Having problems with port size... how big is far too big?
My floorstander speaker project is going along fine. The Neo3 tweeter is an amazing thing - it brings some electrostatic sound to the speakers. I'm using a pair of old Focal 5N421 mids in a D'Appolitto configuration.
So far, things are looking fine. Except for the vent. I miscalculated in every possible way the size of the box (imagine - I was aiming for 55 liters and ended up with 34 - yikes). And the vent I'm using is the 4" 5/16ths from Parts Express... in a cabinet roughly measuring 4 feet tall, 1 feet deep, 0.6 feet wide.
Anyway, the vent is almost flush with the side panels, and 1 inch away from the rear panel.
I tried the Woofer Tester 2 Vented box analysis with the speakers, and QL was around 0.4. Even after using caulk everywhere, I could only get as high up as 2.8. And, no matter what I do to the port, frequency response is terrible - I have two maximum impedances at 30 Hz and 80 Hz (the box should be tuned at 60 Hz).
I *think* my port's diameter is too big, and the box is behaving as if it had a massive leak instead of a tuned vent... but I'm not sure. Do you think this is my problem?
If so, I'll simply block the port (the speaker will actually look better, since I'll add some interesting design to it) and model the thing as a a sealed box. Since I'll be using a ported sub, I'm not that worried about losing too much bass...
:bawling: And I thought I was good at very basic math. :(
The vent is too close to the rear wall for a start. Secondly, vent resonance is roughly at the one point of minimum impedance between the two peaks.
I'm not all all sure what you mean when you say you caulked the box and got something at 2.8. Ql should not be anywhere near that high nor should it be affected so massively by some caulking.
Thanks for your answer.
According to the Woofer Tester manual, QL should be higher than 3, and close to 7, in order for results to be meaningful. Maybe they meant "0.3" and "0.7" instead! I'd have to re-do my tests where QL was 2.8 or so, then.
If resonance is at the point of minimum impedance (I thought it was the other way around, actually), then I may be closer to 60 Hz than I originally thought.
Still, I wonder whether such a huge vent can pose significant problems during day to day listening. If I separated the vent, say, 2" from the back wall, would I have problems with the length of the vent in the midrange? How can I be sure about that?
I think a sealed box is looking a lot better now.... I feel like such a newbie (which I am, BTW).
I wouldn't write off vented because of this experience, if your volume is vastly different than what it is supposed to be you are probably going to be struggling to make it sound how it should, though provided you tune it properly and aren't too far off, you can probably get a reasonable result, albeit with a higher F3 point (ie bass won't go as low). so don't blame the vent, blame your mis-calculation ;) (my Vifa 10" drivers state on the data sheet that they are suitable for cabinet volumes between 50 and 100L!
I'd be modelling the driver in the smaller enclosure and seeing how it behaves, you might be better off with a different tuning freq.
Also a smaller diameter vent needs a shorter length for the same tuning freq, so that would be a way of fixing the problem of being too close to the back wall (of course you then have the problem of filling the bigger hole!)
The other option is to put a bend in your vent though I have never particularly liked this idea myself.
You are right about Ql being 3-7, I mistyped. I should have said I can't see how you got such a low Ql as in your first test as it should not be that low.
I think your main problem is that simply you have made your box much smaller than it should have been. Make sure there is always 3 inches or one vent diameter between the vent end and any walls.
You might be able to salvage your design a bit by cutting the vent inside the box down a bit to give some better clearance then adding a 90 degree elbow and some pipe extension. The idea being to tune a bit lower. However, I think given that your box is relatively small, sealed might be a better bet. Or just build another box the right size :)
Hello fjhuerta and others
I think part of the confusion is the difference between box and driver losses. In your case it sounds as if the box (or port) loss is pretty high. This could be caused by the port being too close to the back wall, too much stuffing, air leaks, or box walls that need stiffening.
But this may not be the only issue. For example, your driver may have a vented pole piece and simply placing the driver on a table top is an oops! Basically the air in the VPP and the sealed dust cap make a very tiny sealed box! Anyhow, we have been adding this and other information to the woofertester.com FAQ (check the FAQ).
I noticed you are also using the simulator/overlay tool. Though this can be used to design a simple vented or sealed system, the intention is to help analyze final system configurations, or if you are a re-cone shop, to adjust the driver mass and stiffness (to tune the driver back into spec). For example, if you enable the simulated and measured phase and impedance plots, in theory they should line up if the correct box size, tuning (and loss) are used.
You may have also noticed that at the moment box loss is not one of simulator inputs. At the moment this is an alpha release 'rough edge' as the feature is put into action. The bottom line is that I need to scratched my head some more and figure out the relationship between Rem, Qts and box Qloss.. and impliment this without making things too much more complicated. So here is what you do for now.
Adjust Vb and Fb untill the two peaks line up in frequency. Then adjust Rem to change the magnitude of the peaks. What you are doing is adjusting the driver loss, but you may recall me mentioning that in the model box and driver losses are combined, so the results are the same.
Another thing to try is sealing box (plug the port) and measuring Vas using the sealed box method. Again, things should line up nicely.. except maybe the impedance peak magnitude (whats up with this?). The reason is that the driver and box springs and losses are now combined. and this shows up (mainly) as the resonance peak magnitude. If loss is *huge* you can also find the combined resonance shifting a bit, but this is pretty rare and it does not effect the Vas computation that much. A good exeriment is to measure the box with and without damping material.
With that said, another easy option for you to try is pulling the port forward so it hangs outside the front of the box. This might look a bit odd, but acoustically the wavelengths are huge, so it has little effect. However you do need to account for the increase in box volume.
