Help with Crossover for MW144 TL

I got the L-Pad fitted a few days ago. While low rate internet streams and poor recordings like a lot of attenuation, I have found just a little is needed when all is well. Perhaps inline with the data sheet discrepancies. I'm yet to actually measure what have done. Some Kef 2005.2 set satellites and a Mission sub appeared from somewhere. I think it was the HiFi fairy again. Though a 12" subwoofer for the car might of been the ebay beer monsters work. Apparently a Muse amp is on it's way and there are notes on my desk about a desktop set to assemble. When the ball gets rolling, it certainly picks up some bits.

The Kef eggs are quite remarkable. Cheaper than my Morel BI-2 set and while nowhere near as chunky sounding, they can put instruments in the room more convincingly than anything else I own.

As you can tell, I'm a little sidetracked.
 
Picking up from where I left this last week ?

I just received some used IPL M4 crossovers. They use the same MW144 with an MDT39 tweeter in a 5L sealed box. They incorporate a Zobel network (say the instructions).

The LF section uses a 1.04mH choke followed by a 10uf to ground that sits in series with a 1ohm resistor.
The HF section starts with a 6.8ohm resistor followed by a 4uf filter with a 0.53mH choke across the tweeter.

I have no idea what this means, but it looks to be around a 2.5khz X-over by my reckoning.

Both my amps are out on loan, so I can't try them yet.
 
hehe long week eh 😉

I ran those values through the same sim as I did before. Same drivers same offset just the new crossover values. The IPL is a bit smoother I suspect the drop between 5 and 10Khz will probably be pleasing. Black is the IPL curve... also slightly less peaking in the 1-2Khz range.

What I would really like to know is the real offset between the tweeters and the MW144's I went with 12mm... the IPL crossover seems to have not very good phase matching.. maybe they did this to get a smoother overall result. Perhaps the phase matching is not as important as what I think it is (certainly my own experience tends to say it does make a difference).. though revisting I certainly didn't get great phase matching either, perhaps the constant non matching of the IPL is better than my crossing phase...

Actualy I just put in some target curves, it looks like the IPL is a 3rd order butterworth at around 2.5Khz Tweeter doesn't match the curve fantastically but it looks like that is what it is which is why the phase doesn't match... it is not supposed to!! should be 90deg out and it looks like they have done a good job with that.

2nd attachment is IPL's phase, and 3rd is the phase of the crossover I originally put forward. 4th shows 3rd order butterworth targets at 2.5Khz 83db

Tony.
 

Attachments

  • friendly_IPL.png
    friendly_IPL.png
    20.4 KB · Views: 133
  • IPL_phase.png
    IPL_phase.png
    23.6 KB · Views: 140
  • friendly_tony.png
    friendly_tony.png
    23.8 KB · Views: 133
  • friendly_IPL_target.png
    friendly_IPL_target.png
    25.5 KB · Views: 132
Last edited:
Interesting the Russian crossover is quite similar to what I had come up with. here is the comparison.

I think I'll trace the russian final curve so I can compare it to the sim and get an idea how close the traced curves of the drivers are to reality!

Tony.
 

Attachments

  • friendly_russian.png
    friendly_russian.png
    20.5 KB · Views: 146
  • russian_schematic.png
    russian_schematic.png
    7.3 KB · Views: 103
  • tony_schematic.png
    tony_schematic.png
    7.3 KB · Views: 96
Quite a difference between the sim (using traced manufacturer curves) and the Russian result! Note he is using a vented enclosure so that explains the low frequency boost, but quite a difference in the upper frequency response! black is the sim, green is result traced from russian site.

I guess this shows why having some real measurements of your actual drivers in the box is the only way to go.

Tony.
 

Attachments

  • comparison.png
    comparison.png
    20.8 KB · Views: 91
And another 14 months go by.

I switched my speaker terminal plate for a biwire one. Giving me the space to fix a board to it's rear, where I then transferred the ipl parts. Rather than the x-over living in the lines throat. Which I treated to some ground rubber carpet underlay, while I pulled out the $5 cables. Swapping them for runs of silver anniversary, to the new x-over location. Down by the floor. It all went well, with the ipl parts having long enough legs that they all joined up nicely. Without compromising component positioning.

It certainly sounds different. A lot thinner in the x-over region. Loosing some authority or shoutyness. But gaining in most areas, as you might expect.

The dome of the tweeter is just a few MM forward of the mw144 center cap. Only because Ive not rebated the tweeter, but rather just surface mounted it. I havn't had time, you know how it is. Properly fitted, the tweeter aligns with the surface mount driver. It's such a small inaccuracy, it's overshadowed by how far off axis I'm usually listening.

My listening has been cut short, as between trialing one crossover, and then fitting both properly... I built some tg9fd speakers. I needed to learn how to use my router, but the speakers produced have sat my down. Literally. To listen to everything I own again, then start streaming. I'm chopping wood for the 12" JBL subs they need now. I will come back to this T-line project though.
 
