Advice needed on 4 Way loudspeaker

[andy19191]

I bought these used from a third person. I had communicated with the designer and he was helpful enough to pass the plots to me.

Does this mean the plots shown earlier were not of your speaker but of some other speaker built to the design that was supposed to be followed in your speaker?

 

[diyaudnut]

Does this mean the plots shown earlier were not of your speaker but of some other speaker built to the design that was supposed to be followed in your speaker?

Thats right Andy.
As i said i will run some measurements based on the discussions here to help see whats going on

 

[diyaudnut]

Good evening Jojip, I just bumped into this thread (as I am not very active on diyaudio). I believe these speakers are from Selah Audio (Rick Craig). In all my interactions with Rick, I have found him to a reasonable gent, and I would address him as he might have the best answers to some of your challenges.

After all his has pictured your speaker (towards the bottom of his DIY page)
DIY ? Selah Audio
http://static1.squarespace.com/stat...d/1406777673664/RC4CurvedBack.jpg?format=500w

Now back to some of the challenges.

1. As far as the tweeter is concerned I would consider the NeoX tweeters from Fountek as they share the same mounting plate and can be crossed low and do not need high order XOs (they have lower distortion than the NeoCD series).
Products_Fountek Electronics Co.,Ltd
Products_Fountek Electronics Co.,Ltd

I say this because I too have built similar speakers (see link below) and also rear speakers using the NeoCD3.0 tweeter with the Seas W16NX woofer.
see link
Pictures of Selah designed Tempesta Extreme and RC4 (no subwoofer)
The Yung SD500 amps died and I have since replaced them with Kiega KG ND-52100-SP amplifiers. Another choice could the Hypex. I use a Yamaha S2000 for the main speakers (8" 3 way).

2. That tube behind the mid might be causing some issues but I cannot see these be as severe as you describe. Did you check the polarity of the tweeters?

3. Before getting the Selah audio I auditioned many speakers like the B&W CM8, KEF R700, Focal 836W, Paradigm Studio 100, Monitor Audio GX300, etc... and none of them satisfied. So I went back to the drawing board and looked at speakers that had satisfied me in the past - Sonus Faber Cremona, Vandersteen 4A, Martin Logan Sequel, Klipsch's Belle Klipsch and Klipschorn combo, Wilson Audio Sophia, Tannoy DC10T ... and I could not afford any, not even used (Audiogon). Speaking of Audiogon a seller called Solstice (or something like that) had a pair of these on sale some moons ago - cant remember when.

I am not putting Selah's creations in the same league as the later list, but they are at least equal to the former list, and they get there at roughly half the price (even less in my case as I built the cabinets locally hence did not pay freight, duties and taxes on the cabinets). This is the thread I had some years ago http://www.diyaudio.com/forums/multi-way/208311-selah-audio-ex-diyer.html

So 2 pointers:

Ribbons by nature are more directional on the vertical axis but I find that when seated this is not a huge issue. The horizontal axis is more critical.

The hard cones (Mg) of the SEAS drivers are very good within their pass band and terrible outside it. Bob Richards has already said this. If the crossover posted by Dissi is correct then the 15cm mid might be your culprit. It's electrical slope is too high for a hard cone (and even combined with the roll of the of the midrange itself) and maybe a lower crossover point combined with a tweeter that can go low enough may be a place to start. Troel's small monitor uses this woofer W15CY001-OWI

Make 1 change at a time so you can tell the effects of the change.

Best of luck.

Thanks Nevin. I did come across you excellent build thread when considering these speakers , was one of the factors which convinced me to go ahead with the deal.
I did audition these at the sellers place. Didnt sound as great as i was expecting, but wasnt willing to pass on the great drivers and wood work

 

[diyaudnut]

......
3. Throw out existing XOs and go full active. Cost of this tuning , additional amp channels, DSP XO etc sound daunting at this time. Ruling this one out.

Hi Jojip,

I feel for you Jo, you are so close yet so far and trying to make sense of all the well meaning advice on here must have your head spinning....Remote re-design via a group of DIY dudes is, er challenging!

I would urge you to take a day off from all this complex passive crossover / new driver integration / speaker re-design and go and have a listen to a high end Yamaha AV Amp / processor playing some music via good loudspeakers.....
If you haven't heard a recent (2013 onwards) Yamaha work its magic you will be gobsmacked.

