Sub response, Opinions please?

[stegalv]

I made a disc of test tones downloaded from the net and did the graph pictured below.
The amp volume was set to 00,the crossover 80hz.
The volume on the sub was about 40%, the crossover at full range and the phase set to 0deg.
During the test i had to change the spl meter range up and down between 80dbl and 110dbl to keep the needle within the scale.
How does the graph look and have a carried out the test correctly,where the response goes off the graph the highest reading was 108dbl.
i would be gratefull for any help

P.S sorry about the quality of the graph it is the largest it would allow me to post.

 

[Ian J]

The graph is way out and I would guess that you haven't calibrated the speakers using the test tone in your amp which should be done first. Start off by calibrating all of the speakers including the subwoofer to 75db then have another try.

 

[Nimby]

May I ask which website your 1/6th octave test tones came from?


Nimby

 

[stegalv]

Thanks for the replys.
Nimby, the test tones i think were from the bfd setup guide site
Ian,all of my seakers have been setup using the video essentials disk to 75dbl,i disconnected the centre speaker and front floorstanders during the test,what next?

 

[GaryG]

What did you set the port tune to in the end? The huge peak looks to be in the mid 20's although we can't be sure as it's off the scale. I would suggest a lower port tune to counteract the effects of your room. If possible take your sub out into the garden and stick it on a table or chair as far away from anything that's like to reflect the sound as possible and measure the response again to get the response of your sub in free space. That wil give you an idea of what the port is doing in relation to the driver. Once you've optimised your port length you will need to experiment with room placement to get the best tradeoff between flat response and sound quality. After that you will need to buy a BFD to flatten out the curve. Good fun DIY in it? Don't despair you're nearly there. If you can't be arsed to do the free space measurement you could just stick a BFD on it as it is but I would persevere and do the job right.

 

[stegalv]

Hi again Gary,bet you thought i was in the distant past but here i am still struggling along,
As you know this is my first diy and first decent sub.
Until i did the response test i thought it sounded ok to my ears but that is the problem i have never heard a decent sub so i am not sure what qualities of sound i am listening for.
What am i trying to achieve by adjusting the port lengh is it a flatter response on the graph or as you say the sound quality of the sub or are they the same thing (but then i am back to the problem of what to listen for,can it be explained in words)

 

[GaryG]

All I mean by sound quality is that you may have to trade that against frequency response, the best location for a flat response may not give you the best sound quality. Unfortunately you cannot locate you sub in a central position between you main speakers which will give you the best integration so you will be compromising wherever you locate the sub.

As you have such a massive peak I think it's important to do the freespace measurement to see if the port is bumping up the response, if it isn't you know it's room induced and will want to concentrate your effort on location and equalisation with a BFD.

On the other hand if the freespace plot shows the port to be boosting the response in the mid 20's you know you need to lower the port tune.

 

[stegalv]

Hoorah i understood that bit,now for more education.

Why is it a bad thing to have a high peak in certain freq ranges,would it not be better if all the freq were so high.

I am sorry to harp on about this but is there any way you can explain when "you" would be happy with the sound of a sub, is there music or a film that you listen to for certain characteristics?.

Run to the hills screaming it's the only way to escape me.

 

[GaryG]

The trouble with such a peak is you get 'one note bass', the bass seems boomy because of the narrow width of the peak. Any signal falling into that range is massively boosted and therefore sounds boomy.

The idea of a flat response is that all frequencies are amplified equally so they don't stand out. It doesn't matter whether you lower the peak or bring the rest of the response up to match the peak, the end result is the same, an even response for all frequencies that the sub is going to handle.

Having said that, the human ear does not have the same sensitivity at all frequencies. It is less sensitive at the lower and upper extremes of the hearing range. Therefore I tend to equalise the sub with a gentle boost as the frequency drops. I like to see the response rise from around 25 hz and below, around 6db or a bit more if I want to have the crap scared out of me on action films.

Once you've discovered what's causing the peak (room or port) we can move on to what's required to fix the problem.

EDIT:

A quick way to determine what the port is doing without doing a freespace check is to block the port by stuffing a towel into it and running your sweep again and comparing the response with your previous plot.

 

[GaryG]

Just a thought are you using the amp with the 6db bass boost?


http://www.partsexpress.com/pdf/300-796.pdf

The response curve on page 3 looks very familiar!

