XSim: 3rd and 4th order blocks?

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Hi Everyone,

Sorry if there is a rule of thumb I'm missing. I use Xsim a lot and notice that there's no 3rd or 4th order circuit blocks.

Is there an easy rule of thumb to combine 1st and 2nd order filters into higher orders? Like, if I wanted a 4th order LR do I add a pair of 2nd order somehow?

Thanks


Erik
 
Eriksquires,

I don't know about XSim, but functionally, I think it should be the same as what I'm using: Active Crossover Designer. software

Even if you don't want to use the software, download the ACD zip so that you can read the manuals, there's a table that answers your question and more:

"LR4 crossover requires two second order filters, each with a Q=0.707. The filter corner frequency is equal to the crossover point for LR type crossovers". Quote, Charlie Laub.
 
Oh I see, that would be convenient. :)

I just think it's super nice of Charlie to have all that software and information available to us. Before I discovered his site, I didn't know if that for Bessel, or BW4, or LR8 etc, just add these xyz etc filters. I love the table showing what all the Q needs to be, etc.

Does XSim do the same thing, crate Biquads that can be used by miniDSP or the like?

Cheers,

Alex
 
A 3rd order crossover is only a second order one preceded with or followed by a first order one. A 4th is two scond order ones in cascade. They aren't offered as separate blocks because beyond 2nd order there's no simple way (like a value for Q) to specify them. So you can configure topologies (at least) by making them from 1st and second order blocks.

The apparently obvious choices of letting you specify as Butterworth, Chebychev, Bessel, etc. wasn't done mostly because crossover filters are always applied to drivers that already have multiple orders of filtering already in their responses - specify whatever electrical named response, you won't get it really since the driver is already filtering before you do. So it didn't seem worth much of anything, not even considering the significant effort it would take. (Though has anyone noticed that you can create your own Circuit Blocks in Xsim if you can get a formula for the component values? ... gets pretty ugly for complicated things, though).
 
Bill,

Thanks so much for getting back to me. This makes sense, but I'm wondering if I am missing something. I get what you mean about crossover topologies used with real-world drivers not making a lot of sense. My main question was, am I missing something if I try to create a filter of order n>2? For instance, given an ideal driver, if I wanted to start with a 3rd order filter at 2kHz, do I just use a 1st and 2nd order Circuit Block with the same corner F?

Looking at the handy dandy online calculator from V-Cap for 4th order LR filters for instance, it seems that cascading two identical 2nd order filters is not what I should do. I seem to be missing part of the recipe after the 2nd order filters.

My main goal is really to use the Circuit Blocks as long as I can before I start to tweak individual values. Having 2 circuit blocks to tweak is still much better than 4 individual components every time I want to nudge the corner F up or down.

Thanks!


Erik
 
No, the component values will interract too much for nice rules of thumb to help much, sorry. But I'd place the first order section (if odd order) and 2nd order sections all down and play with nominal impedance, frequency and Q values (don't take the ohm, Hz and Q values literally then, just use them as knobs on each section to move things around with a little more stability than tuning RLCs individually) till you get somewhat close...then convert to separate components and hand tweak those).
 
No, the component values will interract too much for nice rules of thumb to help much, sorry. But I'd place the first order section (if odd order) and 2nd order sections all down and play with nominal impedance, frequency and Q values (don't take the ohm, Hz and Q values literally then, just use them as knobs on each section to move things around with a little more stability than tuning RLCs individually) till you get somewhat close...then convert to separate components and hand tweak those).

So, alternatively, could I examine the impedance for the new circuit, and then create a new block based on that?

For instance, I add a cap to a tweeter. The impedance at 2kHz now goes from 8 ohms to 32. Now could I add a 2nd order filter, with an impedance of 32 ohms at 2kHz? Would that get me in the ballpark?

Best,


Erik
 
not really, its a filter. Impedance at one frequency doesn't specify a lot. Easiest and probably best is to just put the structure down (assuming you know what it looks like) and twiddle knobs to get near what you want. Easier than trying to calculate or approximate, you'll quickly get a "feel" for it. Then after converting to parts, you can also insert resistors in series or across Ls and Cs to smooth out usually.

Circuit Blocks are subnetworks for which component values can be defined by the designer via formulae of other parameter values. You can put yourself in circuit block development mode to play if you want (the tool isn't very forgiving of errors though, but I was able to use it to make the blocks that are included in Xsim - not for the fainthearted and the tool could stand a lot of polishing).
 
Founder of XSA-Labs
Joined 2012
Paid Member
PCD lets you select with a clicker button various orders and named filters starting from BW1 etc as "Target" functions. The targets are overlays on your real time response. Then the combination of the intrinsic filter function of the driver and whatever circuit blocks you start adding or tweaking the drives you to the target function. In reality I hardly ever find that a true electrical LR4 is needed to achieve an electro-acoustic LR4. You will use perhaps only electrical BW2 and combined with driver you get LR4. The reason the textbook targets are nice is because if you can match them, they follow the behavior of an LR4 (LR2, BW3, etc) as predicted.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.