system7's 4-way.
I used his values, in my schematic.
I try not to use resistors. I know they are sometimes a necessary evil. I also know they are part of Mr. Frieds ideas. Sorry.
Picture 1) The schematic.
Picture 2) What it looks like.
Picture 3) My wife doesn't come in the basement any more. She says I need therapy, maybe even stronger meds?.
Picture 4) Todays midrange trace. I left the 3-way, 1.5mH trace up for comparison. It's the afternoon, so house voltage may be down a bit?.
Picture 5) Tweeter. We cross around 3800Hz.
Picture 6) Mid Woofer. We cross around 2600Hz.
Picture 7) Woofer. We cross around 160Hz.
Let's go back to Steve's pictures and see how close they are?
I used his values, in my schematic.
I try not to use resistors. I know they are sometimes a necessary evil. I also know they are part of Mr. Frieds ideas. Sorry.
Picture 1) The schematic.
Picture 2) What it looks like.
Picture 3) My wife doesn't come in the basement any more. She says I need therapy, maybe even stronger meds?.
Picture 4) Todays midrange trace. I left the 3-way, 1.5mH trace up for comparison. It's the afternoon, so house voltage may be down a bit?.
Picture 5) Tweeter. We cross around 3800Hz.
Picture 6) Mid Woofer. We cross around 2600Hz.
Picture 7) Woofer. We cross around 160Hz.
Let's go back to Steve's pictures and see how close they are?
Attachments
If you have that kind of budget, I would consider the Accuton C173-6-090 as a lower mid. I will be using them in a project very soon.I'll be using the Accuton Diamond Tweeter and the Accuton 2" Diamond Midrange in a speaker next year. I think the best crossover parts are also in order.
As for a woofer, I would love to try the S280-6-282, but my audio budget is exhausted for a little while.
Let's make system7's crossover a 2nd order balanced system.
Picture 1) New schematic. The balancing bits are all attached directly to the drivers.
Picture 2) New woofer trace. I left the purple trace from the 1st order measurement.
Picture 3) Still the woofer, but I increased the 68uF caps to 150uF. The drop in the low end is not what we want. This might imply we need a larger inductor. The inductor we're using has R of 0.53 Ohms. With a larger guage wire, this resistance would drop. A lower resistance inductor might also change the slope. The lower output after x-over is good. This might imply the mid woofer is playing more of the music it's suppose to.
Picture 4) The mid woofer with the 150uF capacitors.
Picture 5) The mid woofer with the 'proper' 68uF. I feel this value pushes too much energy back to the woofer. I need to do more waffling.
Picture 6) Mid woofer with the blue trace removed (Less cluttered).
Picture 7) Midrange.
Picture 8) Tweeter.
Have a great weekend everyone.
Picture 1) New schematic. The balancing bits are all attached directly to the drivers.
Picture 2) New woofer trace. I left the purple trace from the 1st order measurement.
Picture 3) Still the woofer, but I increased the 68uF caps to 150uF. The drop in the low end is not what we want. This might imply we need a larger inductor. The inductor we're using has R of 0.53 Ohms. With a larger guage wire, this resistance would drop. A lower resistance inductor might also change the slope. The lower output after x-over is good. This might imply the mid woofer is playing more of the music it's suppose to.
Picture 4) The mid woofer with the 150uF capacitors.
Picture 5) The mid woofer with the 'proper' 68uF. I feel this value pushes too much energy back to the woofer. I need to do more waffling.
Picture 6) Mid woofer with the blue trace removed (Less cluttered).
Picture 7) Midrange.
Picture 8) Tweeter.
Have a great weekend everyone.
Attachments
Bill, it is time to take a break, and get some fresh air and exercise. Really.
Your more complex filter just doesn't work in any sensible way.
I can only show you what it does with my choice of drivers. A complete mess.
Back to the drawing board when you have had a rest. Your nice simpler filter will still be there.
Your more complex filter just doesn't work in any sensible way.
I can only show you what it does with my choice of drivers. A complete mess.
Back to the drawing board when you have had a rest. Your nice simpler filter will still be there.
Attachments
If you have Excel, have the ability to take measurements, and know how to calculate offsets, it doesn't get much simpler than this.
http://audio.claub.net/software/jbabgy/PCD.html
http://audio.claub.net/software/jbabgy/PCD.html
You can't just use the same values for first and second order filters. Are you using these formulas to calculate the components?
2nd order series:
C = 1 / (2 * π * f * (Z / √2))
L = (Z / √2) / (2 * π * f)
Mark
I remember a 2nd year advanced calculus class. This was in 1982, some 33 years ago. The University of Western Ontario, London, Ontario. Geophysics was my major.
In that class we started dissecting equations with multiple variables. I can't remember the exact name of the type of calculus.
The teacher told us "there are an INFINITE number of combinations". Did you see the word "infinite", Mark?.
On that date, 33 years ago, we could solve FOUR 'types'.
The solving of these 4 types of equations didn't come from 'solving'.
4 Different men 'back engineered' the 4 equations. They applied something (an algorithm?) which helped us solve the equation.
