John Curl's Blowtorch preamplifier part II

Status
Not open for further replies.
In days of high naivete, circa 1996, with computer audio I thought it the ideal opportunity to do look-ahead and processing to optimize things. I soon learned that no OEM customer of Harman's was interested, even remotely.

This is something my boss and I at Nicolet did back in the early-mid '80s. We were phase-precorrecting CD audio by use of FT and convolution with the anti-aliasing and reconstruction filters. It looked great on a scope, but no-one could hear the difference, so management told us to stop ****ing around and get back to our jobs.

edit: With inverses of the anti-aliasing and reconstruction filters. I need more coffee.
 
Dispersion in loudspeaker lingo deals with off axis behaviour of drivers or horns and is determined by geometry.
The context in use is frequency dependant delay, where dispersion has a strict meaning separate from any other.

You are right in the sense that in the xover region frequency dependant delay may play a role, although in audio you would still not use the word 'dispersion' but rather 'lobing' to describe the effect this may have.
Xovers aren't strictly dispersive but they do exhibit group delay. See what I mean..............?

Mechanical vibration is often dispersive, perhaps speaker cones for example, where the cone or its suspension flexes and has mass.
 
Yes, i know... that is why I put the two (your idea and this paper) together here. :)

I hope they take it to commercialization.


THx-RNMarsh

All the tools are already out there.

Playback with convolution engine such as JRiver Media Player.

Kirkeby inverse transfer function implements above cited frequency domain transformations and may be applied to virtually any IIR system to smooth it out.

Here is 70Hz square wave response from 5.25" woofer in sealed 18 liter enclosure with and without Kirkeby based correction:

485954d1433093154-10f-8424-rs225-8-fast-ref-monitor-70hz-squarewave-comparison.png


Similarly this is GD plot for a system fully equalized via Kirkeby:

485023d1432609198-group-delay-questions-analysis-gd.jpg


The above is from this project:

http://www.diyaudio.com/forums/multi-way/269936-cardioid-sum-monopole-dipole-speakers.html

Such a speaker is only realizable with convolution based DSP.
 
Last edited:
Dispersion in loudspeaker lingo deals with off axis behaviour of drivers or horns and is determined by geometry.

You are right in the sense that in the xover region frequency dependant delay may play a role, although in audio you would still not use the word 'dispersion' but rather 'lobing' to describe the effect this may have.

In cables it is used in the longitudinal sense, is different.

I understand how you use the term. Dispersion in some of my systems will result in a 4 millisecond arrival difference between the highs and lows. However due to atmospheric absorption limiting the HF response the real limit is around 2 milliseconds.

It is also easy to correct for it. But that gets into secret sauce.
 
On the subject of velocity in materials this is getting into one of the areas that I have had some real questions but not the equipment sensitive enough to measure. I have mixed multiple materials in my cone and have wondered how this has actually affected the distortion numbers, whether the multiple materials is lowering the "Q" of the nodal breakup of the surface of the cone due to a skewing of those breakups with the multiple physical properties of the materials? I can't speak about the actual resin system that I use to consolidate the composite material. I developed that myself and a friend of mine who is also a speaker design did the actual acoustic testing as I changed that compound along with the combination of fibers and mass density until I achieved the lowest measurable distortion. This resin was combined with both carbon fiber and Kevlar fiber and was found to give the best result. But I have never been able to measure without something like the Klipple analysis system what is actually happening to the physical membrane at these nodal breakup modes, how it actually affects the flexing of the diaphragm. I have always had the intuition that the speed in the different materials has had different velocity profiles and would affect the flexing of the diaphragm but have not had the equipment to prove or disprove this concept.

Does anyone have relevant technical data or theoretical articles about the speed through different solids and how this can be applied to a composite material? I would find this very interesting. This is a multilayer composite material but I can also have custom weaves produced where the materials are interleaved rather than just discrete layers of materials. The resin system is also a combination of material properties so that also adds more complexity to this problem.
 
