Another 604 MLTL build (ping GM)

[omodo]

GM, as per our initial contact via PM, I am looking at building another variant of the 604 MLTL design you did for Jay Fisher. Jay has kindly provided me with a copy of his cabinet plans and an overview of the design you did for him (CSA, Length, port dimensions, driver location, and so on).

At this point I have done a quick sketchup of the basic form of the cabinets I'd like to end up with, based on the dimensions from Jay with inspiration from the Latour 604 cabinets.

Internal:
Width: 664mm (26.125")
Depth: 411mm (16.187")
Length: 1080mm (42.5")

CSA: 422.04"^2

External:
Width: 702mm
Depth: 455mm
Height: 1118mm

(based on single layer 19mm ply for sides, top, bottom, and back, and 25mm ply for the baffle)



as drawn the cabinet is raised 82mm giving a total height of 1200mm.

The driver is located on widest face, located equidistant down from the top and sides (351mm). This places the centre of the driver at 332mm (13") from the top interior or ~850mm (33.5") from the floor. The original dimensions had the driver located 500mm (19.625") from the top interior putting the tweeter height at 700mm from the floor (27.5"), as shown below.



The port is located on the bottom as per the original design, and has the equivalent area of the original 6" diameter by 2" long vent.

I'd appreciate it if you could provide some guidance on:
- adjusting the design parameters to suit my room/system (details below/next post)
- effect of changing the location of the driver relative to the top of the cabinet
- ideal port size/length/location, cabinet height off ground, and so on.

For the driver, I am still deciding between either the GPA 604-8H-III or perhaps Wolf von Langa's A604WVL field coils drivers (which have T/S parameters closer to 604E), which will affect the design somewhat but if you can at least get me to the point of a solid model in the MathCad worksheets I should be able to adjust to suit.

[omodo]

Room dimensions are: 4250 x 6360 x 2590mm (14 x 21 x 8.5')

it is a reasonably difficult room as one end of the room opens out into the main open plan living area of the house (approx 25' x 60'), and the end where the speakers will be located has 4 full height windows close to the corners. Placing the speakers at the other end of the room facing back towards the windows is not really an option.

The room is drywall/plaster sheet over pine stud construction. The external walls (top-left, top, and right looking at the images below) are brick veneer, the southern wall adjoins our dining area and is drywall/plaster sheet on both sides. The walls have insulation batts in the wall cavities. Flooring is carpeted.

Placement of speakers relative to side and rear boundaries can be determined by the model/room loading required to get a good response, so will leave that up to the model (to extent). I expect the windows will affect the room loading somewhat, I'm also not sure how to model this so appreciate any guidance on this.





Listening position as drawn is 4.3m (17') from the rear wall and 2.1m (7') from the side wall. Speaker placement as drawn is indicative only as mentioned above.

In your PM you asked for details of electronics, which I assume was related more to power amps than source/etc. My current power amp is a 45 SET, which might be a bit on the low power side for these drivers, but I have access to an EL84 push-pull amp which I will use in the interim. I don't mix and match gear/room setups a lot, once I get it dialed in it tends to stay put.

[atilsley]

Omodo

We should partner on this!

I like the look of the design...as I was just about to start on a 9 cubic foot box for the 604-111's.

Your design reminds me of the Shindo -

http://www.shindo-laboratory.co.jp/English/speaker.html

THIS pic shows the Shindo in Palesander veneer...which I used recently on an Altec Stonehenge...220 litres (about 8 cubic feet). Nice veneer, with hand polish Bees Wax. See pic attached.

I love the GPA 604-111...hence project #2.

Andrew.
Sydney.
0411865553

[GM]

As mentioned, I've had a shot at modelling this in MathCad, but am a bit unclear as to how to model the bottom port in the worksheet, plus the effect of cabinet height off floor, and room loading to the end SPL response..........

==============


Greets!

The inner opening of the vent location would be 'zport', i.e. L = 42.5", Lport = 2", but ~0.75" is in the bottom plate, so only extends up ~1.25", ergo zport = 42.5-1.25 = 41.25".

I don't have MJK's current worksheets, so you will have to ask on his forum how to do in-room modeling.

GM

[GM]

The driver is located on widest face, located equidistant down from the top and sides (351mm). This places the centre of the driver at 332mm (13") from the top interior or ~850mm (33.5") from the floor.

Greets!

FWIW, based on MJK's math, driver ideally needs to be down from top ~14.83" i.d. and mine puts it at ~19.63" and offset horizontally in a golden or acoustic ratio which JF chose not to do. Never having compared the two, don't know if what differences one sees in a sim is audible and/or whether one sounds better overall. You can view the changes in MathCad and how stuffing density affects it.

The driver position relative to the floor will probably have more more impact due to floor bounce though and why I like very tall cabs, so I can use my calculated offset and still get it up to seated ear height or a little higher which in turn pushes floor bounce lower to where it's not so audible due to room modes and our falling hearing acuity.

His design assumes being hard up against a solid wall (pi space) and driven with a ~matching impedance amp with a relatively short listening distance IIRC, ergo if spaced well away from any boundaries a larger cab to lower F3 is dictated for a given Fb and if driven with a low output impedance, then some baffle step compensation (BSC) will probably be required along with CD horn EQ for the HF as well as more power to compensate.

Cab/floor height isn't an issue until it's quite close.

GM

[omodo]

Andrew, sounds like a good idea, I'll send you an email

Omodo

We should partner on this!

I like the look of the design...as I was just about to start on a 9 cubic foot box for the 604-111's.

Your design reminds me of the Shindo -

http://www.shindo-laboratory.co.jp/English/speaker.html

THIS pic shows the Shindo in Palesander veneer...which I used recently on an Altec Stonehenge...220 litres (about 8 cubic feet). Nice veneer, with hand polish Bees Wax. See pic attached.