The class revolved around recognizing which of the four types was presented, and applying the correct algorithm. I got 100% on that mid term exam. 96% was my final mark.
Let's look at the weather in South Western Ontario, Canada. This includes the North East section of the United States of America.
There are 14 variables which dictate the weather in our region.
There are dozens of weather forcasters on our TV's and radio stations.
On any given day, with the exception of when a high pressure system is in place, and all we're getting is sunshine, they can't predict the weather outside of looking out there window.
There is not a model on the planet that can deal with 14 variables.
If there was, there is not a computer that could do this inside a week, for tomorrows forcast.
Think about my 4-way crossover.
In a first order crossover there are 6 crossover components. Variable 1 is inductance (from inductors). Variable 2 is capacitance (from capacitors).
Each inductor has an internal resistance. Variable 3.
For the time being let's forget the resistance of the wires. Let's forget the output impedance of the amplifier.
The speaker driver has a resistance associated with a frequency. Variable 4.
For the time being let's forget the inductance of the driver. Let's forget the capacitance of the driver.
4 variables makes this a complex calculus equation.
I hope 'n', 'f', and 'Z' of your above equation are part of my 4. Otherwise we are a 5 or 6 variable complex equation.
I was TAVES (Toronto Audio Video Entertainment Show). I went into every room. Not one speaker tricked my brain into thinking there was a live band in the room.
The calculations we use today are wrong.
I am going to 'back engineer' my fr graphs.
I am going to put forth a 'simpletons' way of building a 4-way.
I'm going to leave the math to the genius'.
SYSTEM7, my 'measured' frequency response curve's don't appear to be a mess.
Bill, this isn't an argument. I'm just trying to keep up with what you are doing.
For sure your 3rd/4th order filter has lost the flat impedance.
The more regular way of making a 3rd order series filter might look like the below. The red components are the merest guesses.
Even so, impedance is not flat. But tidier than your series LCRs on the mids.
You can add more to make it fourth order, but I haven't got time right now, and I've forgotten what the component ratios should be.
For sure your 3rd/4th order filter has lost the flat impedance.
The more regular way of making a 3rd order series filter might look like the below. The red components are the merest guesses.
Even so, impedance is not flat. But tidier than your series LCRs on the mids.
You can add more to make it fourth order, but I haven't got time right now, and I've forgotten what the component ratios should be.
Attachments
Steve;
Your second 3.3mH inductor is in the correct position.
Your second 4.7uF cap needs to be moved to the right. Past both 4.7 Ohm resistors.
Your .60mH should be a capacitor (not two inductors in a row).
Your 33uF cap should be an inductor. (not two caps in a row).
Let's start with that.
Recap from page one of this thread;
This is a first order series crossover. That's it. Nothing more.
The 'balancing' components ensure the driver sees the exact same pull up of the wave form, as the pull down. Please review my thoughts on 'balancing'.
The balancing components do not make this a second order crossover. At least not in the traditional sense. The greater slope is simply an added positive byproduct.
Your second 3.3mH inductor is in the correct position.
Your second 4.7uF cap needs to be moved to the right. Past both 4.7 Ohm resistors.
Your .60mH should be a capacitor (not two inductors in a row).
Your 33uF cap should be an inductor. (not two caps in a row).
Let's start with that.
Recap from page one of this thread;
This is a first order series crossover. That's it. Nothing more.
The 'balancing' components ensure the driver sees the exact same pull up of the wave form, as the pull down. Please review my thoughts on 'balancing'.
The balancing components do not make this a second order crossover. At least not in the traditional sense. The greater slope is simply an added positive byproduct.
No, No, No.
Your original filter is very good. I like it. But it's important to look at it in a simple way.
It's a second order series filter with a novel and economical topology. Let's look at the three way version and what it does.
Next, I've done something similar with a parallel circuit. And throw in the SEAS Kit 503 to show how mainstream this is.
You can of course do steeper filters like SEAS-3-Way-Classic, and maybe it helps. But simple filters can be very listenable if done right at every level including driver selection. Even two ways. http://www.diyaudio.com/forums/multi-way/147632-classic-monitor-designs-25.html#post4428857
Your original filter is very good. I like it. But it's important to look at it in a simple way.
It's a second order series filter with a novel and economical topology. Let's look at the three way version and what it does.
Next, I've done something similar with a parallel circuit. And throw in the SEAS Kit 503 to show how mainstream this is.
You can of course do steeper filters like SEAS-3-Way-Classic, and maybe it helps. But simple filters can be very listenable if done right at every level including driver selection. Even two ways. http://www.diyaudio.com/forums/multi-way/147632-classic-monitor-designs-25.html#post4428857
Attachments
-
SEAS Classic Kit 503 Troels Gravesen.PNG13.8 KB · Views: 574
-
system7 simple 3 way Parallel FR.PNG18.7 KB · Views: 488
-
system7 Simple 3 way Parallel.PNG10.1 KB · Views: 338
-
Cousin Billy series filter simple 3 way FR.PNG18.9 KB · Views: 3,588
-
Cousin Billy Series Filter simple 3 way.PNG10.3 KB · Views: 1,382
Hey Steve
Nice impedance on both. Nice slopes on both.
