Clarity on Seas Thor Kit

I am going to be building the Seas Thor kit soon but have a few questions which I would like clarified

1. In the AudioXpress article I have noticed that in some places the woofers are refered to as W18EX units and in others they are refered to as W18E units. Which is it and why this discrepancy?

2. What are thoughts on cabinet wall thickness? The standard design documents recomend +- 18mm for all sides except the front - which recomends +- 25mm. What differences are likely to be noticed with thicker / thinner wood? What is the maximum thickness one would go for, there must be a diminishing returns rule applied to this.
 
Hi, Ross

I don't think you'll find 25mm locally. I've just finished a set of Thor boxes for a friend, and I had to resort to 22mm. Just keep the internal dimensions the same. Thicker walls will result in more rigid sides. However, I don't think it is practical to increase the wall thickness, as it is already of a respectable thickness, yielding a fairly "dead" enclosure.

The boxes without drivers are already quite wieldy, so you may regret adding more mass...
 
Hey - Thanks Shaun

Being from SA - you should welcome heavier speakers. Theives are generaly lazy and wouldn't want to carry em out of your house. I plan to add a cavity to mine and fill it with concrete. Steal that ya bastards :devilr:

I know what you mean about the 22mm vs 1 Inch. Why can't things just be standardised. I would be able to get 32 mm and as I say weight isn't an issue for me, but it is ALOT more expensive. Think I'll stick to the 22. Was your buddy happy with what he got?
 
Count yourselves lucky guys -here in the UK, the maximum size MDF I can find is 18mm.

I know the feeling about thieves. A couple of years back, I lost most of my kit to the loathsome little maggots, including my much-loved LP12. I reflect on occasion that they might best serve the cause of the Earth by being buried under 6' of the stuff.

Speaking of sizes, I seem to recall from other threads that the Thor box is actually shallower than is ideal (I could be wrong there, though). If you're confident enough to draw up your own plans, it would probably be worth your while to use Martin King's MathCad worksheets to work out an optimum for these drivers -the Thor was based upon transitional work and TL tables; Martin's software is much more powerful, and much more accurate. It's relatively easy to simulate two drivers with it too, especially as they're identical. I haven't modelled the Thor in it yet, but I might plug the dimensions and T/S parameters of the drivers into it shortly and see what it comes out with.

Cheers
Scott
 
Hey Scotmoose.

It sounds really strange that a bunch of those sort of folk would make off with a turntable. Scumbags.. Anyway I agree with your sentiments.

Mathcad - "Hissss Hissss - Non free software Hiss Hisssss" ;). Damn don't have it - and if I did would likely not know how to use it.

If you could pop something together for us that would be great - I would be interested to see the results. Apparently the depth was cut down slightly but that it no severe effect subjectivly. I would like to prototype both the Thor as it is and the Thor as it should be and listen for myself.
 
Thanks Scott - will check it out.

JDev - Thanks. My goodies arrive next week sometime - then it's just a matter of finding a bit of time to do some building. I am busy building 12 channels of the Leach amp clone at the moment - so that will take a bit of time to complete.

I wish there was a bit more of a diy community in Durban - it feels very isolated out here hearing about the dudes in JHB & capetown. It must be nice not to have to build everything just to hear it.

Looking forward to lots of pics etc.
 
D'Appolito Fudged

Thor was designed by D'Appolito based on Augspurger's work. In defense of Augspurger's alignment tables -

In the AudioXpress article, D'Appolito states that he'd reduced the size of the enclosure arbitrarily.

Based on building many TLs myself, and modeling many more, I would anticipate that reducing the enclosure in such a way would reduce output in the deep bass. In other words, I would expect the bass to sound 'thin.' There are ways to reduce a TL, but arbitrarily shrinking the box ISN'T the best solution.

I agree with the other posters; redesigning the Thor Enclosure in Mathcad is a great idea.

John
 
Sigh. I was hoping you wouldn't ask us this. Because, frankly, I think these drivers are a downright pig to get the best out of. Unless I'm feeling unusually stupid today, which is always a possibility. I've been fiddling in MathCad, trying to see what gives. Firstly, for your viewing pleasure, I present Dr. Joe d'Apollitio's Thor, as modelled by Martin King's MathCad software.
This is approximate, but the uncorrected, anechoic reality will be close.

It's not as bad as I anticipated, as from the driver specifications I was less than optimistic that they would work at all in a conventional TL. Credit where credit is due for getting a reasonably decent response from it. However, Thor does have a number of problems. That rising response (add 6db overall for the drivers being wired in parallel by the way) is going to need a wad of correction to bring it down into line. I assume it has this built into the crossover, which I have yet to look at. Either way, no wonder many users find it clear sounding! Note also that this is not falling away in lock step with room gain -you're going to get a boost somewhere no doubt, but this is hardly going to be consistant. I also think that 0.78lbs/feet of stuffing is way too much. It should be less -around 0.3lbs per foot is the max that should be used in my view; if you need more, the TL is incorrectly designed.

To be fair to Thor's great (and he is) designer, this is one of the better 'traditional' TLs I've seen, and he only had aprroximate, transitional tables to work with. Also, many people would probably like this sort of sound -it's going to be crisp and highly detailed whatever you do with it. However, I too strongly disagree with his notion that reducing the size of the enclosure doesn't make a difference. It does, and I'm highly dubious about the measurements / calculations that suggest it doesn't. I'm still trying to get something that I would consider satisfactory out of these drivers in MathCad, and it's not easy. I don't personally think they suit the conventional TL taper, though I could have simply missed a magic combination. A straight pipe looks similar to the conventional taper. So far, I've had the most promising results out of the good old TQWT, but still nothing I'd build yet. I'll let you know later when I hit on something that works.

Any other thoughts or cabinet dimensions guys?
Best
Scott
 

Attachments

  • thor fr.jpg
    thor fr.jpg
    62.4 KB · Views: 6,453
Hey Scott - Is that the response of just the driver or does it take into account augmentation by the rear wave.

D'Appolito measured the Thor's f3 @45 Hz which is supported by the graphs where the graphs show combined response with augmentation. Where there is no augmentation the responses look similar to your predicted response. The question is is this an anechoic response or "room response". I would imagine it should be anechoic.

The Seas Thor article is not very clear when it comes to what kind of measurements these are. Very naughty.:smash:

edit
I should read things before opening my mouth / moving my fingers :headbash:

Also just for reference this is the commentary on Depth in the article.

The internal depth, d, is then computed as follows:
To get the external depth you must add the thickness of a 1' front baffle, a 0.75' internal baffle, and a 0.75' rear panel for an overall depth of 18.75��. This number was considerably deeper than I wanted and would lead to a rather large and heavy enclosure.

At this point I made a number of arbitrary decisions. I chose a line taper of 3:1 and limited the overall depth to 13.5'. This led to the internal layout of the line shown in Fig. 1A. Placing the interior baffle at an angle produces the desired taper. A side, but important, benefit of the interior baffle is that it adds
greatly to enclosure rigidity, effectively clamping the side panels together and largely eliminating side-wall vibration. The resulting layout has a throat area of 61.875 in2, which is roughly 1.6 times the combined diaphragm area of the two drivers, and an exit area of 20.625 in2. As you will see, this departure
from Augspurger's recommendation has little effect on f3.
 
Byrd said:
Hey Scott - Is that the response of just the driver or does it take into account augmentation by the rear wave.

D'Appolito measured the Thor's f3 @45 Hz which is supported by the graphs where the graphs show combined response with augmentation. Where there is no augmentation the responses look similar to your predicted response. The question is is this an anechoic response or "room response". I would imagine it should be anechoic.

The Seas Thor article is not very clear when it comes to what kind of measurements these are. Very naughty.:smash:


Oh no, this isn't the driver response curve, you can get that from Seas. Nope, this is the real deal: MathCad predicted system response curve (anechoic). In other words, it's the combined driver and cabinet frequency response up to 1KHz (above that, only the drivers contribute).

Note that phrase arbitary. In his defence, the tables are not great, and were quickly superceeded by Martin's work, but I'm a little surprised at Joe with that -never mind he should know better, he DOES know better. There is actually an improvement with the larger cab, though it admittedly is small. When I modelled Thor, I actually also modelled the original dimensions. Things look a little smoother, but the basic shape of the curve, with the dip below 100Hz remains -it's a characteritic of the cabinet/driver combination.

In room response plots I take with a grain of salt, and view them only as of curiosity value -evey room is different, so it might be perfect in one room, and lousy in another, like any other speaker. You just never know. OK, so a speaker isn't going to be placed in an anechoic listening room either, but it is the only way that shows exactly what the speaker itself does, without external influences. The graphs refered to in the Thor article are not up to much in my view, as they are scattered around, rather difficult to read, and it's unclear as to how they were made / measured.

Thor could, and I'm sure does, work well in some circumstances (a 25' - 35' long room is going to help here), and I reckon that lots of people will like the sound it gives. Rather than redesigning the cabinet (though I'd cut out the nonsense of 0.78 lbs to the cubic foot of stuffing for a start and take it back to 0.3 lbs) I've just taken a look at the crossover, and I notice there's a 1mH inductor in the hot lead to the bass-driver's and their zobel. It might be worth playing around some with the value of this, as we regularly do for compensation circuits using single-driver systems. You could also try adding a resistor in parallel with it. As Thor is just a 2 way, it shouldn't be too tricky to ballance it out.
Cheers
Scott
 
I have never bothered to model the Thor enclosure, I criticized the design in the past without actually doing the homework, so I am very interested in Scott's plots. But I have a couple of questions/comments for Scott.