If you have the money ($1,500 ish) here is a great solution....

You can buy a fantastic used / ex demo Yamaha for around $500, these have the same DSP as the $5,000 top end Yamaha models.
This amazing software needs no skills or knowledge to use, just plug ' n measure with the mic provided and the software guides you through the simple process.
It will integrate and balance all your drivers and subs beautifully and Eq them to perfection with your room acoustics.

You will need to budget for approx $300 for a Mini DSP active crossover plus power amps at $200 per channel if you buy used / clearance.

The end results will be far superior to any passive crossover solution regardless of how much time and $ you spend.

Hope this helps and all the best
Derek.

Thanks Derek.
Its not lost on me that active is a great option, considering the tweaking and tuning that would be needed in a passive design to get the best out of these great but complex drivers.
Are you referring to the professional DSP processors listed on the Yamaha AV site. Will lookout for used ones.

 

[diyaudnut]

W15CY001 E0015 SPL/freq to discuss steep ~2Khz Xover ideas.

L-R-C ideas to remove metal cone breakup.

LineSource, your location shows as silicon valley. I am from there too.
Would you be up for auditioning this to get an experts opinion of whats going on, if things are close by and convenient ?

 

[diyaudnut]

I am using Jeff Bagby's set of simulation spreadsheets.
These are capable of simulating cabinet volumes and their
FR responses in bass region ( Woofer box and Circuit designer),
then there is PCD for simulating XO filters and Response Modeler
to manipulate response curves by adding baffle step response
and merging curves.

In its nature these crunching numbers software are accurate if the
input data is accurate as well. Input data in my case were manufacturer's
measurements of FR and impedance which I obtained by using SPL Tools,
a program that turns pictured graphs into textual files ( frd and zma)
which can be loaded to any simulator.

With additional data on acoustic centres of the drivers in the baffle and
their position, it can be observed how the performance changes over
different angles ( listening position).

My simulation is only accurate sufficiently to have a taste what one might
expect. For better accuracy, measurements of the real drivers in the real
box are mandatory.

Squeeze the last drops of juice these units can give you before you decide
to open your wallet again.

Personally, I would modify this speaker of yours to be a 3 way box with 22EX's
in series filtered passively at 300 Hz and somewhere around 3,5-4 kHz.

The inner volume for the woofers should be in the range of 100 liters vented
to properly play bass notes.

There is always the learning curve whatever you do. Maybe you will consider
to get rid of the speakers and try an alternate route.

Thanks. Yes i may have to let these go, but after some more investigation.
I am starting out on audio diy and this appears to be a great opportunity if the time, money and end results are reasonable. I do understand the complexity of getting a full passive design from scratch right the first time, considering that the original designer is very experienced.

But there are incremental improvements i am prepared for. For example changing the tweeter and just redoing the mid/high XO and leaving the woofer as is after understanding the exiting XOs performance.
Also cleaning up the baffle mounting to the extent possible.

To decide, i need to get some more information across to you guys here. Will do that as soon as i find some more time.

 

[diyaudnut]

That would be unusual for a pair of 8 ohm woofers. If the crossover was designed for the woofers to be in parallel then an incorrect series connection would give about 6dB less woofer output. Might be worth checking.

Andy, yes i was surprised too on connecting 2 8ohms in series. But that's what seems to be done. Will check with the designer if that was his original intent.
Though he understandably hasn't been willing to share details of his proprietary design.

 

[Lojzek]

However you look at it, connectiong 2 of the W22's in series
is optimal solution. Midrange alone can output some 85 dB
in a bandpass XO filter type, one W22 with baffle step can play
some 85 dB too. If you have two of them in parallel, you get
a huge peak in spl which is extremely high without impedance
correction. A series connection with RCL to compensate for
impedance peak brings woofers exactly where they should
be, in level with midrange. A correct value of the LP inductor
for the woofers would be around 12 mH for a 300 Hz XO point.

 

[5th element]

However you look at it, connectiong 2 of the W22's in series
is optimal solution. Midrange alone can output some 85 dB
in a bandpass XO filter type, one W22 with baffle step can play
some 85 dB too. If you have two of them in parallel, you get
a huge peak in spl which is extremely high without impedance
correction. A series connection with RCL to compensate for
impedance peak brings woofers exactly where they should
be, in level with midrange. A correct value of the LP inductor
for the woofers would be around 12 mH for a 300 Hz XO point.

This is incorrect the W15 is 86dB nominal, when you put it into a bandpass you actually get some gain out of this increasing its effective SPL up by 1-2dB. How much gain you get is governed by the xover points to the drivers either side of it. You can see an example of that here.

Given suitable xover points you could get the W15 up to 88dB nominal. This is of course if you cross it over so that none of the baffle step losses area incurred by the mid range.

The W22EX001 has a nominal half space sensitivity of 88dB too. Wire two in parallel and you'll get 94dB sensitivity. For a large loudspeaker like this, meant to be run free from any side walls, I'd want to dial in 5dB of BSC. This brings the W22s sensitivity back to 89dB, add in insertion losses from the series pass inductor and we're going to be back down to around 88dB.

Even if the W22s ended up yielding a lil bit more than this due to the bottom woofer being mounted close-ish to the floor you'd only end up with a slightly downward tilted balance, ie a bit more bass and most people tend to prefer this kind of curve.

Selah Audio likes using ribbons and has built many a loudspeaker with hard cone midranges that leave me scratching my head. The W15s mandate a low xover point if you wish to use them properly, many ignore this at their peril.

 

[Lojzek]

It is correct in the way I simulated it and I used the manufacturer's plots.

 

[andy19191]

The W22EX001 has a nominal half space sensitivity of 88dB too. Wire two in parallel and you'll get 94dB sensitivity. For a large loudspeaker like this, meant to be run free from any side walls, I'd want to dial in 5dB of BSC.

If the woofers are wired in series in a speaker with a specified sensitivity of 86 dB how much BSC is it likely to have?

 

[Overkill Audio]

Yamaha domestic AV receivers

Hi Jo,

You can use the domestic AV receivers, they are much cheaper than the Pro sound versions.
Not sure about the exact models in the USA, but here are a couple of typical UK models.
You can find great deals on end of line / clearance models.....

Around $750 will get you a beast of a product with a massive power supply to feed 7 channels of good power amplification (plus 9 channels of Pre outs to upgrade later with dedicated power amps).
7 channels of ESS DAC's, a stereo A to D, every type of connectivity / WiFi / USB / Android / Apple / NAS sources / phono / line level etc plus analog pre-amp and a full suite of top drawer video processing / up-scaler....All with full remote control from your phone or tablet / lap top to match JRiver or whatever media source you use....Plus the killer DSP!!!

RX-A740 - AV Receivers/Amplifiers - Yamaha - UK and Ireland

RX-A1040 - AVENTAGE - Yamaha - UK and Ireland

One of these plus a mini DSP and a pair of budget power amps and you have a killer system that will absolutely maximise the performance of any speakers / drivers / subs / room you have now or in the future and give you plenty of upgrade options.

All the best
Derek.

 

[Juhazi]

I just can't see how these AV amps could help Jo. This is off-topic.

Older AV amps did have 6ch analog inputs - they could be used as amps for multiway active stereo speakers with minidsp or similar doing the xo job.

 

[5th element]

It is correct in the way I simulated it and I used the manufacturer's plots.

Then your simulation is wrong. For a start it doesn't include the woofer or tweeter responses, which is where bandpass gain comes from and is necessary to see it in action, but SEAS responses include a measured bafflestep, which is why the response falls away and is not indicative of this driver in the threads cabinet. SEAS measurement assumes a narrower box pushing bafflestep losses higher into the w15s passband and knocking some sensitivity off the top too.

 

[DrDyna]

Older AV amps did have 6ch analog inputs - they could be used as amps for multiway active stereo speakers with minidsp or similar doing the xo job.

I've done this, it works remarkably well, it's like a 6 channel integrated amplifier with individual level controls, and many can be had for very little money.

Here's the one I used briefly for this purpose.

Edit: And it doesn't complain one bit on 4 ohm loads, either.

 

[sreten]

Hi,

If things fall apart its generally due to distortion issues not response.
RLCing a drivers peak does not fix the distortion peaks it causes.
I agree with 5E that the mid and treble are unworkable.

rgds, sreten.

 

[Greebster]

I've done this, it works remarkably well, it's like a 6 channel integrated amplifier with individual level controls, and many can be had for very little money.

Here's the one I used briefly for this purpose.

Edit: And it doesn't complain one bit on 4 ohm loads, either.

Just returned a Onkyo TX-DX787(b) that I was asked to checkout. Owner used this 7.1 av reciever some twenty years ago, replaced by krell monoblocks. Owner wanted me to buy for $150 which happens to be what I am shy in funds for my custom power supplies to drive my own amps. Informed him that the reciever using all 8 channels would only do ~35w vs the 100w rating into 6ohms "stereo" power rating. Power supply limited. Other than that it was in mint condition nary a speck of dust, in the original box, w/ manual and remote.

He's going to list it on ebay.

Also has a pair of Opus 3 surround speakers he'd like to ditch. Only minor cracking of the piano finish on the backs. Other than that mint condition, packed in the original socks in original box.

Sorry /ot

 

[DrDyna]

Informed him that the reciever using all 8 channels would only do ~35w vs the 100w rating into 6ohms "stereo" power rating. Power supply limited. Other than that it was in mint condition nary a speck of dust, in the original box, w/ manual and remote.

Yeah, that's true of most I spose. I didn't worry about it too much, though..driving actives means at least 2 channels are high frequency drivers, which can't take crap for power anyway, unless you're driving some 2" compression mids or something

 

[diyaudnut]

Hi Jo,

You can use the domestic AV receivers, they are much cheaper than the Pro sound versions.
Not sure about the exact models in the USA, but here are a couple of typical UK models.
You can find great deals on end of line / clearance models.....

Around $750 will get you a beast of a product with a massive power supply to feed 7 channels of good power amplification (plus 9 channels of Pre outs to upgrade later with dedicated power amps).
7 channels of ESS DAC's, a stereo A to D, every type of connectivity / WiFi / USB / Android / Apple / NAS sources / phono / line level etc plus analog pre-amp and a full suite of top drawer video processing / up-scaler....All with full remote control from your phone or tablet / lap top to match JRiver or whatever media source you use....Plus the killer DSP!!!

RX-A740 - AV Receivers/Amplifiers - Yamaha - UK and Ireland

RX-A1040 - AVENTAGE - Yamaha - UK and Ireland

One of these plus a mini DSP and a pair of budget power amps and you have a killer system that will absolutely maximise the performance of any speakers / drivers / subs / room you have now or in the future and give you plenty of upgrade options.

All the best
Derek.

Hi Derek.

I am a bit confused here. So I will still have to use a mini DSP for the XO functions and use the Yamaha as a multichannel DAC, preamp and power amp? Does it have at least six channels for digital inputs which map to the amplifier channels?

thanks
Joji

 

[diyaudnut]

Here are current options and steps as they stand.

I owe these two to folks here to truly understand what’s going on and precedes undertaking any real passive XO rework
1. Measure in room responses of the current speaker/drivers
2. Try to unmounts crossover boards and reverse engineer. I remember that most component values were not legible

Folks, please do advice if I need to make impedance measurements as well? Don’t currently have the jig for that.

Passive 1:
1. Change from ribbon to seas dome tweeter
2. Redesign the high/mid XO
3. Leave woofer design as is

Passive 2:
1. Change from ribbon to seas dome tweeter
2. Redesign the high/mid XO
3. redesign the woofer XO

I am an electronics engineer by training and profession, so do understand most of the theory, but have 0 practical experience designing XO, let alone for such a complex 3 way + sub. (btw, leaving the subs alone).
Further don’t really have a library of these expensive passive components and measurement meters to tweak and tune.

Active:
1. Get rid of both passive XO
2. Procure MiniDSP for high/mid/woofer XO
3. Procure Multichannel DAC
4. Procure Multichannel preamp/power amp
5. Return expensive stereo amp
6. Return expensive stereo DAC preamp

The active option also involves lots of measurement and tuning.

I am between a rock and a hard place and either option is going to be a slow development due to limited spare time.
Option 3 is to get rid of the speakers, not sure if I can find a buyer at my original cost.
Then again the gorgeous wood work and drivers are too hard to part with