Putting a bend in the port has also been mentioned and is a good idea. The 'port volume' contains a 'plug' of air that loosely speaking acts like a mass, attached at the back through the box air spring and finally to the driver, so in theory nothing has changed. What you need to watch out for is small diameter pipes (and now a bend) that may add turbulance. Or maybe a bend helps to break up turbulance (I actually dont know)? Anyhow, if the port makes 'choofing' noises, you might want a bigger vent.
Hope this helps
Thanks for your answers!
Building another box would be quite difficult. The current one is already detailed (maple laminated, etc.) and it was a pain to build (with 2" MDF and extensive bracing). I think I'll keep this one.
I've been modeling the box using BassBox, adding stuffing, sealing the box, etc. Although I do get a bit more bass extension by tuning it at 40 or so Hz, group delay increases significantly. Besides, frequency response looks identical for both boxes right until 45 or so Hz.
Given that I'll be using a big sub, and that the sealed box will have a roll-off that will mate better with the sub's plate amp, I think I'll go sealed this time.
Still... it's quite funny. You think you thought about everything during your first project, and then you see your math went way off (what happened was that when using Bassbox, I mistakenly typed exterior dimensions as interior ones!) :) Miss Blanch should have made me repeat the first grade!
2" MDF! Did you actually find 2" MDF or did you layer it? One of my tricks is to build a box from 3/4", trim it with a following straight edge router bit, put on a second layer, and then trim again. Internally this produces a zig-zag that is almost impossible to not have sealed. Needless to say those were heavy suckers!
I might as well also mention that cone displacement in sealed and tuned boxes are differerent. Well below resonance the piston force is simply V/R, producing a constant force and a constant displacement amplitude. Acoustic output is proportional to velocity, and this drops off at 1/F or 12db/octave. This can be EQ'd, but the displacement (and power) now grows proportional to F^2. On the other hand the vent in a vented box acts like a spring/mass that at resonance works against the cone motion decreasing displacment amplitude (look at the displacement dip in the simulator). This can help the driver quite a bit untill you go below Fb, at which time the only stiffness working against the motor is the driver Cms (the box is now open baffle). The other thing as you mention is that the output is a combination of the driver and the port producing some phase and group delay issues. So is your goal rumbling bass or tight and well defined? An interesting question that comes down to personal preference and taste.
I too have a big 18" sub but the 7cuft box it is presently in was adopted rather than designed (the spousal acceptance factor of a 11 cuft box being the issue). So rather than flat down to 18Hz or so, I am for now living with a 30Hz peak. This may change soon.
I wanted to be sure I tried everything posted here before answering...
First of all, Keith, thank you very much for your posts. They are very insightful, and I'm slowly learning what I need to build a great pair of speakers. The Woofer Tester is an amazing little device. I can't imagine of a better buy for building peakers.
About QL... besides my speakers havng a vented pole piece, I didn't realize I was using two speakers in series, so calculations are way off! Is there a way to tell the Woofer tester that there are multiple drivers inside a box? I couldn't find the option. I assume this is why my QL values are way off.
Now that I tried a sealed box, the graphs correspond perfectly to my simulations. I need to go back to vented because I lost way too much midbass output - but I learned a lot about measuring a speaker in the process.
As to the box - it's 2" MDF in the front baffle only. A friend of mine is an old time carpenter, and he built the thing. It was amazing to see him cut perfect circles without any kind of jigs :D And yeah, they are very heavy, but I really hope it'll help in keeping the vibration to a minimum.
You mentioned something about rumbling or tight bass. I now have a serious issue regarding midbass output and SPLs. The sealed box has a Qts of 0.279, so I guess it's way too overdamped. I need some bass back. But, how can I be sure I'll not go overboard when tuning the box? Is Group Delay the parameter I need to keep in check? What's an acceptable value for it?
One other thing - after listening to sealed, controlled bass, I found out my Adire Audio Tempest (Adire Alignment) is not as "quick", "controlled" or "taut" as I once thought :) I'd rather listen to my overdamped speakers.
Hello again fjhuerta
Glad to hear your are making good progress. It is also a good thing you replied when you did. I was about to go on vacation for a couple of weeks!
The WT can test multiple drivers and the procedure along with an example is set out at the www.woofertester.com FAQ. Eventually this example will go into the help file. This keeps the SW complexity to a minimum.
Your question regarding QL seems to be two questions in one. Almost every TS parameter, including driver Q's, will be affected if you plug a vented pole piece during the test. But I dont think this was your question. Taking a look at this example you will see that the big difference in measurement is not in the driver Q parameters, but rather Vas.
QL (box loss) is a seperate issue and is a loss that is external to the electro-mecahnical driver characteristics. You will note however that the simulator that was added to our test version of software (ver 1.02) does not yet have a QL input. Luckily, driver and mechanical loss will (at least in this simulator) model nicely with adjustments of the drivers mechanical resistance, Rms.
When I mentioned 'rumbling' -vs- 'tight' bass I was referring to how a 'boom box' might be sonic crud, but on the other hand, it will defintely shake up a party. The problem with 'acceptable' is that this comes down to preference.
As an example, you should have a look at the BagEnd audio products. In this case what appears to be a *way* too small sealed box is used. The result is a -3db frequency roll off that looks almost useless. But if you look at its 12db/octave rolloff and phase characteristics you will see that... if combined with the right kind of audio processor... will produce an essentially flat response with excellent phase characteristics right down to zero. That is, if the driver can handle a monsterous amount of power from an astoundingly big amplifier! This is interesting but usually not too practical.
|All times are GMT. The time now is 03:40 AM.|
vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2015 DragonByte Technologies Ltd.
Copyright ©1999-2015 diyAudio