Optimal bass driver size for sound quality - is bigger really better?

Status
Not open for further replies.
Jump to Latest
#81
thoriated said:
I attenuate LF Out of band information with a large series polypropylene capacitance at the input to the ported Iron Law Breakers (tuned to about 30hz). As a side bonus, it interacts with other xover elements and the speaker's natural ported upper impedance peak to neutralize that peak at the cabinet terminals and gain another couple dB of output in the half octave centered around that frequency. when I write 'neutralize', I mean the peak is gone since the impedance at the Iron lawbreaker terminals is 18 Ohms +/-20% from 40 hz to over 200hz without any Zobel type networks. The impedance drops to a minimum of 12 ohms around 31hz before rising below due to the series input C. A very friendly LF load for my DC Coupled OTL.
Click to expand...

Just out of interest what is the capacitance value?? I use that technique with a large sealed Kicker auto sub ( 15inch / 200 liter box ) and had to wire multiple caps to get the 1400uF + calculated for the nominal 4R load
 
#82
pos said:
I was thinking of using a HP just under that low tuning frequency to prevent overexcursion.
Click to expand...

But now it's not an apples to apples comparison. At any rate, its all room dependent at these frequencies, so there isn't a right or wrong answer - it all depends. Honestly, I don't obsess about LF speaker design. Use anything and "correct it in the mix". Then start worrying about what does matter - everything else.
 
#83
Moondog55 said:
Just out of interest what is the capacitance value?? I use that technique with a large sealed Kicker auto sub ( 15inch / 200 liter box ) and had to wire multiple caps to get the 1400uF + calculated for the nominal 4R load
Click to expand...

Perfect application for JBL's "Charge Coupled" technique. This is series electrolytics with a (typically) 9V bias applied to the center terminal via a battery and high value resistor. This is intended to linearize operation of the caps and allow good performance from practical size+cost capacitors for bass applications (in particular).
 
#84
I'm using 600 uF. Being that it's a nominal 16 ohm system, the capacitance value is more or less half what would be needed for an 8 ohm version. An Ebay seller is still letting 600uF 500Vdc polypropylenes go for $10.00 each or better & those would be useful for this application - in fact I picked up a bunch already for future speaker projects. I was hesitating at first because they're pretty sizable, but even with the 500Vdc rating, they turn out to give nearly as much capacitance per unit volume as a 330uF 250Vdc Solen polypropylene does (at nearly $100/pop), which is one of the caps I use in the Iron Law Breaker xover (in parallel with 2 100uF and a 70uF) for the series input capacitance.
 
Last edited:
#85
badman said:
Perfect application for JBL's "Charge Coupled" technique. This is series electrolytics with a (typically) 9V bias applied to the center terminal via a battery and high value resistor. This is intended to linearize operation of the caps and allow good performance from practical size+cost capacitors for bass applications (in particular).
Click to expand...

I understand this concept, but doesn't it fail to work if the voltage swing > +-9.0 v? I am guessing that the idea is that at these voltages you don't much care since the distortion is masked.
 
#86
have to look for that cap - lots easier than paralleling 50x12uF airconditioner caps - that can be a rough load for some amps. Graham Maynard has a huge thread on a transfomer boost circuit for open baffle but I don't see how he can get a free/painless lunch unless the impedance dip frequency is centered at the driver/baffle Z peak. Can't find that cap - any hints?

re:best bass - bandpass vents can have small area - can mid-upperbass BP be made which sound good for home use in conjunction with LF BP?

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

An externally hosted image should be here but it was not working when we last tested it.
 
Last edited:
#87
Hi, Freddy -

Good to hear from you. In Ebay, you can type 'ASC' and '600' in their search box and it'll get you there. I've got over 2 dozen, and they all measured out in spec, but I haven't put any in circuit yet. I saw a post somewhere where somebody said they cut and sanded the top mounting stud off of them without problems & they'll also chassis mount with the same hardware as a 3" diameter computer lytic cap


I plan to use a bunch in HT subwoofer modules and some for my 'ultimate' speaker where I was also planning to use JBL's 'charge coupled' idea, but I was thinking of going 'green' by using miniature solar panels (like for solar powered calculators) instead of 9V batteries for the bias voltage🙂
 
#88
I understand this concept, but doesn't it fail to work if the voltage swing > +-9.0 v? I am guessing that the idea is that at these voltages you don't much care since the distortion is masked.
Click to expand...

