What to take care of in fine-tuning a BR enclosure?

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Hi all.

I'm trying to design my first BR speaker, and I've read all those warnings about how a BR box needs to be tuned carefully, or else it will sound like a true boombox.

So I tried to look at my box model in Unibox to see what happens if my enclosure volume goes up or down by 10%, and if my port length changes by a few centimetres. I can see the curves changing shape, but I can't seem to find anything dangerous in any of those curves. They all seem nice, gently drooping, and very usable to me. (Sheesh! Bloody newbies!)

Here's my starting point: the parameters for my driver in Unibox. These parameters have been derived from the measured parameters, based on a series connection of a pair. (Re, Le, Vas and Sd are double that of a single driver.)
An externally hosted image should be here but it was not working when we last tested it.


And this is my desired "alignment" (I think that's what it's called for BR boxes, right?) It has a Vb of 60.7L and Fb of 30Hz, giving me a port length of 16.7cm. The gradual droop in the curve before the knee has been discussed in another thread:
An externally hosted image should be here but it was not working when we last tested it.


When I tried to change the Fb, this is how the graph changed. First, at Fb=32Hz and port length of 13.8cm:
An externally hosted image should be here but it was not working when we last tested it.


With Fb=28Hz, I got a port length of 20cm and the following graph:
An externally hosted image should be here but it was not working when we last tested it.


So, as far as I can make out, all three graphs look just fine to my inexperienced eyes. And I can't imagine misalignment more than this... my port length has been varied from 13.8cm to 20cm. That's a 7cm range of lengths.

I then tried changing the box size up and down by about 10%, keeping the Fb at 30Hz. This is the graph for 66L box, with an Fb of 30Hz:
An externally hosted image should be here but it was not working when we last tested it.


And this is the graph for a 55L box, with the same Fb:
An externally hosted image should be here but it was not working when we last tested it.


Once again, all the graphs look fine to me.

So, what should I be careful about when tuning the finished box? What am I missing???? Basically, I want to know what part of the tuning process converts an innocuous BR speaker into a rampaging boombox? What should I measure and what should I match with what else? :confused:
 

GM

Member
Joined 2003
Greets!

'Boombox' performance denotes a strong peak in its FR, typically in the 60-120 Hz BW for consumer gear and to be avoided at Fb also in DIY designs. Consider too that the room dominates down low, so even a T/S aneochoically max flat alignment will sound 'boomy' if placed near/at a corner and rolled off if well away from boundaries.

Rolling off the response as much as you have will sound 'tight'/'fast', but weak in the LF unless corner loaded if there's not a lot of room gain, requiring considerable BSC. Still, it's a good one to start with and you can always shorten the vent to raise Fb, and if any audible peaking occurs you can damp the vent to tune it out.

Other issues are vent compression and vent harmonics due to excessive length, but I assume Unibox gives minimum acceptable vent diameter based on power handling and max acceptable length would be ~SoS/(20*Fb). If longer, then it will have to be damped to prevent its harmonics from possibly comb filtering with the main's output.

GM
 
Originally posted by GM
Hi, thanks for the quick response. :)

'Boombox' performance denotes a strong peak in its FR, typically in the 60-120 Hz BW for consumer gear and to be avoided at Fb also in DIY designs. Consider too that the room dominates down low, so even a T/S aneochoically max flat alignment will sound 'boomy' if placed near/at a corner and rolled off if well away from boundaries.
I understand this (I think!) but this should apply to sealed and BR, both, right? What's so specially difficult and dangerous about BR tuning?

Rolling off the response as much as you have will sound 'tight'/'fast', but weak in the LF unless corner loaded if there's not a lot of room gain, requiring considerable BSC.
I understand the early rolloff and weak LF, and that may be the right thing for these speakers... I'm building them for a friend who will probably have to keep them in the corners of a fairly small portion of her living room. But I didn't understand your remarks about BSC... are you saying that if I use corner placement and early rolloff of LF, I'll need heavy BSC?

Still, it's a good one to start with and you can always shorten the vent to raise Fb, and if any audible peaking occurs you can damp the vent to tune it out.
Yes, that's what I figured. But I got worried because it began to seem too easy, and I couldn't see any real critical sensitivity of box response to my changing these parameters. That's what got me worried.

