Active Statements

[Benji]

Soooo..... I know this may ruffle some feathers, especially since this is my first post (hi all!) and especially since I know next to nothing about speaker design or electronics.

Actually that's a lie, I know nothing about speaker design or electronics.

The problem is I've been reading... lots and lots of reading. I've spent hours and hours, in fact days on Elliott's (ESP) website and at the same time I happened across the Statements.

So having lived a life of below average 'hi-fi' I always planned to buy myself a decent system at some point in my life and go 'all out' as much as I could afford. Of course as with so many other things in life and with the power of the internet I see that DIY is the way to go, making our money go further provided the individual is willing to put the effort in!

So here is the plan, note that there are a lot of bits still not ironed out and this will initially concentrate on the front pair of Statements to be used in a HT setup but at the same time, has to be good for music.

Each 'Statement'(?) will be cast from concrete, yes yes I know, it's insane, it's impractical and they'll weigh a freaking tonne (each) but I've always liked the idea of a virtually inert cabinet material which doesn't need all the extra bracing internally. Apart from removing the unnecessary bracing the cabinet design will remain identical (internally at least) unless otherwise advised.

Each Statement will contain:

ESP P09 - 3way crossover with balanced XLR input, directly connected to...
3x ESP P101 - Three power amps, low power version. Possibly use P03A's to reduce the cost per speaker.

And throw in the necessary power supply(s) for it all (P05B?), soft-start and muting circuits (P39 and P111?).


I realise I'm probably massively over complicating things, especially since I don't have a clue and how much will certain things make a difference? Who knows, building them out of concrete is just my time and I'd have to spend just as long learning wood work as I would pouring concrete.

Having the amps directly connected to the speakers I was trying to work out whether or not they should still be connected in series or parallel to keep everything in balance.

As for the rest of the system, the centre will be the Statements Centre and I'm not sure whether or not to go for Mini Statements or Monitor Statements for the surrounds.

The subs will be IB subs, either a 4x15's, 2x2 of 18's or 2x4 18's all in a line array, driven appropriately and tied into the main system. This has yet to be figured out on the IB cult forum but I see ThomasW is actually part of this forum!

This will all be driven by either a Marantz AV7005 Pre/Pro (7.2) or if budget allows an Integra DHC-80.3 Pre/Pro (9.2), fed by an HTPC bit streaming to the Pre/Pro.

A lot of good things are said about:

The Statements
Active crossovers
Tri-amping
Concrete enclosures
IB subs.

So I wanted to put them all together and see what I got. Overkill? Maybe, but I'm open to your thoughts/suggestions/changes/criticism/anything and I sincerely hope that Curt/Jim do not take offence on my suggesting doing the already brilliant Statements differently.

I figure going 3-way active crossover with 3xP101's will push the cost of *each* Statement >$1000, as for the Mini/Monitors I'm not so sure. Switching to P3A's should reduce the cost a lot and I'm not sure how much benefit would be gained from going to P101's anyway.

Packaging is also a concern, integrating everything into each speaker with 1xbalanced XLR input, 1xpower socket and managing all the heat that comes with it.

I want to go *low*, real low and loud as effortlessly as possible so to not loose control. I'm happy to use the IB subs with the Statements mains to back them up whilst listening to music as well.

Anyway, enough of my rambling.

[5th element]

My first thought upon reading...

I know next to nothing about speaker design or electronics.

Is that this is probably going to be too much. Although on the face of things it probably seems a like a lot of 'simple' things bolted together to form a system, each individual part has the potential to throw issues at you, issues that you wouldn't have expected or even thought existed. This is less of a concern with a simple pair of loudspeakers, but when you involve electronics there can be a lot more involved then what initially meets the eye.

I'm not saying this to put you off, I'm saying this because I've been there and know what it's like. Managing to do something this ambitious without any prior experience is probably going to take you a very long time and end up costing way more then you'd initially think.

A concrete cabinet isn't something to take lightly, least of all because of the weight, and manufacturing a mould from which to cast it will be far more trouble then you probably realise. You mention that you will learn concrete pouring instead of woodworking, which isn't exactly how it works. You will need to make the mould out of wood or something similar, from which to cast your statements from. You will have to learn woodworking to make your concrete loudspeakers.

As a suggestion I would advise you to start out with a small pair of loudspeakers. These don't have to be expensive, but you need something to get your feet wet with and even an MDF pair of statements, to someone with no prior experience, or probably tools for that matter, is an extremely ambitious project.

Even though loudspeakers are said to be the weakest link in any hifi, you may find it easier to start with the electronics. This should require less initial outlay in tools, but the electronics side of things does bring about issues of electrical safety which cannot be over looked.

Then again, if you're after something that will give you much better sound quality then the loudspeakers are the obvious way to go.

Because you've got so many different things that you want to achieve, I am genuinely having a hard time coming up with an idea about where the best place to start is. For any pair of loudspeakers, either big or small, you're going to need tools, so maybe there. Buy a decent router, a load of suitable clamps, a straight edge and some glue, get some MDF and try building things. You will need a solid surface on which to work so you may need a decent work bench and the router will create loads of dust, so unless you have somewhere set aside to work in you will need a router with good dust extraction and a decent hoover. You will also need ear defenders and a face mask.

[richnen]

Concrete enclosures for your first build? Wow that is ambitious!

I have made 14 granite enclosures and twice that many baffles and can tell you from experience that there is a high probability that your first attempt will not be your final product.

I think you will have an especially hard time of it if you are planning concrete baffles as well. I would highly recommend that you reinforce the cabinets with mesh, not sure how you will deal with the baffle as the thin amount of material left between the drivers is going to be a concern.

Are you planning to core the driver holes out after or build them into your forms? Same question for the open back, port and binding post openings.