On the subject of velocity in materials this is getting into one of the areas that I have had some real questions
For what i have seen with various materials, composite, multilayer etc... there is no miracle to hope from the material itself. You can have a (very) little extended response in the treble, but always to the price of higher Q resonance when it breaks. And if you try to go that high, you will hear the material in a way less pleasant than paper membranes. I'm thinking to Focal membranes, by example, with their" thumb thumb" character.
 
Christophe,
Yes I agree with you that most of the composite cones are problematic. I am not doing what they are doing but can't give exact details to what I am using. One of the mistakes that I see is that they are attempting to make the most rigid of materials and there in lies the problem, no real difference than what happens with the metallic cone attempts also. Maximizing the stiffness of the materials is not the answer, this leads to that sharp snap when these materials finally do flex, not a pleasant sound. All the foam cored composite cones have attempted this same thing, stiffness is not the only factor that needs to be looked at. I agree that the modern paper compounds, not truly a single material but also a composite of multiple materials is part of the answer.

At the same time think about your compression drivers and the metallic diaphragm inside that driver. We see the same kinds of breakup modes in those diaphragm at high output, of course the extreme loading on those diaphragms across the surface helps to control some of this but it is still there. There is also the stiffness to mass to think about and the difference between say Al and Be and even somr of the Ti diaphragms. Long ago we even had the glass/phenolic diaphragms in the old large scale Altec drivers but they had to much mass and size to produce much in the upper frequencies.

I see these mistakes made all the time, it is a monkey see monkey do type of thing.

ps. I am speaking of bass/mid drivers when I am talking about the cones I am working on and not a high frequency device. I think this goes with the discussion of the preferences in dome tweeters also between metal domes and soft domes such as silk and some of the polyester types of materials. In your case you are using a compression driver and horn to cover that range and then we have to trade efficiency for size of a horn/driver combination vs a much smaller type of speaker that may sit on a stand or even your desk, your not going to be able to use a horn in those instances. A waveguide and dome tweeter is a completely different animal.
 
Last edited:
Kindy,

At 1/2 wavelength spacing of two acoustic source points the 6 dB down point is at 45 degrees. So to determine if your cones are breaking up, one crude method is to use two drivers placed next to each other (or spaced a bit for LF) and fed 1/3 octave noise. Start at the lowest frequency of interest and determine the angle for - 6 dB. Doing this for each band until you reach the top of the range should give you an estimate of cone break up.
 
Member
Joined 2002
Paid Member
Kindhornman;4347469 Does anyone [/QUOTE said:
Steven
If you can come in contact with an UT NDT technician who has the bare basic UT A- scan equipment , then he would be able to do a survey for appreciable velocity variations within the cone material .
I doubt there will be any. I expect the longitudinal velocities to be somewhere btn 2200m/s and 2700m/s and the local variation not more than 5%
The ultrasonic velocity variations will closely reflect the acoustic frequencies velocity variations. It is the absorption that differs a lot with frequency.
https://www.olympus-ims.com/en/ndt-tutorials/thickness-gage/appendices-velocities/
http://www.journalamme.org/papers_vol20/1451S.pdf
http://www.journalamme.org/papers_vol20/1451S.pdf

I would suggest to investigate mechanical-acoustic properties variations within your cone material with a different inspection technology. Pulse decay test, resonance sweep, impedance sweep, ect., btn 500HZ and 500kHz.
Ask the NDT guys for ZETEC S-9 Sondicator or ZETEC MIZ-21R test equipment or something similar (general term: Bond testing equipment).
If you think you need more details, pm me.

George
 
Kindhornman, look at it this way. A pressure wave cannot travel faster through a material than its specific speed of sound. Tables with the speed of sound through a large variety of materials can easily be found. As a general rule, the harder and lighter a material, the faster the speed of sound through it, with beryllium as the clear winner. So, the dustcap moves, and after (the speed of sound of the cone material times the distance from the dustcap), the rim moves. The speed of sound through air is slower than that through any known solid by a large margin. All these different points on the cone move with some slight delay, but the delay of soundwaves through the air from the dustcap is much larger.

