Series resistor for horn driver

[fpitas]

I'm using TD-2002 drivers in DIY speakers using an active crossover. They were connected directly to the horn amp, but I noticed that the horn seemed to "stick out" a little at times, especially when reproducing female vocals. I tried a small series resistor from the amp to the horn drivers, and with about 2.7 ohms that fixed the problem. I'm happy that the solution was so easy and inexpensive, but I would like to understand the theory. My guess is that I'm "tuning" the decay time (waterfall) of the horn drivers to better match the woofers in the crossover region.

[Earl K]

I'm using TD-2002 drivers in DIY speakers using an active crossover. <<<<snip>>>>
Specifically, which Lansing related company makes this driver ?

:)

[HCSGuy]

By adding a series resistor to the horn driver, you simply dropped the level of the horn driver - same as if you turned it down from your active crossover or the gain control on your amplifier. A resistor by itself has no time component.

[4313B]

Specifically, which Lansing related company makes this driver ?

:)Altec Lansing aka James B. Lansing Sound aka Technical Audio Devices Laboritories. ;)

Someday someone should draw a family tree of this stuff. :rotfl:

[Ruediger]

... especially with the different Q factors, Qe, Qm, Qt and Re.

The resistor modifies the Qt of the driver, makes it greater. If that is SIGNIFICANT here is another question.

The modified Qt lowers the driver's bandwidth. If that is SIGNIFICANT here is another question.

Ruediger

[toddalin]

Instead of a series resistor, how about trying a 20 ohm resistor in parallel with the driver like JBL does it?

[frank23]

Instead of a series resistor, how about trying a 20 ohm resistor in parallel with the driver like JBL does it?

This only has an effect if the driver is crossed over passively. Then a parallel 20 ohm resistor levels out the impedance peaks of the driver and lessening their impact on the crossover function.

Connecting it in series though renders the output impedance of the amplifier useless I gues, lessening the dampening effect of the amp, thereby changing the sound?

[fpitas]

Specifically, which Lansing related company makes this driver ?

:)

Point taken, although I understand it's based on a Bart Locanthi design. Plus, I didn't know where else I'd get an answer.

[fpitas]

By adding a series resistor to the horn driver, you simply dropped the level of the horn driver - same as if you turned it down from your active crossover or the gain control on your amplifier. A resistor by itself has no time component.

I've been tweaking the horn level to be constant as I experiment.

[fpitas]

Instead of a series resistor, how about trying a 20 ohm resistor in parallel with the driver like JBL does it?

I added an 18 ohm resistor directly across the driver, just so the amplifier doesn't see such a widely varying impedance over frequency. Other than having to change my series resistor value to get the same Thevenin impedance seen at the driver, I didn't notice much change from that addition, but it makes me feel better :)

[Earl K]

I'm using TD-2002 drivers in DIY speakers using an active crossover. They were connected directly to the horn amp, but I noticed that the horn seemed to "stick out" a little at times, especially when reproducing female vocals. I tried a small series resistor from the amp to the horn drivers, and with about 2.7 ohms that fixed the problem. I'm happy that the solution was so easy and inexpensive, but I would like to understand the theory. My guess is that I'm "tuning" the decay time (waterfall) of the horn drivers to better match the woofers in the crossover region.


- I'm aware of the opinions ( typically offered up ) over at DIY ( & H.E. ) of the necessity of adding a series resistor to TAD drivers ( when bi-amping ) .

- FWIW, IME it is desirable ( when biamping ) to do most level matching( & EQing ) in the passive domain . Doing so, helps assure ( that dynamically ) the separate components will play as one ( ie; work, in a much more coherent fashion ) . That circuit resistance alters system dynamics is far from universally accepted ( oddly enough ).