How are you going to mount the drivers? In my most recent granite baffles I cut the hole in the granite to the outer dimension of the speaker frame and epoxied a plywood ring in the hole to attach the driver.

I look forward to seeing your progress. Good luck!

[AdelaaR]

Not even touching the subject of how to properly make concrete cabinets in which drivers can be mounted succesfully ... I'm wondering why you would change the existing and proven design of the passive crossovers.
What is the added benefit of an active crossover in this particular case?
If you are going to use an active ESP crossover ... how will you know it sounds right?
How will you make sure the phases are correctly alligned as in the original design?
I made a plan myself for customized statements and will soon be building them, but I sticked to the original, proven design for the most important factors.
I think that you may be getting way in over your head on this one.

[5th element]

It's not that converting it from passive to active is a difficult thing to do either, the statements are a well documented design and simulating the transfer functions of the actual crossover sections isn't actually that hard. But! And here's the key point, that's if you know what you're doing.

Now there are lots of very good reasons to go active and I'm a strong proponent of doing so, but in this case I do feel that building a standard pair of passive statements would be for the best.

As a plan Benji here's what I would do.

1) Learn how to build stuff out of MDF.
2) When you think you're good enough, build a standard pair of passive statements out of MDF, but build them so they can be tri-wired, tri-amped.
3) Investigate electronics design and I would advise you build some simple battery driven circuits first of all.
4) Start building the simpler of the Elliot designs that you will need in your finished system, that is softstart, loudspeaker muting protection etc.
5) Build yourself a two channel amplifier, this can then be used to drive the statements.
6) If successful build more channels and bi-amp then tri-amp your statements.
7) One you've got this far you will need to investigate loudspeaker design and how to go about simulating the transfer function of the statements passive crossover. You will then need to convert this to an equivalent active circuit. This isn't trivial and is extremely critical to the design (in fact it is the design) get it wrong and you destroy all of Curts original design work. This will involve the use of CAD design software, free or otherwise and then require you learn how to measure the actual transfer function of your constructed active xover to make sure it's doing what it should.
8.) Remove the passive xover from your statements and introduce your active crossover using the previously constructed 6 channels of amplification.

Going about it in a fashion something similar to that will give you rewards every step of the way, rather then it requiring everything all at once.

[Benji]

Thank you for the feedback and like I said, the majority of it is probably unnecessary. Just having searched through the Statements threads it's clear they are very capable.

Is it a lot to take on, yes. Perhaps I should have not given off the impression I'd attempt everything all at once without first starting small and 5th Element is on the money.

Let me scale it back a little, just a two channel system with Statements made from MDF... if you did want to go active crossover + tri-amped, here is where I am most likely oversimplifying things but given a stock cabinet design provided the active crossover... crosses over at the same points and the slope of the crossover is the same then all should be well?

[5th element]

If the electrical transfer function of the active and passive crossovers is identical then yes all will be well. Managing to do this can be quite a challenge though.

[Benji]

If the electrical transfer function of the active and passive crossovers is identical then yes all will be well. Managing to do this can be quite a challenge though.

Electrical transfer function? <googles.....>

http://en.wikipedia.org/wiki/Transfer_function ?

I just want to educate myself so I'm not talking complete tosh.

Here is project 9:

http://sound.westhost.com/project09.htm

I'm wondering if you are referring to things like phase delay?

[5th element]

The transfer function basically describes what the filters do to the signal. In other words the passive filter to the tweeter does a specific job with respect to frequency and amplitude. You would need to simulate what the filters transfer function is for the passive crossover, then you would need to replicate this precisely with an active crossover.

[benchtester]

The transfer function basically describes what the filters do to the signal. In other words the passive filter to the tweeter does a specific job with respect to frequency and amplitude. You would need to simulate what the filters transfer function is for the passive crossover, then you would need to replicate this precisely with an active crossover.

I would suggest a different perspective. (While in the end the results might be similar, this tactic would simpler.) Set a goal to mimic the acoustic response of the passive crossover, then design the active crossover accordingly.

Determining the transfer function of a passive crossover is quite complex (please excuse the pun). The speaker driver has changing impedance as a function of frequency and displacement. So characterizing the transfer function is pretty difficult. Duplicating it actively is also challenging.

I would just try to duplicate the crossover frequencies and slopes in an active system. (Acoustic targets: 350 hz LR2 and 3700 hz 3rd order Butterworth.) Hopefully the phase will track in a similar manner, especial if you use the same electrical orders as the original crossover.

My only other advice, much like what others have written, is to take on one part of the project and see it to conclusion. I would say build the Statements passively, but mount the crossover externally (with terminals for each driver). Once they are sounding good, then attack the active crossover. Having the passive crossovers will continue to be a useful reference point while developing the active crossovers. The ESP board might be a reasonably good starting point; to accurately duplicate the statement crossover it will need some tweaking.

[5th element]

Well surely the passive filters transfer function includes the drivers own natural impedance I Agree with you that emulating the target acoustic response of the individual drivers is the way to go, but I think that it would be pretty hard to achieve unless one is provided with Curt's original driver measurements.

The filters transfer function can be simulated quite easily however by using the drivers impedance and the provided crossover schematics. SPL trace on the impedance plots provided by the manufactures should be more then enough for this too as the impedance at the crossover frequencies is dominated by the voice coil + gap, rather then anything the box will do.

There are several ways one could go about this. Perhaps discussing this with Curt would be the best idea, he has the LspCAD design files after all and knows what his original acoustic targets are. Coming up with an active version should be child's play and take very little time. When designing with LspCAD myself I tend to do an active version of all my passive designs as the first step anyway as they tend to be easier to work with, maybe Curt has done something similar