I see what you mean.
Monday I'll pick us cheap high value caps and continue measuring.
My circuit will be using a 5.6mH woofer inductor.
Any chance you could do a quick sim with that value? (on the first of your three crossovers)
Nice impedance on both. Nice slopes on both.
I see what you mean.
Monday I'll pick us cheap high value caps and continue measuring.
My circuit will be using a 5.6mH woofer inductor.
Any chance you could do a quick sim with that value? (on the first of your three crossovers)
But what does this mean? An impedance peak at the resonance of a woofer is an indication that less energy is needed to produce the result due to mechanical assistance.Nice impedance on both.
A crossover often has a resulting impedance peak due to such things as the reduced radiation impedance of the two drivers working together at the crossover. The filter resonances also play a part. A flat impedance at crossover is not necessarily an indicator of perfection due to the complex nature of the contributing effects.
You also need to consider shunting currents in a given crossover topology where the impedance will not clearly represent the consumed energy used in the speakers themselves.
Remaining impedance variations that might have a side effect on the performance of the amplifiers can just be conjugated separately.
Now this from Lynn Olson's Beyond the Ariel thread. Lynn remarks that the original simple second order filter Ariel works much better with valve amplification. IMO, it was significant that he tuned it by ear, probably with a valve amp.Why is that whenever a test yields answers that we don't like we condemn the test?
Why is that "different" sounding means "better" to the person who is doing the test?
I was once talked into using some guys tube amp on my speakers. Admittedly it sounded different - way different. I was, naturally, curious, so I investigated. Turns out that the high output impedance of the tube amp completely changed the speakers frequency response, and not in a good way. So was this "different", "better"? To me "better" was the way the speakers were designed to work.
It is a fairly easy thing to say that things are "different", its a whole magnitude more difficult to say what is "better".
And I am certainly not a "golden ear" (if that was implied). I have never trusted my own hearing - I've done too many psychoacoustic tests to know that isn't reliable. I rely, for the most part, on measurements. If this were somehow wrong, then I believe that after a decade or more the "error of my ways" would become apparent, but, alas, that has not happened. As things measured better they simply sounded better over the long haul.
Impedance isn't a huge factor with solid state amps. But with a valve amplifier with lower damping factor, as Dr. Geddes remarks, frequency response changes greatly where impedance varies. The bass tuning with valve amps is different too. The bass becomes underdamped and slightly peaky, easily enough fixed with a bigger cabinet, as it goes.
But given the choice, I would always prefer flat impedance for a myriad reasons.
Bill, do you mind if I duck your question? It's just tinkering really. Better to sim it yourself.
Downloads
Oh, FWIW, if you want to simulate the effect of a valve amp on a speaker, just add a resistor of 1 ohm for a feedback type, or 6 ohms for a SET type in front of the filter. Not rocket science!
Last edited:
I'm pretty sure I agree with Steve and Earl, just providing an insight.
By the way, when I first discovered the (admittedly quite minor in some cases) effect of conjugating impedance around a crossover, which I particularly was doing because of the low/mid kHz frequencies involved, I was working with a high feedback chip amp. I've also noticed no audible benefit in doing it with some amps.
These days I'm bi-amping with one SS and one valve amp each with a Zo of around a few ohms, but that's getting a little off topic..
By the way, when I first discovered the (admittedly quite minor in some cases) effect of conjugating impedance around a crossover, which I particularly was doing because of the low/mid kHz frequencies involved, I was working with a high feedback chip amp. I've also noticed no audible benefit in doing it with some amps.
These days I'm bi-amping with one SS and one valve amp each with a Zo of around a few ohms, but that's getting a little off topic..
I was mulling over driver selection for this clever little crossover of Bill's.
We were thinking a smallish paper fullranger initially.
But why reinvent the wheel? Bud Fried did a lot of work on precisely this shallow type of slope. He chose a 5" polycone mid. Like this sort of Peerless 830860:
Peerless HDS PPB 830860
The Peerless falls away to nothing above 5kHz. So some of those breakup problems that you can see happening at 7kHz with the fullranger won't be there. Oh, BTW, you must add 6dB to the mids in my sim, since I have twinned them, to get an accurate picture of midrange level.
IDK. Just thinking aloud.
We were thinking a smallish paper fullranger initially.
But why reinvent the wheel? Bud Fried did a lot of work on precisely this shallow type of slope. He chose a 5" polycone mid. Like this sort of Peerless 830860:
Peerless HDS PPB 830860
The Peerless falls away to nothing above 5kHz. So some of those breakup problems that you can see happening at 7kHz with the fullranger won't be there. Oh, BTW, you must add 6dB to the mids in my sim, since I have twinned them, to get an accurate picture of midrange level.
IDK. Just thinking aloud.
Attachments
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
- Home
- Loudspeakers
- Multi-Way
- 4-Way Series Crossover. Accuton and Acoustic Elegance Drivers