Clarity on Seas Thor Kit

Scottmoose said:
....
Why are the simulated and measured outputs different? Well, the curves are basically very similar. Did you remember to apply to the MathCad charts the +6db boost gained in system efficiency for having paralleled drivers when comparing to the measured data?

Regards
Scott

All response graphs are similar - IE they are curtailed at both bottom & top end ;)

But seriously I still don't see what you see. I am not talking here simply of amplitude but a fundamental difference. Look at the 30 Hz Dip in Driver output & 35 Hz Peak in Port output as predicted by the sims. There is no evidence of this in the measurements taken by D'Appolito.

There is something causing this difference. Either the sims are wrong (No offence intended here please, just object discussion), or they are being fed different parameters to what actually exist in reality, or the worlds * "foremost expert of speaker measurement" has fudged the results.

I wish D'Appolito were here to state his case.

* Wherever this title comes from I don't know but it has been bandied about often enough.
 
As do I. It would be interesting to get his views on the matter. Oh -no offence taken by the way. Life would be very boring if we all went around taking umbridge at every objective observation! Besides, I don't claim to be infallible by any stretch of the imagination -I might have been off with the vent dimensions, but I don't think it was by more than a few mm -not enough to make a major change. The best way of ascertaining this is to try it yourself -the worksheets are on Martin's site, along with a free version of MathCad Explorer 8 if you don't own the full version (works just the same, you just have to cut & paste into something like Word if you want to save).

I think the important thing we need to keep in mind when considering all this, is, whether you go with the simulated or the measured data, Thor looks poor, especially considering its high price. If I wished to build a speaker in that price range, and I saw only the measured data for example, I would be very wary, especailly when I saw the crossover.

OK, why the differences, if we're assuming for the moment that my simulation is in the ball-park? Well, firstly, keep in mind that the measured data has clearly been heavily smoothed -speakers are not that flat in real life, and this disguises response aberrations. Looks like 1/3 or 1/2 octave smoothing to me. Note that the dip in woofer output circa 30Hz is still there on the measured data -it's simply been flattened out, which confirms this, at least to my satisfaction. Ditto the peak in vent output, though apparantly on the measured data it occurs a little lower, suggesting a slight error on my part on the vent dimensions -I hold my hand up on that one. It's not that significant though in the overall scheme of things. Staying with the vent output, the remaining climbs, dip and peak just after 100Hz are all still there on the measured plot, albeit flattened out, as is everything post that, though it's not easy to compare at first glance, with such odd values being used in the measured data graph. The rapidly rising response predicted by the sim to 1KHz will be mitigated slightly by the presence of the correction applied to the woofers in the crossover, which is not enough to negate the overall rising trend indicated -there's peaking at circa 1KHz in both sim and measured data for example.

Below 100Hz in room, you won't get a huge performance boost. It'll be better than the anechoic response of course, but the vent output is too low, and being mounted high up, and firing horizontally rather than vertically it can't drive the room as well as a low mounted vent a la Lynn Olson's Ariel. Worse, you'll still have the 3-4db baffle-step losses Martin mentioned, which the crossover does not correct. And just to add a further fly in the ointment, the cabinet is full to the brim with a very heavy quantity of stuffing, which itself will cause the bass to sag further. Whichever way you look at the data, Thor will be significantly down at low frequencies. Only by listening to it in your own room however, will you be able to determine if it works well enough for your taste in your home. Depending on what you listen to and how large your room is, an alternative speaker you might like to look at is Bob Brines FTA-2000. It uses a single Fostex F200a full-range driver, and is pretty much flat in room from 30Hz - 20Khz. It'd cost a fair bit less than Thor too. You'll find it here: http://www.geocities.com/rbrines1/Pages/FTA-2000/Main.html
Perhaps not ideal for thrash metal or a very large room, but most anything else, it's a winner. The Ariel is also very good -no deep bass, but Lynn always stated it didn't, which is nice and honest of him, and it simulates and measures well. Both are cheaper, which means you could add a decent subwoofer from REL, or perhaps a pair of Linkwitz Thor (I kid you not!) monopolar subs, or the open baffle subs from his Phoenix project. That'd be what I'd do anyway!

Cheers
Scott
 
Improving Thor

I have been following this thread with much interest.

After several months of research I descided that the Thor TL was the speaker for me.

Now I'm having doubts, but I admit I am somewhat confused. Please forgive my ignorance as I am a neophyte.

Is the performance issue that Scottmoose is questioning related to the actual driver selection in this particular TL geometry or is it a combination of the driver(s) and geometry ?

I ask this question because I was going to change the design and substitute a Seas 27TDFC for the Excel tweeter, and replace the drivers with Extremis 6.8's bi-amped with an active XO at 3.5 or 4.0 kHz. The Extremis doesn't have a cone breakup mode (at least I don't hink so), and the poly cone should soften the sound a bit for the type of music I listen to.

I used the original article as the basis for my substitutions. Plugging in the extremis specifications into the calculations, the Thor dimensions actually work out to the "arbitrary" resized dimensions, which should result in the "full" bass response.

So my big question then is this:
Will my driver substitutions result in exceptional speakers or is the Thor design just pre-destined to be an "OK" speaker no matter what.

Feel free to ridicule my assumptions, maybe I need a dose of reality here, or do I actually have some potential as a DIYer.

Thanks.
 
Forget ridiculing. An active crossover is always a good step forward; pity they aren't used more often, but of course, they mean another box, more amps etc., which is probably why they're relatively rare.

When designing a TL, you have to consider the geometry, dimensions and driver T/S parameters as a whole, and whether simulated or measured, I find Thor wanting, so I suppose it's your second point -combination of drivers and enclosure that I dislike. Basically, as you know, you start with the drivers you wish to use, and design a cabinet specifically for them using alignment tables or the available software, and that's where it all goes pear-shaped for Thor in my view, bacause the tables / theory it's based upon are decidedly rocky. There's much better information and software available, like Martin's, which will yeild superior results. As you seem to have settled on the drivers you wish to use (I don't know them myself), I'd suggest you plug their parameters into his worksheets and design an optimised cabinet for them that way rather than the alignment tables Joe used when designing Thor, which aren't overly accurate. Also, if you haven't already, read some of Bob Brines guides to designing a TL. While he sticks with single drivers (as I usually do), the same theory is equally applicable to multi-ways, and Bob is a master at refining designs; they make Martin's MathCad sheets easier to understand, and give you a good starting point rather than trying things more or less at random.

Cheers
Scott.
 
I don't believe it. You know I mentioned in previous posts that a TQWT enclosure seems to produce the best response with these Seas Excel drivers. Well, try this on for size. You know what the dimensions are? Groan. They're the same as Martin's Project 2 ML TQWT, but with a 3"x2" (WxL) port. That's it. So=1Sd, Sm=7Sd. TL length is 60". Driver height is half the TL length. I've mentioned the port dimensions.

It's got a slightly rising response that'll need bringing down somewhat, but nothing like as bad. As before, add 6db overall for the drivers being paralleled. There's a slight peak at cut-off which will add a little weight to the sound. You could try either a folded or version or straight. Remember to make allowance for baffle-step (unlike the original). Not a small enclosure, but the best TL response from these drivers I've seen so far.

Scott
 

Attachments

  • untitled.jpg
    untitled.jpg
    61 KB · Views: 2,347
Well this is comforting.

Seeing you're having so much fun :) with this simulation, would
you care to post the results using the Extremis parameters ?

I'm curious to see what the difference would be.

Thanks


Extremis 6.8

Nominal Impedance: 8 Ohms
Usable Bandwidth: 30-5000 Hz
Rated Power: 100W RMS
Peak Power: 300W
Rated Efficiency: 86 dB
Moving Mass: 23.3 g
VC Diameter: 40mm
Compliance: 1.2 mm/N
Linear Travel: 26mm p-p
Maximum Travel: 30mm p-p
Driver Weight: 2.5 kg

Inductance: 0.13 mH
DC Resistance: 7.02 Ohms
BL: 9.0 N/A
Resonant Frequency: 30 Hz
Cone Area: 140 cm2
Vas: 33.2 L
Qms: 2.54
Qes: .383
Qts: .332
 
As in, in the Thor cabinet? It shall be done. If I'm feeling generous, I'll see what I can come up with as optimised dimensions for them as well.
Give me half an hour or so. (and in case anyone is wondering why I've so much spare time at present, I'm currently in limbo: I'm waiting on my supervisor to send back his suggestions on the final draft of my Master's thesis.)
Cheers
Scott
 
frugal-phile™
Joined 2001
Paid Member
The Extremis & the SEAS is a good combo. The Extremis is certainly not bass shy (as i was preparing the chicken wings at the latest DIY-fest, they were playing some classical with some serious bass in it... my room is not small, and the amp was an EL34 PP, but i was being made quite nauseous as the waves of bass rolled into the kitchen.

You can consider the folding geometry of the Thor, but i would discard the actual dimensions. From the Scottmoose sim, that cab does embarrassingly poor at getting the bottom out of the Extremis that is possible -- it also has a large amount of ri[pple. I'll be very intersted in the more optimum sims -- i have 4 extremis sitting here and a TL in my queue of things to come up with. He comes up with dimensions i'll whip up a working drawing.

Do note that the mids of the Extremis -- thou reasonable for a 7" driver -- are no match to a WR or FR125, Fostex FE127 or FE108, or even an Aura.

Since you are bi-amping you might think about HF options that go lower (my 1st thought was any number of the 3" FRs) -- or at least do the design to make this kind of thing realizable leater.

As an example, I ended up with the left-overs from the North American Monsoon service depot and have some 40-50 subs with good drivers under my deck (check my AVB thread for details). I was going to do an experiment with 4 of the 8" units and a set of "non-existant" 3" TBs in a 3 box MTM XOed at ~300 Hz (ie just seal off the hole from the plate amp with a piece of ply that ties 2 woofer boxes together with the TB in its 1/4 litre cube sandwhiched in between...

dave
 
Hey Scott

Scottmoose said:
I think the important thing we need to keep in mind when considering all this, is, whether you go with the simulated or the measured data, Thor looks poor, especially considering its high price. If I wished to build a speaker in that price range, and I saw only the measured data for example, I would be very wary, especailly when I saw the crossover.

It looks fine in terms of measurement - very flat - and the measurement matches the claim that f3 = 45Hz

Scottmoose said:
OK, why the differences, if we're assuming for the moment that my simulation is in the ball-park? Well, firstly, keep in mind that the measured data has clearly been heavily smoothed -speakers are not that flat in real life, and this disguises response aberrations. Looks like 1/3 or 1/2 octave smoothing to me. Note that the dip in woofer output circa 30Hz is still there on the measured data -it's simply been flattened out, which confirms this, at least to my satisfaction. Ditto the peak in vent output, though apparantly on the measured data it occurs a little lower, suggesting a slight error on my part on the vent dimensions -I hold my hand up on that one. It's not that significant though in the overall scheme of things.
[/B]

If as you say smoothing has been applied how do you explain the peaks & troughs in port output at higher frequencies? (Octave Frequency band widths increase as the frequency increases). Also look at the driver output, for evidence of this. The graph produced by the mathcad sims looks alot more "smoothed" in this region than the measured result. What accounts for this discrepancy?

If we assume the above to be true then we might say something in the model is not working correctly unless of course the model applies smoothing as well.

Scottmoose said:
Below 100Hz in room, you won't get a huge performance boost. It'll be better than the anechoic response of course, but the vent output is too low, and being mounted high up, and firing horizontally rather than vertically it can't drive the room as well as a low mounted vent a la Lynn Olson's Ariel. Worse, you'll still have the 3-4db baffle-step losses Martin mentioned, which the crossover does not correct. And just to add a further fly in the ointment, the cabinet is full to the brim with a very heavy quantity of stuffing, which itself will cause the bass to sag further. Whichever way you look at the data, Thor will be significantly down at low frequencies. Only by listening to it in your own room however, will you be able to determine if it works well enough for your taste in your home."
[/B]

Doesn't room response usualy slope upwards @ 12db / octave? This is quite significant in terms of response. I have had experience of listening with my own ears to a speaker in an anechoic chamber. I can tell you that in that instance - low frequency output was dramaticaly less than inroom response. When I say dramaticaly - almost non audible.

