diyAudio

diyAudio (http://www.diyaudio.com/forums/)
-   Full Range (http://www.diyaudio.com/forums/full-range/)
-   -   Advice on "12 Fullrange Rebuild Needed (http://www.diyaudio.com/forums/full-range/203094-advice-12-fullrange-rebuild-needed.html)

T101 24th December 2011 12:16 PM

Advice on "12 Fullrange Rebuild Needed
 
8 Attachment(s)
Hello and happy holidays! :)

I have undertaken a rebuild of a pair 1966 AlNiCo fullrange drivers.
They will receive new original cones, ultra soft suspension, new original HF cone (whizzer) and a new voice coil.

The main question that I am unable to answer scientifically is the number of layers of the voice coil.
The doubt is between 2 and 4 layers. With 4 layers we will be able to achieve BL above 10. Usually with 2 layer VC those drivers have a BL of about 9.2 to 11.
The gap is approximately 2mm wide and we are able to make a voice coil that has 4 layers for more BL.

I have read opinions stating both that multi layer voice coils are anti audiophile and that there is no difference. I am not able to form a well backed statement in either of the directions, so please throw the science at me!

Given parameters:
- Voile coil diameter/gap diameter: ~37.5 mm (1 1/2 inch)]
- Magnet gap height ~ 9mm (~0.35 in).
- Flux density: unknown, but gives BL ~10 with two layers on the outside with total height of approx 15-16mm and 8 ohms with Re 7.2 and Le 0.4... - I would use any approximation based of this from anyone enough experienced to make such a statement.

Project Objectives:
- Fs of 20 or less Hz;
- Moving Mass - typically those have Mms of 27 to 37 grams (around one ounce);
- BL preferably above 14, but everything above 10-11 will be good;
- Qms above 4;
- Qes - as low as possible - see BL;
- Xmax - 2.5mm one way;
- Pmax - 15-20W or whatever drives it to Xmax in sub 0.707 Qtc sealed or aperiodic (breathing) enclosure;
- Le 0.15 or less, preferably under 0.1 mH - for the variant with 4 layers easily achievable through parallel connection of layers;
- Re and Z, having in mind the ultra soft suspension and that the maximum power will be limited purely by X-max, and having in mind that a greater damping factor is preferable in this case, we will be aiming for 16 ohms impedance;
- Kapton voice coil former;
- Qts - we are aiming for less than 0.4, but we'll see.

Project objective: Goodmans Axiom style speakers which do 20 Hz @-6 db, look great and sound beautiful. Approximate enclosure volume: 120 to 160L. Intended finish: stained, oiled and etc. Birch plywood... - you get the picture, It will be a present for my father :)

Again, my only or let's say main doubt is the number of layers of the voice coil.

Maximum HF extension is not an objective, because they will be backed by an AlNiCo magnet ribbon super tweeter. Or maybe I coud use a pair of paper cone 4in HF drivers that also sound beautiful and also have AlNiCo magnets and are more vintage... either way there will be a super tweeter.

VaNarn 24th December 2011 07:57 PM

If the main purpose is to get the best high frequency performance from the 12'' speaker,then it is essential to keep the voice coil mass at a minimum and this implies an aluminium two layer coil for the dimensions that you have listed.It is also wise to make the voice coil former as stiff as possible and avoid a long neck.With reference to a four layer version,the Bl factor increases, but then you will have a higher inductance.The suggestion of a parallel connection of the 4 layers is only giving you the equivalent of using two thinner wires with the same cross sectional area as a single wire.I would suggest that if you are striving to get down to 20Hz then given the speaker driver that you propose to modify,that a vented box,unfortunately a much larger enclosure than your 160L upper limit,is required.

T101 25th December 2011 10:07 AM

1 Attachment(s)
Hi, Merry Christmas!

Thanks for the input.

The main aim is to get the best LF extension. In the high frequencies it would be sufficient if they go to 8-10 kHz, because I have two great alternatives for super tweeter - in the photos above. Please tell me if Kapton is not suitable for 10 kHz operation.

I made some simulations and interestingly, too big BL is not giving me better LF extension, actually it's the opposite. See the attached simulation. I get similar results with slightly different sets of parameters. For instance with Vas between 600 and 900L and Fs of 15-18 Hz it is almost the same. Not to mention that 160L Aperiodic should be equal to twice it's size sealed.
Also the difference in the footprint between 160 and 285 L is not that big. 280L is 1150x550x470mm outer dimensions.

When I dial allot of BL, Qes falls and thus Qts and the driver becomes not bass friendly. With Qts round 0.3 I get the best results. Interestingly, with BL of 11-12 the sims look the best?

It is not an actual mod of drivers, because all parts except the voice coil will be original, only we'll use the two softer suspension parts types available.