 

[stegalv]

no, this amp does not have the bass boost

 

[Nimby]

panny

If you don't want to drag the sub outside which means feeding it with a signal and controlling the PC somehow:

I'd lower the volume to the sub drastically. By at least 20dB and try again. Turn off all your speakers to avoid confusion.

Go to the BFD linked page for tones and load tones into your computer. Use only round figures 10 Hz apart starting at 120Hz going down to 20Hz or 10Hz. The tone files should end up in your "My Music" folder where you can click on each file easily to play the tones at will.

Set the volume to your sub so you start at 70dB on the meter (C-Slow) at 120Hz. Then don't touch the volume again until you've finished testing. Now work downwards from 120Hz recording your measurements off the SPL meter.

Move the sub a good couple of feet and try again. It doesn't matter exactly where you put the sub because it's just temporary for the trial.

It shouldn't take long if you start with the fixed frequncies you're going to use written down the left side of a bit of lined paper.

Mark the paper out in vertical columns so you can enter the reading off the meter one column per trial. With each new trial and sub movement just use a new column.

Using well spaced frequencies will save you time. 120,110,100, 90 etc down to 20. You aren't trying to get a perfect graph you're looking for a 'family trend' in the figures off the meter.

Your present graph may be right on the edge of maximum output. It may be effected as a result of your very high testing level. It must have been deafening!

I found 70-80dB bad enough while testing. I stuck my meter on a camera tripod and placed it at the listening position on its side (using the pan and tilt head and screw hole on the back of the meter). So I could read it from a distance. You could lay yours on its side on top of something stable where it won't fall off if you don't have a camera tripod.

To start with I was running back and forth to read the meter and then back to set off a new tone on the computer. So being lazy I just made a playlist of fixed falling tones. Then I could just sit between meter and computer and record off the meter and read the frequency off the computer screen as each tone was played automatically.

It should be easy to see if the peaky trend is repeated regardless of room position by looking across your columns to see how the figures vary after 4-5 runs of recording the meter readings.

All this probably takes longer to describe than to do in reality.

Nimby

 

[stegalv]

Thanks nimby i will try your test next although i will have to put the test tones on a disc to play back in my dvd player, i keep my computers in the loft.
I have already tried the test that gary suggested by stuffing the ports,the result is below, is it any better?

 

[GaryG]

That's a pretty good plot!

You now have a decision to make.

1. The response of the sealed sub looks good. The sealed sub has a low Q with a better transient response than the ported sub. It should sound very good now. If you have enough amplifier power for your needs and are not 'clipping' then leave it as it is. A BFD will flatten it out very well.

2. The port is clearly tuned too high, if you have the patience and want to tailor the sub for the ultimate in low frequencies you need to increase the port length and then do another sweep. Reiterate this process until the hump is below 20hz. The final result will be extreme low bass but won't have the transient response of the sealed sub.

Your original goal was for use with HT more than music, option 2 is the way to go. Personally, I like option 1.

You pays your money....

 

[stegalv]

Thanks gary,the ports are nearly at their full extension so what are my options now and why is it tuned to high.
The ports are about 14"long and angled

 

[GaryG]

It looks like you've reduced the volume for the second plot. Taking a second look at your first plot, it's possible that the bit that we can't see isn't actually that much of a peak, we can't tell from the graph. Post you SPL readings between 20 and 30 hz for the first plot so that I can better understand what's going on.

With regard to the ports, what do you estimate the cabinet volume to be with the bracing, driver and amp in place?

Are both ports tuned to the same length?

Lastly what did you think of the sound quality of it with the ports plugged?

EDIT: Make it the readings between 20 and 36hz.

Your ports are the 260-389's?

 

[stegalv]

I did lower the volume to keep it within the limits of the graph.

The raw readings were 96, 100, 102, 105,108 dbl. (20 to31hz)

with compensation 103,106,107,109,111 dbl

BOth ports are the same length by eye not measured.

I think the volume is around 7.3ft^3

I hav'nt listened to anything with the ports plugged yet,what would the best material to plug them with.

Sorry to ask again How do i judge better sound quality if i have never heard what a top sub can achieve.

 

[GaryG]

The biggest change in the response was between 25-28hz. I've plotted the theoretical response using WinISD (alpha).

If you take a look at the graph you get for a ported sub at 7.3^3ft and a sealed sub at 7.3^3ft there is around 6db difference at 25 hz. If you then model a third plot for vented and change the port tuning to 15hz (don't forget to set the dia to 3.375) you will see that plots a line fairly centrally between the other two plots.

