of view and hopefully(??) an amps point of view?

I think I'm just having last minute paranoia again - the drivers are now with me and I have had them all out of the boxes at the same time to go 'oooooh, pretty' and I don't want to kill them! lol

I was thinking that I could put them all on one channel and give them 625wRMS each instead, a la carp, rather than the 1kw+ that popalock gives his with his Clone amp, but I'm not quite sure how I'd wire 4 D2 drivers to show a 2 ohm load to the amp - by my reckoning I could only wire them up to show 1 Ohm (not going to happen) or 4 Ohm (reduced power) so I think I have given myself no choice but to go with 2 drivers per 2Ohm channel...

I think I will have to buy a MiniDSP (and no doubt require a bunch more help choosing the right one!!) and then see if I can employ a LT to shape the curve at the lower end - my initial playing with WinISD to gain a 0.707 curve at the bottom end reduces excursion between 30Hz and the very bottom end, and reduces power to about 1000w per driver rather than 1250w, so it may 'waste' the amp power but should hopefully mean safer drivers... (OHF is still worrying me lol)

If by some miracle I could wire them up to have 625w each then it looks like I drop from 121dB@20Hz and 109dB@10Hz, to 118dB@20Hz and 106dB@10Hz - which I imagine should still be adequate with room gain, hopefully?!

Thanks again for putting up with my noobness lol

Quote:

Originally Posted by **LTD02**

"Man, my head hurts"

you complicated it too much. :-) before my head starts hurting, let's try to sort it out.

voltage squared/resistance = power; e^/r = p

95^2 / 3.5 = 2,578 watts

to calculate current...is impossible because subwoofers have an impedance that varies with frequency. you can see how this affects power and current on the "apparent amplifier tab" in winisd.

but if you want to i = v/r, so 95 / 3.5 = 27 amps where the impedance is at a minimum. obviously the driver isn't going to last too long at that rate as at least one member found out. even a fraction of that, sustained, will cook it.

as the driver gets closer to xmax, the bl field rolls off and even more power is required to move the driver by some amount. so its not a linear relationship as winisd makes it appear. winisd just takes the small signal performance and scales it up, which is fine for a first order approximation, but it won't be exact.

"Man, my head hurts"

you complicated it too much. :-) before my head starts hurting, let's try to sort it out.

voltage squared/resistance = power; e^/r = p

95^2 / 3.5 = 2,578 watts

to calculate current...is impossible because subwoofers have an impedance that varies with frequency. you can see how this affects power and current on the "apparent amplifier tab" in winisd.

but if you want to i = v/r, so 95 / 3.5 = 27 amps where the impedance is at a minimum. obviously the driver isn't going to last too long at that rate as at least one member found out. even a fraction of that, sustained, will cook it.

as the driver gets closer to xmax, the bl field rolls off and even more power is required to move the driver by some amount. so its not a linear relationship as winisd makes it appear. winisd just takes the small signal performance and scales it up, which is fine for a first order approximation, but it won't be exact.

Apologies I didn't thank you for this immediately, LTD - I was attempting to digest it and make sense in my head (I still need to get all this stuff clear in my head lol) but was failing, and have been wiped out with illness over Christmas!

Going from those calculations, and the fact I have a CV5000 to power them at 1250wRMS each (2 drivers drawing 2Ohm per channel), working back I'd get:

Sq Rt of (1250 watts x 3.5) = 66.14378 volts

66.15 / 3.5 = 18.89 amps

which seems safe from a volts point of view and hopefully(??) an amps point of view?

I think I'm just having last minute paranoia again - the drivers are now with me and I have had them all out of the boxes at the same time to go 'oooooh, pretty' and I don't want to kill them! lol

I was thinking that I could put them all on one channel and give them 625wRMS each instead, a la carp, rather than the 1kw+ that popalock gives his with his Clone amp, but I'm not quite sure how I'd wire 4 D2 drivers to show a 2 ohm load to the amp - by my reckoning I could only wire them up to show 1 Ohm (not going to happen) or 4 Ohm (reduced power) so I think I have given myself no choice but to go with 2 drivers per 2Ohm channel...

