Yes. We've shared a "coldie" here...Originally Posted by Ian Mackenzie
And, sorta - this is obviously not the 52-5140 card - but the graphed comparison is clearly close...
Yes. We've shared a "coldie" here...Originally Posted by Ian Mackenzie
And, sorta - this is obviously not the 52-5140 card - but the graphed comparison is clearly close...
bo
"Indeed, not!!"
When I was working on my 3155 project with Giskard's help, he posted this: http://www.audioheritage.org/vbullet...2&postcount=32
It clearly shows the expected ~3dB bump from using a Butterworth filter. I suppose in the "real world" the curve may be quite different due to lack of output from the 2122H.
It is interesting that the factory recommendation was for an 18dB Butterworth filter set at 290 Hz and the GT supplied voltage drive was for a 12dB Butterworth set at 320Hz.
Widget
Yes it is interesting.
Somewhere in a thread and I dont know where are some Spice generated voltage drives for the full set of filters into respective 8 ohm loads.
That may shed more light on this discussion.
I might look for that later on.
If you have a copy of Vance Dickasons Cookbook on page 90 (4th edition) there is very interesting paragraph on the combined response o even order filters....
In short when even order filters are combined the two sections sum flat when the level of both filters is 6 db down at the crossover point.
Normally they are -3 db down at the crossover point for Butterworth filters when phase of high and low pass filters is coincident and the summation is +3db which is +6db gain at the crossover point (+3 minus -3 is 6 db ).
As Dickason points out though if the in phase relationhip is relaxed and the crossover points are spread by a given facter of about 1.3 the filter will sum nearly flat when both high and low pass filter sections are -6db down at the crossover point.
It would be interesting to measure this in practise.
Ian
Linkwitz-Riley filters are -6dB and the sum to a flat on axis output... here is an interesting interview of Siegfried Linkwitz in Stereophile. Look at the middle of this selected page for a discussion about this topic:Originally Posted by Ian Mackenzie
http://www.stereophile.com/interviews/503/index3.html
I am not sure how accurate DEQX is in it's emulation of L-R and Butterworth filters, but I have found that the L-R filters sound and measure better. When I have compared the same slope and frequency on Project Widget and measured them with CLIO.
Widget
When Bo started talking about this I was wondering this they were LR , pure or modified Butterworth filters as I mention above fitted to the Bryston.
If you were keen it would be interesting to test out these various types of filters on the fly with the DEQX plugged into Bo's system.
At any rate it would seem the stock 3145 passive filters were almost entirely empirically tuned and probably with an eye for flat on axis response.
I do like the approach of using Allpass (3rd order fiters) throughout the network.
For those not familiar Allpass filter have an Allpass Phase characteristic. At a basic level what this refers to is the phase relationship (correlated and uncorrelated) of the highpass and low pass filter sections where phase is a function of the filter slope and Q. All odd order networks exhibit and high and low phass phase relationship that constantly 90 degress out of phase and this is termed phase quadrature (no The Who is not invent this term!).
Because the outputs are 90 degrees out of phase at all frequencies the phase is said to be uncorrelated, and the filter outputs will sum together flat when the level of both filters is -3 db at the crossover point. The advantage of the 3rd order filter is that it sums with a flat magnitude response and it attenuates the output of the driver quickly in the stop band. Second and Fourth order LR filter also belong to the Allpass Group of filters, however as Mr Widget indicates, they are -6 db at the crossover point and also provide flat magnitude response.
The distinction between the Third order Butterworth filter and the Second order LR filters as Allpass filters beyond this discussion becomes more complex when we start to look at what is termed Power Reponse.
Power response refers to the off axis amplitude response in the crossover region when we are talking about the distinction crossover filters. Second order LR filters (even order) which sum flat when their high and low pass outputs are -6 at the crossover point do not have a flat power response but have a dip of -3 db at the crossover point. Odd order filters however which have an uncorrelated phase response on axis response and have a flat power response.
The audible differences of these various filter types is depeds on the specific design of the loudspeaker and is open to wide debate.
Might be a good idea if this whole filter thinghad its own devoted thread with a link from here and from the 4345 Club thread.
These sort of discussions are a myar and can drag on endlessly.
I was about to roll out the self powered 2405 crossover thing being a Greg Timbers recommended 43XX modification and I would like to maintain continuity. If there is anything conclusive regards alternative active filters we can paste them back in here.
Ian
Bo,Originally Posted by boputnam
Might be a useful opportunity for you and Ed to cross paths.
I know Ed spends time in the Bay area, suggest you send Ed a pm as a courtesy.
Ian
Giskard told me it was a very iterative process. They would start with a planned topology, listen, tweak, listen some more and tweak some more.Originally Posted by Ian Mackenzie
bo
"Indeed, not!!"
Yep,
I got the same impression in that email about the box tuning..quite amazing. I impresses me that it was not done with a computer model. They obviously have great ears!
Ian
Bo,
Have you done any more comparions or A/b's ?
I am assembling another project for my system so I won't have any feedback on the filters for 2-3 weeks
Ian
Originally Posted by boputnam
Some parts arrived so I should be able to start on this project soon.
The bags of capacitors are polystyrene, hard to get but excellent for signal filters.
Ian
Re: Active Crossover / External Auricap Crossover/4344 Review (True Love)
Hi, Ian. I got the active crossover back from Bo last Friday. I don't think he ever had a chance to audition it. I reinstalled it in my system on Sunday. I've got to tell you that after not listening to my 4344's (can I call them that now?) for the last two months, I am just utterly blown away. The system sounds so damn good. Just absolutely amazing. I think I had been making so many improvements over the last year I had started to take the system for granted. But after listening to my other JBL's for the past two months (Performance Series, 240Ti's, 120Ti's, 4412's) and then coming back to these, I realize just how special these are. Full, rich, articulate sound that can't be fully duplicated by the smaller boxes. It's like meeting up with an old love and remembering why you fell in love in the first place. I continue to be amazed. Thank you once again for the sharing your skill in the crossovers.
Ed
Nice to hear you so happy!
Can you post us another picture please mate?
Ian
Originally Posted by porschedpm
It'll take me a day or two but I'll take some pics and get them posted.
Ed
Have you tried switching back and forth between the 2420s and the 2425s?
If your 2420s were up to spec, I'd expect you'd find them a bit less aggressive sounding yet retaining all of the detail.
Widget
Having spent counltess hours stuffing around with nine volt batteries I will now focus on this project. I have a populated board, heatsink and PS. All that is required is the 18 db slope filter.
Hope to post some details soon, it should be fun.
That why we are all here.
Ian
http://www.audioheritage.org/vbullet...&postcount=266
There are currently 1 users browsing this thread. (0 members and 1 guests)