This JFET opamp circuit was developed in diyAudio threads as a clone of the circuit used in Nelson Pass Xovers. The Pass SuperSym amps favor +/- inputs, so an additional JFET opamp with single ended-input and dual ended outputs would need to be developed. LTSPICE schematics would allow team development and data comparison.
Attachments
This is what I mean. The following circuit combines my driver resonance management, a notch filter, baffle/dipole compensation and a second order filter all in one for a flat mesured in room response to within +-2dB from 25Hz to two octaves above the crossover point (800Hz).
There is also a volume control and it does not affect the frequency response at any position in it's travel.
There is also a volume control and it does not affect the frequency response at any position in it's travel.
Attachments
Re: SL's response to O'Brien ASP
Publishing an alternative XO in the public domain will encourage people not to pay SL's design fee. So we shouldn't do it without real permission. I think the alternative XO is more likely to happen if we face the ethical issues head-on.
Personally, I'm not against lots of op-amps, but it would be very cool to see how to combine filter stages.
Perhaps SL and Nelson can work something out -- they actually live fairly close together.
David
Tosh said:
Publishing an alternative XO in the public domain will encourage people not to pay SL's design fee. So we shouldn't do it without real permission. I think the alternative XO is more likely to happen if we face the ethical issues head-on.
Personally, I'm not against lots of op-amps, but it would be very cool to see how to combine filter stages.
Perhaps SL and Nelson can work something out -- they actually live fairly close together.
David
Re: Re: SL's response to O'Brien ASP
All you need to know is the target transfer function (expressed in the Laplace notation), then try to approach**) that with stages with interacting/non-ideal networks, looking at the mag/phase errors. Probably best split into minimum-phase and allpass sections. From what I see on SL's infos about the Orion (and the Phoenix) the transfer functions are already there, albeit not in direct Laplace notation (not a problem, though).
**) that is, empirically optimize, after using first order (non-interactive) guesses.
See andy_c's Website for details of this technique (on a way less complicated example), using LTSpice.
http://andycpublic.50webs.com/phono_preamp.htm
You can nest parameter sweeps three levels deep, thus you can quickly see how close you can get with a certain parameter shift and how sensitive it is. It will most certainly be a lot of trial and error work...
Just my 2 ct., Klaus
It has been said (here) how to get there.davidespinosa said:
All you need to know is the target transfer function (expressed in the Laplace notation), then try to approach**) that with stages with interacting/non-ideal networks, looking at the mag/phase errors. Probably best split into minimum-phase and allpass sections. From what I see on SL's infos about the Orion (and the Phoenix) the transfer functions are already there, albeit not in direct Laplace notation (not a problem, though).
**) that is, empirically optimize, after using first order (non-interactive) guesses.
See andy_c's Website for details of this technique (on a way less complicated example), using LTSpice.
http://andycpublic.50webs.com/phono_preamp.htm
You can nest parameter sweeps three levels deep, thus you can quickly see how close you can get with a certain parameter shift and how sensitive it is. It will most certainly be a lot of trial and error work...
Just my 2 ct., Klaus
two paths to progress and another idea
The path forward seems clear:
1) Define the transfer functions, then come up with circuits to approximate them. Unless the transfer functions are plundered strictly from the public information on the website, I would be, and it sounds like others may also be uncomfortable using the information. Unfortunately, it is much more difficult to design the network to approximate a more general function.
2) Develop discrete op-amps to plug into the circuit - either as drop-ins for op-amps in the original circuit or for specific functions in a modified circuit
3) And now the "new" idea. You could run the circuit at line level rather than from the output of a volume control pre-amp. This reduces dramatically the required dynamic range of the signal processor. Place the volume control on each output signal. If it is single ended, the volume control would only be 50% larger than a balanced-stereo volume control implemented on circuits such as the P1.7.
Unless you try to implement some of the transfer function in the op-amp internal circuitry, the first two are independent. The third would allow you to move the relative level adjustments to the end of the processing chain and may change how the transfer functions are defined.
Jeremy
The path forward seems clear:
1) Define the transfer functions, then come up with circuits to approximate them. Unless the transfer functions are plundered strictly from the public information on the website, I would be, and it sounds like others may also be uncomfortable using the information. Unfortunately, it is much more difficult to design the network to approximate a more general function.
2) Develop discrete op-amps to plug into the circuit - either as drop-ins for op-amps in the original circuit or for specific functions in a modified circuit
3) And now the "new" idea. You could run the circuit at line level rather than from the output of a volume control pre-amp. This reduces dramatically the required dynamic range of the signal processor. Place the volume control on each output signal. If it is single ended, the volume control would only be 50% larger than a balanced-stereo volume control implemented on circuits such as the P1.7.
Unless you try to implement some of the transfer function in the op-amp internal circuitry, the first two are independent. The third would allow you to move the relative level adjustments to the end of the processing chain and may change how the transfer functions are defined.
Jeremy
To be perfectly honest, having played around with the my NaO circuits at length, and having built the Orion with both original and simplified active crossovers I don't think there is that much to be gained through this. As I said previously, it doesn't hurt the system to simplfy the circuits, but I'm not sure much was gained other than a cost reduction. And I would suggest that it is possible that a redesign could just as well take something away from the system.
If you want to improve on the performance of the Orion, maybe the ASP isn't really the place too look. Maybe you should look at the Orion overall and determine what its limitations are and then start with a clean sheet of paper addressing all those factors.
If you want to improve on the performance of the Orion, maybe the ASP isn't really the place too look. Maybe you should look at the Orion overall and determine what its limitations are and then start with a clean sheet of paper addressing all those factors.
From a technical standpoint and particularly a sound quality standpoint, you are probably 100% correct.
However there are still two things to be gained: the knowledge and the fun of the exercise.
We are talking Do It Yourself, after all. I mean we do this for the fun of it, right? If it wasn't fun, why would we bother?
eL
However there are still two things to be gained: the knowledge and the fun of the exercise.
We are talking Do It Yourself, after all. I mean we do this for the fun of it, right? If it wasn't fun, why would we bother?
eL
Why not incorporate coils aswell?
http://www.steinmusic.de/audiofino/bauteile-spulen.html
Or would perhaps trading opamps for ferrite in the signal path be a bad swap.
http://www.steinmusic.de/audiofino/bauteile-spulen.html
Or would perhaps trading opamps for ferrite in the signal path be a bad swap.
I for one would find the process very educational.
Several things are very clear to me.
Any acceptable design would be plug compatible to the ASP.
Second, I started audio in the 741 eara, so my intuition is that a simpler, discreet implimentation would be both estetically and sonicly superior. YMMV
My goals are first to learn, second great sound.
I have a 3 way OB on a DCX, and would very much like to have an option to migrate to a analog active crossover.
Thanks in advance.
Doug
Several things are very clear to me.
Any acceptable design would be plug compatible to the ASP.
Second, I started audio in the 741 eara, so my intuition is that a simpler, discreet implimentation would be both estetically and sonicly superior. YMMV
My goals are first to learn, second great sound.
I have a 3 way OB on a DCX, and would very much like to have an option to migrate to a analog active crossover.
Thanks in advance.
Doug
For my bi-amped system, I use a BOSOZ which drives a line level analog crossover consisting of coils and capacitors..the crossover terminates into Aleph J's...
This all replaced an active Marchand crossover which was heavy on Op Amps.
I much prefer the sound of the simple coil and capacitor filters. As diyers, we can do this because we know what the impedences of the stages before and after the filter.
The BOSOZ is very cool in this application because the BOSOZ lets you adjust the gain of the high and low channels without changing the output impedence of the preamp channels. (the analog filters would be affected if the output impendance of the BOSOZ changed as you adjust levels) So you can change levels of the high channel and low channel to suit your taste.
The sound is simply natural and exquisite.
This all replaced an active Marchand crossover which was heavy on Op Amps.
I much prefer the sound of the simple coil and capacitor filters. As diyers, we can do this because we know what the impedences of the stages before and after the filter.
The BOSOZ is very cool in this application because the BOSOZ lets you adjust the gain of the high and low channels without changing the output impedence of the preamp channels. (the analog filters would be affected if the output impendance of the BOSOZ changed as you adjust levels) So you can change levels of the high channel and low channel to suit your taste.
The sound is simply natural and exquisite.
Surely amplifiers of different power amplifier gains could be used rather than using the atten/gain stages of the ASP - http://www.linkwitzlab.com/images/graphics/asp-blocks.gif
Yes, but assuming plug and play of speaker system without measurement/adjustments requires matched gain.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.