Page 2 of 4 FirstFirst 1234 LastLast
Results 16 to 30 of 54

Thread: M2 diy passive crossover thread ---- from down under

  1. #16
    Senior Member
    Join Date
    Feb 2004
    Location
    Central Coast California
    Posts
    9,042
    That Lizzy's a looker!

    The spiky bulge on the back of her neck looks like one of your speed bumps measurements.
    Out.

  2. #17
    Senior Member
    Join Date
    Apr 2003
    Location
    Australia
    Posts
    7,942
    I have attached below some driver impedance comparisons below. If time permits and Scott and Rob are able to send frequency response measurements of the 2451 and the 0476 driver I will post those also.

    Notes:

    The purple plot of the 2430 on the M2 with parallel voice may not be accurate because my LMS analyser PC was not functioning correctly at the time.
    I will make another impedance measurement of the weekend.

    As can be seen there are three impedance peaks on these curves. What these impedance curves they mean and how they impact on the design of a passive network is complicated.

    A key point is that lower impedance curves will result in a higher sensitivity driver than the high impedance curves.

    Ian
    Attached Images Attached Images  
    Last edited by Ian Mackenzie; 06-17-2022 at 11:29 PM. Reason: Edit of impedance curves

  3. #18
    Senior Member
    Join Date
    Apr 2003
    Location
    Australia
    Posts
    7,942
    Quote Originally Posted by Titanium Dome View Post
    That Lizzy's a looker!

    The spiky bulge on the back of her neck looks like one of your speed bumps measurements.
    Hi Dome,

    We don't do flat around here. Flat is boring. Nothing to listen to and gripe about.

    Leads to dementia...Lol.

    Flat is the death of Hifi because people stop buying stuff.

    Because the Lizzy says its so!

  4. #19
    Senior Member
    Join Date
    Apr 2003
    Location
    Australia
    Posts
    7,942
    Ultra Simplistic Passive M2 Crossover

    There's always room for improvement. This idea was discussed when I was collaborating with Nelson Pass and Greg Timbers for a simple passive CD EQ to be used with active crossovers.

    I have taken the approach further to realise a full working network. Attached are simplified schematics comparing the two high pass passive networks for comparison.

    This is a very simple passive crossover and it has very few vices. Its relatively immune to driver and network impedance interactions.
    Its reasonably inexpensive to make and not a lot to complain about in terms of the response variations. The divisions are 5 db.

    It you are running a valve amplifier or what ever this network has very smooth impedance curve .

    Best of all the diy user can very easily tweak the network to their liking with REW.


    Edit:

    Ultra Simplistic Passive Crossover Description

    The easiest way to describe how this network functions is to look at the raw 2430 frequency response (top curve).

    The first stage to the left is a 4th order LR passive filter which combines with the driver's acoustic slope to form a 6th order high pass slope.

    Next there is a passive voltage divider with dual duties. The divider reduces the overall voltage to the driver and level shift the load impedance up the following passive EQ and the driver to deliver an amplifier friendly load.

    Looking at the raw driver response there are two main areas that require equalisation. A peak at 850 hertz and a broad hump centred at 3,000 hertz.
    There are a couple of approaches to achieving the EQ. The role of the EQ is to pull down the voltage to the driver in the area of interest. The most common method is to shunt a series LCR resonance circuit tuned and defined by the centre frequency, bandwidth and attenuation. At resonance the LCR conducts current which is limited by a resister stopper R. This approach is limited by the minimum acceptable load impedance.

    The second approach is a parallel LCR tank circuit. The tank circuit is connected is series with the load to reduce the voltage to the driver in the area of interest. Like the series LCR circuit the parallel LCR tank resonance circuit is tuned and defined by the centre frequency, bandwidth and attenuation. At resonance the LCR reduce the current conducted which is limited by the stopper resister R. This circuit forms a high series impedance and causes a voltage drop in the area of interest. Two tank circuits are used to flatten the raw driver response.

    The next stage has the task of equalising the fall in the driver response about the mass rolloff point of 4,000 hertz. The most common approach to is to shunt a capacitor across the preceding voltage divider. The disadvantage of this approach is that it loads the impedance down at high frequencies. The alternative approach is to use an LR shunt circuit which at frequencies below its resonance shunts current to ground limited by the stopper resister. This caused a voltage drop in the area of interest across the previous voltage divider limited by the stopper resister R. By empirical adjustment and optimisation the HF Level and Tilt in the response can be accurately set by adjustment of the L values. The R value acts as a level shifter at frequencies below resonance forming a tilt on the response up or down above the crossover frequency.