- A (very) primitive formula that I use to index transducers ( according to their "dynamics potential" ) is ;

DPI = ( BL2 / Re ) / ( Sd / Mms ) x .5

where;
Sd is in sq inches
Mms is in grams
.5 is simply an arbitrary constant added into the formula to keep the derived numbers under 100

( for example ; a 2235H is @ 32 while a 2226 sits around 51 and a 2220H comes in at @ 88 )

- Once one gets used to using this formula , one may start to understand how a transducers dynamic behavior is impacted by something as simplistic as resistance ( hence this answer to your question ) .

- So simply put ( IMHO ) , level matching ( & EQing ) a 110 db efficient horn-driver combo to a 95 db woofer should be done passively ( even when bia-amping ) if one wants to rebalance the dynamics of the components being used .


:)

[fpitas]

Thanks Earl.

[Hans Bleeker]

- So simply put ( IMHO ) , level matching ( & EQing ) a 110 db efficient horn-driver combo to a 95 db woofer should be done passively ( even when bia-amping ) if one wants to rebalance the dynamics of the components being used .

That Throws overboard the biggest bennefit of active filtering IMO...
I made an all passive version of the 4355 filter, sounded great, but to check where it realy ended up I went back to a full active version, same components, night and day difference, more dynamic, cleaner less constraint.
So for now I prefer the methode of using a way bigger amp on the lowest eff driver, most of the time the low unit, in my case a 2235. Second option, using the volume controls on the amps.
But im very well aware of the difficulties to get a horn matched properly to a cone driver , and it might be the reason I go back from a 2352 horn to and exponential version so I dont need any more passive stuff in front of it to correct it :)

[fpitas]

I have to think the driver designer assumed the driver would be connected to a resistive attenuator with an output impedance of a couple ohms, and planned on that.

[Ruediger]

A (very) primitive formula that I use to index transducers ( according to their "dynamics potential" ) is ;

DPI = ( BL2 / Re ) / ( Sd / Mms ) x .5

where;
Sd is in sq inches
Mms is in grams
.5 is simply an arbitrary constant added into the formula to keep the derived numbers under 100

( for example ; a 2235H is @ 32 while a 2226 sits around 51 and a 2220H comes in at @ 88 )



What is the motivation or reasoning for this formula?

curious
Ruediger

[4313B]

Point taken, although I understand it's based on a Bart Locanthi design. Plus, I didn't know where else I'd get an answer.Yeah, after all these years I've finally come to the conclusion that it really doesn't matter who signed their paychecks. Altec, JBL, TAD, whatever...

[pos]

What is the motivation or reasoning for this formula?

curious
Ruediger

It looks like it is loosely related to the mass breakpoint frequency formula:

f = BL^2 /(Pi*Re*Mms)

This is the frequency above which the SPL will rolloff at a 6dB/oct rate if a constant directivity is maintained.
It is typically around 3kHz for a compression driver, and in the realms of 500Hz for midbass woofers.

It only impacts high frequency response though (think 2234H vs 2235H)

[Earl K]

What is the motivation or reasoning for this formula?

curious
Ruediger


Hi Ruediger ( Pos ),

The dynamic indexing formula ( as I see it ) is a product ( of sorts ) for displaying potential energy transfer ( & really, it's not very different from simply using Sensitivity ) .

I use this formula ( as well as the EBP product, ie; mass break point ) to identify & compare different woofers to the known entities that I've used for decades( such as 2225 & 6s ) .

I realize that for most, these numbers might be meaningless / but for those of us with a few decades experience ( with specific Pro products ) the numbers help me relate to how JBLs' competitors ( such as all those Italian speaker manufacturers ) might perform ( if purchased ).

Futzing about with EBP & this DPI index, also allow me to quickly execute a bunch of "what if scenarios" ( for altering a systems over-all voicing ).

Running these numbers do show ( for instance ) why someone ( GT ? ) bothered to create separate 2108 ( and 112 transducers ) as well as the 2 distinct classes of Le5-x, 5" mids ( I'm sure you can guess which have the lower dynamic numbers ) .

:)

[frank23]

I will have to re-read this formula thing a few times. It seems intriguing.