Rick,
Here is more accurate internal volume
Sub 14.5 16 9 1.2 cuft
woof 21 16 9 1.75 cuft
mid 6 16 9 0.5 cuft
Sub seems to match you number
I will adjust the other two volumes to your numbers
Here is more accurate internal volume
Sub 14.5 16 9 1.2 cuft
woof 21 16 9 1.75 cuft
mid 6 16 9 0.5 cuft
Sub seems to match you number
I will adjust the other two volumes to your numbers
Sounds good. Please provide these, i will try out the cheaper values and then replace with better ones.
As I pointed out earlier the peak at 2.8K is about -40db from the reference level (provided that it's purely due to breakup and not a diffraction from the baffle or the tunnel loading from the subenclosure). If you look there's more than one well-respected softer cone driver of similar size that have distortion at that level or even higher at frequencies much lower than 2.8K, yet they are often used with crossover points in the range that I used. The ribbon's 3rd harmonic is -40db or lower over the range I used, and I'm not the only one to use ribbons of this size and stiff cone drivers to have good feedback on the sound.
Joji,
Those volumes are much better than what you originally estimated. I modeled them without the crossover and compared them to the original design volumes. Some slight differences in the response, but not enough to explain your full-range measurement or the midrange nearfield in comparison to my graphs.
Rick, So what could be the source of the problem then?
Are there any other investigations that you want me to run?
Now i am getting worried if the drivers have been tampered with.
Joji
Very unlikely the drivers were tampered with as that would be pretty obvious to see. The best way to troubleshoot problems is to use the process of elimination. Are the inputs single or bi-wired?
There are two binding posts for biwiring. But i am using a single wiring and the two posts are bridged with some wire jumpers which were left there by the original owner.
Also during the measurements i used the posts individually for woofer and mid/high measurements. For separate mid and high measurements i disconnected the appropriate driver manually after opening the speaker.
Also during my dismantle the XO boards did seem to be connected correctly to the binding posts
Could a Cap have failed. Caps when they fail can short. I know I am just shooting in the dark but just trying to cover all bases however mad they may be. BTW Solen caps have failed on me before (and yes both channels in one speaker pair) and I got them direct from Dennis in the old days.
Jojip, how handy are you with a soldering iron? Might even be a cold solder but then why would both speakers measure the same.
Last edited:
I'm guessing that the great majority of people on this forum are designing their own speakers and sourcing them from the distributors in their area. If I'm correct, then this thread has little value to increasing my personal business.
What is your interest in seeing how this turns out? Is it simply to disagree with anyone who has an opinion that is similar to mine? How would you know what my motives are? I help people quite often when no financial gain is involved. If there was more time I would do it here but that's just not possible. This one thread is a good example.
Yes, posting in an open forum has its' risks including dealing with people who always have to have the last word, think they're the smartest person in the room, etc. Unfortunately there are some DIY guys who feel that taking potshots at those like myself in the audio business is mandatory when we post in a DIY forum. I have no problem with being accountable and if someone is obviously posting just to make a sales pitch then they should be called out. But I've also seen good people (and extremely knowledgeable ones) leave or avoid forums because of all of the junk they have to endure from a few idiots.
For you to insinuate that this isn't a "wise design" perhaps I should further explain to you the history behind how it came about as well as the other details involved. Actually, that would really be more important to the owner or anyone here that's truly interested. Now do you think that if it's as bad as you suggest the owner would've purchased the pair after listening to them?
Last edited:
Ok that all sounds fine. I would next check the polarity of the W15 / W22 crossover point. Basically overlay two curves, reversing the polarity on the second one. 1.5M distance with the microphone on axis with the tweeter. 1/12 octave smoothing too.
Could be the caps. Wish i had decent gear to quickly measure the components out and annotate on the crossover schematic.
I am fairly handy with the soldering Iron. from my inspection, i hadn't noticed any bad joints. Will check again.
thanks
Joji
Last edited:
Not following this at all. The third harmonic distortion @2700 is generated from 900Hz, not 8000. To suppress this peak limiting it's range to less than 900 would be in order.
Navin mentioned the caps - very low probability but easy to test. The Arta jig would allow you to measure the impedance of the speaker which would show if there's a issue somewhere. That's the best place to start before trying to measure at the crossover component level.
Then that soft cone is a well respected mediocre driver, nothing wrong with that, there are lots of well behaved soft coned drivers with basic motors that people enjoy.
The point here is that in the case of the W15 you've paid all that money for the advanced motor design and this gives you very low distortion, typically better than -55dB for normal listening levels. The large peak, causing the third order distortion to rise by an order of magnitude, is an anomaly and wide discontinuity in the drivers, otherwise excellent, performance.
You are misusing the driver if you crossover high and allow this peak to break through to such a degree. And it might be that at low listening levels the peak remains below -40dB, but turn up the volume and it no longer will.
This is not what the graph is saying. The graph is saying that the third order product of a 2700Hz stimulus is exaggerated vs the rest of the drivers performance. In other words when you put 2700Hz into the driver the motor and soft parts generate a certain level of third harmonic distortion at ~8kHz. Now as the cone rings like a bell at this frequency it starts to resonate in sympathy with the third harmonic and amplifies it considerably. The only way to reduce the effect is to reduce the level of the 2.7kHz stimulus, which means crossing low and steep as a means of avoiding it.
The third harmonic peak is generated from ~900. IM third subharmonic is what you are referring to and that is not what the measurements show or are measuring. This can be easily verified by exciting the driver by sweeping 800-1000. Now do the same 7k-9k.
His post merely confirmed my brief reading into schem, comments and overall design of the speakers that most posts are just misleading and didnt touch the real issue of the speakers.
I want to help but there are too many helping already and Rick is on board with this and i dont see anyone better to assist than the designer himself.
I am sorry but you are simply incorrect on this.
The way ARTA displays distortion data is not how you are thinking. If the graph shows a peak in the third harmonic distortion at 2.7kHz then it is saying that the stimulus of 2.7kHz has its third harmonic elevated to the level of the peak.
The mechanism of why hard, rigid, cones exhibit peaks in the third harmonic (and at other harmonics at their respected frequency divisions) at 1/3rd the frequency of their primary bell like resonance I already described. But once again to reiterate, the motor and other parts of a drivers construction generate non linearities.
If under the excitation of a 2kHz signal the motor produces a third harmonic, at 6kHz, of -60dB, then if coupled to a suitable cone would show -60dB worth of distortion, or 0.1%. If however you decided to use a cone that rings like a tuning fork at 6kHz, such as some sort of aluminium based cone, then the cone will resonate in sympathy with the third harmonic produced by the motor and thefore amplify the third harmonic product. 20dBs of amplification isn't unusual, turning your -60dB third into -40dB and from 0.1% up to 1%.
You can see this clearly in action if you look here at some of the metal cone drive units.
Zaph|Audio
The W15CY001 is included.
If you want I can hook up my W15CY001s and show you an FFT spectrum of the driver excited by a 900Hz tone and a 2700Hz tone and you can see what the third harmonic is generated from.
Alternatively here is my STEPS measurement of the driver, taken some time ago now, but you can clearly see the relationship between the frequency response and the harmonics generated. The driver is flat as a pancake at 900Hz and is operating as a pure piston, there is literally nothing within its construction that could generate a spike in the third harmonic of a 900Hz stimulus. However a 2.7kHz stimulus and its respective third harmonic? Well the third harmonic of 2.7kHz lands at ~8kHz and what do we have happening at that frequency? Oh right the cone is ringing like the liberty bell.
Attachments
- Status
- Not open for further replies.