Other issues are vent compression and vent harmonics due to excessive length, but I assume Unibox gives minimum acceptable vent diameter based on power handling and max acceptable length would be ~SoS/(20*Fb). If longer, then it will have to be damped to prevent its harmonics from possibly comb filtering with the main's output.
Yes, Unibox warns me about vent area, which in my case is low enough to barely touch 5% of SoS. Unibox doesn't warn me about max duct length, but I think the lengths I'm seeing are quite okay. Unibox is telling me the freq of the first port resonance, though, and it's in the hundreds of Hz. With an Fb of 30Hz, I'm getting about half a metre as max port length as per your formula. All my port lengths are in the range of 10-25cm, so I guess I'm well below that limit.

What I'm really confused about is all those stories I've heard/read about BR box tuning. People seem to talk about measuring the impedance of the driver in the box, noting the frequency of the valley between the twin peaks in the impedance curve, and then cutting the port to a length exactly matching something (what???) w.r.t. this frequency. People also talk about putting a mike at the port egress, measuring the exact freq of the peak in the port's output, and then trimming the port length to match something (what???) What's all that about? Or am I just misinformed? I don't remember seeing any clear descriptions in Dickason 5/ed.
 
Just another Moderator
Joined 2003
Paid Member
Hi TCP/IP I too have seen plenty of posts warning people off BR, I was also at a bit of a loss as to why :) without access to simulation tools then yeah I'd say trial and error with BR is a recipe for disaster, but with simulation I don't see a big problem.

Also have a look at the driver displacement graphs, and step response in unibox for your different designs, that could give you more information to base your final decision on.

You may also find that a different driver/box combination may be more critical.... hard to say.

Tony.
 
diyAudio Editor
Joined 2001
Paid Member
You also may find that the different tunings aren't as drastic as people imply.

My experience is that with simulation, its not difficult to get decent response. Then it is a matter of that trial and error and/or measurement to work out the best choice-a bit disappointing, no?

. Still, it's a good one to start with and you can always shorten the vent to raise Fb,

That's what I'm up to at the moment. I have enjoyed the tight/fast bass, but I'm wondering if I can get a bit MORE bass without too much compromise... It really is a bit rolled off. I can hear the low notes, they just don't have a lot of volume impact. That's why I suggest the adjustable port- I sure wish I had left mine adjustable. I might end up having to return them to the current state.

Tcpip: try wildly adjusting the size of the box and ports to create peaks. Then you will learn how they occur and how to avoid them.
It takes quite a bit of misadjustment to get this. Small boxes as cheap/mass market speakers have make it easier to mess up

Which slope is best for you depends on the room- at least you can determine a range of reasonable port lengths to try.
 
Variac said:
Tcpip: try wildly adjusting the size of the box and ports to create peaks. Then you will learn how they occur and how to avoid them.
Actually, I've done it. I've tried setting Vb to 80L for a driver whose Vas was 30L, and then I've tried changing Fb. In that case, I've seen that there are peaks and (mild) dips. But really serious peaks seem to occur only if the box itself is smallish, and then you tune the port to the same freq at which the box itself peaks. That sort of stupid alignment will not be created by any diyer with half a brain and a half-way decent simulation software, so I don't know what Dickason 5/ed was talking about. (In fact, I actually wonder whether the BR chapter of Dickason 5/ed is at all a good starting point for anyone designing a BR box. The rest of the book is of course great, but the BR chapter specifically seems divorced from what we do these days. In fact, the entire idea of "start by choosing an alignment" seems so anachronistic.)

It takes quite a bit of misadjustment to get this. Small boxes as cheap/mass market speakers have make it easier to mess up
Yes, I am beginning to understand that the only severely "misaligned" BR boxes seen today are probably those which have been deliberately "mis"-aligned to create a peaky response. I am under no illusion that everyone wants a flat frequency response. (Do you know that in India, you get re-mix albums of popular Hindi songs with deliberate bass and treble boost, specifically sold for being played in the music-systems of auto-rickshaws?) Therefore, I guess if you or I start with the aim of getting realistic, uncoloured low end response, we may not find it all that difficult to avoid "boombox" behaviour. Am I right in suspecting this?