Last edited:
Well, they sound absolutely foul. After an hour listening as I moved about the house, I disconnected one tweeter and the balance was much better, but my ears were still so irritated I needed some silence.



I'm happy with them without tweeters. Obviously loads missing, but lots of promise and nothing wrong. I do read these mw144's are inconsistent, and people talk of a midrange dip. Perhaps I need to turn the tweeters down a lot.



What do we know about these ipl labeled capacitors? The humblehomemadehifi page doesn't list them. Are they highly detailed? I need porridge. Sound you can barely stir with a big stick. I pretty much dislike most speakers attempts at HF reproduction, and anything that emphasizes HF in a bid to fake clarity is rubbish imo. I'm often putting my hand over tweeters and finding things sound much better. Does this offer any clues as to what I may like? I have had these tweeters sounding nice before, so it's implementation.



They're not flushed into the box, but it's not a small difference I'm looking for.



I was happy to see the iirc 6.8 ohm series resistor and thought that would work well. I think I need a couple more though. It's not right..




Is there a db meter app for android I can use on my phone, while I play test tones from youtube? I know the phone will be off, but I can benchmark against other speakers. Perhaps finding a simple response peak I can't hack. I don't have anything else to use really. Though I can dig out a condenser mic for my pc if that were helpful.



I did such a clean job of installing the buggers as well.

20180405_213250.jpg

That's the capacitor
 
Last edited:
Are you running the MW144's full range? They do sound pretty good even just running that way by them selves. I definitely think you need some measurements to work this out.

I know my MW144's bear no resemblence to the manufacturer curves. My DMS37 is pretty close to the manufacturers curve down to about 2Khz after which it deviates substantially. I would expect the MDT30 would be similar to manufacturers data.

I have seen speaker measurement apps for android, haven't tried any though. Just looking back at the sims I posted with the IPL crossover, it has quite a rising response, I can't remember if the intention was to put these close to the wall, and hence not worry about baffle step??

Tony.
 
Hello Tony. I hope you're well.

I'm not running the 144s full range no. I can just like what it is they are doing.



My speakers are near enough Ivan's transmission lines. Placement is varied. They like a bit of space and a big room. Listening from 4-5 meters away they're at their best. Right now, they are laid down on top of head height furniture, firing at a wall just 3 meters away. In a very damped room. Kept off the furniture with foam rubber isolators. A bad situation, but the furniture is shelf like, so isn't a baffle extension.



I'm going to check the response curve again. I thought this set of crossovers was like washing my hands of the job and handing it over to someone else. I didn't expect them to be the worst yet, and I've tried a 'LOT' of variations.



Perhaps this thread bump will attract the attention of someone that knows a little more about phone apps.



Now Ive a bi-wire plate, adding resistors is at least a lot easier. If blocking HF getting to the tweeter pushes it through the 144s, I don't imagine that will be a bad thing.



Maybe I just don't like the tweeters. They were streets ahead when I swapped them into my old T-Line TDL speakers, using my A400. Ultimately it was taming the TDL's that made me switch to a Arcam 8r though. The TDLs were a mail order mistake. I expected a pair of 6.5" in a transmission line to dig in, but they just sounded like tweeters. The most tissy speakers I ever owned.



I just noticed your plots are on this page. There is so little between them all, it's like the finishing touch's to your enclosures lacquer, while I'm still at the banging nails in stage.



Maybe in 20 years, when I'm about deaf, I will like them lol
 
I did some new measurements on Sunday, and wanted to have a go at doing a new crossover, at a lower frequency. I'm using vituixcad for the first time and wanted to try the off axis measurements but something is not working out right. I assume that's where you saw the plots.

If I make that crossover and it works out as good as the sim I will be very happy indeed! I've been mostly happy with the current crossovers for the last 6 years or so, but have always thought I could probably do better. The funny thing is that the current crossover involved weeks (or maybe months) of trying things in the sim, the one I just did took me a couple of days, and is heaps better.

I saw your stuff in speaker workshop when I started on my crossovers, and at first I was thinking what is this mdt30 doing in here, then your post popped up and it made sense 😉

Tony.
 
Can you post the schematic for the crossover you finally went with. There were a lot of sims and the mention of the IPL's and I'm curious to see which crossover worked out badly.

I'm also wondering about the baffle step. I don't remember the details (and couldn't divine it from browsing back through the thread, but I wonder whether the problem is the need for more baffle step compensation.

I had a bit of a look, and one of these may fit the bill http://www.daytonaudio.com/index.ph...device-calibrated-measurement-microphone.html They are available here in aus for $29.95 and if used with a decent app should give you a much better idea of what is going on!
 
I found I did have the IPL crossover in speakerworkshop. Not sure what you have tried but my first suggestion would be to try upping the 6.8 ohm resistor to 12 ohms and see how it sounds. I think (once again assuming that the factory traced plots I have bear some resemblence to your drivers) that that would be a whole lot more listenable.

blue is original IPL crossover values, black is with 12 ohms instead of the 6.8 ohm.

