Would be interesting to ge both your comments on this Bolserst and yours Joe as the filter in this version is the conclusion of the MK5-filter + Bolserst-Mod + Joe-comments-on-Bolserst-mod. It has some drawbacks and some advantages as I concluded the many comments (maybe pretty soon a year or so back).
Cheers, /H
Ok,
Great to understand, as a number of vented box calculators say the same (wrong) size compared to what actual sizes of the vents that are used in this thread. Such as 90IDx105mm. Good to know as it is measuring actually 35% bigger than what the calculators say.
The characteristics of the drivers will I guess be reflected in the curves for the speaker even if the box itself will have the same tuning, and mostly be tuned by the vent itself.
Cheers
/H
Done and working well, but also intersting.
The top most driver was perfect to use for this, but not for looking. Using the wave generator it was not easy to see but very easy to feel the resonance frequency. The cone motion at both 32 and 33Hz was almost none at all. At 31 and 34 Hz the motion was more and similar to each other, as so on. So 32,5Hs it is - at least by this method.
Tried the same on the other drivers too but it was more difficult as they were moving a lot more and did not actually dicrease in motion a lot at any point. The lower position of the the driver the more movement and less dicrease of cone movement. The conclusion is that there are different working environments for the different drivers depending on the postion in the box. Interesting.
Absolutely. I meant to suggest use of the top woofer, but forgot.
Basically with the the vents at the bottom of a tall enclosure the acoustic behavior starts to act a bit more like a TLine and a bit less like a classic Helmholtz resonator. Properly modeling this cabinet requires distributed acoustic parameters rather than the lumped parameters used in most vented enclosure calculators. Each of the woofers is placed in a different location down the line, so will behave slightly differently. The largest acoustic impedance changes for the primary enclosure tuning occur at the end of the line furthest from the ports.(ie the top of the box)
Joe alluded to this behavior in his description of the enclosure and damping in post#1695.
No prob, then I just did get to learn something by myself for once.Absolutely. I meant to suggest use of the top woofer, but forgot.
Ofcourse... that was actually something in my mind.
That's right. Those calculators work OK in a straight rectangular box with internal surfaces lined, but the Elsinore enclosures don't fit this description and I noticed very early on that about 75 Litres looked like over 100 Litres and worked with that. But I also did a mini-monitor design around the same time, under ten Litres, where I filled it with a Dacron like material (wool and polyester) except near the entrance to the port (any damping there just seem to make the port act like resistive/aperiodic type vent) and was able to drop Fb by about 15% without changing the port - that meant Vb in effect went up. I must admit that it surprised me a bit at the time, but that the quality of the bass really did improve, more open and more breathier, deeper in tonal solidity - and have continued to use this idea since. The judiciously filled vented enclosure sounds better to me than a lined one. Sort of more the kind of quality that TL lovers like, where the TL is filled and not lined.
Cheers, Joe
Yes, the driver's suspension effectively couples to the resistance (or lack thereof) from the walls of the enclosure.
Here is a discussion I had with a forum member on his project in this respect:
http://www.diyaudio.com/forums/full...x-fe-108ez-project-part-2-a-2.html#post874667
The better method with "fill" is to suspend it within the cabinet - free from all surfaces (..like an open basket containing the "fill" suspended in the enclosure). (..and of course still keeping it reasonably away from the driver as well.)
I think Martin Colloms mentioned something about this at one point.
Last edited:
I think you will get a different result if you drive all drivers in series parallel if you have left the other three drivers unconnected in the box and only driving the top one? If that is the case, driving four drivers will give a higher Fb I suspect.
But yes, the position of the driver in a non-rectangular box can lower Fb and seemingly break the rule that the tuning is independent of the driver, but here the distance between the back of the driver and the entrance to the port may actually give an effect similar to lengthening the port? Or... if there is a filled rather than lined strategy used and also be the time (slowed velocity) from the back of the driver to the entrance of the vent making Vb looking larger, thus changing Fb. Bur whatever, the standard calculator makes no allowances for any of these sort of things and the designer simply has to deal with it in a real world situation and make the appropriate choices.
Cheers, Joe
To test the single drive case, you would need to remove the bottom 3 woofers and replace with wood covers. Otherwise, the unconnected woofers will act like passive radiators and load the driven woofer contaminating the experiment.
See post#1251 & 1255.
http://www.diyaudio.com/forums/multi-way/97043-elsinore-project-thread-126.html#post3132884
http://www.diyaudio.com/forums/multi-way/97043-elsinore-project-thread-126.html#post3124178
I don't recall seeing a shift in tuning as more woofers were added to an enclosure when I was experimenting with vented/TLine columns.
