This is somewhat off topic, appologies in advance for that but I thought it was a good point.
I was chatting with Hugh of AKSA amplifier fame about the impact on my AKSA amplifier from driving low and high impedance drivers. During the conversation the point was raised about correcting impedance with a zobel, cap and resistor in series to correct and flat the phase further more throughout the passband in order to improve imaging and also reduce distortion in the voltage feedback to the amplifier.
We were specifically talking about the Percieve 2 which is an active design.
Here's Hugh's comments:
"Anthony, I should mention that impedance correction of the drivers is a VERY good idea on active systems. Most systems do not address this problem at all, with attendant problems with spatial presentation. Find some literature on this if you can and drop a Zobel, a cap and resistor in series, typically 6.8uF and 10R for a woofer, across the driver connections. This corrects for the inductance in the voice coil, and pulls back phase angle to near zero across the pass band. In turn this does not distort voltage feedback to the amp, permitting the very best imaging performance. This is important!!"
What are peoples thoughts on such a practice? To be honest it never even entered my head! But I can see reason from his words, would the benefits warrant attention? I'm not so sure about adding in passive components when I went all out to go with a digital active design.
I was chatting with Hugh of AKSA amplifier fame about the impact on my AKSA amplifier from driving low and high impedance drivers. During the conversation the point was raised about correcting impedance with a zobel, cap and resistor in series to correct and flat the phase further more throughout the passband in order to improve imaging and also reduce distortion in the voltage feedback to the amplifier.
We were specifically talking about the Percieve 2 which is an active design.
Here's Hugh's comments:
"Anthony, I should mention that impedance correction of the drivers is a VERY good idea on active systems. Most systems do not address this problem at all, with attendant problems with spatial presentation. Find some literature on this if you can and drop a Zobel, a cap and resistor in series, typically 6.8uF and 10R for a woofer, across the driver connections. This corrects for the inductance in the voice coil, and pulls back phase angle to near zero across the pass band. In turn this does not distort voltage feedback to the amp, permitting the very best imaging performance. This is important!!"
What are peoples thoughts on such a practice? To be honest it never even entered my head! But I can see reason from his words, would the benefits warrant attention? I'm not so sure about adding in passive components when I went all out to go with a digital active design.
ShinOBIWAN said:
now that is interresting and totally counter intuitive, but maybe it is something you could test in a program such as lspcad, as you can add passive components onto an active design , and that it takes accound of driver inductance. but maybe not, i think that considering the cheapness of the considered components, it is certainly worth a try, especially for a guy that is having such a wonderfull active system
ShinOBIWAN, it is a reasonable suggestion to add such impedance compensation in woofers and mids and you can just try it out.
I would not choose values to flatten their impedance to their minimum value though, I would just choose values to gently flatten it from, say, half an octave before their crossover point and over.
I generally dislike every form of fearless power consumption in the loudspeaker-crossover era
And though I’m not a components junkie, I would prefer there some cheap mkp caps, and don’t forget to have some good wattage for your resistors.
Regards,
Thalis
I would not choose values to flatten their impedance to their minimum value though, I would just choose values to gently flatten it from, say, half an octave before their crossover point and over.
I generally dislike every form of fearless power consumption in the loudspeaker-crossover era
And though I’m not a components junkie, I would prefer there some cheap mkp caps, and don’t forget to have some good wattage for your resistors.
Regards,
Thalis
pinkmouse said:
I'll give it a go just for the hell of it. Nothing to lose and maybe something to gain.
Measuring the driver electrically and the phase looks OK in the operating ranges I use although its not perfect. I'll have to try a little impedance correction as Hugh suggested and see if things can be improved.
I've measured acoustic phase in room and it looks a little surreal with DRC enabled; +/- 30degree down to 300hz, below 120hz its somewhat ragged but that's impossible to fix with DRC and needs big physical treatments. Turn off DRC and it looks very rough with large phase swings below 1.6Khz. When I tested the speakers outside(no boundaries) things looked good also.
Hi Black cat,
you are referring to a different Zobel.
