You may know me as a bit of a purist when it comes to speaker design.
However it occurs to me that there may be some advantages in adding a low-pass filter (switchable of course) to the Nonsuch 4s outer drive-units.
Such a network could both act as a baffle step compensation circuit and as a way of virtually eliminating the comb effects that make the Nonsuch 4 a 'stand-up-and-you-lose-some-of-the-high-frequencies' speaker.
I'm not an electronics designer and my knowledge of crossover circuitry is basic so I thought I'd post this to get views on whether this will cause more problems than it is supposed to solve. I've attached my initial schematics below.
I quite like the capacitors approach as it allows two switch positions for more flexibility in baffle step compensation. Also, when it's switched off there's nothing whatsoever that can adversely effect the performance of the speaker.
However, as the drive units are nominally 4 ohms, the overall impedance would be only 2 ohms at high frequencies. Also, the output from the inner (full-range) drive units would be variable above the low-pass point.
With the inductive approach there is less flexibility as the filter would effect both outer drivers at the same time. The impedance problem would be solved. My concerns are that the inductor would add some phase shift/time delay to the outer drivers and also my network theory is not good enough to be sure what would be the effect on the output of the inner, full-range drivers above the low-pass point.
What do you think?
However it occurs to me that there may be some advantages in adding a low-pass filter (switchable of course) to the Nonsuch 4s outer drive-units.Such a network could both act as a baffle step compensation circuit and as a way of virtually eliminating the comb effects that make the Nonsuch 4 a 'stand-up-and-you-lose-some-of-the-high-frequencies' speaker.
I'm not an electronics designer and my knowledge of crossover circuitry is basic so I thought I'd post this to get views on whether this will cause more problems than it is supposed to solve. I've attached my initial schematics below.
I quite like the capacitors approach as it allows two switch positions for more flexibility in baffle step compensation. Also, when it's switched off there's nothing whatsoever that can adversely effect the performance of the speaker.
However, as the drive units are nominally 4 ohms, the overall impedance would be only 2 ohms at high frequencies. Also, the output from the inner (full-range) drive units would be variable above the low-pass point.
With the inductive approach there is less flexibility as the filter would effect both outer drivers at the same time. The impedance problem would be solved. My concerns are that the inductor would add some phase shift/time delay to the outer drivers and also my network theory is not good enough to be sure what would be the effect on the output of the inner, full-range drivers above the low-pass point.
What do you think?
Attachments
Either circuit will work... you are essentially making it a 1.5 way speaker. You do have a phase issue as each of the outer drivers rolls thru 90 degrees above their XO point and the 2 inner drivers don't -- this is one of the reasons i'd mount them on the back
dave
planet10/just call me bi-pole dave
dave
planet10/just call me bi-pole dave
Latest iteration of the network
Below, is the latest version of the network. This version seems to solve the question of varying power at the middle (full-range) drivers.
There is some question about whether the circuit is stable or whether there could be oscillations.
Also 'BPTL Dave' and others, what happens in the time domain (phase, delays) to the filtered outer drivers below the low-pass frequency when they are working at full power? This is where I think things are most critical and I wouldn't want to compromise the coherence of the speaker.
Also, can anyone tell me whether I can split C1 to shunt both drivers so I could vary the capacitance and the frequency to each?
Below, is the latest version of the network. This version seems to solve the question of varying power at the middle (full-range) drivers.
There is some question about whether the circuit is stable or whether there could be oscillations.
Also 'BPTL Dave' and others, what happens in the time domain (phase, delays) to the filtered outer drivers below the low-pass frequency when they are working at full power? This is where I think things are most critical and I wouldn't want to compromise the coherence of the speaker.
Also, can anyone tell me whether I can split C1 to shunt both drivers so I could vary the capacitance and the frequency to each?
Attachments
Re: Re: Latest iteration of the network
1. To partially compensate for the diffraction step.
2. To cut down the comb effects at high frequencies (over 10kHz), allowing 'stand-up listening'.
3. To boldly go where I've never gone before with these speakers but always with a safety net.
I have to play with the circuits and see if there is any audible improvement. Any perceivable loss of clarity, coherence or transient response and they're OUT.
This has basically come about through postings on this forum of people who I respect as designers, although I don't necessarily agree with all their views. Generally, I'll stick to my basic design principles and fight my corner but I do listen to other views and I will try alternatives if it's reasonable to do so. I'll let my ears and the trusted ears of my wife and certain friends be the final judges.
PS: What's a BS compo filter
The LR (number 2.) is a simpler version. The purposes of both circuits (when switched in) are:planet10 said:
1. To partially compensate for the diffraction step.
2. To cut down the comb effects at high frequencies (over 10kHz), allowing 'stand-up listening'.
3. To boldly go where I've never gone before with these speakers but always with a safety net.
I have to play with the circuits and see if there is any audible improvement. Any perceivable loss of clarity, coherence or transient response and they're OUT.
This has basically come about through postings on this forum of people who I respect as designers, although I don't necessarily agree with all their views. Generally, I'll stick to my basic design principles and fight my corner but I do listen to other views and I will try alternatives if it's reasonable to do so. I'll let my ears and the trusted ears of my wife and certain friends be the final judges.
PS: What's a BS compo filter
Steve,
refering to Your schematic No 1 "LCR lowpass filter for D1 D4":
when You`d leave out R1 You have a pure second order low pass with the according phase shifts.
For "smoother" phase response You can play with "reasonable" values of R1 (too high and it has no effect - too low might cause amp stability problems as C1 shorts D1 & D4 at high frequencies) in order to get something between (almost) a first (6db) and a second order filter .
This circuit is more flexibel as You have more possibilities to tune it (freqeuncy AND Q) and it might turn out to be the better solution (as compared to a first order filter in schematic No 2) and I wouldn`t reject this completly.