Unfortunately a tune of 15hz gives us a length of 26" can you accomodate that? if so you will need to find something to extend the length of port (drainpipe?).

With regard to sound quality stick on a DVD with some deep bass effects and see what you think. Does it shake the room, do you like it? Does it hit you in the chest? If not we'll need to persevere and get that hump in the ported response down to something more reasonable.

 

[GaryG]

Just took another look at the WinIDS plot. As luck would have it, if you block just 1 port the tune comes out at about 14hz with the port length around 14". This is close to what we want to test so you can go ahead and run a sweep with just 1 port blocked and see what the response comes out like.

Doesn't matter too much what you block the port with, towel, foam, ball, whatever.

Obviously running with just 1 port may cause 'chuffing' to be audible but we're not bothered about that at this stage as we're just trying to find the optimum tune for your sub at it's current location.

Once again, chin up, it'll be worth it when you finally nail it.

BTW, you don't need to bother running the sweep above 40 hz we know what that looks like and the port isn't going to have much effect at above that.

EDIT:

Make the upper cutoff for your sweep 45hz, that's the low point of your dip.

 

[Nimby]

Reflex ports certainly don't have to be straight. Take a look at one of Ron (SVS) Stimpson's DIY designs:

http://www.hometheaterhifi.com/volume_6_3/diy-11-sonotube-subwoofer-september-99.html

I believe a bent port length should be measured on the shortest side of the inside of the tube, whatever the angles, or number of bends.

Hopefully bends are readily available in the same size tube as the commercial reflex port flares.

Nimby

 

[stegalv]

Thanks nimby but my ports are angled and adjustable,and while trying garys test with one port blocked i also moved the sub a couple of feet as you suggested,it is the pink line.

Gary the one port blocked test is below in blue, the pink is also one port blocked but moved a couple of feet as stated above.
Yes it did shake the room when needed and sometimes i could feel a suttle thump in quieter moments or music.but my biggest problem is i dont know how good it should be.

The volume on the sub may have been slightly higher than last time
I plugged the right port and noted the readings then the left,the readings were only about 1 dbl different on the odd freq

 

[GaryG]

Looking at the plot of the sub sealed and comparing it with the plot of one port blocked the sealed sub has more output at 20hz, not what you would expect. 5db difference between 20 and 22 hz on the one port blocked plot is strange. Also, the difference between the peak at 31.5 and the dip at 45 is around 20db on the sealed sub compared with around 10db with one port blocked. These are odd readings. Were the conditions exactly the same for the tests?

Having said that, the port doesn't seem to be adding much to the low end, if you've got the patience it would be worth doing another sweep with the single port shortened to move the tune up a couple of hertz. Key point is any further sweeps should be done under the same conditions, don't move the sub around, all the work you've done up to now was to optimise the sub for where it was.

Bear in mind, if you get fed-up, the sealed sub plot looks good and a BFD would smooth it out nicely (assuming the dips can be equalised, if they're true nulls then there's nothing you can do about it). The beauty of designing for a ported sub is that you can always change your mind and run it sealed because of the larger cabinet volume, if you had gone for sealed you wouldn't have been able to do the reverse because there wouldn't be enough cabinet volume.

 

[Nimby]

The point of moving the sub between tests was simply to eliminate the room from the response equation

Perhaps I should emphasise that ONLY ONE change should ever be made between tests.
Otherwise too many (unknown) variables are introduced.

It then becomes pure guesswork which change actually made the difference. For better or worse. This is true of most things in life. Not just testing subs.

This sub is certainly capable of putting our some dBs! But it would be nice to see a raised response at the bottom end to flatten the curve relative to 31.5Hz.

Is there an obvious room mode at ~31.5Hz Gary? I've lost track of panny's room dimensions.

Nimby

 

[GaryG]

Certainly looks that way, as it's a bank holiday weekend I think I'd be tempted to experiment with the port tune to see if it will bring up the bottom end, once the best tune is found we're in BFD territory.

 

[stegalv]

Nimby my room is 18x10x8 ft,as you can see in the photo i am limited to positioning of the sub due to size of it, at the camera position there are three recliners near the back wall.

Gary the test with one port was carried out under the same conditions as both ports plugged and open.
I only moved it after these test, the only difference may have been slightly higher sub volume setting.
I may try the sub outside and also adjusting port length over the weekend.

How do you know what the freq the sub is tuned
to?
Thank you both for your time