Beyond the Ariel

[mige0]

Originally posted by nickmckinney

Its not the VC material, its getting the heat from the VC to the magnet structure that is the problem.

From a reliability stand point I agree – not when modulation is concerned.
Actually ANY modulation whatsoever = distortion.


--------


Lets do a quick calculation.

Lets assume the voice coil wire is made of 10m / 0,2mm Copper
Then the resulting resistance at 20° is around 5,5 Ohm

Lets assume that the same voice coil has got 120°C warm
Then the resulting resistance at 120° is around 7,5 Ohm

Obvious the resulting current for the same input is dropping at around 35% - SPL will drop by 35% consequently.

This is a 35% !! DISTORTION - nothing in the range of (or far below) 1% as for THD or IM like we are familiar with.


120°C is nothing THAT unusual.


--------


Lets do another quick calculation on how much energy is needed to heat up that same wire by 100°C

The weight of that VC wire is slightly below 3g
To heat up this <3g of copper we need around 1 Watt for 1 sec for each Kelvin increase (no cooling assumed here)

From this we only need 10 sec of 10W input to get our 100°C increase – or 1 sec of 100W input.

In other words – if you want to enjoy the fortissimo at 115dB SPL and have a 95dB sensitive speaker - well after 1 (!) sec you have reached 35% of distortion - having a 105dB sensitive speaker you can enjoy your 115db SPL fortissimo a fantastic 10 sec until you end up at 35% distortion.

Quite revealing – no?

Of course there is heat leakage due to cooling in real world speakers, but don't expect them to cool down a 100Ws input immediately.
Heating up the VC is an immediate process – cooling down isn't.



-----

Lets do last quick calculation.
Lets assume a 50mm = 2" VC with a height of around 13mm (one layer wound to be optimistic )

This will provide a heat radiating area of around 40cm2 from front AND back.

If we – optimistically – assume to keep the magnet structure at 20°C while the VC is at 120° the heat leakage due to radiation is somewhere at around 3W

This means the decay from our 35% distortion fortissimo would last around 30 sec
Unfortunately – if the VC is cooled down to 70°C (half way down) then there is only about 1W cooling through radiation left (still optimistically assuming the magnet structure to be at 20°C), further stretching the time of distortion decay.
In fact decay would be rather asymptotic / exponential than linear of course.



Sure it might not really be 35% of distortion.
Sure there is also convection as cooling mechanism but if pianissimo follows fortissimo you only have the additional cooling during fortissimo.

But anyway - a decay time in the minute range to come down to reasonable distortion levels should be a very good argument for developing better working materials for VC wires – no?

Anybody ever considered temperature decay matching of drivers?

Originally posted by nickmckinney

Aluminum and/or copper surrounding the magnetic gap as close as possible is your best friend for many reasons.

The AIC paper draws a slightly different picture as I already have pointed to.


Greetings
Michael



PPUUhh – hope I haven't mixed all that formulas and constants

[chrismercurio]

Earlier in the thread the Beyma TPL-150 was mentioned and considered....

Beyma will be releasing a horn for this Heil variant soon, which looks to improve a few areas.

http://profesional.beyma.com/newslet...reliminary.pdf

I tried to attach the pdf, but even after shrinking it it was still to big.

C

[Magnetar]

Originally posted by Variac


Well, so far the 18 sound Wavehorns meet my criteria for many things in life:

Good looking
Cheap
Available

Well- we seem to agree an what really matters!

I'm practically cheating on my wife with the XT120's.......

[mige0]

Its NOT bad weather today – its lots of DUST from the SAHARA in the middle of the ALPS today.

Crazy world!

Greetings
Michael

[gedlee]

Originally posted by mige0

But anyway - a decay time in the minute range to come down to reasonable distortion levels should be a very good argument for developing better working materials for VC wires – no?

Anybody ever considered temperature decay matching of drivers?

I appied for a patent on a VC wire material that had about 10% of the variation of Re with temp. A few % nickel added to copper does the trick, but it raises the resistivity of the wire. However at some temperature the two slopes cross so the NI-CU wire works better for drivers that tend to be warm and the CU works better for cooler situations.

[nickmckinney]

Originally posted by mige0


From a reliability stand point I agree – not when modulation is concerned.
Actually ANY modulation whatsoever = distortion.


--------


Lets do a quick calculation.

Lets assume the voice coil wire is made of 10m / 0,2mm Copper
Then the resulting resistance at 20° is around 5,5 Ohm

Lets assume that the same voice coil has got 120°C warm
Then the resulting resistance at 120° is around 7,5 Ohm

Obvious the resulting current for the same input is dropping at around 35% - SPL will drop by 35% consequently.

This is a 35% !! DISTORTION - nothing in the range of (or far below) 1% as for THD or IM like we are familiar with.


120°C is nothing THAT unusual...............

Hi Micheal, I snipped the last half or so.

Let me explain what happens internally with my TD and Apollo drivers:

1 - My magnetic gap is tall, 18mm of steel plate on the outside compared to the 6-13mm of regular drivers.

2 - My solid copper pole is tall, about 100mm give or take of height

3 - My solid copper pole is thick, about twice as thick as I have seen done on other drivers

4 - My steel pole is ~100mm tall and solid.

5 - The clearance between the voice coil and the above "heatsinks" is about 50% less than most other drivers.

6 - The VC former is made of both Aluminum and Kapton, the Aluminum runs the entire height of the former which is roughly 60mm give or take.

Steel is a poor conductor of transient heat, but you can sink a lot of heat into it. Copper is the opposite and in this case the copper sleeve is pressed tight onto the steel pole. Then this is in very close tolerance to a tall heat conductive voice coil former. This conducts the heat into the steel pole better than most any other solution I have seen.

