New DIY Mid High (90deg) - AKA PM90

[Boyd Williamson]

Peter Morris, how does the RCF MB12N351 compare to the B&C 12NDL76 in the PM90?

At a guess I'd suspect the RCF would do a little better up near the crossover to the BMS, whereas the B&C might have better displacement limited power handling due to higher xmax (I'm not sure if this would be any benefit at 100Hz with a HPF?).

I'd eventually like to upgrade my tops to PM90's. My system is used mostly for DJ orientated gigs, which seem to suit having a little extra in the kick bin frequencies, for better integration with the usually somewhat louder subs. The cost savings of the B&C are tempting, although I'd be reluctant to make significant performance compromises for the sake of cost either.

 

[Andy Turner]

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

Thanks Peter. I wasn't aware of that. Need to look into these details to get things right.

 

[Peter Morris]

Peter Morris, how does the RCF MB12N351 compare to the B&C 12NDL76 in the PM90?

At a guess I'd suspect the RCF would do a little better up near the crossover to the BMS, whereas the B&C might have better displacement limited power handling due to higher xmax (I'm not sure if this would be any benefit at 100Hz with a HPF?).

I'd eventually like to upgrade my tops to PM90's. My system is used mostly for DJ orientated gigs, which seem to suit having a little extra in the kick bin frequencies, for better integration with the usually somewhat louder subs. The cost savings of the B&C are tempting, although I'd be reluctant to make significant performance compromises for the sake of cost either.

Hi Boyd,

Not sure exactly how they compare - I have the RCF in my PM90s and the B&C in the PM60s. The 60 degree box has a slightly longer horn and I have never swapped the drivers to do a direct comparison. All things being equal I SUSPECT that B&C will produce a little more LF and the RCFs will go a little louder and are slightly more articulate.

 

[Guest]

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

Wich is the horn in the last picture ?

 

[Peter Morris]

Wich is the horn in the last picture ?

Its a proprietary design - build in accordance with these patents -

http://www.google.com/patents/US7044265

https://www.google.com.au/patents/US8494815

It is not unfortunately available for sale.

 

[Andy Turner]

Thanks to all of you helping me so far.

I think I have tried almost all possible ways of equalizing this box, but no matter what I do I am really not happy with the result.
I could use some advice.



The enclosed near field measurement of the HF horn I think leaves me with two approaches of dsp'ing.

One being to gain the rather big dip in the 5khz area?? and at 2.2khz'ish area. The other method would be the opposite, lowering the 3Khz, 1Khz and maybe 10Khz.

Both methods sounds bad to my ears. Even when the results of measurement 1M from the horn is within +-2.5db the sound is harsh and canned?, boxy?.



Also to be honest, - I think the BMS and horn combination has one of the most non-linear responses I have ever worked with.
To a point that actually wonder if something could be wrong with the drivers or the passive cross over?

Also, what do you think would be a good starting point for gain differences of 16Ohm BMS and 4Ohm(2x8) RCF?

Your help is very much appreciated, thanks

 

[Peter van Gils]

Hi Andy,

Strange that you have this much trouble in getting the BMS to sound good. Did you check the resistance of the BMS HF and VHF? Maybe it's 8 Ohms and that could explain the rather big dip you have at 5-6 kHz.

No air leak between driver and horn?

Are you sure that your measurement mic is ok? You might be correcting a defective mic...

I have the 8 Ohms BMS and it is running at -8 dB compared to the 12" RCFs. So, I would expect that you would need -8 + 6 = -2 dB on the BMS since you're using the 16 Ohms version (if your amp channels have the same gain!).

I only need -4 dB @ 3-3.5 kHz PEQ on the BMS to get a decent sound and a rather flat response (+/- 3 dB or so)... Hmm.

 

[Peter Morris]

Thanks to all of you helping me so far.

I think I have tried almost all possible ways of equalizing this box, but no matter what I do I am really not happy with the result.
I could use some advice.



The enclosed near field measurement of the HF horn I think leaves me with two approaches of dsp'ing.

One being to gain the rather big dip in the 5khz area?? and at 2.2khz'ish area. The other method would be the opposite, lowering the 3Khz, 1Khz and maybe 10Khz.

Both methods sounds bad to my ears. Even when the results of measurement 1M from the horn is within +-2.5db the sound is harsh and canned?, boxy?.



Also to be honest, - I think the BMS and horn combination has one of the most non-linear responses I have ever worked with.
To a point that actually wonder if something could be wrong with the drivers or the passive cross over?

Also, what do you think would be a good starting point for gain differences of 16Ohm BMS and 4Ohm(2x8) RCF?

Your help is very much appreciated, thanks

Hi Andy,

Have a look at post 315 https://soundforums.net/forum/low-earth-orbit/diy-audio/11601-new-diy-mid-high/page21
and posts 23 (double 10 box) for the passive crossover version of a standard 4594 / HF950 https://soundforums.net/forum/low-earth-orbit/diy-audio/13224-double-10-and-horn/page2

Something is wrong with the results you are getting …

Suggestions:
- Check the crossover is correct – what value are the two capacitors? What impedance is the 4594?
- Check the time alignment between the low section a HF – the HF should be delayed about 1.5ms.

Peter

 

[Andy Turner]

Hi Andy,

Strange that you have this much trouble in getting the BMS to sound good. Did you check the resistance of the BMS HF and VHF? Maybe it's 8 Ohms and that could explain the rather big dip you have at 5-6 kHz.

No air leak between driver and horn?

Are you sure that your measurement mic is ok? You might be correcting a defective mic...

I have the 8 Ohms BMS and it is running at -8 dB compared to the 12" RCFs. So, I would expect that you would need -8 + 6 = -2 dB on the BMS since you're using the 16 Ohms version (if your amp channels have the same gain!).

I only need -4 dB @ 3-3.5 kHz PEQ on the BMS to get a decent sound and a rather flat response (+/- 3 dB or so)... Hmm.

Hi Peter van Gils.

I have checked the DC resistance on my old Fluke77:
VHF: 6.8 Ohm
HF: 8.5 Ohm

I have tightened the screws between driver and horn within reason to begin with. I could do an audible test maybe.

I have also tried a different mic and even though the result varies a bit the overall tendency is the same.

Thank you for the suggested gain. I have settled with +2-3 db on the low mid.

Thanks.
Andy

 

[Max Warasila]

First I think you need to verify that you are getting a reasonable measurement. The best way to get bad results tuning the DIY is to take bad measurements. You might want to try flipping polarities just to see if that's why you're getting a strange null in the crossover region. I'd honestly suggest downloading the Smaart DI demo just to be sure it isn't an issue with REW, as I remember I had issues with it working well for me in the past and getting consistent measurements.

Best of luck!

 

[Andy Turner]

Hi Andy,

Have a look at post 315 https://soundforums.net/forum/low-ea...id-high/page21
and posts 23 (double 10 box) for the passive crossover version of a standard 4594 / HF950 https://soundforums.net/forum/low-ea...and-horn/page2

Something is wrong with the results you are getting …

Suggestions:
- Check the crossover is correct – what value are the two capacitors? What impedance is the 4594?
- Check the time alignment between the low section a HF – the HF should be delayed about 1.5ms.

Peter

Hi Peter Morris.

First I wan't to thank you for your time, and for sharing your knowledge and experience.

I did try the DSP suggested in post 315, - but I was a little reluctant as these are for the 3way all active setup?.

Anyway I tried it on a Powersoft dsp amp after converting Bandwith to Q, but the sound was not right.

I also tried the double 10 settings, which looks very much like the ones I figured out myself, but again the sound is harsh and has quite a bit to much treble for my taste.

I have 1.5 ms delay on the BMS.

One thing I have learned is that a good FR is is never any better than the measurement.
When measuring on axis and adjusting to be within even +/- 2db, the sound is no good.

I also tried 'tailoring' the FR to a d&b Q7 including some PEQ to my taste, - just to copy the 'sound design if you will, - but still no good. I know it can not be done in that way because there are many many other factors, but still, - just to see if it was pure tonal issues that could be fixed in that simple manner. But no.

It is not possible to read the values on the caps without de-soldering. I have attached some very experimental impedance measurements, -maybe they can give an indication. I have also measured the DC to be VHF: 6.8Ohm and HF: 8.5Ohm.

Imp: BMS including passive X-over:



Thanks again.

Andy

 

[Jason Joseph]

Andy if you want to be adventurous i can give you a few settings to try..
Just need to know what amps you are using for each section..

 

[Andy Turner]

Andy if you want to be adventurous i can give you a few settings to try..
Just need to know what amps you are using for each section..

Jason,

Yes please I would love that, thank you

I have two options at the moment:

DBX driverack PA+ in front of a Digam LQ2804

Powersoft K8 and K6 with DSP.

Maybe I should add that it is the HE version of BMS 4594, if that matters.

- Andy

 

[Andy Turner]

First I think you need to verify that you are getting a reasonable measurement. The best way to get bad results tuning the DIY is to take bad measurements. You might want to try flipping polarities just to see if that's why you're getting a strange null in the crossover region. I'd honestly suggest downloading the Smaart DI demo just to be sure it isn't an issue with REW, as I remember I had issues with it working well for me in the past and getting consistent measurements.

Best of luck!

Max.

I have tried the flip polarity on the HF and VHF but it only got worse.

I tried another mic but you are right in that I should of cause also try another measurement app. I did actually try smaart, but didn't have time to learn how to make traces. Good suggestion thank you.

 

[Mark Wilkinson]

Next are rew traces of the HF section, then then VHF section. Sorry I couldn't put them in one panel, but I'm still learning rew.





I thought the curves looked pretty much the same between smaart and rew.
1/12 oct on all, but scales don't match so doesn't mean much.

What I would suggest, is that you measure each section without x-over and see if you get similar traces.
You should, or rew is probably giving you false meas. Are you using the UMIK mic?
All the above was with same mic...non umik, and Not spl calibrated obviously...

Good luck...

 

[Andy Turner]

Mark,

Thank you for the measurements and input.

During the measurements I found out that the calibration file for my second mic(UMC8000) was not correct. After changing it I found that my measurements started looking a bit more like the ones on the BMS site and yours. (Thanks)

My old RTA is maybe broken. Here is the difference between the RTA and UMC(VHF=Green):



UMC mic in Smaart:



With the new measurements in place, I made some quick settings with Smaart and found the sound a lot more natural.

Adding 3-4 db in the 4.8K area, Lowering 3K and 11.5K 3db.

Next step is trying to copy the B&K 'optimum curve' https://www.bksv.com/media/doc/17-197.pdf - reason being that I think the sound with a linear FR is very bright.
- And a new 'better' mic with a correct calibration file :0)

Comments are very welcome.

- Andy

 

[Max Warasila]

Andy, a few thoughts:

I assume you're using the HF950, correct? Reason being is that I normally see a peak around 9k with the XT1464 that resembles the one you got around 12k, so I wanted to be sure it's just a coincidence.

I've tuned the same boxes using both REW and Smaart. The Smaart driven settings were nearly always better.

Lastly, I would personally not suggest tuning to that curve. I think you'll get a much better sounding box overall with a linear response, even if it does sound a tad on the bright side at first. I also believe, though I do not intimately understand why, that you'll get a better phase response in the end this way as well.

 

[Mark Wilkinson]

Andy, my guess is the vast majority of people think a flat freq response sounds too bright...well, at least those that still have their HF hearing left.
There are many studies floating around that tend to agree there is widespread preference for highs that roll off about 1dB per octave at a minimum, starting around 1kHz. And of course, many like bass boost if it stays clean....
A negative slope of 6-10dB, lows to highs, is not so uncommon I think....

I do like to initially tune for flat though, as far as setting up speaker processing.
That simply lets me start with a known benchmark when tuning the speaker to the environment / room.

Then, I normally just use a series of shelving filters at the mixer, both high and low, to get the overall tone I like.
Parametric eq's mainly get reserved for killing room modes, etc...

 

[Andy Turner]

Max, Mark.

Thank you both for your input. I get your point about aiming for a flat FR. It really makes sense, though it is tempting to establish the 'nice to the ear' setting as the baseline.

Max: Yes I am using the HF950, but to be honest I don't really trust the readings in the high range, mainly due to the mic. Thank you for the advice on using Smaart. I agree with you that the result is better so far, - and also faster to work with.

Mark: What subs are you using?

I found another strange thing is that I need to gain +6db on the 12" horn even though they are 8Ohm's in parallel and the BMS driver is 16Oms?

The measurement below is made indoor in my summer house, with the mic 60 cm from center of HF950 with the box sitting 1 meter from the floor and 60 cm to back wall and 2 meters to the side wall.

PEQ's on the HF and VHF (passive XO) are:
4.75k +3db Q=3.42 and
3k -4db Q=4.41
EQ's on a 1/3 Oct. GEQ: 800 -2.5db and 2k -1.5db
+6db on LF(12")




Even though the sound is a little rich in the low mids, and also a little on the dark side lacking a bit of presence and treble, in return is very relaxing and no ear tiring tendencies or harshness. Not at all my prefered tonal balance, so needs some work. I will make some new settings measuring outside using Smaart to try to get it flat without the walls.

Any recommendations for a measurement mic in the 100USD price range?

Thanks.

 

[Mark DeArman]

Did anyone shoot a unprocessed ground-plane measurement of the RCF LF section? I was looking at Andy's Impedance measurement up there a couple posts, and interested to see SPL Frequency Response.