Theoretical Design - Current Feed Active Dipole Speaker

[Nelson Pass]

As a practical matter, one does not need a pure current
source. The F1 for example has an 80 ohm output impedance,
which gives it a maximum gain figure with no load. 80 ohms is
enough to act like a current source in terms of the improvements
that we are looking for, and from actual experience, you typically
will want a smaller figure, say 20 to 50 ohms.

 

[phase_accurate]

80 ohms is enough to act like a current source in terms of the improvements

That would mean that we don't need pure current feedback, which would make things easier.


Thorsten

I do not oppose the ideas of current drive. I definitely think that it actually has it's merits. I just want to warn about the possible pitfalls.

One reason B&M and Silbersand are using current-drive in their active speakers is the compensation of the pole that is introduced by the driver's Lvc, and thus making the application of MFB easier. But it actually pushes the problem form the MFB loop design to the amp! For a given drive-current through the coil we will now have an increasing drive-voltage with increasing frequency (while we have a falling drive current for a constant drive voltage with rising frequency for voltage-drive). The response would in fact look as if we used voltage drive and a shelving filter in front of the amp, with the difference that we don't have to care about determining the necessary shelving filter transfer function => current drive does that automatically !
But that's the reason why we would have to take care about TIM.

The series connection of an inductance and a resistor is definitely a lowpass regarding current through it, if we introduce this into the NBF loop (which is now a current loop) it definitely reduces the phase-margin. If this is enough to destabilise the amp has to be checked for each application seperately.

Regards

Charles

 

[ThorstenL]

Konnichiwa,

One reason B&M and Silbersand are using current-drive in their active speakers is the compensation of the pole that is introduced by the driver's Lvc, and thus making the application of MFB easier.

They use it also on drivers that are not subject to MFB.

But it actually pushes the problem form the MFB loop design to the amp! For a given drive-current through the coil we will now have an increasing drive-voltage with increasing frequency (while we have a falling drive current for a constant drive voltage with rising frequency for voltage-drive). The response would in fact look as if we used voltage drive and a shelving filter in front of the amp, with the difference that we don't have to care about determining the necessary shelving filter transfer function => current drive does that automatically !
But that's the reason why we would have to take care about TIM.

Hmmm. If we want a flat response from a current drive application to a conventional speaker we need to incorporate a shelving filter. In the end, the rising response needs to be attended to ever which way and thus TIM is no more an issue as with voltage drive. I would have thought that this (and the answers to many of the other points raised ) would be blindingly obvious to any modestly "skilled in the art", as they say.

Sayonara

 

[phase_accurate]

They use it also on drivers that are not subject to MFB.

The secret lies in the word "ONE".

In the end, the rising response needs to be attended to ever which way and thus TIM is no more an issue as with voltage drive. I would have thought that this (and the answers to many of the other points raised ) would be blindingly obvious to any modestly "skilled in the art", as they say.

That does not definitely have to be so in every case. If one wants to extend upper end response by compensating for said pole then he doesn't want to sacrifice it again by another lowpass, does he ?

Regards

Charles

 

[ThorstenL]

Konnichiwa,

That does not definitely have to be so in every case. If one wants to extend upper end response by compensating for said pole then he doesn't want to sacrifice it again by another lowpass, does he ?

Depends. If your driver has a flat response under voltage drive you will have to compensate the tilted up response which results from driving the drivers voicecoil with current, unless of course you do not wish for a flat frequency response. The solution to the problem may be found in using MFB or equalisation. Equalisation is in effect what you refer to as "lowpass".

I would add that compensating the voice coil inductive pole is pretty low on my list of things that I feel make current drive worthwhile. Indeed, it would really come into play only in the context of MFB, a simple open loop driver does not really care how many poles are in the response and you can compensate each of them by open loop pre-equalisation to a great degree.

So, I still fail to see how TIM comes into this any more than with a voltage drive speaker. Again, just as the oscillation in the (wildley unlikely) case of open loop operation you need to drag TIM over by it's hair, kicking, screaming and resisting.

So please, what is actually your point?

Sayonara

 

[phase_accurate]

I just wanted to raise the awarenes of possible pitfalls, nothing more and nothing less.

BTW: The inductance-related pole of the driver that I use is definitely higher than it's total (i.e. mechanical and electrical transfer functions combined) cutoff frequency so I would not end up with an upwards tilted response but a less steeply falling one. So this would be ONE of the benefits for using current-drive with MY speakers.
Things may of course look different in any other case.

Regards

Charles

 

[ThorstenL]

Konnichiwa,

BTW: The inductance-related pole of the driver that I use is definitely higher than it's total (i.e. mechanical and electrical transfer functions combined) cutoff frequency so I would not end up with an upwards tilted response but a less steeply falling one. So this would be ONE of the benefits for using current-drive with MY speakers.

Well, the example posted here and supposedly discussed in this thread happens to use drivers where the voice coil inductance will be influential in the actual operational range of the driver. So, in the context of the thread which was aiming to discuss a CONCRETE EXAMPLE of how one might implement current drive you remarks are in effect lacking relevance, as the proposed configuration "re-compensates" the pole removal due to current drive in order to keep the midrange drivers response flat. The tweeter is another story, however again, if the usual precautions against TIM are taken (a input lowpass filter at a suitable frequency) the result is still the same for TIM, regardless of voltage drive pre-equalised for flat response or current drive pre-equalised for flat response.

