MJAO G 098 - The Anagram - Crossover revision

Polarbear · « **Reply #60 on:** 23 Mar 2009, 03:46 am »

Thanks Erling. I see.

Bjørn

Mr Content · « **Reply #61 on:** 23 Mar 2009, 06:32 am »

Hi Till, welcome to AC, Nice baffles, have done a few very similar, and in white as well. I have a pair of white ones in stock now, and I am thinking of bringing them home for a change from my veneered ones. Nice job

Mr C aa

THWO · « **Reply #62 on:** 23 Mar 2009, 05:34 pm »

Thanks for your nice welcome.

Hi Erling, you are right, the L8 do not have a frequency correction at the moment. Let´s see how it will sound when the Alphas will replace the Ciares. These are run actively at present by two Hypex DS 2.0 modules. I plan to try an entirely passive modus then. There is a horntweeter-backup on the back side of the panel, smoothly cut off with 1.33 mF:

Hi dmiller, I will be happy to report after the exchange of the bass.

Hello Josh, the engraved stripes (by router) indeed "function" just as an optical support to slimline the overall shape. You asked about the flat back around the back of the midrange; yes it was done just by router. As there is no router allowing digging such a wide and flat shape with a diameter of 38 cm (15") and depth of about 3.5 cm (1.4"), I did it by starting with an outer circle and, while decreasing the diameter step by step, I went deeper and deeper, creating a structure like in an opencast mining field:

After that: sanding and sanding (you will know...)...

The back of the panel after receiving it back laquered, before assembling the rest:

Hi Mr. C., I like your vaneered ones, too! Very nice idea with the diagonal design of the "curtains".

Till

intermission · « **Reply #63 on:** 24 Mar 2009, 12:34 pm »

Quote from: scorpion on 22 Mar 2009, 10:06 pm

The flat section at the top is not there by mistake it helps minimizing diffraction. A rounded top would have been worse.

Erling (and for that matter everyone that has been posting information about their projects here,) I've found it all very interesting and am contemplating putting together an OB along the lines of The Anagram at some point.
Would you be kind enough to explain to a relative simpleton why the above is so? Would a square or angled baffle be 'better' than an gently arched top then acoustically? I'm sure there's probably a very obvious reason but it's useful to understand it (mainly so I can explain it to the wife when she starts commenting on the appearance!)
While I'm here, could I ask you what the useful lower frequency range is? I'm guessing somewhere around 40 Hz?
Thanks,
Gary

scorpion · « **Reply #64 on:** 24 Mar 2009, 08:21 pm »

I'll try to illustrate with an example with the Edge program, that illustrates diffraction effects of the baffle. Edge is here: http://www.tolvan.com/edge/ .
First a baffle like the MJAO but with a rounded top and the midrange placed in the middle at the same height as in the MJAO.

resulting in diffraction effects from the baffle like this:

You can do guite a bit better by manipulating baffle shape and speaker placement. You will however always have some kind of dipole effect with a hump in response. But this rounded baffle top is not the best you can do.

It should also be pointed out that this is pure theory and real life measurement in a listening room can very well prove it wrong.

/Erling

intermission · « **Reply #65 on:** 25 Mar 2009, 09:39 am »

That is a pretty mind-mangling program! (I've been up working all night so perhaps it will make more sense after a sleep)
Any chance of a pic showing the cross-over or back of the baffle please?

scorpion · « **Reply #66 on:** 25 Mar 2009, 05:24 pm »

You wouldn't like to see that at the moment. However things will change and I will have a presentable LS at our Swedish yearly Hi-Fi Forum DIY-show April 18th in Gothenburg, where I hope to make an impression. Then you will have the picture. I am among the few Swedes advocating OB but have plagued my fellows with the Ciare CH 250, the Volks-OB, the B200 and now the MJAO and quite a few are converting.

I also have taken some distortion measurement of the MJAO . These are just one meter closer to the speaker (right one) and mic is pointing to somwhere in the middle of the baffle a little above woofers. Frequency response is a bit different from the sweetspot measurments. Also German Klang & Ton in a rewiev of ARTA stated that distortion mesurements should be taken more near the loudspeaker.This is the picture, 95 dB avarage SPL:

-40 dB is equal to 1% distortion - 50 dB is equal to .3% distortionm and -60 dB is .1 % distortion

/Erling

intermission · « **Reply #67 on:** 25 Mar 2009, 07:10 pm »

Cool. Look forward to that, interesting this is.

scorpion · « **Reply #68 on:** 3 Apr 2009, 04:34 pm »

Eventually I have received some Dayton Metallized Polypropylene 60 uF capacitors from PE and I have played around with the mid crossover for some days now.
The result is that the PCA capacitor has just been a try. It is not as good as either the Obbligatos or the Daytons. I have played both with 90 uF (Daytons + 30 uF Obligattos) and 130 uF (Daytons + 70 uF Obbligatos) and it is funny but the effect on crossover frequency has been quite low if I believe my measurements. But the midrange comes out with a new very pleasent authority - like this is the way things should sound. It would certainly be interesting to hear how other users find the Dayton caps. Tweeter is now crossed with a standard Mundorf M-cap 4.7 uF.

/Erling

scorpion · « **Reply #69 on:** 15 Apr 2009, 07:05 pm »

I do run the MJAO with 130 uF (60 uF Dayton + 70 uF Obbligato) now to the midrange. It is a bit on the low side for heavy bass, but the sound is the better. I should perhaps try to design a 12 dB crossover for MS-100CHQ to protect it from heavy low-bass. But it seems to cope very well so far. Stay tuned if interested.

