A technical discussion of the Borresen ironless woofer.

[smowry]

Let me compare the Borresen Woofer topology to the Purifi woofer topology. We know that the Purifi woofer is no dog and is highly regarded.

Now compare the following.

1. Voice Coils
2. Voice coil to diaphragm coupling
3. Surrounds
4. Spiders
5. Wheelbases (distance between surround roll and spider)
6. Length of Voice Coil bobbin that is cantilevered (spider to windings)

Things just don't look right in the Borresen sectional illustration. It could be a puppy dog.

 

[smowry]

Okay now please rate the 6 categories from 1 to 10 for each transducer. 10 being highest score and 1 being lowest score.

Purifi Woofer vs. Borresen Ironless Woofer

1.
2.
3.
4.
5.
6.
Totals

 

[Snickers-is]

Al and Cu (no one would use silver or gold) have the permeability of air, Almost air core. Placing high conductivity materials between the magnet will realize and induction heater. That heat(er) will encourage the failure mode of thermal demag. Then there will be the high electromagnetic drag. So now I understand why the Qms was identified as, Qms < 1, ouch! I was taught by a EM scientist from MIT in 1995 to minimize mechanical losses, moving coil audio transducer efficiency is already very low, ~1% efficient. Transducer engineers have design rules. Is that an inventive step anyway? Yes a step backwards. At best it's a claimed improvement to Almost air core but I claim it is not an improvement and it introduces a potential failure mode and reduces efficiency.

I am quite surprised that you, after having worked with transducer design for so many years, seems to have no clue how shorting rings work, or even loss.

Firstly, you do not get high loss by using low resistance shorting rings. Yes, you will induce current, but even if this current is 10A, the copper ring would bave maybe 10mohms of resistance. This translates to 1W peak, and as I am sure you know, the average crest factor is around 8, meaning we are looking at around 0,125W nominal at high levels.

The reason heating in the copper ring may occur is because of the induction between the motor and the surrounding conductive materials. With no copper there, you will generate almost the same amount of current within the magnet rings instead as they are also conductive.

The heat loss in the coil is at a completely different level. If you replace the copper ring with non conductive material, you will lead less heat away from the magnet rings, the coil will dump more heat into the magnet rings, and the overall ability to dispose heat from the motor will go down.

The ring does work as an induction coil. Normally this leads to slightly increased efficiency compared to not using any shorting ring. Not reduced efficiency as you stated.

The resulting Qms typically go up, not down, as you add shorting rings. This is because shorting rings have less resistance and then also less loss than steel and magnets.

Adding a shorting ring does indeed improve this design. The more questionable part is if this is a good design for a woofer to begin with.

Someone posted in this thread that they tested the Borresen transducer and the voice coil reached a temperature of 350 degrees (no unit posted C or F) and the magnet temp only increased 2 or 3 degrees. So if the magnet spacers were Al or Cu then I will assume that Santa Claus or the Tooth Fairy was cooling the magnets!

I guess you refer to my post. As I stated, this was testing done on a 21IPAL from B&C, powered by an 8,5kW amplifier with pressure controlled PID servo. It also uses a real power sensor for RMS limiting. It has a neo motor, 48mm winding height and 18mm gap. It was mounted in a 30Hz horn cabinet for the test. The driver was flipped with the magnet out for the test to give better access to the motor.

I used contact thermometers in 3 different locations on the motor. I also used a thermal camera monitoring the voice coil through some gaps. The temperature changes rather quickly, but I recorded in excess of 350 degrees celcius during the test. After about 30 minutes of testing we recorded an increase in temperature of between 2 and 3K on the motor. The readings were stable across all three sensor locations.

Neither Santa Claus or the Tooth Fairy was present at the time, but two other engineers from two different companies were present.

1. Steve Mowry
2. Guy Lemarquand et al.
3. Mathias Rémy
4. Michael Borresen

The list is short to say the least. Can anyone find any additional work on this topic? I am still interested in this topic and it would have the potential to add information to this thread.