Aksa was talking about correcting the speaker driver impedance to make the speaker look more like a resistor with respect to the driving voltage coming from the amp.
You are referring to the high frequency dummy load on the amp output that ensures the amp sees a sensible load if it ever gets disturbed into producing an output in the hundreds of kHz.
What Aksa is proposing is an extra Zobel.
Thinking about the Aksa proposal, it seems that the correction will reduce the load value seen by the amp by a small amount particularly at the higher end of the passband of the speaker. This will increase dissipation in the output stage. However the increase in dissipation when driving a reactive load is enormous sometimes approaching three times the power. It appears obvious, to me at least, that the power dissipation benefit by using the extra Zobel will be considerable.
Is there a reason built into this analysis that could affect the sound quality coming from the speaker?
Asking the amp to deliver LESS current may help improve the sound and as Aksa points out driving less signal back into the feedback circuit should also improve things.
Seems like a new thread on this series of experiments is warranted.
you are referring to a different Zobel.
Aksa was talking about correcting the speaker driver impedance to make the speaker look more like a resistor with respect to the driving voltage coming from the amp.
You are referring to the high frequency dummy load on the amp output that ensures the amp sees a sensible load if it ever gets disturbed into producing an output in the hundreds of kHz.
What Aksa is proposing is an extra Zobel.
Thinking about the Aksa proposal, it seems that the correction will reduce the load value seen by the amp by a small amount particularly at the higher end of the passband of the speaker. This will increase dissipation in the output stage. However the increase in dissipation when driving a reactive load is enormous sometimes approaching three times the power. It appears obvious, to me at least, that the power dissipation benefit by using the extra Zobel will be considerable.
Is there a reason built into this analysis that could affect the sound quality coming from the speaker?
Asking the amp to deliver LESS current may help improve the sound and as Aksa points out driving less signal back into the feedback circuit should also improve things.
Seems like a new thread on this series of experiments is warranted.
Hi all,
my sealed subs are actually powered w/o any x-over from my power amps.
These units are based on Adire Shiva, each coil is connected to 100/8 channel (so 4 channel/2 amps/2 subs), just to clear connections.
The signal is actively LPFd before entering the amps, to let them play under roughly 50Hz.
Would you recommend a Zobel net across the subs? Any hint on wattage/values?
Thanks for sharing,
Stefano
my sealed subs are actually powered w/o any x-over from my power amps.
These units are based on Adire Shiva, each coil is connected to 100/8 channel (so 4 channel/2 amps/2 subs), just to clear connections.
The signal is actively LPFd before entering the amps, to let them play under roughly 50Hz.
Would you recommend a Zobel net across the subs? Any hint on wattage/values?
Thanks for sharing,
Stefano
Hi,
NO.
Why:- you are limiting the output from your source to little and falling signal above 50Hz.
The bass speaker will be operating close to it's box induced resonance at these low frequecies. It will never see a significant signal at frequecies where inductance of the voice coil becomes a problem.
I think there is little (nothing?) to be gained by smoothing the upper frequency impedance of the sub-bass speaker.
NO.
Why:- you are limiting the output from your source to little and falling signal above 50Hz.
The bass speaker will be operating close to it's box induced resonance at these low frequecies. It will never see a significant signal at frequecies where inductance of the voice coil becomes a problem.
I think there is little (nothing?) to be gained by smoothing the upper frequency impedance of the sub-bass speaker.
It entered my head long ago and I tried it in an actice system.
I put a zobel on the power amp of active-crossed parallel bass-mid drivers covering the range from 60Hz to 500Hz. This drivers had quite some impedance increase at higher frequcencies (upwards about 500Hz or so) due to voice coil inductance.
I corrected that with said zobel RC-network after the power amp.
There was a very, very small difference in sound in this case.
It was too small in order to decide if it was for good or worse.
This does not mean it couldn`t make a bigger (or maybe even big) difference in other amplifier/crossover/driver scenarios.
Just try it, it`s easy and as the zobel is parallel the amps outputs You can just put it in and out (clip leads or a relay) while the amp is running and listen wether it makes for a difference or not.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.