Value for input cap in special hi pass application

Status
Not open for further replies.
I am looking for some advice regarding a fairly specific application of the LM1875 module attached.

Basically I am using three of these per channel in the modified form as shown (the circuit is basically the silicon chip 20w amp) these are built point to point smd and are used to run ambience tweeters and a high efficiency full range driver on an open baffle system.

Bass from 160hz down is handled by a separate sub amps and 4 drivers so a non issue.

In essence I want to optimise the modules for the drivers and cut out lower frequencies, so you could say active amping except I don't need the sharp cutoff of such as system and in essence I feel one cap might be cleaner than several op amps and caps and power supply of the active system for this application.

The whole system runs on battery power so I wish to keep the design as clean as possible and the crossovers for the tweeters are adjustable at the tweeter from 6000 to 1200hz, the full range driver runs open but cut off at 120Hz via a cap. I am currently running these amps and the sound is sweet but I wish to refine further.

All I want to do is lighten the load on the amp so it can better serve its intended driver.

So here is what I wish to do. On the tweeter amps cut the input frequencies below say 300hz and on the full range one from say 50Hz down. 6db per ocatave is fine.

So what value caps then and what type/brand is likely to be the best in this particular application?
 

Attachments

  • 20w amp.jpg
    20w amp.jpg
    54.3 KB · Views: 145
If you want a –3 dB point of approximately 300 Hz (actually a bit under) you’d use an input cap of .033 uF with your circuit. For a –3 dB point of 50 Hz you’d use a .18 uF cap. I personally would recommend a film and foil type, and feel they sound better than the metallized type. For whatever quirk there is about my hearing I personally prefer polyester to polypropylene, even though the polypropylene has better electrical performance. But that’s just me, opinions will vary.
 
If you already have a passive cross-over, the amplifier already sees a high impedance = low load below the cross-over frequencies, i. e. below 120 Hz for the full-range and below 1200-6000 Hz for the tweeter. So you won't lighten the amp's load significantly further.

Additional filters before the amplifier, whether active or passive will add up to the existing filters. The frequency response and phase behaviour could be affected in a way you don't like. Either filter before or after, but don't do both, because that makes things too complex.

If your amplifier is built according to the schematic, you have a 12 dB filter. 6 dB introduced by 2,2 µF and 18 k with a roll-off at ~4 Hz and 6 dB introduced by 22 µF and 1 k with a roll-off at ~7,2 Hz.

You can calculate the roll-off frequency according to the formula f = 1 / (2*PI*R*C)

Also keep in mind that those two filters should be combined in a reasonable way. You may have heard of Butterworth, Bessel, Tschebyscheff, Linkwitz-Riley, etc. filters. Choose the correct type for your application.
 
Pacific, just a quick Q

All drivers are operating through L Pads, these are used for actual system volume control with the amps running on a constant input level, ie no preamp. I assume the Lpads would have some effect here as they should mean the amp see a constant impedance (ie 8 ohms).

I realise this is a kind of odd set up but the sonic results are terrific, anyhow...

This may actually change things a little re the effect of the filters on the inputs, what is your opinion.
 
An L-pad does not provide a constant 8 Ohm load for the amplifier. It would only do that, if the speaker was a resistive load, which it is not. An L-pad will have an influence on the frequency response that changes with the wiper position. It also introduces it's own parasitics, e. g. inductance, due to being made of two wirewound resistors.

If implemented correctly, the L-pad should not have a big influence on the cross-over frequency. It will however have an influence on the T/S parameters of the full-range driver, especially the Q factor. That is, where you can counteract, if you choose the Q factor of the amplifier's input filters accordingly. But it will still be easier and better to use only one cross-over, either before or after the amplifier.
 
Status
Not open for further replies.