Which slope is best for you depends on the room- at least you can determine a range of reasonable port lengths to try.
Absolutely. And that's when my next doubt comes up: room-induced peaks can affect even sealed boxes, so why blame just BR boxes?

All in all, all your responses have made me feel a lot less worried about what I'm about to build. :)
 
wintermute said:
... but with simulation I don't see a big problem.
Good to hear from you. :) Yes, I'm beginning to feel more confident that it's not such a terrible risk.

However, a new "risk" has sort of reared its head. I now realise that to get good in-room results, I should be prepared to set up the speakers where they will finally be installed, and then, over a few weeks, take SPL measurements, do listening tests, and play with the port length to see what alignment complements the room gains. I'm now beginning to feel that this will be necessary for a sealed box too, but then with a BR box, the port is a handy way to change the low-end SPL curve.

Also have a look at the driver displacement graphs, and step response in unibox for your different designs, that could give you more information to base your final decision on.
Actually, the excursion graphs are useful, but the step response graphs are something weird: they all look the same. I have two completely different drivers whose BR enclosures I'm modelling in Unibox. They have each and every parameter different, including box dimensions. Yet the step response graphs in Unibox look exactly the same. I tried to change some parameters like Vb, Fb, and so on, and I couldn't see any change in the step response graph. I don't know what's going on. Am I just too careless to see small changes in the graph? I've given up on that graph for the time being.

You may also find that a different driver/box combination may be more critical.... hard to say.
Yes, I've seen that if you have a driver with a high Q, the curve you get for low Fb is more uneven. I may be wrong...
 
diyAudio Editor
Joined 2001
Paid Member
the port is a handy way to change the low-end SPL curve.

Exactly. On a consumer speaker, no one wants weird adjustments - they want to believe that the speaker is "good" . With an adjustable port, you can at least try to fix some room problems. Of course more adjustments can lead to more frustration, but keep in mind:

You can do the listening/and or measurements in your room after the speakers are installed and you are enjoying them. You can just play around with ports of various lengths cut from PVC pipe if you don't have measurements. Strangely I don't find this toofrustrating as at least you are listening to music at the time (albeit jumping up like a monkey every 2 seconds, driving your wife mad)

I think you are doing the right thing by not going for "flat" room response, but depending on room size AND CONSTRUCTION, bass lift can vary from non-existant to excessive, If your room is not too big and concrete, you will get some..


Some reasonably highly regarded speakers came with foam plugs to put in the ports if the bass was too boomy. Back when I was a consumer, this seemed that they were making a compromise and "how could a speaker be a bass reflex OR a sealed box". BUT it is possible to get a volume that works for both- if I recall right it is a reasonably big box.....

I must be some kind of cloth eared idiot, but actually prefer a well tuned BR to a sealed box in most cases...

:D Of course the CORRECT thing is to work out a code to write on each recording specifying the crossover setting, port length, distance from the wall, so you can change the parameters. Of course you will jump out the window half way through this process:D
 
Just another Moderator
Joined 2003
Paid Member
tcpip said:

Good to hear from you. :) Yes, I'm beginning to feel more confident that it's not such a terrible risk.

Actually, the excursion graphs are useful, but the step response graphs are something weird: they all look the same. I have two completely different drivers whose BR enclosures I'm modelling in Unibox. They have each and every parameter different, including box dimensions. Yet the step response graphs in Unibox look exactly the same.

Hi TCP/IP must have missed this before, I'm just doing some modelling of my vifa 10" drivers and did a search on step response and this turned up ;) I think that the step response graph isn't updated unless you click on the button on the graph itself!! it just stays at whatever it was before you made changes, even if you change drivers!!

below I have attached a step response graph for my vifa woofer, the blue is in the mentioned 75L box. the other colour is in a 50L box with 38Hz tuning freq, as you can see the step response on the smaller box is different.... what I'm having difficulty with is which one is better ;) I'm leaning towards the 50L one because the decay rate is quicker, but the overshoot?? on the 75L one is less..... makes me want to go look at a TL again ;)

oh and in case you are wondering, unibox doesn't let you do comparisons like that, I photoshoped it ;)

Tony.
 

Attachments

  • vb step response vifa m26wr09-08_3.gif
    vb step response vifa m26wr09-08_3.gif
    30.6 KB · Views: 110
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.