Then -

We install a solid chunk of aluminum onto the top of the pole to act as a phaseplug/heatsink. The 15" drivers have one around 100mm tall and 50mm thick. You are welcome to drive one as hard as you want and touch the phase plug with your bare hands.........

And that is a standard TD motor, now lets look at the Apollo:

Add another solid aluminum ring about 150mm in diameter and 18mm thick at the thickest point, 7mm at the thinnest, and spaced as close to the voice coil as the steel plate.

Add yet another solid aluminum ring below the magnetic gap on the outside of the voice coil, almost the height of the magnet stack, and as close to the voice coil former as a regular speaker would have its clearance.

Now I am not saying these drivers don't have power compression, but lets just say I will put them up against any other design with no worries, especially when it comes to covering the 20Hz-1Khz range at the same time.

The above is talking only about heatsinking BTW, the copper sleeve and aluminum rings also work to help keep the magnetic flux field stationary, which is a source of distortion in ceramic magnet speakers.

[John_E_Janowitz]

Originally posted by gedlee


I appied for a patent on a VC wire material that had about 10% of the variation of Re with temp. A few % nickel added to copper does the trick, but it raises the resistivity of the wire. However at some temperature the two slopes cross so the NI-CU wire works better for drivers that tend to be warm and the CU works better for cooler situations.

The problem with this is not only that nickel raises resistivity, but that nickel is also magnetic. Using any kind of magnetic material in the coil or form introduces a force factor that resists movement of the coil, acting like there is a parking brake on and constantly attempting to pull the coil back to center. It will also cause the coil to be pulled laterally towards whatever edge is closest and can rub much easier. Even a small percentage of nickel causes a significant effect. If this wasn't the case, you would be able to use ceramic coated 27% nickel clad copper wire (kulgrid 28) with ceramic adhesives. It would virtually be a burn-up proof coil. The wire is good for continuous operation up to 1000degrees F and the adhesive is good to even higher temperature. Unfortunately the Nickel required for the ceramic to be put on the wire doesn't allow this to work.

John

[Lynn Olson]

Another clever trick used by the prosound drivers is winding the voice coil on both the inside and outside of the VC former, so each side can radiate infrared directly onto the pole-piece and top plate. The VC's of audiophile speakers don't usually use a double-wound VC, which is an indirect comment on their peak headroom.

Magnetar and Variac, I dunno which horn will sound better. The LeCleac'h, with its trademark rollover, has a lot softer edge termination to the circular mouth than the sharp-edged and rectangular-mouth XT1464. To the XT1464's credit, there is thankfully no diffraction "pinch" in the throat region. I've suspected for a long time the reason the Altec 811 and 511 sectoral horns sound so dreadful is the diffraction "pinch" in the throat region - take a good look at one some time, they have a really weird profile that you can't see until you look at the back portion.

Build quality of the black molded-plastic XT1464 is pretty good as well - but users of old-style 3-bolt Altec 288's will have to drill their own set of holes in the mounting flange, which only has the modern 4-bolt pattern. I suspect clever users of the XT1464 could always fill the thing with open-cell foam and see what effect that has on the time-domain response.

If you think the Azurahorns are expensive at about $600/pair (shipping is around $90 the pair), you should see the prices of other custom-made horns. The Music Concretes are about half again to twice that (they are heavy to ship), and US-made wood horns seem to be about twice that again. That gets into boutique beryllium-driver territory - OK for the hardcore horn fanatics who already know what they want, but too much for prototyping and just fooling around with various set-ups and measurements.

[mige0]

Originally posted by nickmckinney

Hi Micheal, I snipped the last half or so.

Now I am not saying these drivers don't have power compression, but lets just say I will put them up against any other design with no worries, especially when it comes to covering the 20Hz-1Khz range at the same time.
.

Hi Nick, yes lots of effort to transmit heat.

Nevertheless you wont get rid of the thermal transients this way – just reducing it by maybe a factor of three or four (my rough guess only) to common designs.
Not bad at all – but by no means comparable for what could be realised with better wire composits.
Good thing is, your speakers do not relay on the dust-cap-air-pump-cooling-effect that is heavily program dependant.


The core part is the VC - anything else is just to keep down the temperatures of the VC ambience to allow for good heat transport.
You may have noticed that I assumed in my example a constant 20°C for the VC ambience.
NO real world speaker will EVER provide that (at usual room temperatures) - yours wont as well.

Sure the aluminium part of your VC helps to enlarge the radiating area of the coil itself – but the aluminium material usually is too thin to allow for massive heat transport (I leave that quick calculation to someone else )


--------------


Earl, yes looking at Konstatan (?) I had the same idea with Nickel , but....
JohnE_J already has shown the weak point...
Maybe there is a workaround?



--------------

Originally posted by Lynn Olson
Another clever trick used by the prosound drivers is winding the voice coil on both the inside and outside of the VC former, so each side can radiate infrared directly onto the pole-piece and top plate.

Lynn, yes – this I already assumed in my quick example

Greetings
Michael

[nickmckinney]

Originally posted by Lynn Olson
Another clever trick used by the prosound drivers is winding the voice coil on both the inside and outside of the VC former, so each side can radiate infrared directly onto the pole-piece and top plate. The VC's of audiophile speakers don't usually use a double-wound VC, which is an indirect comment on their peak headroom.

I have to be honest in that I have reconed a few thousand speakers and never saw a driver that did that. I imagine it would be a pita to keep the dimensional tolerances correct and to shim the gap correctly when assembling. We do ours with a layer of Nomex between the windings and the former to lower the outgassing problem that comes with extreme high power. IMHO someone would have to teach me why this is a better way.