The down side to adding series resistance to horn drivers

[AllenB]

In regards to using series resistance to tweak the response of a horn for a boost around the impedance maxima, or where a high Zout amp is used... since adding this resistance works against the low Qes and high EBP wanted from a horn driver, will it contribute to poor performance/sound?

[CLS]

There would be at least two impedance peaks in most driver-horn combos within their operating range. Given that most amps have very low output impedance, any additional resistance in series with driver (in horn) forms a voltage divider.

So the frequency response is inevitably altered. Can't say it's good or bad by that alone, though. OTOH, many had critisized the effect of series resistance being 'sucking life out of the sound'. I've heard that effect too and don't like it, either.

Interestingly, 2~3 Ohm of Zout (typical tube amps) doesn't affect sound quality as those external resistors in series with drivers. Maybe mental effects, I'm not 100% sure.

In applications of midrange and above, I think it's OK if means are taken to flaten the impedance peaks first. For bass, that would be a huge waste of efficiency.

[AllenB]

Yes, sucking the life out of it is a term I wish I could quantify. I thought about the change in driver damping causing a change in the horn effect, so I took a case I'm working on and looked at the horn gain versus direct radiator response, each with and without a good series resistor...the horn gain was the same in each case (assuming hornresp handles this accurately).

Which may be unfortunate for the other thought I had about shorting across the driver to increase damping (maybe with a resistance the same as Rs). Efficiency is a secondary issue, although it's always good, with 105dB/W/m there's room to move.

It's not as if I can't do it all actively, but I'm also concerned about using a suitable power amplifier.

[mashaffer]

Might want to model it in hornresp. I played with it a little bit on bass horns with the idea of seeing how a current amplifier would drive horns but in the designs I tried it on the large increase in impedance peaks and valleys led to very ragged response. Based on admittedly limited modeling and no real life builds I was beginning to come to the conclusion that in general open baffles or quasi-infinite baffles like current amps but horns probably wouldn't.

[Helmuth]

I did test a very nice autotransformer to attenuate the level. Sound was good only the low end became influence by the impedance peaks.

now I am using my carbon resistor divider again it performs much better then the autotransformer because when level is brought down for 15dB the parallel resistor to the driver becomes that low you get a good damping factor, and still lively sound with out of loss of dynamics.
I use multi parallel 1W carbon resistors to get higher power and less induction, so to reduce distortion by XL.

I would not recommend a series resistor to attenuate that will reduce damping and impulse behavior, and also the thiele-small parameters will change.

I did test with metal-film to is also good with wire-wount sound becomes distorted perceived is brighter and more aggressive so do not use wire wound types HF-drivers are very sensitive to that.

[KSTR]

Allen,
a few posts you might want to check out :
MMs / BL = dynamics?
What are the unique challenges to making a "baffle-less" speaker?

When I'm evaluating a design (I do active designs exclusively) and driver candidates for it, I tend to completely discard datasheet Qes and hence Qts (as a result in voltage drive mode), as those are of little importance to me and in fact those are the *only* parameters that the user can change almost arbitrarily without touching the driver's construction, provided that the original values is too low (the other way round is cumbersome, dealing with negative amplifier output impedances and such). There is no such thing like a "too strong" driver in this regard.

As I wrote in the above posts, the electrical damping only affects the fine-print of a driver's non-linear and large signal behaviour, and the response to events external to the system.

Note that I found it absolutely mandatory to EQ two systems to the same identical response to fractions of a dB, otherwise you will not able to reliably judge the effects of different electrical damping of the same driver, let alone comparing different drivers.

- Klaus

[KSTR]

Originally Posted by mashaffer

Based on admittedly limited modeling and no real life builds I was beginning to come to the conclusion that in general open baffles or quasi-infinite baffles like current amps but horns probably wouldn't.

I simply EQ'd a small experimental tapped horn, which is a true hell of an impedance inferno, which was pure current driven, with most excellent results, because it was an excellent driver

A more general rule : if the motor is more linear than the suspension, make use of the feedback in the driver. In other words : if there are unstable system resonances, use voltage drive. If you have medium-Q but stable resonances (Helmholtz-, Quarter-Tube-, etc) current drive is OK.

So in fact it often is just the other way round, for driver-only resonances as they appear to be dominant in OB, current drive may result in chaotic behaviour, notably with high Qms drivers.

When the driver's resonance is damped down by air load, current drive is not a problem (but make sure the undamped resonance of the driver in situ is far away from any acoustic impedance peaks.

- Klaus

[AllenB]

Originally Posted by mashaffer

Might want to model it in hornresp.

I have a screenshot to show that the series resistance works well in this case to at least flatten the response. It is as large as Re.

Originally Posted by Helmuth

impulse behavior,

and impulse looks better with the flat response.

KSTR, these look like a good read thanks.

[sreten]

Hi,

It might flatten response but generally you will be boosting output
in a region where distortion is relatively poor, making it worse.

Excursion issues are a major point in these regions.

Flattest is not good, if you targets are say 2nd order L/R acoustic c/o.

rgds, sreten.

[AllenB]

Sreten, given the frequencies involved I am prepared to take that into account so that I can keep everything to a reasonable domestic size. With proper box tuning and some intended active high passing below the passband I think it will be OK.

We choose drivers with an EBP greater than 130 for horns. There seems to be so much misinformation out there. I'm thinking that if a low Q driver is needed to balance with a horns loading to produce critical damping then I suspect going too far can produce an overdamped result which might justify some Rs.