I would expect slew rate induced distortion to arise in carelessly designed MFB systems, where the inertia of the driver and the available force is not acounted for correctly....

Also, I would suspect that even with your driver the voicecoil inductance is large enough to cause response alteration in the operational range, assuming it is a midrange driver or tweeter....

Sayonara

 

[phase_accurate]

assuming it is a midrange driver or tweeter....

No, it is a wide-range driver with an Lvc of 18 uH.

Regards

Charles

 

[johnnyx]

I think it's best to try it and see how it behaves. If the amp needs extra compensation, or if you need a response - compensating filter, whats the problem? It's no different to other systems in that respect, and you get the advantages of current drive. It's fun trying something different, and always educational.🙂

 

[Bill F.]

Charging paddles.... Clear!

*Bump*

😀

Thorsten, Kudos on the design. I think perhaps you've so thoroughly outlined it that few have yet thought of more to contribute, ergo the silence.

Just thinking from the hip here: One way to open up the options for usable bass drivers might be to use a U baffle with higher-Qm drivers and stuff the back to taste.

You could even use an H baffle and stuff both in front and behind the driver. This would take some more experimentation as it would also supply an acoustic low-pass.

 

[Svante]

We can use "The Edge" (how about stuffed crust?) to get a general idea what the SPL distribution will be like, the recent version can handle 4 Woofers on a narrow baffle, but still does not account for the floor image.

"The Edge" is here:

http://www.tolvan.com/edge/

Sorry, I saw this post a bit late, and admit that I have not read the whole thread. I have two short comments:

Playing with the output impedance of the amplifier is a classic way to modify Qts, and for the open baffle a positive resistance probably would be beneficial in many cases, since this raises Qts. It does not have to be infinite (as in current drive), using a combination of current and voltage feedback. In the case where you want to lower Qts, you could even make the output resistance negative by using positive current feedback.

Enough about that. The floor image can actually be simulated in Egde, simply by creating an image. Ie, make the baffle twice as high and symmetrical around the "floor level". Something like this:

An externally hosted image should be here but it was not working when we last tested it.

The floor would be at y=1000 mm. The part below 1000 mm is the reflected image of what's above 1000 mm. The actual baffle is 600x1000 mm.

 

[ThorstenL]

Re: Re: Theoretical Design - Current Feed Active Dipole Speaker

Konnichiwa,

Playing with the output impedance of the amplifier is a classic way to modify Qts, and for the open baffle a positive resistance probably would be beneficial in many cases, since this raises Qts. It does not have to be infinite (as in current drive), using a combination of current and voltage feedback.

But we would LIKE to have an infinite output impedance, to take advantage of the various improvement in performance this brings....

Enough about that. The floor image can actually be simulated in Egde, simply by creating an image. Ie, make the baffle twice as high and symmetrical around the "floor level".

Yes, but IMHO this overstates the amount of LF gain.

Sayonara

 

[454Casull]

resurrection is complete

 

[annex666]

Re: Re: amp stability

If the "Wankel Motor" (rotary engine) where a superior design, offering less fuel consumption with more power from a much smaller and cheaper to make engine, with much less operational noise and vibration surely everybody would be using it?

...I'm not so sure.

The engine DOES infact offer the gains that designers claim (provided it is correctly designed), however when the design was first proposed metalurgy was not as advanced as it is now and the lobe tips would wear excessively (reducing reliability and increasing maintenance costs). This problem is avoided in a conventional (reciprocal) engine by the use of piston rings that are sprung to maintain a constant contact force with the cylinder lining - and thus preventing a complete engine strip and rebuild every x thousand miles.

Recently Mazda has (re)introduced the rotary into its newer models and have received aclaim for doing so, no doubt the fact that it is not as widespread as it could (should?) be can be put down to the design rights held by Maxda (I'm not sure how widespread their "ownership" of the design is) and consumer acceptance.

..or maybe this is a moot point as hybrids are threatening to take over the world (some time in the next millenium) - of course that's assuming we haven't harnessed a more powerful energy source by then - zero-point energy anyone?...

 

[Giarsun]

...the CA18RLY as 4pcs Woofer.

The woofers have a Fs of 40Hz and a Qe of around 2. In other words, fed with "constant current" they will exhibit a 4.5db lift at 45Hz. This lift can be used to compensate part (or all) of the dipole rolloff, below 40Hz however the SPL will drop like a stone at a rate faster than 18db/8ve.

Using 4 Drivers may give enough displacement to use a Linkwitz transform circuit to shift the LF cutoff down, however I feel I would suggest as preferable a sealed box subwoofer to be cut in,

Beyma 10BR60 has almost identical parameters : Fs = 45 Hz, Qms = 1.9, Qts = 0.47.

A single unit would provide approximately the same Vd as four CA18RLY units. Maybe with two 10BR60 one could avoid the need for the sealed box subwoofer ? Baffle would be somewhat wider but could be a little bit shorter.