If it was there, it was not a big effect.
Or it may just modulate the Vb and make it look larger and that is why he seems to have gotten a slightly lower Fb than I would have expected. But those suspensions could go into some kind of passive mode and as you say, it could be mild or not - who knows - but it will contaminate the experiment. It seems the obvious and easier thing is to do is drive all four drivers and check where Fb is, if that is the aim.
Cheers, Joe
The Peerless driver were 'designed' to be used both ways and had an edge profile to suit that. But the SB drivers do not and is safest and visually best to flush mount them in a recess. Without doing that there could well be a wrinkle in the response mid-frequency. The fact they don't have truncated frame also makes it easier than the Peerless.
The modelling will be based on measurements where the edge will be flush. The time alignment is relatively non-critical because of the way the Tweeter works, compared to Butterworth which is extremely critical. Recessed, the top three drivers focuses slightly further away, that's all.
Cheers, Joe
I did drive the four drivers in series parallell as designed. I think this will give the correct working environment for the drivers, and the aim was to find the Fb in "real" conditions. The other three drivers were not acting in any other way than the top most i think. It was just much harder to determine at what freq the smallest movement was. I did feel a small difference but not enough to state any clear result. But at least it did not say that it would be totally different.
The fill that I have used. A combination of glass wool around the inner panels and sheeps wool in the back compartment are known from earlier contructions to slow down the air in the cabinett. Especially the sheeps wool. And as a major part of the "TL-like" air transport area is in contact with the sheeps wool it might actually make the box "bigger".
Thanks for all the input.
Still waiting for any comments on the filter, where I whish to go for the MK5std + Bolserst mod + Joes-Bolserst-mod-comments (as of my drawing). Maybe I will put in some switches for turning discussed parts on and off to hear the difference.
Cheers
Still waiting for any comments on the filter, where I whish to go for the MK5std + Bolserst mod + Joes-Bolserst-mod-comments (as of my drawing). Maybe I will put in some switches for turning discussed parts on and off to hear the difference.
Honestly, I think you should consider just building the standard Mk5 crossover and live with it for a bit of time and get to know the sound profile. After that, you can introduce the various mods and get a better idea of what they do in terms of changing the overall sound presentation of the speaker.
Personally, I stopped after adding the 33R resistors. I am extremely happy with the sound for my listening space. That said, I have a fairly deep room, so my speakers are focused a bit more straight back (less toe-in) to help extend the sweet spot if you will. The resulting sound and imaging is outstanding.
Hi Guys
Just letting you know a huge amount of work is being done on the Mark 6 right now. All the testing and data has been converted into complex transfer files and modelling of the Crossover looks very good - but as always there are trade-offs - but will point these out. For example, there is an elevation centered around 100Hz which is a proximity effect and there is a depression between 1.5KHz and 4KHz - but because the modelling is based acoustic on data that was collected at 2 Metres, there is a cancellation effect between the two bottom drivers (due to difference in travel to the listening position), but because most of us will hardly be listening at 2 Metres, that depression should become less at you move away from the speaker.
When the Crossover is built and tested, this will be the real test.
Cheers, Joe
Just letting you know a huge amount of work is being done on the Mark 6 right now. All the testing and data has been converted into complex transfer files and modelling of the Crossover looks very good - but as always there are trade-offs - but will point these out. For example, there is an elevation centered around 100Hz which is a proximity effect and there is a depression between 1.5KHz and 4KHz - but because the modelling is based acoustic on data that was collected at 2 Metres, there is a cancellation effect between the two bottom drivers (due to difference in travel to the listening position), but because most of us will hardly be listening at 2 Metres, that depression should become less at you move away from the speaker.
When the Crossover is built and tested, this will be the real test.
Cheers, Joe
Thanks fpr you input jdJAke,
As I at that time followed the thread discussions thinking that the Bolserst mod with Joes adjustments was the final stage, that is what I did build.
I could unconnect the 33Ohm resistors but that is from what I read unwise as it seem to be a good improvement. And as Joe pointed out the R and L values on the conj link closest to the bass drivers should be changed while adding 33Ohm resistors over the speaker terminal.
Going back from would my current version will then only be possible by unconnecting the three hidden components in the Bolserst conj link for the midbasses. The ones that also concerns Joe the most. That I also will do with a switch. So there will A-B tests on that.
Last edited:
Please see the CSD plot below, this this the Peerless Nomex driver (measured of the Hamlet box) at 2 Metres and 15 degrees off axis (the recommended listening axis) and note when off axis there is no problem with the 4KHz peak that in on axis. There is also no severe ridge that would indicate a significant resonant behaviour.