Scottmoose said:
Depending on what you listen to and how large your room is, an alternative speaker you might like to look at is Bob Brines FTA-2000. It uses a single Fostex F200a full-range driver, and is pretty much flat in room from 30Hz - 20Khz. It'd cost a fair bit less than Thor too. You'll find it here: http://www.geocities.com/rbrines1/Pages/FTA-2000/Main.html
Perhaps not ideal for thrash metal or a very large room, but most anything else, it's a winner.
[/B]

Sorry - dont like full range. Also couldn't find any response plots here. Have they been measured? Could you point us to one? I am sure though that the upper frequency reponse of these will certainly not be as flat to 20k as the Thor even though it may be a good attempt.

Scottmoose said:
The Ariel is also very good -no deep bass, but Lynn always stated it didn't, which is nice and honest of him, and it simulates and measures well. Both are cheaper, which means you could add a decent subwoofer from REL, or perhaps a pair of Linkwitz Thor (I kid you not!) monopolar subs, or the open baffle subs from his Phoenix project. That'd be what I'd do anyway!
[/B]
D'Appolito didn't claim Thor had deep bass which is also quite honest of him.

I think the thing we are forgetting about here is that generaly response below 40 Hz in real music is not all that common. I would prefer to have a speaker that does everything well over most of the range, instead of one that does so-so over all of it. I have heard the Seas excel drivers in a different configuration and love them in terms of higher frequency qualit and impact.

I will build the Tthor in its original config as well as possibly attempt the mathcad software sims and build a prototype of what I get there.

Thanks for you input this has been an interesting discussion.
 
Byrd said:


That still looks like a response plot without augmentation by the rear wave exiting the TL. Take a look at the delay. In fact it looks very similar to the response plot for the W18 driver measured anechoicly alone. These measurements I have been told are taken by placing the driver in a 6 litre sealed enclosure. Surely a T/L no matter how badly designed or mismatched should best this.

Take a look at this graph from the AE article. From what I understand - this shows the measured response simulated to anechoic response of the Thor, it combines both port and driver output. It does not include room gain.

If I have misread these please could you explain.

There is nothing simulated about the Seas measurement of the Thor. They put the speaker in the middle of the anechoic chamber with the mic about a meter away and then they run a simple frequency sweep and record the SPL. No MLS or quasi-anechoic windowing tricks are needed. Joe's measurements are more suspect because he's doing a lot of mathematical tricks to try to filter out the room and splice the port's response in.

You can look at the measured response of the Trym kit to see that Seas is capable of measuring a speaker with good bass.

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

That said, there should be more bass in a room than in an anechoic chamber so the anechoic measurements don't tell the whole story.
 
catapult said:


There is nothing simulated about the Seas measurement of the Thor. They put the speaker in the middle of the anechoic chamber with the mic about a meter away and then they run a simple frequency sweep and record the SPL. No MLS or quasi-anechoic windowing tricks are needed. Joe's measurements are more suspect because he's doing a lot of mathematical tricks to try to filter out the room and splice the port's response in.

You can look at the measured response of the Trym kit to see that Seas is capable of measuring a speaker with good bass.

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

That said, there should be more bass in a room than in an anechoic chamber so the anechoic measurements don't tell the whole story.

"The enclosure uses a newly developed transmission line configuration, derived both from sophisticated computer modeling, and extensive experimentation. The line is tapered, and filled with Dacron stuffing. The two W18E001 woofers excite the line at slightly different points, smoothing the response and increasing the range of bass output. The transmission line produces 4 dB of bass lift from 20Hz to 110HZ, with less than 1 dB of ripple. The -3dB point is 45Hz, with roll off of 12dB per octave below 45Hz. Usable in-room bass response extends well into the
low 30Hz range."

I am not sure if you noticed but on Thors main page that leads to the graph you linked to the fact that the above is stated which does not lock up with the graph that they have displayed for the Thor. Is this statement on their site incorrect?

