AVForums South Africa

DIY => DIY Audio => Topic started by: The Godfather on April 18, 2012, 01:56:30 PM

Title: "Small" Thor Measurements
Post by: The Godfather on April 18, 2012, 01:56:30 PM
Some measurements I have taken of the "Small" Thors I built some time back. I finally got around to using the measurement rig I ordered and am starting to get the hang of things.

Spot any problems, speak up.
(https://lh4.googleusercontent.com/-tflB-yzyDiU/T46pEvcCi-I/AAAAAAAAAHk/4ah1EaACrRU/s890/1.+Nearfield+Woofers+Summed.png)

(https://lh4.googleusercontent.com/-VOPRDx-upR4/T46pEvKKuuI/AAAAAAAAAHg/IJG0Ciaf49A/s890/2.+Nearfield+Port.png)

(https://lh3.googleusercontent.com/-2z51DtqE2aQ/T46pFPmAsAI/AAAAAAAAAHo/xP-i-rL4qAM/s890/3.+Nearfield+-+Port%2BWoofers.png)

(https://lh4.googleusercontent.com/-JMBsJCj5tco/T46pGUF90DI/AAAAAAAAAH8/vVbAPVAxq7g/s890/4.+Farfield+Frequency+Response+-+Gated.png)

(https://lh6.googleusercontent.com/-5XXB2WTiG04/T46pHLZbdRI/AAAAAAAAAH4/AbgHXMllnmo/s890/5.1+Integrating+Farfield+%26+Nearfield+Responses+-+No+Box+Diffraction+Applied.png)

(https://lh6.googleusercontent.com/-3BCjzNvm6fE/T46pIXwu02I/AAAAAAAAAIE/FhNdLz0lJjc/s890/5.2+Integrating+Farfield+%26+Nearfield+Responses+-+Box+Diffraction+Applied.png)

(https://lh4.googleusercontent.com/-EZeJYlBzmp4/T46pJbckzjI/AAAAAAAAAIU/cEXNeJBtnfQ/s890/5.3+Integrating+Farfield+%26+Nearfield+Responses+-+Box+Diffraction+Applied.png)

(https://lh4.googleusercontent.com/-QMmMNCgDj-w/T46pJeKnt4I/AAAAAAAAAIQ/T7_qr237YIM/s890/6.+Freefield+Response.png)
Title: Re: "Small" Thor Measurements
Post by: DRNB on April 18, 2012, 02:12:18 PM
They seem to perform better than I thought on the Low end. Interestingly I see the whole response Tilt down towards the HF side, is that the way the x-over was designed? From what I remember the standard x-over didn't have much BSC? Are you still using the standard or the modified version? would be interresting to see the vertical off-axis response, because to my mind the x-over frequency on the original design is too high for an MTM. Although I have never heard a set myself. Based on the dip between 2k and 3k I assume that is the x-over point, so still original?

Looks good overall, and no, I can't spot anything wrong.

Title: Re: "Small" Thor Measurements
Post by: Shonver on April 18, 2012, 02:16:04 PM
Measurements look good. FR dip just above 10kHz does look strange, though I don't think it will sound like there's a dip.
Title: Re: "Small" Thor Measurements
Post by: DRNB on April 18, 2012, 02:19:37 PM
Measurements look good. FR dip just above 10kHz does look strange, though I don't think it will sound like there's a dip.

saw that too, but assumed it might be mic.  :-\
Title: Re: "Small" Thor Measurements
Post by: The Godfather on April 18, 2012, 02:37:05 PM
They seem to perform better than I thought on the Low end. Interestingly I see the whole response Tilt down towards the HF side, is that the way the x-over was designed? From what I remember the standard x-over didn't have much BSC? Are you still using the standard or the modified version? would be interresting to see the vertical off-axis response, because to my mind the x-over frequency on the original design is too high for an MTM. Although I have never heard a set myself. Based on the dip between 2k and 3k I assume that is the x-over point, so still original?

Looks good overall, and no, I can't spot anything wrong.