    The overall circuit is elegant in its simplicity but requires considerable thought in its execution.

    Ian
    Attached Images Attached Images     
    Last edited by Ian Mackenzie; 06-17-2022 at 02:18 PM. Reason: Additional information

  5. #20
    Administrator Robh3606's Avatar
    Join Date
    Apr 2003
    Location
    Rocinante
    Posts
    8,170
    Whoops! The green curve is a 435Be on the M2. That's what happens when you are not careful labeling things as you go!

    Rob
    "I could be arguing in my spare time"

  6. #21
    Senior Member
    Join Date
    Apr 2003
    Location
    Australia
    Posts
    7,942
    Hi Rob,

    No worries. I will edit and remove.

  7. #22
    Senior Member
    Join Date
    Apr 2003
    Location
    Australia
    Posts
    7,942
    I received some files from Scott this morning. Thank you.

    I have had a look at the 2451 measurements.

    I think the best way forward is to put it out there and compare with the 2430 with the 2451 based on the data available. I checked several files and the results are consistent. When Robert is able to send his 0476 files we can do more comparisons then.

    Notes
    This comparison was for a high level look at the response measurements by REW. We don’t want anyone jumping out a window unnecessarily yet…Lol. I understand Scott’s measurements are close or near field. My measurements are at 1 metre.

    There are obviously some differences that need further thought and investigation. The 2430 green curve and the 2452 green curve are attached below.

    You can see the 2451 curve has more of a bump at 3Khertz while the 2430 has a wider and larger hump. The 2430 extends lower with a peaked output at 800 hertz.

    Once l have the data from these curves in Leap 5 with the high pass crossover objective line overlayed l will be able to make more definitive observations.
    Attached Images Attached Images  
    Last edited by Ian Mackenzie; 06-18-2022 at 03:53 AM. Reason: Normalised curves

  8. #23
    Senior Member
    Join Date
    Apr 2003
    Location
    GTA, Ont.
    Posts
    5,109
    Hi Ian ( Rob ),

    Here's sebackman's comparison of the 2430K ( on an M2 ) vs the 2451SL . The 2451sl is the darker top trace.

    Name:  sebackmans'_Comp D2 & 2451SL on M2.jpg
Views: 514
Size:  105.2 KB

    While sebackman's windowing makes things somewhat blurred below 1500hz, the overall response trend is still very visible >> and that is the 4" diaphragm has quite a bit more output below 2K ( which is consistent with the posted experiences of most others ) .

    Here's RobH's 476mg measurement ( raw ) on the M2 horn ( top black trace ).

    Name:  200-20K SPL 476Mg on M2.jpg
Views: 535
Size:  156.3 KB

    The 2451sl and 476mg measurements are more similar than different ( to these eyes ).

    Also, the "Q" of the 3K bump is consistent in all 3 traces ( which makes some sense, since it appears to be an artifact generated by the M2 horn )

    I think there's something is wrong with Scott's measurement ( and/or the fitment of his diaphragms ).
    - He might want to look at the distortion profile of each individual measurement to try to exclude various user-induced problems.

    A rubbing//binding diaphragm usually shows an obvious frequency-dependent spike in the 3rd harmonic trace ( typically getting higher than the 2nd harmonic trace ) while also ( in the worst cases ) hindering the generation of the driver's lower range of frequencies.



    PS Ian; within REW, use the little camera icon to generate custom pics for posting ( if you desire a cleaner way to post pics ).
    - It's found to the top-left of the main display field.

  9. #24
    Administrator Robh3606's Avatar
    Join Date
    Apr 2003
    Location
    Rocinante
    Posts
    8,170
    Hello Earl

    Thanks for posting Rob's comparison measurements between the 2 drivers. I forgot that he posted them. I agree that the posted 2451 measurement looks a bit out of family. Even when testing the 476Mg on the PTH1010 there was quite a bit more low end on a much physically smaller 100x100 waveguide. Same thing with the 2453H so we will have to sort it out.