You can safely add me to the list as well, but I know I am not the only one.

A healthy discussion around ironless motors will not take place in this thread as it focuses on your anger towards Børresen for several pages. You should start a new thread without any of that.

PS. What do we want between the magnets if not Al or Cu? Ideally we want air!

That is like stating that demodulation rings do not work.

 

[smowry]

Here's an example of loudspeaker engineering at the highest industry standards. This doesn't mean that this represents my approach to transducer and/or system design. I follow the GENELEC approach with a focus on directivity control. Now with regard to Laurence Dickie, he is approaching legend status or maybe he's there already. He left B&W but now he has surpassed them in my opinion. Dickie is for real, 100% snake oil free.

 

[olieb]

Firstly, you do not get high loss by using low resistance shorting rings.

Actually you do. The (change of) flux generates an induction voltage. The lower the resistance the more power that is converted to heat. That is the reason why induction heaters do not work on low conductivity materials.
With 10A (in the voice coil) you will get a lot more current in the copper ring as this eddy current tends to neutralise the field change.

That is like stating that demodulation rings do not work.

I do not understand what a demodulation ring is supposed to do here. There is very little flux modulation to begin with as this seems to be big part of the incentive for an ironless motor.

 

[smowry]

Okay, it looks like we are going to talk about induction now. Undocumented claims are flying around and personal attacks are on the rise.

Kindly let me start with a transducer model without any secondary shorting ring. The first two models are AC with frequency f (Hz). Note the AC sources in both models, force and voltage respectively.

Using super position analysis, I will show the induction model separately below. The voice coil drives the moving assembly above, while the same voice coil and the shorting ring et al. form a transformer below. Note that the shorting ring is a single turn coil. The voice coil being the primary and the shorting ring being the secondary,

I understand that this is a simplification; however, my points will be generalizations as will my colleagues and adversaries. I have no meaningful values and the objective is to illustrate inductance. So fundamentally and with respect to conservation of energy, the shorting ring cannot increase efficiency.

Unlike the dynamic top model, the induction model is static and a current density is induced in the shorting ring, that is related to (αI)^2 R) heat losses (W).
Where α is the loss factor and 0 < α ≤ 1. Then if we assume the induction process is static and lossless, the model can be further simplified.

The highest temperature magnet wire used in the loudspeaker industry is rated at 240 degrees C and utilizes polyimide based insulation. This wire is not popular because in requires the use of an induction heater during winding where the polyimide is heat activated. The typical magnet wire rating is 180 degrees C where the bonding adhesive is pre-applied and typically activated with alcohol and the finished voice coil can be baked. 350 C is possible but only short term. Temperature readings need to taken at the steady state level.

More later, but do shorting rings work in an ironless woofer motor? No! An analogy is like pouring water on a fire that has already been extinguished. Are there heat losses related to induction? Of course there are. Can shorting rings increase efficiency? How, by tooth fairy or Santa Claus?

 

[Snickers-is]

Actually you do. The (change of) flux generates an induction voltage. The lower the resistance the more power that is converted to heat. That is the reason why induction heaters do not work on low conductivity materials.
With 10A (in the voice coil) you will get a lot more current in the copper ring as this eddy current tends to neutralise the field change.

But this is a loudspeaker, not an induction heater. The purpose is to form a magnetic field to create motion, and the type of loss we look at is not pure resistive loss, but mechanical loss and derived loss that interfers with this motion.

When you add shorting rings to an existing design, Qms normally goes up a tiny bit. It is unlikely that this provides us with an explaination for the very high loss of the Børresen driver. One explaination could simply be that what he showed me (the DATS plot) was in fact not just the driver.

10A in the voice coil of a 4,5 inch woofer is highly unlikely unless it is a very low impedance design. For a 5 ohm driver, that is 500 Watts... It probably can not take 20W in the bass region. Inside the shorting ring (as I stated in the post you quoted from) it is a realistic peak figure.