However messy this sounds, it doesn't matter very much. All these slightly delayed waves from different parts of the cone combine perfectly to produce a wavefront of approximately the size of the cone, but with the acoustic center somewhere between rim and dust cap as Ed already mentioned. The wonders of sine addition.
 
Last edited:
At the same time think about your compression drivers and the metallic diaphragm inside that driver. We see the same kinds of breakup modes in those diaphragm at high output, of course the extreme loading on those diaphragms across the surface helps to control some of this but it is still there.
We agree so often :)
What you say about compression is true. But it looks like 1" diaphragm is a good compromise for Hifi levels (not PA). We can design the horn down to 350Hz (700Hz of Xcross) . But they are a little short in treble. I Just wonder why they put the coil at the edge of the diaphragm. If we designed it with the coil the other side (on the back) somewhere near the middle of the diaphragm, we could expect an extended range, not ?
 
Last edited:
diyAudio Member RIP
Joined 2005
This is something my boss and I at Nicolet did back in the early-mid '80s. We were phase-precorrecting CD audio by use of FT and convolution with the anti-aliasing and reconstruction filters. It looked great on a scope, but no-one could hear the difference, so management told us to stop ****ing around and get back to our jobs.

edit: With inverses of the anti-aliasing and reconstruction filters. I need more coffee.
What prompted me was an absurd claim (by a notoriously manipulative person at Compaq iirc) that a competitor had some breakthrough amp that leapt tall buildings at a single bound etc.

Taking this as a challenge I suggested how look-ahead could be used to provide signal information that would allow the electronics of a powered speaker to be yet cheaper :)

I did incorporate some ideas along those lines in a patent, for bus-powering of equipment with less-than-100% duty cycle demands, which Harman may not even be paying maintenance on by now and never used---although some imagined that it had been. IMO a massive blunder, but as usual politics prevailed over common sense; and then there was the absence of someone to champion the applications.
 
Thanks everyone for the responses. George I do know someone if they are still there who did our NDT testing of composite parts, that was done with ultra-sonics and a water stream. My guys build the holding fixtures for those tests and I spent time in that department.

We have done waterfall testing of the speakers and the decay is very clean and had no noticeable peaks to deal with. It was easy to create a composite cone with just the opposite effects with peaks in the waterfall plots by increasing the stiffness of the cone, it became obvious with a worsening FR and waterfall plots this was not the correct direction to go.
 
Christophe,
I'm not so sure that moving the voice-coil to the first nodal breakup point on the back of the diaphragm would really help. Take a look at the Focal dome tweeter and that is what they do and as far as I can see from the work I have done it didn't help those inverted dome tweeters in the least, they still have horrid breakup modes up high. One of the reasons I think so many find those dome tweeters so fatiguing. This isn't something that can be easily transferred to the discussion of a compression driver but I think it may be similar. I myself only used 1" exit compression drivers for the designs I have done for home use, I never thought that any of the larger exit compression drivers had good off axis response up high, it just became a beaming high frequency uncontrolled by the horn and the added mass was of no help. You have to think about how a compression driver is actually working in that very small air space in front of the diaphragm and between the phase plug. There is rarefaction happening there in that small space. Don't forget the average compression driver has 10 to1 ratio of diaphragm area to phase plug entry. There are many other areas of the compression driver that I think make more difference than a change in the placement of the voice-coil position.
 
Kinhornman, a question, if I may. How do you feel about 10" bass drivers covering frequencies up to 800 Hz, 4" cone drivers taking it from there to 3,500 Hz with a 1" titanium dome finishing the lineup?

In general of course, as a principle. What kind of crossovers would you want to see built in, aside from the obvious, that they use quality materials?

FYI, the bass driver uses Son Audax' version of aerogel cone, and is claimed to be good for 2 kHz; also aerogel cone for the midrange, who response is declared as 400-8,000 Hz.

Just wondering.
 
Status
Not open for further replies.