But the main question is whether there is any negative influence on the sound from a 4 layer voice coil. Does it sound more muddy, blurry or with less clarity when compared to 2 layers?

VaNarn 25th December 2011 07:07 PM

For low frequency performance it does not matter much if the coil is 2 or 4 layer,but it is more to do with what best fits the magnet assembly gap.In the specified case of your simulated model,Fs is shown as 12Hz.Given the Mms is 38 g,then this implies a floppy suspension and this in turn normally leads to an unreliable speaker.It is best to consider a woofer design and contemplate a mid-range unit added to your scenario.For a Fb of 20Hz this means possibly double the Xmax that you have considered.An aperiodic box is basically a sealed enclosure with high frictional losses and you have to compensate for this, if you wish to extend the response lower.The room influence is an unknown factor but in general it will create some useful gain below 30Hz, and in addition two speakers allow a significantly lower Q to be a better proposition under real world conditions.

T101 26th December 2011 11:02 AM

I know that for LF it does not matter how much layers are there in the VC.
But the drivers will be required to cover at least the 20-10000 Hz range and thus any downside in the sound quality arising from the number of VC layers should be considered.
The sluggish suspension is intended for several reasons:
- we have reached the conclusion that general listening is conducted at power levels in the sub 1W region;
- once I was at the national radio and there were those funny looking 2 way speakers with ~5 inch LF drivers... They were reaching effortlessly down to 30 Hz with their funny small cones. I found an engineer and had a talk with him about those speakers. Full bandwidth was retained only until 0.25 or 0.5 W, their suspension was soft to the point where it had only cone supporting and VC alignment functions... But you could tell for the bass guitar stings texture from their sound...;
- considering the above two, why not sacrifice power handling for maximum extension? They should be able to maintain flat FR @ 93-96 db @ 1m full bandwidth with 2.5-3mm xmax one way. Thus actual power handling in Watts is a parameter of secondary importance.

As a fan of the slightly irrational sophistication, I consider two irrational extras...
- acoustic blanket that could be applied on and off through a handle from outside - this should add resistance and thus limit excursion if greater power levels are desired at the price of LF extension;
- a 6mm paper shutter attached to the back of the cone surrounded by two optical semiconductors, forming an xmax overdrive warning system together with the respective peripheral electronics.

This ia apart from the planned 2 crossover options for the super tweeter - classic two way and super tweeter mode... And apart from the planned tweeter level adjustment.

Best regards!

T101 18th February 2012 01:06 PM

And the results:

Qes - 0.26
Qms - 1.7
Fs ~23 Hz
Re - 11.8
Le ~1.2 mH - oops
BL.... 16.6-17 - big OOOPS
Mms - 42g
Qts ~0.216... ooops and ouch... would be happier with 0.34... but...

What we have here is a slight overkill...

Bandwidth up to 12 KHz
Linear up to 8-10 KHz

Intended use: up to 7.5-8 KHz with 3rd order low pass forming 4th order LR slope and crossed to a ribbon driver with LR-4 high pass @ the same frequency.

Simulations return 20-25 Hz @-12 db which is out of target performance, but a 30 mH + 32 ohms correcting network returns almost flat 20Hz to Fo SPL curve...

Any ideas for what to do with such driver are welcome!

Preliminary listening impressions: Maaan! They are fast!

Best Regards!

T101 19th February 2012 07:56 AM

Problem solved, an 8 ohm added resistance will raise Qes and Qts will become ~ 0.35 while lowering BL to under 14, this will allow ~20 Hz @ ~-6db while lowering sensitivity as well which is good, because the 8 ohm ribbons have under 90 db sensitivity... Great!

There will be a Zobel and LF impedance equalization circuit as well.

And by the way, VAS is ~385L and with such a motor, 25Hz tuning frequency in a 185L enclosure is achieved with approximately 66x25mm port... ;)
... and group delay will be under 8ms above 30Hz ;) and almost nothing above 45 Hz... with a peak of 15-16 ms around 25 Hz ;) *very happy*

Best regards and thanks for the input!!!

T101 22nd February 2012 08:15 AM

1 Attachment(s)
Here are simulations of the drivers with 8 ohms added resistance in a 185L vented cabinet.

Everything looks pretty satisfactory.

I consulted an engineer, whether driving a 22 ohm load would be a problem for most amplifiers and the answer was that there will be no problem except that the power supply should be able to put out greater voltages.

Thus I am ordering the enclosures very soon. :):):)

Best regards!


All times are GMT. The time now is 02:52 PM.


vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2014 DragonByte Technologies Ltd.
Copyright 1999-2014 diyAudio


Content Relevant URLs by vBSEO 3.3.2