These are one of the reworked Thors, an MLTL. The low end seems to follow quite closely to what the predicted response should be in King's worksheets for it. This is, as you have correctly guessed, the original crossover. There was much of speculation that D'Appolito didn't do his homework on the crossover WRT the blimp in the magnesium woofers, so I was looking for that but as you can see there is nothing evident. ITO off axis response .. mmm will get back to you on that one :)

There are two things I suppose you should consider when looking at the X-over point. The seas graph for the Millenium shows it under a filter at 2.5k. I wonder how ragged it starts becoming if you lower the XO point? Then there is also power to consider as it is rated at 90W under those data sheet conditions (From 2.5K upwards). Who is willing to test what the power rating is from say 1.6k  :EGrin:

Measurements look good. FR dip just above 10kHz does look strange, though I don't think it will sound like there's a dip.
Looking at the seas graph for the driver you can see the dip it just doesn't look as severe. Their graph isn't to the same scale though but they also don't specify their smoothing which makes a difference. Overall though it matches pretty well so I am happy enough for my first attempt. But i was surpised to find the dip there and it took me ages to convince myself it wasnt me making an error.
(http://www.seas.no/images/stories/excel/tweeters/images_big/f_seas_excel_loudspeaker_tweeter_e0011_t25cf002.jpg)
Title: Re: "Small" Thor Measurements
Post by: Shonver on April 18, 2012, 02:46:27 PM
Look at the CSD plot on Zpahaudio.com, and you'll see an anomally above 10k. So it looks like there's a real issue there.

http://www.zaphaudio.com/tweetermishmash/compare.html
Title: Re: "Small" Thor Measurements
Post by: DRNB on April 18, 2012, 03:29:53 PM
There was much of speculation that D'Appolito didn't do his homework on the crossover WRT the blimp in the magnesium woofers, so I was looking for that but as you can see there is nothing evident. ITO off axis response .. mmm will get back to you on that one :)

No I don't think there's a real issue wrt the cone breakup, as is evident in your measurements.
The problem I have is that the bottom to top distance of the midranges is too much for the x-over at 2.5k But I am a little anal about that, and if the tweeter is at the correct height, there should be no issues.

There are two things I suppose you should consider when looking at the X-over point. The seas graph for the Millennium shows it under a filter at 2.5k. I wonder how ragged it starts becoming if you lower the XO point? Then there is also power to consider as it is rated at 90W under those data sheet conditions (From 2.5K upwards). Who is willing to test what the power rating is from say 1.6k  :EGrin:

Th Tryms x-over at 2.6k, and I'm currently crossing my bookshelves at 1.7k.... no issues so far. :nutter:  I have to admit though that I am a little cautious, and would like to add a waveguide at some stage to help ease off the load of the tweeter. Their waaay to expensive to take chances with  ;D
Title: Re: "Small" Thor Measurements
Post by: JimGore on April 18, 2012, 03:35:44 PM
Linkwitz uses a 4th order electrical crossover frequency of 1.4k to 1.5k on the Orions (Seas Millenium) with no issues on power handling.

Whether this is a good idea is another story however.

Ian.
Title: Re: "Small" Thor Measurements
Post by: The Godfather on April 18, 2012, 04:04:31 PM
No I don't think there's a real issue wrt the cone breakup, as is evident in your measurements.
The problem I have is that the bottom to top distance of the midranges is too much for the x-over at 2.5k But I am a little anal about that, and if the tweeter is at the correct height, there should be no issues.

Th Tryms x-over at 2.6k, and I'm currently crossing my bookshelves at 1.7k.... no issues so far. :nutter:  I have to admit though that I am a little cautious, and would like to add a waveguide at some stage to help ease off the load of the tweeter. Their waaay to expensive to take chances with  ;D
It would make me sad if I were to damage mine  :'( Budgetry re-allocation is taking place so not much place for audio budgets.