Tony.
 

Attachments

  • ipl_comp.png
    ipl_comp.png
    20.3 KB · Views: 128
Thank you Tony. Your a Star. That is just about what my intuition is telling me. I think the crossover point is good, and the tweeter. I just need about 3db taking off. As I did by disconnecting one tweeter and listen to the tonal balance from afar. I have been scratching about and only have a single 6.8ohm w/w. I might get creative with a few 1w carbons and see where that takes me 🙂


I chucked in a tenner today, for all the help you have given me. If you ever need feet on the ground in the UK, I'm here.



I will let you know how it goes
 
When I first started doing crossovers for my MTM, I went out and bought a bunch of 10W non-inductive power resistors in the 0.5ohm to 10 ohm range so that I had plenty of values on hand to make up whatever I thought was necessary to tune things. Maybe you should put the tenner towards that 😉

But do check out that mic I linked to, this Android app looks the business Impulse Response - Apps on Google Play It's big brother is around $11 but has heaps more functionality. doesn't appear to be a demo version for trying out though.

Oh and on the idea of connecting a condensor mic, it is worth a try (use holm impulse for the gentlest learning curve), see this post of mine from way back... http://www.diyaudio.com/forums/mult...orkshop-microphone-invaluable.html#post273443

Tony.
 
Tony and I always look in on Morel driver threads. The woofers have a fascinatingly good natural 12dB rolloff even for small drivers. Big voicecoils and damped cones and dustcaps and about 0.7mH Le inductance, evidently... 😎

They really shouldn't need a lot of crossover to get sounding OK. I thought a simple bass coil ought to do for the woofers, though here we are nearer 3kHz with the 6" Morel unit:

312484d1353176495-help-xover-please-morel_test_crossover-jpg

Need help with a xover please...

Tweeter polarity can be a bit suck it and see, but easily swapped around on Morel tweeters. Below a couple of images. The current crossovers look like about 2.5kHz. I'd aim for about 3.5kHz here.

From what I see, these are closed box drivers needing upwards of 10L box. TBH, I'd give them a listen with just the woofers connected with a 1.5-2mH bass coil. If they sound OK, splicing on a tweeter is just quite simple then, IMO.
 

Attachments

  • Morel MW144 FR.JPG
    Morel MW144 FR.JPG
    70.1 KB · Views: 136
  • friendly1uk crossover.PNG
    friendly1uk crossover.PNG
    6.4 KB · Views: 131
  • MW144 MDT30.PNG
    MW144 MDT30.PNG
    3.8 KB · Views: 139
Hi Steve! Simple provided the mw144's are actually similar to the factory specs 😉 not sure if you remember what mine look like... I did these measurements on Sunday, as I am having a crack at a new crossover for them.

It's looking promising, would you believe I got them (in the sim) to within +- 0.75db from about 160Hz to 13Khz...

In my case I found getting a good phase match essential to getting a really good sound with these drivers, and without some actual measurements in friendly's case I'm not confident of getting it.

Tony.
 

Attachments

  • mw144s_polar.png
    mw144s_polar.png
    83.8 KB · Views: 84
Hello Steve.



Ive gone old school. As I turn the volume up, the fist thing I hear is tiss tiss. It's this percussion noise I dislike. Like a boy racer just went by with too many tweeters. At a distance it's the last thing you hear before I'm too away to hear anything. Symbol like noise. Where you might hear the bristles of a metal brush on a symbols brush. This I believe is around 6-10khz. Where I find the graph shows the speakers are the loudest. So this was to be expected. The 144s is barely doing anything at this point. It's quite certainly the tweeters that are overbearing. With this in mind, I reckon the resistors will sort it out 🙂



I did have a selection or w/w resistors but one by one they went into projects. Ive done the indecent thing and engaged the tone controls for reasons other than more bass, and it's still sharp but then I am listening to internet radio, playing underground tunes mastered in someones bedroom perhaps. I do myself no favours sometimes 🙂
 
Here's hoping a bit of attenuation on the fabric tweeter floats your boat. 😱

It didn't for me with the 1" Morel CAT 298 tweeter. I really worked hard at the acoustic butterworth crossover below, but all I got was spit and sizzle from it. I think it's what soft domes do. Maybe a ring radiator is better, like the SB acoustics or Vifa XT25 models. The theory says that fixing the middle of the dome stops the uncontrollable resonance around 7kHz.

All in the game. Me, I like cone tweeters for their musical sound.
 

Attachments

  • KEF_104AB_Crossover (2).JPG
    KEF_104AB_Crossover (2).JPG
    42.2 KB · Views: 88
  • KEF_Acoustic_Butterworth_Crossover.jpg
    KEF_Acoustic_Butterworth_Crossover.jpg
    82.3 KB · Views: 90
  • Sony SCD-M77 CD Player.jpg
    Sony SCD-M77 CD Player.jpg
    73 KB · Views: 93