Just to put it into perspective:

In the upper right diagram, blue curve shows power, and red curve shows current. I set it to 10A at Fb, which gives us about 700W and 24mm one way excursion. The DUT is an old Peerless 5,5 inch driver, so a 4,5 inch Børresen woofer would not even do this well. This is in any way totally unrealistic numbers.

I do not understand what a demodulation ring is supposed to do here. There is very little flux modulation to begin with as this seems to be big part of the incentive for an ironless motor.

It will provide a minor stabilizing effect, but it will have a more pronounced effect on cooling the voice coil. The heating of the voice coil will contribute far more to lowered Bl than the heating of the magnets. We can expect at least 10K rise in the coil for every K of rise in the magnets.

 

[Snickers-is]

I understand that this is a simplification; however, my points will be generalizations. I have no meaningful values and the objective is to illustrate inductance. So fundamentally and with respect to conservation of energy, the shorting ring cannot increase efficiency.

It does not increase efficiency, but your statement was that it contributed to a low Qms, which is a totally different thing.

The relationship between the shorting ring and the coil is not exactly as you describe above however. The rings do not work at DC, and hardly at all at low frequencies. At higher frequencies they counter act the inductive component, lowering the impedance at higher frequencies. This is where the energy loss comes in. It is important to distinguish between this type of loss and the mechanical loss that lowers Qms.

The highest temperature magnet wire used in the loudspeaker is rated at 240 degrees C and utilizes polyimide based insulation. This wire is not popular because in requires the use of an induction heater during winding where the polyimide is heat activated. The typical magnet wire rating is 180 degrees C where the bonding adhesive is pre-applied and typically activated with alcohol and the finished can be baked. 350 C is possible but only short term. Temperature readings need to taken at the steady state level.

Not sure where you are going with this. This is one of the most power-oriented drivers on the market, and we know it uses high temp rating coil wire. That is why we did run it at around its curing temperature. It sure did smell. But it gave us some important info that you seem to be missing, and that is the heat generation that happens in the motor, even including the transfer to the motor, totally dominates what happens in the motor, or is even transferred to the motor. This is a bit different in high frequency designs, but not significantly different.

But increasing the coil cooling will have some benefits in a design like this.

 

[smowry]

Where's the thermal analysis? Did you review Buck's paper?

https://pearl-hifi.com/06_Lit_Archive/14_Books_Tech_Papers/Mowry_Steve/Bucks_Induction_Driver.pdf

Here's an example of thermal analysis with the quantity of interest being steady state temperature.

Do you have formulae?

 

[smowry]

Snickers-is,

Could you kindly tell us what your contribution to this thread has been?

I certainly can't make you leave; however, I can suggest that you leave for the sake mankind.

 

[smowry]

Snickers-is,

"Not sure where you are going with this."

I am questioning the credibility of your claimed data.

So here's my position. No one but you have made any personal attacks on the thread participants. This discussion is about transducers and not trying to one up the participants. So it's either you or I. If you stay, I go; if you go, I stay. My mission here from the get go was to share. Unfortunately, I cannot figure out what your mission is and I have come to dislike you.

 

[olieb]

But this is a loudspeaker, not an induction heater.

Yes sure, but Maxwell's equations do not care how it is called.

The purpose is to form a magnetic field to create motion, and the type of loss we look at is not pure resistive loss, but mechanical loss and derived loss that interfers with this motion.

Hm, I tried to give my input to an argument about heat in the motor (from an induction shorting ring). The argument was, that a good inductor will not produce much heat (as opposed to a part with lower conductivity). And like in an induction heater it is the other way around.
The effect of some copper in the motor on mechanical properties will be very small if not zero (without current the ring does basically nothing).

10A in the voice coil of a 4,5 inch woofer is highly unlikely unless it is a very low impedance design.