As they are identical drivers and as You still have the benefits of a line source at lower frequencies (which is the main advantage IMO) I wouldn`t worry about coherence too much.
When the circuits are beneficial soundwise (what I believe they are) the next step could be a NONSUCH 5 with a single center driver playing the high frequency range alone and with two different lowpass circuits for the "outer" pairs for even increased vertical listening area at high frequencies .
Your idea of making the circuits switchable is a good one IMO!
So people could choose beteen line-source sound (BTW: I think You`ll need baffle step circuit for this anyway also) or towards "point"-source sound.
I`d be VERY interested in what You find the differences are between this two approaches.
7V,
seems like what i am looking to build is a bit like your 4x JX92 idea only differnces being (a) I am looking at going open back for the mids (there will be sides but no back) and (b) I hope to build a dipole sub (sida facing) for the bottom part of the speaker,
the reasons are because the subs would be required in HT mode, they would be used in stereo mode since they are there but I feel in HT mode the need to create floor shaking bass is more essential. in stereo mode bass quality (ot quantity) is essential as poor quality bass (one note, boomy, resonant) actually detracts from the music.
i know i am acting difficult but what i am trying to do is build a speaker that can:
(a) be simple preferably 1.5 way
(b) quality mid and HF as good as a good mini monitor
(c) bass that is authoritative at moderate levels (95db)
(d) have a small foot print
so i figured 4 x full ranges (JX 92, W3 871 etc..) would cover 80/100hz+ and the dipole bass would cover below 100Hz.
seems like what i am looking to build is a bit like your 4x JX92 idea only differnces being (a) I am looking at going open back for the mids (there will be sides but no back) and (b) I hope to build a dipole sub (sida facing) for the bottom part of the speaker,
the reasons are because the subs would be required in HT mode, they would be used in stereo mode since they are there but I feel in HT mode the need to create floor shaking bass is more essential. in stereo mode bass quality (ot quantity) is essential as poor quality bass (one note, boomy, resonant) actually detracts from the music.
i know i am acting difficult but what i am trying to do is build a speaker that can:
(a) be simple preferably 1.5 way
(b) quality mid and HF as good as a good mini monitor
(c) bass that is authoritative at moderate levels (95db)
(d) have a small foot print
so i figured 4 x full ranges (JX 92, W3 871 etc..) would cover 80/100hz+ and the dipole bass would cover below 100Hz.
navin said:
Actually, I use 4 x Bandor 50mm drivers. They're only 2" (50mm) so are sufficiently close that I only get any combing over 10kHz.
I think that open back is a great way to go. It's relatively cheap too. All my sophisticated enclosure design is just to make a sealed box sound like an open back.
I believe that open backs take up a little more space as they can't go too near the wall. They are also a little more room dependent. Is this right?
Sounds good
You're even madder than I am. Someone lock him up, quickly.navin said:
Navin
It's just hit me that if you're using anything larger than the 2" Bandor drive-units (mounted very close together) you will need to use a low-pass filter with your full-range drivers.
Either you could low-pass filter all of your full-range units and provide a high quality tweeter or you could low-pass filter all but one of the center drivers, leaving the unfiltered driver to handle the treble
Without this filtering, comb effects of the line array will cause a significant drop in the high frequencies.
As you had posted on a thread entitled "Low-pass filter network for mini-array", I'm sure that you have realized this anyway. Still, I thought I'd better make sure.
Steve
It's just hit me that if you're using anything larger than the 2" Bandor drive-units (mounted very close together) you will need to use a low-pass filter with your full-range drivers.
Either you could low-pass filter all of your full-range units and provide a high quality tweeter or you could low-pass filter all but one of the center drivers, leaving the unfiltered driver to handle the treble
Without this filtering, comb effects of the line array will cause a significant drop in the high frequencies.
As you had posted on a thread entitled "Low-pass filter network for mini-array", I'm sure that you have realized this anyway. Still, I thought I'd better make sure.
Steve
Sad I am, to hear this. I've built a small TL tower using 4 TB's (with Steve's help) sonically similar to a pair of 1.5's. The clarity and SPLs suggest we're on to something for $50, but I notice a lack of "snap" in the treble. Must be the comb cancellation?
I've been playing with tweets and small horns to fill it out, all are a slight improvement, even a TB as a tweet. About the 'sweet spot" listening position problem, the TB's are very good off-axis, and this doesn't seem to be a problem that I've noticed.
So I was looking forward to your experimental findings. Oh well...
Yes, Bandor.navin said:
Is that Steve meaning me?x. onasis said:
If you move left and right while listening to (or measuring) the Tangbands and compare that with moving up and down from the mid-point, you should get a good idea of the comb cancellation effects at high frequencies.
Is the treble performance ok if you just use the two centre TBs? Can you disconnect the others and test this? If this would work for you, then you could use a filter circuit like filter 1 - Post #5.
Otherwise, you have to apply a filter such as this one to all four of your TBs and use a tweeter.
There is a significant difference between the 50mm diameter Bandor drivers that I use and the larger TBs.
x. onasis said:
hey relax. if i can find a design that deals with comb and still provides decent SPLs using TBs I am all for it. the TB are great value for money. Only they might have limited use as a line array.
7V said:
well bandor is not distributing much anymore is it? me thinks to eliminate comb one would have to use 3-5 TBs with ONLY ONE going full range.
How low do the Bandors go? what SPL can one expect from 4 Bandors at 100Hz. If I am looking at a modified line array anyway (bandor or TB) where all but one driver is rolled of then the TBs are still usable.
Maybe there are other drivers (planar etc...) that can be used to 100Hz. I am not married to TB, Bandor, or any other make.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.