I bought a MiniDSP last year in order to be able to attempt different crossover configurations. I have attached it, but while sweating and biting my nails. Once i have ensured that the MiniDSP isn't going to destroy my speakers with the off surges I will do some more work on the x-o.
Title: Re: "Small" Thor Measurements
Post by: Shonver on April 18, 2012, 04:15:38 PM
OK. Found a study on MTM crossover frequency and off-axis response.

http://www.birotechnology.com/articles/VSTWLA.html

FWIW, I think that 2kHz is a safe XO frequency to start experimenting with for this class of tweeter. If it starts to honk at you, you know you've gone too low.
Title: Re: "Small" Thor Measurements
Post by: Timber_MG on April 18, 2012, 04:26:02 PM
The Millenium, 27TDFC and derrivative tweeters both don't mind 1.5kHz crosses. WIth waveguides i've done 1.3kHz with loud levels without issues. The other Seas tweeters are usually crossed around 1.8kHz or higher.
Title: Re: "Small" Thor Measurements
Post by: The Godfather on April 18, 2012, 04:28:19 PM
The Millenium, 27TDFC and derrivative tweeters both don't mind 1.5kHz crosses. WIth waveguides i've done 1.3kHz with loud levels without issues. The other Seas tweeters are usually crossed around 1.8kHz or higher.
What Xo slopes & power were you driving?
Title: Re: "Small" Thor Measurements
Post by: tangmonster on April 18, 2012, 05:42:41 PM
isn't going to destroy my speakers with the off surges I will do some more work on the x-o.

I hope you are planning to build those linkwitz riley boards you bought from me? :)

If you use the power supply board that has the delay on and off muting you really can't damage it.

Title: Re: "Small" Thor Measurements
Post by: The Godfather on April 18, 2012, 06:17:45 PM
I hope you are planning to build those linkwitz riley boards you bought from me? :)

If you use the power supply board that has the delay on and off muting you really can't damage it.


:thumbs:

The miniDSP is supposed to have muting outputs, however they dont work 100% so that is the plan. I tried doing a quasi work around forcing the Mute to engage before the PSU unloads, but it worries me. Going to do it the proper way at the outputs. I was dreading the 8 + 16 switches that I would have to implement, but Shaun convinced me to only do the 8 for single ended.

The LR boards will be when I have fiddled an xover with the minidsp, in about 10 years time :)
Title: Re: "Small" Thor Measurements
Post by: DRNB on April 19, 2012, 07:13:45 AM
The LR boards will be when I have fiddled an xover with the minidsp, in about 10 years time :)

now you sounds like me..... I can do a lot in a looong time  >:D
Title: Re: "Small" Thor Measurements
Post by: Prince on April 19, 2012, 08:53:21 AM
now you sounds like me..... I can do a lot in a looong time  >:D

-its called forward thinking!!

 :ROFLMAO:
Title: Re: "Small" Thor Measurements
Post by: tangmonster on April 19, 2012, 09:00:32 AM


The LR boards will be when I have fiddled an xover with the minidsp, in about 10 years time :)

If you struggle just call. Or swing by. It really is 30 minutes soldering work. But yes ,you need to be 100% of your crossover frequencies.

I can honestly say that the plain linkwitz riley crossovers with no weird or expensive caps or opamps sounds much better than a DCX2496.

Haven't ever heard a minidsp so can't comment on that.
Title: Re: "Small" Thor Measurements
Post by: Timber_MG on April 19, 2012, 09:05:01 AM
4th order final slopes on the TDFC crossed to two 4.5" driver MTM (the Millenium's slope is a bit easier to cross than the TDFC that low). When crossing to a sub at 100Hz I was positively shocked by the levels possible.

The 27TDFC waveguide was crossed 3rd order passive and didn't compress or complain at levels that were "borderline" for the RS180 beneath. The WG could have gone down lower with a 4th order slope (acoustic LR24 has the lowest crossover potential of all)  but the passive x-o was too involved and with a WG's better power response matching the low X-o isn't needed that much. I am planning another 7" 2-way with RS-180 and some waveguide.....perhaps go for a 1.0 kHz cross just because :-)

The DCX is hard to do proper gain-structure with in most hi-fi setups. If you cannot run +14 dBu in and out, you're loosing out on the capability of the unit drastically. Most hifi preamps do about 0 or +4 dBu, so you're running the unit almost 30dB below it's abilities. If you run it hot enough, it's almost a copy of the BSS unit it's modeled after.
Title: Re: "Small" Thor Measurements
Post by: Shonver on April 19, 2012, 09:24:07 AM
The DCX is hard to do proper gain-structure with in most hi-fi setups. If you cannot run +14 dBu in and out, you're loosing out on the capability of the unit drastically. Most hifi preamps do about 0 or +4 dBu, so you're running the unit almost 30dB below it's abilities. If you run it hot enough, it's almost a copy of the BSS unit it's modeled after.