Maybe, but my point was not about 10A or 2A but about the fact that in the ring you can have up to N times the current in the coil (N being the number of turns). That can be more than 10A and definitely more than the current from the amp.

The heating of the voice coil will contribute far more to lowered Bl than the heating of the magnets.

Again I do not understand as this conflicts with my training in physics (I am not an expert in driver technicalities or magnet structures). Which part of B x L will be lowered, when the voice coil gets warm?

 

[mwmkravchenko]

Mark,

What is your cost estimated to manufacture those 4.5 in woofers in China? My guess is about $100. Borresen is also a price gouger.

That is completely related to quantity. That 3D metal printed basket will be in the hundreds all by it's self. So per driver cost in 500pcs? I would throw a dart in the vicinity of $500. Take away the funky basket that rally is doing nothing of any real value and I'd put it at about $280. Nobody wants to machine copper. The Neodymium will be near $120 for the four magnets. The rest under $20. So a sensibly designed version of this with a nice open frame sturdy basket in 500pcs could be $150 to $170 is fairly modest quantities.

 

[smowry]

When the voice coil heats the magnet(s), B(T) will change moderately with Temperature based on the BH characteristics. The important quantity here is not Bl but BETA, where
BETA = (B(T)l)^2 / Re(T) N^2/W. BETA is the true figure of merit for transducer motors where BETA can be compared to similar transducer without considering Re. Note that efficiency is proportional to BETA and not Bl.

 

[smowry]

Hi Mark,


Thanks for stopping by. You have much transducer knowledge and it's current (pun intended). I could sure use your input to the survey at the top of page 6.

 

[smowry]

An imaginary air gap between magnets will result in the lowest linear permeability possible and lowest conductivity possible that will not change with temperature. This is consistent with the ironless motor design goals. The ideal magnet would have the permeability of air (it has that) and conductivity of air that does not change with temperature. However, the voice coil may overheat without adequate thermal mass and conductivity. This is why nonmagnetic stainless steel is a consideration. It is closer to the ideal than Al or Cu but still offers a heat sink with thermal mass. There are tradeoff decisions at many points in the transducer design process.

 

[smowry]

"A healthy discussion around ironless motors will not take place in this thread as it focuses on your anger towards Børresen for several pages. You should start a new thread without any of that." Then there was, "Relax".

So you are telling me what to do and what I can and/or cannot do. Amir says we are free here but be good when a review thread is in progress. I have had folks message me and thank me for starting this thread and that is why I am still here but if you return, then count me out. Anyway now I am expressing my disappointment in you. You stay, I go. You go, I stay.

 

[mwmkravchenko]

My compendium of pictures from Borreson Audio. I make a living knowing what other companies do, and figuring out how it's done.

So, I need to establish a few baselines and then try and configure some dimensions for the driver simulations.

 

[smowry]

Mark,

That's great and thanks!

 

[mwmkravchenko]

Mark,

That's great and thanks!

So we are going to need to guess about a few things. As I look through and come up with ideas I'll post the reasoning behind each of them. Altering any demension in the simulation is trivial.

The Spreadsheet is only a convenient interface method to call out to the other program components. Libre Office has a Draw program that allows direct DXF generation that I can through a macro import into FEMM4.2 The blocks allow almost any arbitrary shape to be built up very quickly. FEMM 4.2 doesn't see the blocks, it will treat them as a monolithic piece. Green is the voice coil travel area. Dark grey the magnet. Violet is the copper in this drawing.

The voice coil in this simulation is not oriented at the correct height for a symetrical flux. But it gives you an idea of what I can simulate quickly.

Voice coil design page. As you can see, this is not a fully baked design. I would never post that. But again, it shows the design possibilities.

So, over today I will try and figure out some of the dimensions of the Borreson driver. If anyone has any clues, I'm all ears.

I started out life as a Cabinetmaker. And figuring out dimensions from photos and drawings are not that uncommon. I'll try and walk you through the thought process.