/Erling

scorpion · « **Reply #70 on:** 21 Apr 2009, 08:27 pm »

Also 4.7 uF Obbligatos in place now for the treble, another not so small lift in overall performance. I went for our yearly DIY-meeting in Gothenburg with the MJAO. Overall I think people liked what they heard, but I was especially pleased by female response to the speaker, picking it out as one of the very best there, more than two witnesses. Women hear much more than we men do, usually. And we boys are competing more than we will acknowledge.

/Erling

ecir38 · « **Reply #71 on:** 22 Apr 2009, 12:34 pm »

Hey Erling, just letting you know someone is listening. Keep up the good work and thanks for sharing.

Brad

scorpion · « **Reply #72 on:** 30 Apr 2009, 03:33 pm »

Rudolf was right, the crossover was too frugal. I have had a lot of comments from the Gothenburg DIY-show. Most were very favourable but a few insisted that the Alphas played too high in frequency interferring with the midrange. I went home and listend to the crossover. Using ARTA and Pink Noise signals this is easy. However I should have done this long before. But my computer where measurements are generated and the speakers are in different rooms, so I usually ran forward and back to watch measurements so this had slipped my attention. Conclusion is: Alphas do give out much more HF than desirable.

I have read some comment about, either here or at diyAudio, that trying to crossing Alphas looked more or less the same irrespective how you set crossover. I can vouch for this. I have both before and after making changes measurements thats looks more or less the same, and that's not really depending if you use 1st or 2nd order filters. They more or less produce the same result. That means that the Alphas are a tough beasts to tame passively.

I wanted to stay with 6 db slopes so this is my result:

The result is that MJAO G 098 is an even more detailed speaker now. Both the Zobel and the Parallel Notch contribute, you can't do without the other. The notch is calculated at 1500-1600 Hz. Getting rid of those HF-contributions generates an effect that is in fact a bit amazing, there is a kind of Blackness in the presentation that usually is associated with amplifier performance but an effect that I now can associate to the resolution and to 2nd- and 3rd-dimensional presentation.

The small story, you can't leave these changes out, then you will underperfom these good speakers. I hope you can see from the picture the chosen values. If not let's know. The Zobel is 3.9 ohm and 47 uF and the parallel notch 1mH and 10 uF.

/Erling

Polarbear · « **Reply #73 on:** 1 May 2009, 08:30 am »

Great, I have to order some more parts then

You also reduced C1 from 130uF to 100uF?

Good work Erling. Keep it up

Bjørn

scorpion · « **Reply #74 on:** 1 May 2009, 09:11 am »

I still play with the 130 uF combo (I just happen to like the tonal quality of the Obbligato - Dayton combination), but I have tried values from 70 uF and upwards with very little if any difference. So 100 uF should work fine.
Yes, the extra parts will cost some money but they will be worth it !

/Erling

scorpion · « **Reply #75 on:** 1 May 2009, 09:59 pm »

I forgot to mention, if you got a 6.8 uF cap at home it will do the same job as the 10 uF in the Parallel Notch. Both efficiently blocking the big 2 kHz Alpha hump to an enormous benefit for the speaker.

/Erling

nvrgdenuf · « **Reply #76 on:** 3 May 2009, 12:07 am »

Scorpion,
Probably an obvious answer, but what about dynamics comparing using the Monacor's vs the B200. I have received the Monacors but havent opened them, possibly looking to trade them for a pair of the B200's. The Alphas can certainly respond with the dynamics, but can the little Monacors keep up?
Just to fyi presently I use a DDS waveguide and BMS 4552nd x-over at 1k, and have a driverack PA and dcx24/96 to choose from, but I think lowering the x-over point to a better midrange than using the Alphas to 1k may (hopefully),improve things.
thanks

ttan98 · « **Reply #77 on:** 3 May 2009, 12:04 pm »

scorpion,

How do you apply your Zobel to the Alpha 15"? do you measure the impedance first and the apply the calculated Zobel network? The Zobel network makes it easier for the amplifier to drive.

How does apply Zobel makes a different to the sound reproduction?

Cheers.

Telstar · « **Reply #78 on:** 3 May 2009, 12:15 pm »

Hi Erling,

Yeah, I feared that going 3-way would require more complexity in the XO, which was the main reason why I took a different route.

Your project still looks great and will stay in the affordable budget

And now it sounds better too

scorpion · « **Reply #79 on:** 3 May 2009, 01:30 pm »

nvrgdenuf,

If you want to match the BMS sensitivity the Monacor will fall short. I have tested with high bass/low mid high transient content and gone to 100-105 dB SPL with no problem. But that's probably the limit, the midrange will be too stressed going higher. There is a new Monacor unit, which already has attracted some interest in Germany I understand, the SP-8/150PRO: http://www.monacor.de/typo3/index.php?id=84&L=0&artid=7376&spr=DE&typ=full . This might be interesting to you. From my simulations it would be a perfect match for two Alphas on the 40 cm wide baffle.

ttan98 and Telstar,

Well, the Alphas are just 1-way but impedancewise evidently complex enough to deal with.

The Zobel is calculated from measured electrical resistance of the parallel Alphas and Eminence published Le figures. My LC-meter is of the simple sort measuring inductance at a given frequency, in my case 900 Hz, so this won't give you a good measurement but could be used for guidance. I have applied values obtained from a calculator which probably uses some rule-of-thumb formula.

The Zobel helps the crossover coil a bit but it is not effecient enough to get rid of all hissing content in crossover output. To effectively silence that the notch is introduced. Theoretically the Low-Pass should be effective from 300 Hz and now it also starts to sound like that. If you can, listen to your crossovers. Then you will much easier identify things that should not be there than what you can hear only by playing music.

/Erling