I love the GPA 604-111...hence project #2.

Andrew.
Sydney.
0411865553

[omodo]

Hi, thanks that was how I initially entered the values, but was not sure if there were any other 'tricks' due to the exit being from the bottom rather than the front/back

As mentioned, I've had a shot at modelling this in MathCad, but am a bit unclear as to how to model the bottom port in the worksheet, plus the effect of cabinet height off floor, and room loading to the end SPL response..........

==============


Greets!

The inner opening of the vent location would be 'zport', i.e. L = 42.5", Lport = 2", but ~0.75" is in the bottom plate, so only extends up ~1.25", ergo zport = 42.5-1.25 = 41.25".

I don't have MJK's current worksheets, so you will have to ask on his forum how to do in-room modeling.

GM

[omodo]

thanks again for the advice.

with respect to impact of stuffing density on the models, I have seen references to 1" acoustic treatment to 3 adjacent sides rather than using stuffing, in this situation is it better to model using a low stuffing value (e.g. 0.01) or just assume a value close to the default of (0.25 lb/ft^3)

I would appreciate it if you could take a quick look over the following for any obvious input errors, the driver placement appears to have little affect on the far field SPL response

Zdriver = 332mm

Zdriver = 375mm

Zdriver = 500mm

Greets!

FWIW, based on MJK's math, driver ideally needs to be down from top ~14.83" i.d. and mine puts it at ~19.63" and offset horizontally in a golden or acoustic ratio which JF chose not to do. Never having compared the two, don't know if what differences one sees in a sim is audible and/or whether one sounds better overall. You can view the changes in MathCad and how stuffing density affects it.

The driver position relative to the floor will probably have more more impact due to floor bounce though and why I like very tall cabs, so I can use my calculated offset and still get it up to seated ear height or a little higher which in turn pushes floor bounce lower to where it's not so audible due to room modes and our falling hearing acuity.

His design assumes being hard up against a solid wall (pi space) and driven with a ~matching impedance amp with a relatively short listening distance IIRC, ergo if spaced well away from any boundaries a larger cab to lower F3 is dictated for a given Fb and if driven with a low output impedance, then some baffle step compensation (BSC) will probably be required along with CD horn EQ for the HF as well as more power to compensate.

Cab/floor height isn't an issue until it's quite close.

GM

[GM]

OK, essentially there should be no room gain to speak of in the audible BW, so the speakers must carry all the load down low and there's windows that need some seriously dense/heavy drapes, say, made out of thick comforters with weights in the bottom to make them hang tight since windows tend to be more reflective than a wall. Simming software typically assume infinitely rigid/massive boundaries, so can't account for the various construction losses. Speaker internal damping is typically done by ear, i.e. once any perceived box 'hollowness' is attenuated, you're done, ergo any fine tuning of the vent then becomes an empirical/personal 'taste' one, i.e. when there's no perceived 'hangover'/'boom', you're done.

I imagine the 'best' location for the speakers will be with the inner baffle edge being near and on the same plane as the large furniture's face between them and if you make them with the drivers offset in a mirror image, then the drivers will be closest to the side walls. Toe in is typically the left speaker aimed at the far right seating position and vice versa, so a good first approximation.

Depending on its topology, the SET could be up to a matching impedance one and the PP typically < 1.0 ohm, so tuning to Fs is de facto for the SET and when the PP is used, some LF boost via variable DF would be required to ~match it tonally. LF capability then becomes totally dependent on cab net Vb with up to ~ Vb = Vas being required.

GM

[GM]

thanks again for the advice.

........or just assume a value close to the default of (0.25 lb/ft^3)

I would appreciate it if you could take a quick look over the following.........

You're welcome!

I typically use the 1" acoustic fiberglass insulation (now OC 703?) seen in virtually all Altec cabs after Celotex was discontinued and best I can tell, lining one of each parallel wall = ~0.2 lbs/ft^3 in MJK's software, but many folks have gotten back to me that in my cab designs it tends to be too much, so recently I've been suggesting lining just the top and hanging a full height insulation 'blanket' of polyfil or similar diagonally across the cab and shortening it if need be. Several folks said they wound up shortening them all the way back to just below the driver, which makes for a pretty lively looking sim, so Caveat Emptor. If I get a chance, I'll have to try it in my cabs.

One more time:

"FWIW, based on MJK's math, driver ideally needs to be down from top ~14.83" i.d. and mine puts it at ~19.63" and offset horizontally in a golden or acoustic ratio which JF chose not to do. Never having compared the two, don't know if what differences one sees in a sim is audible and/or whether one sounds better overall. You can view the changes in MathCad and how stuffing density affects it."

Right, wrongly or somewhere in-between, I arrived at a way to calculate the technically 'best' driver location as I understand the physics of the situation ~36 yrs before I could sim it and ~15 yrs before I could measure it with a 1/3 octave RTA, which needed to be 1/6 to ~ match what I was hearing, but one makes do with what one can afford. You want more info on how different locations affect in-room response, then knock yourself out doing the necessary 'due diligence' and please share your findings.

WRT your sims, they're unfortunately completely bogus since BL is off by almost a factor of two. Even if BL is published, one should always calculate it when using published specs and why MJK has a BL calculator WS. Unfortunately it would still be off due to the Sd you used being too large for any 604 with 96"^2/619.354 cm^2 being more technically correct, though at a glance it doesn't change the MC sim, so assume it's a moot point.

Regardless, using HornResp or WinISD Pro to 'correct' published specs for inputting in various programs that require more than just Vas, Fs, Qts is a good plan plus it allows more accurate comparisons between the various simming programs.

GM