X max problems? Possible solution

[consort_ee_um]

How many times have you modelled a design in Unibox or Winisd only to find you exceed the X-max limits? My obsession with getting good bass out of small boxes has given me some ideas on this problem. See my blog here
Consort3’s Blog

[gfiandy]

Have you actually tested this, as the values required tend to leave the amplifer on the edge of stability. I notice that your design has a gain of 10 which is the abs min that these chips recomend. The addition of other poles in the response may cause the chip to go unstable.

I have made this work in the past but found that the design was not unconditionally stable and would go off for certain capacitance on the output, it also had fairly poor damping characteristics when fed with a square wave as it tended to ring. Chip amps arn't really designed with this sort of application in mind and most don't have the open loop gain and phase magin to make this work well.

If you are designing your own amplifer it wouldn't be that much more work to add an opamp on the input and do it in a controlled maner.

Regards,
Andrew

[LineArray]

WinIsd AFAIK does not show drivers excursion dependent from
frequency. This is a major drawback, since the excursion's dependency
from frequency is essential for a BR alignment to give useful results.

The main advantage of BR over Closed Box is the reduction of excursion
around fb thereby giving more freedom in the compromise between

undistorted max SPL
lower frequency limit
imulse response
volume of cabinet

A software not showing excursion is simply deficient and may lead to
results where someone tends to "over stretch" the sane ranges of
lower frequency limit. A speaker with Qts too high, fb too low in a huge
cabinet might show a sweet FR graph in some alignments, but will suffer
from excessive excursion above fb .

IMO such a misconstruction cannot be healed by equalizing, you can only
reduce the LF input. To make a proper alignment of the box is the more
effective way.

High pass filtering the input signal to avoid unnesserary excursion somewhat
below fb might be useful to increase dynamic headroom for some
alignments ("subsonic filter").

Kind Regards

[gfiandy]

Hi,

I think WinISD does report frequency dependant excursion. I have attached a JPG of an example of a system running at 50W input with a 2nd order peaking filter High Pass with the peak placed on the port resonance.

You can see how this controls the excursion below resonance and if I included the SPL as well you could see how it extends the bass response.

Regards,
Andrew

[LineArray]

Seems i had an older version ...

[event horizon]

Originally Posted by gfiandy

Have you actually tested this, as the values required tend to leave the amplifer on the edge of stability. I notice that your design has a gain of 10 which is the abs min that these chips recomend. The addition of other poles in the response may cause the chip to go unstable.

Regards,
Andrew

Hi Andrew, i think it could be made stable as long as there was enough current delivery available from the feedback network. This would be something similar to Douglas Selfs use of a resistor & capacitor to bootstrap the input & increase AC impedance whilst keeping noise (& offset voltage) low by using low value resistors. The two resistors on the feedback network just need reducing in value to supply more current & stability should be ok.

None of the poles are within the amplifier itself so that should remain stable as long as the above is taken into consideration. I'll admit i don't know how much the minimum gain is allowed on chip amp "X" but as long as it's above the minimum i see no problem. It'd be just like an opamp filter on steroids

I could be wrong & it wouldn't be the first time, so don't take the above as gospel as i can't test it

Bests, Mark.

Originally Posted by LineArray

Seems i had an older version ...

WinISD pro doesn't half help

[consort_ee_um]

All,
Thanks for your comments. The filter technique is described in ST's datasheet for the TDA2030A. In my experience, the best solution to the problems outlined by Linearray is this:
http://www.xlrtechs.com/dbkeele.com/...Alignments.pdf
Where Keele proposes a 6dB lift for the filter at the port tuning frequency
Winisd actually has that filter built in to the filter options.

[thoriated]

How many times have you modelled a design in Unibox or Winisd only to find you exceed the X-max limits? My obsession with getting good bass out of small boxes has given me some ideas on this problem.

I once built a 4th order BP filter (really a fourth order high pass and a fourth order low pass at a somewhat higher frequency) with a single op amp section using multiple feedback loops. It didn't seem to have any problems with stability, but I began running into limitations with the amount of pass band gain I could squeeze out of the single op amp.

[CharlieLaub]

I am assuming that you are talking about vented boxes (sine you didn't mention it). There are a whole slew of vented alignment that combine a second order HP filter with a vented box (4th order system) to get an overall 6th order system. There are also some "quasi 5th order" alignments using the same approach. Both of these will attenuate excursion below the box tuning frequency, and thus keep your woofer from killing itself.
Quasi Butterworth 5 alignments
A.Thiele Vented Boxes Part I (see Table 1)

A third order HP does not really help much more than second order, and you get slightly worse transient response because of the higher order system.

IMHO controlling excursion below tuning is needed for vented boxes with Fb above about 25-30Hz. Below 30Hz, non-organ music (e.g. pop, jazz) just does not have enough content to really supply any energy there so you probably won't be overdriving the cone. If you are listening to home theater/movie content, that is another story and I would recommend a large sealed box (Vb>Vas) and a large driver (e.g. 380mm/15" diameter) with low Fs, high Xmax, and moderate Qts. This will give you lots of low excursion output.

-Charlie