As long as the total capacitive reactance is significantly lower than of other elements in that current leg, AC voltages quite a bit higher than the bias voltage can be sustained without the bias voltage across either series cap reversing polarity.
 
#89
thoriated said:
I'm using 600 uF. Being that it's a nominal 16 ohm system, the capacitance value is more or less half what would be needed for an 8 ohm version. An Ebay seller is still letting 600uF 500Vdc polypropylenes go for $10.00 each or better & those would be useful for this application - in fact I picked up a bunch already for future speaker projects. I was hesitating at first because they're pretty sizable, but even with the 500Vdc rating, they turn out to give nearly as much capacitance per unit volume as a 330uF 250Vdc Solen polypropylene does (at nearly $100/pop), which is one of the caps I use in the Iron Law Breaker xover (in parallel with 2 100uF and a 70uF) for the series input capacitance.
Click to expand...

Thanks, never considered Ebay for a capacitor source!

Could you explain your " Iron Law Breaker xover "?
 
#90
gedlee said:
I understand this concept, but doesn't it fail to work if the voltage swing > +-9.0 v? I am guessing that the idea is that at these voltages you don't much care since the distortion is masked.
Click to expand...

I believe much of the idea is around reducing crossover distortion (or whatever you wish to call capacitor misbehavior around the zero crossing). It would take a fairly high level of output (with sensitive speakers) to exceed this, as we're talking about a +9V center, so >18V P-P swings. Even then it would seem to be an improvement up until that level was sustained, which wouldn't be the case most of the time for all but high powered PA setups.

Sure seems like in a home environment you'd be nuts to use that much juice... unless you're driving big ATCs or something similarly ludicrously inefficient that can handle that much juice.
 
#91
thoriated said:
I'm using 600 uF. Being that it's a nominal 16 ohm system, the capacitance value is more or less half what would be needed for an 8 ohm version. An Ebay seller is still letting 600uF 500Vdc polypropylenes go for $10.00 each or better & those would be useful for this application - in fact I picked up a bunch already for future speaker projects. I was hesitating at first because they're pretty sizable, but even with the 500Vdc rating, they turn out to give nearly as much capacitance per unit volume as a 330uF 250Vdc Solen polypropylene does (at nearly $100/pop), which is one of the caps I use in the Iron Law Breaker xover (in parallel with 2 100uF and a 70uF) for the series input capacitance.
Click to expand...

Do you have a link?? My search produced nil result, at that price i'd be willing to pay postage from the US
 
#92
Hi -

I wanted to see what I could do with an air core transformer because I'd never heard of anyone using one for an analog application, especially for a speaker xover. I started with a bobbin wound 8mH 12awg air core inductor and overwound it with an approximately 150mH 'secondary' winding using approx. 20awg wire. The 8mH 'primary' is/was connected in a conventional series choke configuration immediately after the 600uF input capacitance.

An essential feature of an air core transformer is that the windings couple only over a fairly narrow range of frequencies which makes them pretty useless for most audio (but great for RF, etc), but I wanted to see if I could turn it to advantage. I found (very empirically), that with a small series RC network in series with the secondary (and with the secondary and this network shunt connected across either the input or output of the primary winding (I forget which) to ground) with the proper polarity tuned to the upper/bass/lower midrange that I could shape the frequency response to compensate for baffle step by causing a mild dip at the baffle step corner and an octave or so above AND boost the response by a couple of db below the baffle step corner. The unique feature is that this shaping appears to be a true transformer action with almost no resistive loss (the series resistor mostly keeps an excessive dip from occuring above the baffle step corner over a very narrow range of frequencies - perhaps a quarter of an octave. So both useful frequency response shaping (of about +2 -4db max amplitude change) and impedance transformation is occuring with this three element network between about 50 hz to 250hz. As I recall, the values for the series RC network i came up was with the C not too far from 5uF and the resistor close to 100 ohms.