1. Did you combine the two drivers into one equivalent driver and then use the actual enclosure dimensions to model the speaker's response? If yes, did you place the equivalent driver at the midpoint/tweeter location?

2. The plot you produced with the sagging low end response is for the speaker radiating into 2 x pi space. Put this design out into a room and the bass/midbass will be an additional 3 to 4 dB down due to baffle step. If you mentally sketch this corrected plot, how does the simulation's low end response compare to the published audioXpress measurement data?

3. If I remember correctly, the audioXpress article had some near field measurements for the driver and terminus, how do these plots compare to the individual calculated responses?

Thanks for running the sims, very interesting results.
 
catapult said:
You guys have probably seen this but, just in case, the Thor was measured in the Seas anechoic chamber.

http://www.seas.no/kit/thor.måling.pdf

Well, that pretty much confirms it.

MJK said:
I have never bothered to model the Thor enclosure, I criticized the design in the past without actually doing the homework, so I am very interested in Scott's plots. But I have a couple of questions/comments for Scott.

1. Did you combine the two drivers into one equivalent driver and then use the actual enclosure dimensions to model the speaker's response? If yes, did you place the equivalent driver at the midpoint/tweeter location?

2. The plot you produced with the sagging low end response is for the speaker radiating into 2 x pi space. Put this design out into a room and the bass/midbass will be an additional 3 to 4 dB down due to baffle step. If you mentally sketch this corrected plot, how does the simulation's low end response compare to the published audioXpress measurement data?

3. If I remember correctly, the audioXpress article had some near field measurements for the driver and terminus, how do these plots compare to the individual calculated responses?

Thanks for running the sims, very interesting results.


Hi again Martin –we seem to be doing this sort of thing quite a bit of late! What was the last –the Hedlund ‘horn’ (shudder) wasn’t it?

Dealing with your queries in order:

1) Yes to all points (it seemed the best way forward –hope I was right in thinking that?). I combined the drivers as per your site article on modelling two drivers. The enclosure dimensions were taken from the audioXpress pdf file –the frequency plot I posted was for Thor as it stands, not Joe’s originally calculated dimensions (which I also plugged in, with fairly similar results), and I took the tweeter location as the assumed drive centre. Only changes made in the driver section was to double vad, halve the impedance and remember to take the +6db for the drivers being paralleled.

2) It looks very close I think, especially compared to the graph our friend has provided a link to. With the audioXpress article, it’s only when you look further that you realise it’s been disguised by the way the graphs are presented. The measurements look far too flat to be taken at face value, suggesting ½ octave smoothing 9applies to the other graph too), and I note that a very wide dynamic range has been squeezed into a very small space –the oldest trick in the book when you want to hide something. The crossover appears to have some compensation built into it, though nowhere enough in my view to sort out the overall rising response. So if we say that cancels out the circa 3 - 4 db lost through baffle-step, and look carefully at the original graphs, what MathCad predicts looks pretty close. Figure 18 pretty much confirms it if you look at the sagging ‘woofer pair’ curve, as does the anechoic chart in the above link.

3) It does, and they’re a joke. The problem with the plots I’m looking at now is that they were all apparently done with different levels of stuffing all over the place, and also different values in the graphs. There’s no consistency at all. Overall, once you sift through the irrelevance, they seem to agree pretty well with the calculated responses in your worksheets, but again, the presentation appears to disguise problems unless you look very carefully.

That said, I’ve been fiddling around with different values, and I still can’t hit on anything TL wise that I’d consider building. A TQWT taper appears to be best, but I simply can’t tame that sagging bass below 100Hz with either a conventional taper or straight pipe, leading me to believe that these drivers simply don’t appear to be suited to a TL. Any thoughts Martin? Perhaps we should do the community a favour and come up with something for these drivers that actually works.

Cheers
Scott
 
frugal-phile™
Joined 2001
Paid Member
Re: Re: Re: D'Appolito Fudged

Scottmoose said:
In his defence, the tables are not great, and were quickly superceeded by Martin's work

Just a historical note: Augspurger's work and Martin's were being done at the same time. When Augspurger 1st presented his paper at AES NY in 1999, i was already working on the 1st published version of Martin's work. There was lots of time for Joe to see & use the MJK worksheets for the Thor.

I've also used both pieces of software... from a development point of view, the MJL worksheets are way more versatile than Augspurger's (if i was a SW developer i'd be embarrased to ask money for something as crude as TLwrx -- windows users might not be as spoiled as i am thou)

dave
 
frugal-phile™
Joined 2001
Paid Member
It may seem from the above that i'm bashung George abit -- not at all. His worl is very much complementary to Martin's, and the 2 validate each other. To alarge extent George has also gotten the attention of a different audience.

Martin on the other hand has made his work much more accessible and has continued his research & interaction with others, whereas George has gone onto other things (leaving rick Schultz to explore the space opened up by the GA SW). Martin is to be applauded for continuing to push the frontiers & to dramatically improve our understanding of box design.

dave