I wouldn't say "Joe" is doing alot of mathematical tricks, but rather the CLIO system he uses. Is CLIO really as bad as all that to show a 45 Hz f3 where an anechoic response shows an f3 of 100Hz? mIf so there is no hope for any DIYers unless they have access to an anechoic chamber. My closest is about 500 Meters away :D
 
I'd rather you gentlemen than me -we clearly have rather different ideas about what constitutes a well engineered and measuring speaker. I suspect we're going to agree to disagree -fine with me, I like a debate, and, likewise, I've thoroughly enjoyed this one. I also sincerely hope you enjoy your Thors when they are built -perhaps you could give us some feedback when they're completed? The more user reports and measurements the better I say. Out of interest, what sort of power-amps will you be using with them?

You're not alone in having experience in anechoic listening -so have I. Anechoic bass should be much less than in room -that's the point. Under those conditions, you're only listening to the speaker. Personally though, I don't like using too much room gain, for that very reason: you're listening to the room, not the drive units. Not always a good plan. How much is applied, and where in the frequency range rather depends upon the room size and its furnishings, which is unpredictable to say the least. One of the reasons I also like open baffles, a la Linkwitz. No mentions that Thor has deep bass? Well, from the number of comments in that article regarding strong output at 25Hz... Draw your own conclusions there I reckon! (Though I don't hold Dr J. D. responsible for that.)

I can understand why many people dislike full-range units -that said; heard a good one of late? Most of the old ones were poor, but the new Fostex and Lowther units are frankly astounding. Not quite as flat as a multi-way speaker, but they have slightly different priorities -the usual trade offs in speaker design: gain one thing, sacrifice another. For your interest, I've got a response graph of Bob's FTA-2000s somewhere -I'll dig it out when I have a moment as you asked. Bob's site has become smaller of late -too many people were pinching his designs I suspect, though I haven't yet asked him.

It's interesting that we actually agree on many points though you might not believe it -I too favour high quality mids and highs over broader, but lower quality output over a wider range, and I like these Seas drivers a lot. My primary concern lies in the fact that you can match and beat this performance though for around half the price with Ariel for example, but we've already discussed that.

Oh, one last graph for your interest. Meet the Thor cabinet -with a single Seas Excel driver. Everything else is exactly as per the standard cabinet Joe designed. I smiled a wry smile. Actually, the ripple you see will actually be less than shown here, as the driver placement needs a fractional modification, which I didn't put in. Blame the doubled Vas of the paralleled drivers. I confess to being impressed (though I still don't like high mounted vents!). This looks much better.

Best regards & enjoy the music
Scott
 

Attachments

  • untitled.jpg
    untitled.jpg
    62.4 KB · Views: 1,749
planet10 said:


<rant>Not Word please.... I'm fighting with one of those as we speak. Seems anytime something gets near Word it screws it up.

I wish there was a real WP available for PC users</rant>

dave

I know the feeling Dave -Bring back WordPerfect, that's what I say. I never had any problem with that. I still have v5, on an 8086 with 40mb HD, 2 5.25" FDD drives and Hercules green-screen monitor! And it still works...

Best
Scott
 
That is interesting. As I say I will build the Thors as they are first - then do some fiddling. Will be over the next 2 / 3 months. (If the drive units ever arrive that is)

I have also mailed Seas on the apperent discrepancy - because that sort of thing drives me nuts as you can tell. The reason probably being in South Africa we are so far away from everything that when other people **** specs up - we pay for shipping and its not cheap or easy.
 
frugal-phile™
Joined 2001
Paid Member
Scottmoose said:
Oh, one last graph for your interest. Meet the Thor cabinet -with a single Seas Excel driver. Everything else is exactly as per the standard cabinet Joe designed. I smiled a wry smile. Actually, the ripple you see will actually be less than shown here, as the driver placement needs a fractional modification, which I didn't put in. Blame the doubled Vas of the paralleled drivers. I confess to being impressed (though I still don't like high mounted vents!). This looks much better.

Almost makes you think he forgot to multiply by 2. So an easy fix for the Thor might be to double the cross-section all the way round :^)

dave