Above that frequency, the air core transformer acts essentially like a conventional series air core inductor which in conjunction with an LC shunt across the woofer VC forms a low order LP elliptical filter which I have previously found that I prefer due to its simplicity and the ability to cancel out of band driver breakup modes very effectively with its null.

The 600uf input cap at the input acts more or less as I described above but appears to interact also with the speaker upper impedance peak and the air core transformer to some extent between 40-120 hz, giving a bit of additional boost in that range, as well as cutting the subsonics to the driver.

This all works out very well with a very low Qts driver like the JBL 2220A tuned to 32hz (the Iron Lawbreaker is about 100liters but is also about 30% filled with activated charcoal in airtight baggies for significant 'compliance enhancement' as per a KEF white paper) since the xover above augments the normally attenuated bass response up to over 100hz by up to 3-4 db while cutting the emphasis above 120 hz due to baffle step and rear wall reflections by a similar amount and keeping the impedance within a narrow range (a somewhat unexpected benefit). The presentation is extremely detailed with good balance and reasonable weight down to the low 30hz region, and due to the high efficiency down to 60 hz or so has no trouble playing very cleanly at live/near live levels in a 17 x 25 foot room with a moderately powered amplifier of 100W or less.

Because the box tuning is a third of an octave below 40hz, there is very little audible BR 'overhang', which I don't like the sound of in a system tuned much higher. If I listen for protracted periods of time, occasionally I think I can start to hear a very slight coloration that might be related to the action of the air core transformer, but it could be a number of other things, also, and considering the unique benefits that the air core transformer provides in this application, it is well worth it.
 
Last edited:
#96
badman said:
It would take a fairly high level of output (with sensitive speakers) to exceed this, as we're talking about a +9V center, so >18V P-P swings. .
Click to expand...

Back-to-back polarised electrolyic capacitors, biased to 9 volts via the JBL scheme, will allow much more than 18 volts P-P across the pair under most conditions without exceeding the capacitor bias.

A polarised capacitor is conceptually two capacitors and two diodes: one capacitor with the rated capacitance and voltage, and one with a much higher capacitance but much lower voltage rating. Conceptually, it looks something like this:


.....+-----|>|-----||-----+
-----+....................+-----
.....+-----|<|-----||-----+


Current flowing one way goes through the rated capacitance and voltage capacitor, current flowing the other way flows through the higher capacitance but lower voltage rated capacitor.

For example, imagine the rated capacitance is 100uF, 50v working. Imagine the "reverse" capacitance is 10,000uF, 0.5 volt working. Assume a load resistance and frequency where the AC voltage developed across the "forward" capacitor is 50 volts peak. On the "forward" half cycle, the voltage across the capacitor will vary between 0 volts and +50 volts. On the other half cycle, the voltage will vary from 0 volts to -0.5 volts. The impedance of the "reverse" capacitor is only 1/100 that of the "forward" capacitor, so the voltage across it will be 1/100 that too. Since the voltage across the "forward" capacitor is 50 volts, the voltage across the "reverse" capacitor is only 0.5 volts, within its rating.

Obviously, you don't want such a nonlinear device in your circuit. that's why you place two such capacitors back-to-back. Applying a bias prevents the "internal diodes" from switching.
 
Last edited:
#99
thoriated said:
As long as the total capacitive reactance is significantly lower than of other elements in that current leg, AC voltages quite a bit higher than the bias voltage can be sustained without the bias voltage across either series cap reversing polarity.
Click to expand...

Good explaination, I understand now. But ther actually are a lot of circuits were the capcitive reactance is the dominate factor, at least in some frequency regions. One would have to look at the situation in some detail.

How do you know when to change the battery?
 
Status
Not open for further replies.
Top Bottom