GC with Active Subtractive XO

[jackinnj]

Originally posted by moamps
Hi,

Maybe is good idea extracting posts about subtracting XO in new thread. Here are offtopic, IMO.
Also, in EW, february 2000 is published article "Adaptable active loudspeaker". It's three way active system with subtractive XO and bass EQ. Amps are TDA1514.

Regards
Milan

Milan, let's just start here.

Originally posted by richie00boy
I too am interested in this subtractive crossover. On first sight it looks as though the two combined signals cannot sum flat, but maybe I need to investigate it more.

Kuei Yang Wang, I am particularly interested in why you recommend 3rd-order as the optimum.

Digressing slightly, can 3-way crossovers be made with this topology?

The article which Milan cites, written by Christof Heinzerling, and he does a very thorough job -- it also incorporates bass equalization. I was tempted, but won't post the schematic as I dont' want to irritate the EW folks.

[phase_accurate]

Why would anyone want an active subtractive crossover ?

Fun aside: It seems as if this has been split from another thread, am I right ? If so, which one ?

Regards

Charles

[jackinnj]

Originally posted by phase_accurate
Why would anyone want an active subtractive crossover ?

Fun aside: It seems as if this has been split from another thread, am I right ? If so, which one ?

Regards

Charles

it was split from the Industrial SMD GC thread.

[phase_accurate]

Within this thread you may find some info about active subtractive crossovers.

Active Subtractive XOs

BTW: subtractive phase_accurate crossovers can be made with any desired order and both, symmetrical and asymmetrical. Though only the asymmetric 1st/2nd, 1st/3rd and symmetric 2nd/2nd seem to make sense. The other ones have too large humps and too much overlap.

Edit: You may of course use the IC-power amp as part of a subtractive x-over.

Regards

Charles

[Nuuk]

That article by Christof Heinzerling makes for good reading for any Gainclone fan is an excellent blueprint for a full-range active speaker system.

It was originally published in the February 2000 edition of Electronics World for anybody who wants to look for it.

[moamps]

Any comments?

[phase_accurate]

While this would theoretically work, you don't gain that much. You still use the LF amp fullrange.
If you want the benefit of active solutions (lower IMD, better cone control, larger power margin etc) you will have to subtract at the input of the LF amp, i.e. you have to use it as a differetial amp.

Better: Use a separate OP-AMP for the highpass, because clipping of the HF amp won't affect the LF channel.

Regards

Charles

[moamps]

Hi Charles,

first off, thank you for your response.

The basic idea here is to build an active two-way system using as few components as possible and, at the same time, keep all the comparative advantages of a subtractive crossover. So far, the test version with 2xLM3875 and the cross-point at 2,5kHz has performed pretty well and encouraged me to start building a simple, hassle-free system that anyone would be able to build and enjoy (despite its minor shortcomings, if any). Having built a few standard two-three-four way subtractive crossovers (a combination of a standard crossover, delay circuits and differential opamps), I am well familiar with them by now and that is precisely why this time I wish to explore other possibilities of crossover implementation.

The LF amp I am using is not current-driven in the HF range so I expect intermodulation distortion to be smaller than were the amp fully loaded in the entire range. Cone control is certainly better than in case of passive implementation due to the fact that there is no passive crossover between the amp and the speaker. Power margin is of little importance here as we're fiddling with only GC chips . Subtraction for LF can hardly be done at the input because the HF output has already been boosted by ca 20dB but I am going to check this in the next couple of days. HF amplifier clipping alone indicates that the HF speaker is dead or has been seriously damaged so there is no point in discussing how much of this distortion affects the LF speaker. Besides, if you look at the harmonic distortion spectrum resulting from HF amplifier clipping, it is quite clear that this distortion has no effect on the LF range whatsoever. However, there is no vice-versa.

Additional opamps would not be welcome here...perhaps some other time.

Regards,
Milan

[jackinnj]

Originally posted by phase_accurate
While this would theoretically work, you don't gain that much. You still use the LF amp fullrange.
If you want the benefit of active solutions (lower IMD, better cone control, larger power margin etc) you will have to subtract at the input of the LF amp, i.e. you have to use it as a differetial amp.

Better: Use a separate OP-AMP for the highpass, because clipping of the HF amp won't affect the LF channel.

Regards

Charles

a good point with the LM3875, LM3886, LM4780 -- I thnk that the SPIKE protection kicks in before clipping, however -- see application note 898.

[swak]

Here is a copy of Christof Heinzerling's article which can be orderd online.

http://www.softcopy.co.uk/mag/defaul...020105&count=1


Is there a schematic for a GC with HP / LP and "rising response" correction? Can this be implemented on the current drive version?