Mods available place a 6-way volume control after the output of the DCX to enable it to run at full scale.
Title: Re: "Small" Thor Measurements
Post by: tangmonster on April 19, 2012, 09:30:19 AM

Mods available place a 6-way volume control after the output of the DCX to enable it to run at full scale.

That could work! makes sense.
Title: Re: "Small" Thor Measurements
Post by: The Godfather on April 19, 2012, 09:40:59 AM
If you struggle just call. Or swing by. It really is 30 minutes soldering work. But yes ,you need to be 100% of your crossover frequencies.

I can honestly say that the plain linkwitz riley crossovers with no weird or expensive caps or opamps sounds much better than a DCX2496.

Haven't ever heard a minidsp so can't comment on that.

Thanks for the offer, but I think you did a fantastic job documenting your project so I should be able to follow along. I am surprised there weren't more takers. I will call on you if I get stuck though.

Setting up the XO is what I think is going to take the time, and finding the time to do it. It has taken me over 6 months since getting all this measurement equipment to put it to use :) Then dips like the 10k which you aren't expecting take up days. Aaarg.

Title: Re: "Small" Thor Measurements
Post by: Shonver on April 19, 2012, 10:47:17 AM
Doing an analogue L-R is clever, as doing so in the digital domain can strain the gain structure (~6-20dB boost). I think analogue L-R and the rest digital would work well.
Title: Re: "Small" Thor Measurements
Post by: The Godfather on April 19, 2012, 12:48:56 PM
Interestingly I see the whole response Tilt down towards the HF side, is that the way the x-over was designed?
Something has been unexplained to me in the x-over for some time and I remember somebody making reference to it before.

Why is C4 there? Surely this would cause a downward tapering type response as displayed? I have also noticed that D'appolito's measurements and Seas measurements look slightly different. Seas shows this downward taper, but D'Appolito's does not.

http://www.seas.no/index.php?option=com_content&view=article&id=27&Itemid=40
http://www.audioxcel.com/audioXpress%20Thor%20Review.pdf
Title: Re: "Small" Thor Measurements
Post by: DRNB on April 19, 2012, 01:42:45 PM
Something has been unexplained to me in the x-over for some time and I remember somebody making reference to it before.

Why is C4 there? Surely this would cause a downward tapering type response as displayed?

You're right with regard to C4, IMO at least. And its not necessarily a bad thing. I would think that's perhaps part of the voicing of the speakers. I would imagine the speakers to be very detailed in the mid, although somewhat laidback? Kind of like my Charios was  :thinking:
Bass would be deep, extended, even detailed, but not a kick in the chess?

Kind off a speaker one can sit and listen to the whole day? Or am I waaay off?
Title: Re: "Small" Thor Measurements
Post by: The Godfather on April 19, 2012, 01:51:32 PM
Lets ask Shaun, he is ruthlessly honest ;D and probably has more experience with comparison.

Room acoustics are a big factor for me and probably overwhelm the character of the speakers.
Title: Re: "Small" Thor Measurements
Post by: Shonver on April 19, 2012, 04:46:13 PM
Lets ask Shaun, he is ruthlessly honest ;D and probably has more experience with comparison.

Room acoustics are a big factor for me and probably overwhelm the character of the speakers.

There is a school of thought that believes that a continuously falling response provides the most balanced sound (ask Rodney what he thinks of a flat response!). But I agree that the room response swamps the overall response. A down-sloping response might be the best for a live room. I also think that keeping the response linear makes it least offensive to the ear; whether rising, flat or falling (within reason).