    My measurements are typically 1/6 octave smoothed but not windowed so you don't loose resolution on the low end of the drivers response curve. I find it difficult to get the driver to driver transitions right if you don't have the resolution. The windowing tends to smooth and shift the curve.

    Odd about that 3K on axis peak??

    Sent Ian the text file for the measurement you posted.

    Rob
    "I could be arguing in my spare time"

  10. #25
    Senior Member srm51555's Avatar
    Join Date
    May 2013
    Location
    Wisconsin
    Posts
    623
    Interesting about the possible distortion. I left the slope filters on in the crown on accident and wonder if this is the issue. Here is a real quick measurement with the filters off compared to the original one I sent Ian. I started the sweep at 450Hz

    Name:  compair.jpg
Views: 556
Size:  88.6 KB

    Thanks,
    Scott

  11. #26
    Administrator Robh3606's Avatar
    Join Date
    Apr 2003
    Location
    Rocinante
    Posts
    8,170
    Hello Scott

    That looks better! Can you take another measurement from 100Hz to extend the roll off into the mud??


    As far as low end on the 4" drivers. Here is a 2453 on a PTH1010 for comparison purposes.

    Rob
    Attached Images Attached Images  
    "I could be arguing in my spare time"

  12. #27
    Senior Member srm51555's Avatar
    Join Date
    May 2013
    Location
    Wisconsin
    Posts
    623
    OK, that’s good to know. I wasn’t really sure where I should be starting and didn’t want to ruin the diaphragm. When I originally aligned these I had the hardest time. Since the drivers were new and a diaphragms were new, I wasn’t sure which one it was but I ended up figuring it out by not using a star pattern for torquing but tightening one side and then the other. I’ll try and post some measurements in about five hours.

    Thanks,
    Scott

  13. #28
    Senior Member
    Join Date
    Apr 2003
    Location
    Australia
    Posts
    7,942
    Hi Earl,

    Thanks for chiming in and posting the comparisons.

    In the absence of readily available comparisons l wasn’t sure what to make of Scott’s measurements.

    Hi Rob, you raise a good point. I have my Windowing open and use 1/6 octave resolution.

    Hi Scott, thank you for sending the measurements. As it turns out we now have an understanding on the data with Earl and Robert’s input. Incidentally l run all my compression driver tests down to 300 hertz and l’ve never had a problem with small level tests. I use 0.07 volts and scale up up the measurement 32 db. If you run your DATS3 impedance tests directly into the driver from 300 hertz with the horn mounted from 300 hertz it will be fine.
    Great link for scaling your measurements
    http://www.sengpielaudio.com/calculator-gainloss.htm

    Rob S is away from the forums at the moment and has kindly given me access to his measurements.

    I have a pair of 2451 driver coming towards the end of June.

    It’s early am here. I’m out paddle boarding on the Maroochy River. I will have a look at the files this afternoon. I’m fairly relaxed about this project. Obtaining consistency with measurements of drivers remotely is always a challenge.

    Talk soon.

  14. #29
    Senior Member srm51555's Avatar
    Join Date
    May 2013
    Location
    Wisconsin
    Posts
    623
    Thanks Ian, unfortunately I only have the capability of running frequency sweeps in REW at this time. Luckily we have the impedance from Rob S which would be more reliable anyway. I still haven’t had time to figure out impedance measurements in REW.

    Enjoy the water!

    Thanks,
    Scott

  15. #30
    Administrator Mr. Widget's Avatar
    Join Date
    Apr 2003
    Location
    San Francisco
    Posts
    9,719
    I have nothing to add, but it is nice to see all of this DIY action in the forum!


    Widget

Thread Information

Users Browsing this Thread

There are currently 1 users browsing this thread. (0 members and 1 guests)

Similar Threads

  1. Passive Monitor Thread
    By Robh3606 in forum Lansing Product DIY Forum
    Replies: 165
    Last Post: 03-13-2021, 11:55 AM
  2. passive crossover JBL 077
    By nfbertone in forum Lansing Product Technical Help
    Replies: 13
    Last Post: 08-08-2012, 02:06 AM
  3. Need help with passive crossover!!
    By rbrandt73 in forum Lansing Product DIY Forum
    Replies: 3
    Last Post: 07-05-2007, 11:43 AM
  4. Passive crossover JBL
    By alnico in forum Lansing Product General Information
    Replies: 4
    Last Post: 10-06-2004, 08:44 PM

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •