New DIY Mid High (90deg) - AKA PM90

Andy,

I think you get reflections from the ground during the measurement that may cause to (partially) cancel out some frequencies in the 3-400 Hz area. Which you are probably adding (or at least not removing) with the peqs.
I just put the speakers on a stick, measured them at different heights and angles at 2-3m distance. In REW I applied a time window of a couple of ms (I think it was 5-10ms) on the impulse response to get rid of parf of the reflections. I ended up with only 2 peq: 3-400 Hz -4dB and 3-3.5 kHz -4dB and it sounds quite good.
I also see that you have a peak in the HF. Did you use the passive filter? I did not have this peak with the passive filter.
My strategy at first is to only remove things that should not be there (resonances). First convince yourself by multiple measurements (different heights/distances) what should not be there and what is really missing. You can easily misjudge a single measurement, is what I learned.


Verzonden vanaf mijn iPhone met Tapatalk
 
Last edited:
Andy, +1 re peter's comments about ground reflections with the box laying on the ground.
I recently tried testing a DIY60 that way, that has a very flat response in air. Laying on the ground, it measured really whacked down low. I took it to prove why we're supposed to dig a hole in the ground for the box to fit flush in...
 
Peter, Mark and Mark, Thank you.

Peter, - I was hoping to get away with just a few PEQ's.
Could I be making things worse adding more PEQ's making the response having more sharp(hi-Q) peaks and dips? I hear different opinions on this subject. - The audibility of resonances?

Yes I am using the passive filter from BMS.
The peak at 10Khz was next on my list :0) I need to investigate if this could be due to a faulty mike?

According to BMS curves, the VHF response with passive filter is a bit high though.

Maybe the VHF needs a few db attenuation? L-PAD before the VHF driver maybe?

The drop at about 5khz, could that be due to problems with phase at HF and VHF XO ?
Should I try to reverse polarity on the VHF driver? I just followed the +/- printed on the filter to the drivers.

I will do both ground and also near field measurements on the low for reference - thanks, but I was hoping to have that one great measurement to get timing/phase alignment and for getting the sound balance between low and high at least 'in the ballpark'.

More measurements. Just waiting for a pause in the rain :0)

 
I get it that you minimize the reflections by measuring ground plane but I'm not understanding what these measurements can be used for? At low frequencies I get weird response partially due to the D'appolito arrangement? The High frequency is off- axis and also suffer from timing issues at XO ? I am bit lost here. Need to study more :0)

Enclosed measurement of speaker standing on the ground and mike 1 meter distance on the ground. (IR window 10ms and 1/24 smoothing. XO at 700/48db LP 100/24.
DIYMH90_ground_mike_ground 1meter.jpg
Is this the best baseline for adding PEQ and so?

I made PEQ's based on measurments made with the speaker laying flat on the back and mike 1m above, and was able to make it relatively flat (+12 db), - but the sound was really no good for me :0/.
 
Andy, one thing we both found out...a speaker on the ground with the mic above it doesn't work....unless the speaker is in a hole in the ground, and then supposedly it works. And I ain't gonna dig no hole !
I tried speaker on ground, mic in air, to try to verify some REW harmonic distortion measurements at high volume. I was rattling my house and thought getting away from the house would help. Frankly, the REW traces looked very alien. I wish I had tried the mic on ground.......

......as Peter and Mark both say, box on ground, mic on ground, works. I've always figured floor bounce would screw up middle thru higher freqs. But I guess the on-ground measurement is more about measuring the low end.
Maybe 1m on the ground, with the box standing up is too close for the two mids, making too much distance variation to them ? Why not lay it on its side so you can center the mic? Or back the mic up? Or tilt the upright box to aim to mic?

I have a pretty good setup for measuring in-air, reflection wise. What I've found comparing inside measurements to outside, is that with the DIY on a sub or something, in the middle of a fair sized room, and as far away from everything as possible...with the mic centered on the horn about 2 ft away...I get results that tie pretty OK. Magnitude and timing will both be off some, but not that bad really for straight on axial.

Also found it was much easier to divide and conquer, ie tune the bms first by itself, then the mid, and then try to put them together...
(It's real easy to toy with the bms alone, inside without reflection probs.)

I'm left thinking in-air works best above some given freq. On-ground helps confirm, or adjust in-air meas below said freq.
Maybe the guys can confirm or kill this idea....
 
Andy Turner said:
made PEQ's based on measurments made with the speaker laying flat on the back and mike 1m above, and was able to make it relatively flat (+12 db), - but the sound was really no good for me :0/.
Click to expand...

Andy, one thing I have noticed about these boxes is that there is a fair amount energy that emits from the rear of the box especially at higher SPLs. This is most likely contributing to the reflections you're seeing with it on its back.

My best measurements so far have been achieved by building a cradle for the box to sit in on its back facing up that can be used on the crank stand getting it about 4 meters high. The mic is on a pole positioned about 2 meters straight above the center of the box pointing down. I try not to turn the pink noise up any higher than necessary to help minimize ground reflections.
 
Mark,
I continued my work today by making a close up measurement of the box standing upright on the ground with the mike on the ground and a few centimeters from the box and used this for tuning the low mid only.
I think you are right that 1 meter is too close to get good results from the low mid's. I did actually try the ground-on the side, but wasn't sure if the result was any good due to the coupling to the ground on both low mid's.

For tuning the top temporarily I then turned to Peter Van Gils method of applying only two PEQ's in total. I added a PEQ at 3.5khz -3db for the BMS and there you go, - sweet sound..
Of cause it could use a lot more tuning and I will still pursue a much flatter response, but it was a very nice experience to finally get an idea of what this speaker is really about.

Next is trying your method of close up measurement on the bms.
 
Don J Davis said:
Andy, one thing I have noticed about these boxes is that there is a fair amount energy that emits from the rear of the box especially at higher SPLs. This is most likely contributing to the reflections you're seeing with it on its back.

My best measurements so far have been achieved by building a cradle for the box to sit in on its back facing up that can be used on the crank stand getting it about 4 meters high. The mic is on a pole positioned about 2 meters straight above the center of the box pointing down. I try not to turn the pink noise up any higher than necessary to help minimize ground reflections.
Click to expand...

Don, Thank you for sharing. I was thinking about using the method that I have seen Peter Morris using, just putting them 4 meters in the air.
 
Last edited:
Peter van Gils said:
Peter, the DIYs in action: I did sound for a local youth band yesterday. A quick iPhone shot:
https://youtu.be/3Tqt-CjbW0o


Verzonden vanaf mijn iPhone met Tapatalk
Click to expand...


Fantastic ... it’s really great to see so many people around the world getting such great results with these boxes.

Even though it’s a youth band recorded on an iPhone in a reverberant space you can still hear how clear and effortless it was :-)

 
Peter Morris said:
Fantastic ... it’s really great to see so many people around the world getting such great results with these boxes.

Even though it’s a youth band recorded on an iPhone in a reverberant space you can still hear how clear and effortless it was :-)
Click to expand...
Yes, was one of the first songs and they were not completely warmed up yet ;-)
During the video I had 15-20dB headroom. At the peak there were about 600 people in venue and I still had approx. 10 dB headroom.

It's great to see/hear/feel the reaction of the crowd when you give a few dB extra during a climax of the vocals and the sound stays crystal clear. :-)


Verzonden vanaf mijn iPhone met Tapatalk
 
Peter Morris,

I've been looking through drivers upon drivers upon drivers as of late, and I was wondering why you chose the 12NDL76 over the 12NDL88. Just looking to understand your thought process a little more.

Thanks!
 
Max Warasila said:
Peter Morris,

I've been looking through drivers upon drivers upon drivers as of late, and I was wondering why you chose the 12NDL76 over the 12NDL88. Just looking to understand your thought process a little more.

Thanks!
Click to expand...

The critical parameters for a driver in a horn are the moving mass and the strength of the motor.

The 12NDL88 Mms is a bit high compared to the RCF MB12N351, which is more efficient and provides a smoother response in this horn; but it does not have as much Xmax.

The 12NDL76 (when you model it) has a slightly smoother response compared to the 12NDL88, is cheaper (much cheaper than the RCF ) and lighter ... and I had 4 :-) ... part of the reason was to test a cheaper option - 12NDL76 with a standard BMS4594

There are a bunch of other drivers including the 88 that I would love to try in the DIY but my $$$ were limited.

FWIW – B&C 12/14/15NDL88 range looks fantastic for a lot of applications - a great compromise between efficiency, Xmax, SPL, weight, cost etc.

FWIW Here is a new design ... it use 2 x B&C 14NDL88s
 

Attachments

  • double 14.jpg
    double 14.jpg
    313.8 KB · Views: 79
Last edited:
Peter Morris said:
The critical parameters for a driver in a horn are the moving mass and the strength of the motor.

The 12NDL88 Mms is a bit high compared to the RCF MB12N351, which is more efficient and provides a smoother response in this horn; but it does not have as much Xmax.

The 12NDL76 (when you model it) has a slightly smoother response compared to the 12NDL88, is cheaper (much cheaper than the RCF ) and lighter ... and I had 4 :-) ... part of the reason was to test a cheaper option - 12NDL76 with a standard BMS4594

There are a bunch of other drivers including the 88 that I would love to try in the DIY but my $$$ were limited.

FWIW – B&C 12/14/15NDL88 range looks fantastic for a lot of applications - a great compromise between efficiency, Xmax, SPL, weight, cost etc.

FWIW Here is a new design ... it use 2 x B&C 14NDL88s
Click to expand...

I played around in HornResp this morning, and I see what you mean. The NDL88 is far less efficient, and the only point to it I can see is a little more breathing room in the lower extremities of the DIY.
 
Peter van Gils said:
After some tweaking I got the response as depicted below. This was measured using REW in a room, so full of reflections. I had to apply IR windows which affect the low freq data somewhat. Quite happy with the phase response now (note: with IIR non-lineair phase filters, DCX2496). Settings:
MID: 100 Hz BES12 HP, EQ -4dB Q=2 352Hz and 676 Hz BES24 LP.
HF: 676 Hz BES24 HP, EQ -4dB Q=2 3.5kHz and 6.5 kHz LR12 LP (BMS passive C8-8 filter)
VHF: 6.5kHz LR12 HP, inverted (BMS passive C8-8 filter).

So, you would expect 90+360+180 degrees phase shift due to the filters. At 100 Hz there is an additional 90 degrees phase shift due to the BR ports. So, this is about the best phase response you can get out of the DIY with IIR, I think. (I don't want to take the risk of damaging the BMS by going to a 12dB/oct x-over instead of 24dB/oct...)

For further EQ tweaking I will need to get the set in open air (or a very large room).


Click to expand...

Peter Van Gils,
I've been working (hard) to get this project to sound good, but with no luck so far.
My measurements keep showing a drop in the 4-6khz range so I thought I would try inverting the VHF as you did. I am using the passive filter(16/16) from BMS. I tried but as can be seen in the measurement below, it doesn't work for me.

Any idea what is going on?

The measurements are indoor on a sub 1.53m from floor and mic at 60cm distance pointing directly into HF950 horn. 10ms window and 1/12 smoothing.

Inverted in red:
DIYMH90_indoor_VHF_inverted.jpg

Phase:
DIYMH90_Passive_XO_VHF_Inverted.jpg
 
Last edited:
Hi Andy,

I think you misunderstood what a meant with inverting: I mean all bands, so low, mid, hf and vhf. If you connect the vhf as indicated by the colors on the filter, then it should be ok. (But when you do this the vhf is in fact inverted compared to the hf, which is ok.)

Next, when measuring with REW you probably have to invert all xover channels, so REW can synchronize on the vhf. (The vhf will then be in-phase.) This should give you the cleanest phase plot (flat in the vhf and increasing phase for lower frequencies).


Verzonden vanaf mijn iPhone met Tapatalk
 
Top Bottom