I think I will have to buy a MiniDSP (and no doubt require a bunch more help choosing the right one!!) and then see if I can employ a LT to shape the curve at the lower end - my initial playing with WinISD to gain a 0.707 curve at the bottom end reduces excursion between 30Hz and the very bottom end, and reduces power to about 1000w per driver rather than 1250w, so it may 'waste' the amp power but should hopefully mean safer drivers... (OHF is still worrying me lol)

If by some miracle I could wire them up to have 625w each then it looks like I drop from 121dB@20Hz and 109dB@10Hz, to 118dB@20Hz and 106dB@10Hz - which I imagine should still be adequate with room gain, hopefully?!

Thanks again for putting up with my noobness lol

just a heads up, it doesn't quite work like that because speakers have a resonance. that resonance creates an spike in the impedance where the current goes down dramatically.

here is the lm ultra in 4 cubic feet, but all drivers in a sealed box have this behanvior.

impedance:

power:

only at the extreme low end and top end where the impedance falls to the dc resistance of the driver does power (and current) rise to what you would calculate based on ohm's law.

in practice, most of the time, you will be drawing no where near what you calculate with ohm's law.

here is the lm ultra in 4 cubic feet, but all drivers in a sealed box have this behanvior.

impedance:

power:

only at the extreme low end and top end where the impedance falls to the dc resistance of the driver does power (and current) rise to what you would calculate based on ohm's law.

in practice, most of the time, you will be drawing no where near what you calculate with ohm's law.

Thanks for the detailed response, as always

I think I have it clear in my head now - so that if I was to draw a horizontal line across the Apparent Amplifier Load graph previously posted at the rated power of 4 SIs (2400w) then everything under the curve but above that line would technically be over the rated power of the drivers?

http://www.diyaudio.com/forums/attachments/subwoofers/385384d1386003100-looking-build-2x18-enclosure-ppsl-18d2x4_15cf_5kw_qi10_qa70_red_qi50_qa80_green_qi50_qa20_blue_qi50_qa5_yellow_-_apparent_amp_load.jpg

But, going from popalock's experience/setup, they*should* be fine up to about 1100w or so (if used sensibly rather than sine-wave-tested to death lol) which means that if I calibrate to reference using my Audyssey XT32-equipped Onkyo AVR, they should only be 'over-powered' below around 25Hz or so (as anything above that point is either below the 2400w line or the SPL graph is way above a calibrated level so will never need full power) and with amp roll off at -3dB at 5Hz, power/current draw should probably max out about 10Hz / 1100w or so? I think??

If the above is correct then great, hopefully all will be fine and I won't have to worry about running at Reference level or even a little beyond!

I think I have it clear in my head now - so that if I was to draw a horizontal line across the Apparent Amplifier Load graph previously posted at the rated power of 4 SIs (2400w) then everything under the curve but above that line would technically be over the rated power of the drivers?

http://www.diyaudio.com/forums/attachments/subwoofers/385384d1386003100-looking-build-2x18-enclosure-ppsl-18d2x4_15cf_5kw_qi10_qa70_red_qi50_qa80_green_qi50_qa20_blue_qi50_qa5_yellow_-_apparent_amp_load.jpg

But, going from popalock's experience/setup, they

If the above is correct then great, hopefully all will be fine and I won't have to worry about running at Reference level or even a little beyond!

yeah, that and this:

"...Indeed Stereo Integrity confirms that the power rating for this driver is based on the mechanical overload not what the coil will handle thermally..."

http://www.data-bass.com/data?page=driver&id=46

"...Indeed Stereo Integrity confirms that the power rating for this driver is based on the mechanical overload not what the coil will handle thermally..."

http://www.data-bass.com/data?page=driver&id=46

Sweet, thank for the reassurance! DIY Noob nerves, that is all lol

Have been playing with Sketchup so will have to upload a picture!

Have been playing with Sketchup so will have to upload a picture!

Just saw this thread as I don't usually take much interest in sub builds. You mention PPSL being a concern. In your case, it is really just a simple sealed enclosure and should act that way except around where the cavity resonance for the "slot" occurs. I haven't tested a PPSL, but the theory behind it is legitimate. Part of the distortion reduction is due to the cavity itself. How much that matters with a home theater sub is questionable, but it won't hurt. How high do you plan to play them?

The one caveat with PPSL is that one of the drivers is inverted and you might run into suspension noise that is usually masked by the enclosure. This isn't an issue when used with low excursion pro drivers, but pushing 20mm xmax can be noisy. If that ends up being a problem, you might be able to run the drivers facing each other but that makes the cavity larger which lowers the resonant frequency and its resultant low pass effect. There is a solution to that if it becomes an issue. Simply attach a small wedgelike piece into the slot which will take up some of the slot volume.

Anyway, it looks like you are on a good path.

The one caveat with PPSL is that one of the drivers is inverted and you might run into suspension noise that is usually masked by the enclosure. This isn't an issue when used with low excursion pro drivers, but pushing 20mm xmax can be noisy. If that ends up being a problem, you might be able to run the drivers facing each other but that makes the cavity larger which lowers the resonant frequency and its resultant low pass effect. There is a solution to that if it becomes an issue. Simply attach a small wedgelike piece into the slot which will take up some of the slot volume.

Anyway, it looks like you are on a good path.

Quote:

Originally Posted by **coctostan**

Just saw this thread as I don't usually take much interest in sub builds. You mention PPSL being a concern. In your case, it is really just a simple sealed enclosure and should act that way except around where the cavity resonance for the "slot" occurs. I haven't tested a PPSL, but the theory behind it is legitimate. Part of the distortion reduction is due to the cavity itself. How much that matters with a home theater sub is questionable, but it won't hurt. How high do you plan to play them?

The one caveat with PPSL is that one of the drivers is inverted and you might run into suspension noise that is usually masked by the enclosure. This isn't an issue when used with low excursion pro drivers, but pushing 20mm xmax can be noisy. If that ends up being a problem, you might be able to run the drivers facing each other but that makes the cavity larger which lowers the resonant frequency and its resultant low pass effect. There is a solution to that if it becomes an issue. Simply attach a small wedgelike piece into the slot which will take up some of the slot volume.

Anyway, it looks like you are on a good path.

Just saw this thread as I don't usually take much interest in sub builds. You mention PPSL being a concern. In your case, it is really just a simple sealed enclosure and should act that way except around where the cavity resonance for the "slot" occurs. I haven't tested a PPSL, but the theory behind it is legitimate. Part of the distortion reduction is due to the cavity itself. How much that matters with a home theater sub is questionable, but it won't hurt. How high do you plan to play them?

The one caveat with PPSL is that one of the drivers is inverted and you might run into suspension noise that is usually masked by the enclosure. This isn't an issue when used with low excursion pro drivers, but pushing 20mm xmax can be noisy. If that ends up being a problem, you might be able to run the drivers facing each other but that makes the cavity larger which lowers the resonant frequency and its resultant low pass effect. There is a solution to that if it becomes an issue. Simply attach a small wedgelike piece into the slot which will take up some of the slot volume.

Anyway, it looks like you are on a good path.

Hey Stan, thanks for your response - I'm glad to hear this thread has piqued your interest a little!

The other chaps on here have been kind enough to help me along greatly with my build plans, which is very much appreciated and I think I have now got to the point that I am happy / confident that:

- WinISD's suggested response (groundplane, I presume) is far more than needed (!) across most the of the range for Reference level playback and drops past 115dB at around 15Hz, at which point room gain should be helping nicely, so I am hopeful I will have comfortable Reference-level capabillity down to well below 20Hz.

- the drivers should only just touch Xmax at max amplifier/subwoofer output, going from the WinISD modelling.

- the drivers have been reviewed by Josh Ricci as pretty quiet in their operation, so hopefully the reversed sub won't create any distracting suspension noises, even at Reference level.

- the SI HI 18s should be safe with the max of 1250wRMS each from the amp due to room gain reducing power needed low down, where most of the power is drawn, and that the WinISD modelling shows the drivers would only reach 1kw or more below around 10Hz, which should drop off again as the -3dB of the amp is 5Hz, so the coils should hopefully not cook.

My plan is to have a more capable system with more headroom than my current dual-12" SVS, so that it digs deeper and is 'tighter' due to being sealed, but I'm aware that these things are subjective and blind testing has not proved the difference is easy to spot!

I don't listen at Reference very often but do approach it on well-balanced soundtracks, such as Oblivion, so hopefully these subs will be coasting along most of the time. It might mean that I don't necessarily reap the reduced harmonic distortion benefits of the PPSL design if I'm not driving them that hard, but when I do I am hopeful that they will sound 'clean' compared to other options

As you suggest, though, I have designed the box so that in a worst case scenario I can revert to running both drivers 'normally', but I did not know about the reduction of slot volume required if I do so thank you for that information I have read about the LPF-effect of the slot but I have been struggling to find any measurements or indications of what to expect - my initial readings seem to indicate that it's not likely to be an issue unless I run the drivers up to really quite high for a subwoofer, something like 200Hz? so fingers crossed all will be fine!

I will attempt to upload / link to a sketchup drawing later this evening - for some reason I can't upload anything to this forum so I may have to host elsewhere again!

Thanks again, everyone, for your assistance with this

It sounds like you have a pretty good handle on the design. I don't see any flaws in your reasoning.

The low pass effect of the slot is not easily calculated. I've modeled it in HornResp, but I haven't measured to see how accurate the model is. Ultimately you would just need to measure if you find the need to run the sub with both cones facing each other. If the dip from the cavity is apparent in the passband you simply need to occupy some of that space. Depending on much aesthetics matter, you could simple attach a wedge made of wood between the drivers which would occupy some space and move that dip up in frequency. It has been a while, but I think I might have also modeled it as a bandpass enclosure with the cavity acting as a very large slot bandpass, but that is not easy to model either.

Basically if you measure the sub and get a frequency response that dips in an area that Ricci's measurements don't show (and you are confident in your measurements, construction, connections, signal chain, etc) it could be the cavity resonance. That is basically the only caveat I see with the PPSL and it is likely moot.

The low pass effect of the slot is not easily calculated. I've modeled it in HornResp, but I haven't measured to see how accurate the model is. Ultimately you would just need to measure if you find the need to run the sub with both cones facing each other. If the dip from the cavity is apparent in the passband you simply need to occupy some of that space. Depending on much aesthetics matter, you could simple attach a wedge made of wood between the drivers which would occupy some space and move that dip up in frequency. It has been a while, but I think I might have also modeled it as a bandpass enclosure with the cavity acting as a very large slot bandpass, but that is not easy to model either.

Basically if you measure the sub and get a frequency response that dips in an area that Ricci's measurements don't show (and you are confident in your measurements, construction, connections, signal chain, etc) it could be the cavity resonance. That is basically the only caveat I see with the PPSL and it is likely moot.

That's good to hear, thank you for the input!

I have been thinking that I might 'fill in' behind the drivers, to reduce the cavity space, but I will have to speak to the guy building the boxes. I am pretty hopeful that I will have enough output to not worry about peaks and troughs at the listening position, I would prefer to split them up for smoother response across the room, but I only have one location so will just have to stack!

Below are some links to the WinISD graphs that illustrate how a LT will help protect the drivers for little loss in output (if I can work out how to do it ) and also a sketchup of my proposed boxes and 1/4 and 3/4 height driver stacking layout!

Sketchup drawing:

http://www.diyaudio.com/forums/attachments/subwoofers/392387d1389146818-looking-build-2x18-enclosure-ppsl-ppsl-sub-boxes-stacked-measured.jpg

SPL with / without LT

http://www.diyaudio.com/forums/attachments/subwoofers/392388d1389146818-looking-build-2x18-enclosure-ppsl-si-ht-18-d2-x4-sealed-15-cu-ft-5000w-qi50-qa20-blue-qi50-qa20-lt-orange-spl-comp.jpg

Excursion with / without LT

http://www.diyaudio.com/forums/attachments/subwoofers/392389d1389146818-looking-build-2x18-enclosure-ppsl-si-ht-18-d2-x4-sealed-15-cu-ft-5000w-qi50-qa20-blue-qi50-qa20-lt-orange-excursion-comp.jpg

Apparent Amplifier Load with / without LT

http://www.diyaudio.com/forums/attachments/subwoofers/392390d1389146818-looking-build-2x18-enclosure-ppsl-si-ht-18-d2-x4-sealed-15-cu-ft-5000w-qi50-qa20-blue-qi50-qa20-lt-orange-amp-load-comp.jpg

Apologies I can't seem to host them here!

I have been thinking that I might 'fill in' behind the drivers, to reduce the cavity space, but I will have to speak to the guy building the boxes. I am pretty hopeful that I will have enough output to not worry about peaks and troughs at the listening position, I would prefer to split them up for smoother response across the room, but I only have one location so will just have to stack!

Below are some links to the WinISD graphs that illustrate how a LT will help protect the drivers for little loss in output (if I can work out how to do it ) and also a sketchup of my proposed boxes and 1/4 and 3/4 height driver stacking layout!

Sketchup drawing:

http://www.diyaudio.com/forums/attachments/subwoofers/392387d1389146818-looking-build-2x18-enclosure-ppsl-ppsl-sub-boxes-stacked-measured.jpg

SPL with / without LT

http://www.diyaudio.com/forums/attachments/subwoofers/392388d1389146818-looking-build-2x18-enclosure-ppsl-si-ht-18-d2-x4-sealed-15-cu-ft-5000w-qi50-qa20-blue-qi50-qa20-lt-orange-spl-comp.jpg

Excursion with / without LT

http://www.diyaudio.com/forums/attachments/subwoofers/392389d1389146818-looking-build-2x18-enclosure-ppsl-si-ht-18-d2-x4-sealed-15-cu-ft-5000w-qi50-qa20-blue-qi50-qa20-lt-orange-excursion-comp.jpg

Apparent Amplifier Load with / without LT

http://www.diyaudio.com/forums/attachments/subwoofers/392390d1389146818-looking-build-2x18-enclosure-ppsl-si-ht-18-d2-x4-sealed-15-cu-ft-5000w-qi50-qa20-blue-qi50-qa20-lt-orange-amp-load-comp.jpg

Apologies I can't seem to host them here!

Just thought I'd add a post here, partly for my future reference as I can never find things on this forum because there is so much stuff (!) and partly because I wanted to check something...

In this post:

http://www.avsforum.com/t/1520779/help-with-10hz-output/90#post_24447510

derrick very helpfully does the calculation for where room gain starts in a 20 foot room (which mine also is, luckily!)

12dB per octave gain would seem to suggest that I potentially could see 125-126dB across the whole bandwidth below 28Hz (!!) which I have calculated just now by plotting manually in MS Paint on the SPL graph linked to above. (Can't upload it right now but will attempt to do so tomorrow.)

Is that really correct??

Even with the 10dB/octave of room gain that Bill FitzM suggests is more realistic, I'm looking at potentially 120dB at 7Hz, assuming that's correct?!

That seems a*lot*...

In this post:

http://www.avsforum.com/t/1520779/help-with-10hz-output/90#post_24447510

derrick very helpfully does the calculation for where room gain starts in a 20 foot room (which mine also is, luckily!)

12dB per octave gain would seem to suggest that I potentially could see 125-126dB across the whole bandwidth below 28Hz (!!) which I have calculated just now by plotting manually in MS Paint on the SPL graph linked to above. (Can't upload it right now but will attempt to do so tomorrow.)

Is that really correct??

Even with the 10dB/octave of room gain that Bill FitzM suggests is more realistic, I'm looking at potentially 120dB at 7Hz, assuming that's correct?!

That seems a

for a medium size room, the gain curve can be roughly modeled in winisd as a linkwitz transform: f0=30hz, q=0.7 ; fp=9hz, q=0.5

it looks something like this.

all rooms are different, particularly with respect to lossiness, but for a medium home room that is good construction, this would be a reasonable start.

it looks something like this.

all rooms are different, particularly with respect to lossiness, but for a medium home room that is good construction, this would be a reasonable start.

You're a star, thanks very much LTD

I will plug those figures into WinISD and see what comes out!

I will plug those figures into WinISD and see what comes out!