I have long held the view that under the 'Jordan knee' being the frequency where the cone ceases to be a piston, that under the knee it is diffraction effects of the box that dominates and above the knee it is diffraction effects within the cone (much to do with cone shape and surround) that dominate, provided there are no severe cone break-ups or other anomalies. I don't see any such here.
So if that 4KHz peak is a problem, toe the speaker further out - that has always been my recommendation. Also, rooms can have a significant influence here as well. I once made a speaker sound right in a very dead room and when it was taken to an average (good) room, it sounded much to energetic in the upper mids and up.
So the Elsinore Mk5 can be made to work without the LCR on the MidBass - just takes a bit of applied thinking setting them up.
Cheers, Joe
I have long held the view that under the 'Jordan knee' being the frequency where the cone ceases to be a piston, that under the knee it is diffraction effects of the box that dominates and above the knee it is diffraction effects within the cone (much to do with cone shape and surround) that dominate, provided there are no severe cone break-ups or other anomalies. I don't see any such here.
So if that 4KHz peak is a problem, toe the speaker further out - that has always been my recommendation. Also, rooms can have a significant influence here as well. I once made a speaker sound right in a very dead room and when it was taken to an average (good) room, it sounded much to energetic in the upper mids and up.
So the Elsinore Mk5 can be made to work without the LCR on the MidBass - just takes a bit of applied thinking setting them up.
Cheers, Joe
Attachments
Early modelling of the Mark 6 Crossover - it looks like this may settle in to close to the end result, see below.
Please note the plateau centered around 100 Hertz is influenced and accentuated by microphone proximity effect of nearfield measurement - so account that it in reality it is a little lower than that.
Cheers, Joe
Please note the plateau centered around 100 Hertz is influenced and accentuated by microphone proximity effect of nearfield measurement - so account that it in reality it is a little lower than that.
Cheers, Joe
Attachments
Last edited:
I think you must have misunderstood. The mods I suggested in post #1072 were: “…For builders that find their listening setup to highlight the excess woofer energy around 4kHz, this modification might prove to be beneficial...” Basically, if you are able to achieve good imaging and tonal balance with the speakers aimed more straight ahead rather than toed in aimed at the listener then the notch filter for the 4Khz is unlikely to be desirable.
My comments below will reference the 3 STAGEs Joe mentions in post #1273
http://www.diyaudio.com/forums/multi-way/97043-elsinore-project-thread-128.html#post3137048
No changes are required to the conjugate R and L values when adding the 33 ohm resistors.
Those changes aren’t made until STAGE 3.
My personal experience is that there are no drawbacks to STAGE 1(adding the 33 ohm resistors), only benefits. You have already discovered while testing the tuning of your enclosure that motion of the woofers at low frequencies is different depending on their location in the cabinet. Ideally you would like each pair of woofers to have identical motion. However with woofers connected in series, if the motion of the woofers aren’t identical the woofer with the larger motion will generate more back EMF than its series mate. This upsets the impedance balance and you will have one woofer producing more output than the other. Adding the 33ohm resistors across each woofer voice coil goes a long way in swamping this difference resulting in better matching between the series pairs. Similarly, any difference in midrange impedance/inductance(including impedance modulation from voice coil current) is also reduced by these same resistors. A side benefit is that this also helps stabilize the mid-woofer crossover point. The end result is improved bass and midrange clarity, particularly if your woofers happen to not be well matched.
You would need to simultaneously connect and disconnect the 10ohm resistor across the L2 for the A-B comparison to be valid.
That is, you would need to connect/disconnect ALL the components in STAGE2.
If you decide you do want/need to tame the excess 4Khz energy in the woofer on-axis response by adding STAGE2, a more interesting A-B comparison might be with the STAGE2 notch filter in place, and connecting/disconnecting the conjugate network Joe suggests in STAGE 3.
Note that my recommendation in post #1072 was to remove the conjugate network(L5, R3, C4) when adding the 4kHz notch filter. Whatever the theoretical advantages of the conjugate network it comes down to what sounds best to you in your setup. The measured response is the same with or without it; only the impedance the amplifier sees is different.
Without the conjugate network the impedance is higher but more reactive.(blue curve in post#1267)
http://www.diyaudio.com/forums/multi-way/97043-elsinore-project-thread-127.html#post3135695
With the STAGE3 conjugate network the impedance is lower but more resistive.(green curve in post#1269)
http://www.diyaudio.com/forums/multi-way/97043-elsinore-project-thread-127.html#post3135699
In the case of tpate’s setup, it definitely sounded better without.
More details in post #1272:
http://www.diyaudio.com/forums/multi-way/97043-elsinore-project-thread-128.html#post3136339
Last edited: