Vas measurements by added-mass method in Speaker Workshop

I am trying to measure Vas for a few mid-woofers that I have. I have already measured the other parameters and have got believable results.

For Vas measurements, I haven't yet gotten around to sticking the drivers in a sealed box, so I wanted to try the quicker added-mass method. Is this method much less accurate than the sealed-box method? I have official data sheets for the driver, and the Sd given there matches with the Sd I'm measuring from middle-of-surround to middle-of-surround. (Everyone tells me that added-mass Vas depends heavily on an accurate Sd.)

I'm placing two five-rupee coins (exactly 10g each) on the cone, and taking the added-mass impedance measurements. They give me nice clean graphs, sometimes with twin-peaks. And they have given me Vas figures of 24L, 27L, 29L and 30L. The published figures in the datasheet are higher: 36.8L. The data sheet is here:

Peerless India 6.5" Kevlar cone woofers

I repeated the tests with three coins instead of two (ie. 30g weight), and the Vas measured was just 5-7% away from the 20g measurements. I believe this is how it should be, within a reasonable range of weights?

The other parameters I've measured tally quite closely with the data sheets. I'm getting 0.39 as measured Qts for three out of the four pieces, and Fs in the range of 42-44Hz (published is 39Hz).

My question basically is: can I trust my added-mass Vas, or should I necessarily do the sealed box measurements too?
 

jomor

Member
2004-02-20 1:36 pm
Athens
i thought Peerless where made in China..... :confused:


i m not sure which method is more accurate, but i would trust the measurement more than the manufacturer's data. make sure your drivers are runned in before measuring them. If you havent done so, create a .wav file ( google for a sine generator ) of 10 Hz, in stereo, with reversed phase on the left (or right channel), and burn it into a cd. 10Hz are not audible (reversing the phase in one channel helps too) and you can play this for hours to run in the woofers before making any measurements. Fs for example will be quite higher if the driver is brand new..
 
I can't say how accurate the added mass method is...I just don't think these issues are all that critical, myself.
But one thing you could do is move on to the next phase, where you select RESOURCE/NEW ENCLOSURE. You can specify box dimensions, insert your driver, and model it's response. Then you could vary the VAS parameter and see how it affects the frequency response. It may not make much difference....a fraction of a dB or something like that.

Actually, I think there's a way to let the SW calculate the optimal box for flattest response. I think after you create new enclosure, you select CALCULATE/ then pick sealed or vented, and the program creates a new box and/or port. You could then change the VAS of your driver, then repeat the process and see what difference SW chooses for your box size. (This is certainly less time consuming than building a physical box just to do the SEALED BOX method)
 

JMB

Member
2004-04-01 8:47 pm
Texas
Measurements are most accurate as you get further away from DC (0 Hz). With the added mass approach, you are decreasing Fs and thereby decreasing accuracy if the frequencies are too low. As long as you are not using frequencies much lower than 20 Hz, you should be fine. Also remember that with the added mass method, you want to decrease the Fs by about 25%.

The closed box method raises the Fs as it's comparitor and is therefore a better choice if the frequencies involved are very low.

The accuracy of each depends upon how accurate you are with your measurements including the weight and how well sealed the box is for closed box measurements.

HTH

Jay
 
I've measured hundreds of woofers using both methods. Delta compliance Vas is generally much more accurate than delta mass. I've determined that the mass inaccuracy does not come from a bad weight measurement of the mass itself, but the way it is affixed to the cone. Any adhesive method using a soft compound effectively adds damping to the mass, greatly reducing accuracy. For example, the "blue-tack with coins" method I used years ago. So, if you have some test boxes handy, delta compliance is the way to go. But if not, there's a better way to do delta mass:

It requires a selection of small and medium size neodymium magnets and some medium density closed cell foam. The magnets must be a known weight. Get them accurately measured if you need to. Basically, to use this method, you place neo magnets opposite each other on both sides of the cone and they grip the cone strongly due to their attraction to the magnet on the other side of the cone. If multiple weights are required, space them evenly arond the cone.

I mention the foam requirement because sometimes the magnets are so strong that they may be difficult to remove or the pressure may cause a mark on the cone if it has a textured surface. A thin strip of foam may make the magnet easier to remove, and reduce the potential for marking the cone. Find a few various thicknesses of cardboard to practice on before trying a real cone.

There is no more accurate way to do delta mass measurements.

Aside from that, don't be surprised if you vas measurements are way off spec anyway. Manufacturer specifications are rarely accurate.
 

JMB

Member
2004-04-01 8:47 pm
Texas
Hi Zaph,

Thanks for your clarification. I have tried both methods (Magnet and Blue Tac) and I agree that the magnet method is more accurate if you can evenly distribute the weight. In my experience it is best to get a bunch of lighter magnets so that you can evenly distribute the weight. I have seen some inconsistent VAS measurements with unevenly placed magnets. These inconsistencies seem to be able to cause even more inaccuracy than the blue tac method, at times. Have you had the same experience?

The blue tac method tends to be fairly evenly distributed in weight if done correctly but I can see your point about the damping effect where the effective weight is influenced by the acceleration. I had just empirically noted that my results seemed better with the magnets, again evenly distributed.

Regarding the box method, I agree that it is generally more accurate assuming you can get enough of a change in VAS for your calculations.

Jay
 
Thanks, all of you. You've given me all my answers.

I'll do a sensitivity analysis... try calculating QB3 box size with my measured Vas, and then with the manufacturer's published Vas, and see how much the box size differs. I'll also see what happens to the response curve if the real Vas is 36L and I build a box as per what would be the size for QB3 for 30L.

And then I guess I'll try to see if I can do the sealed-box measurement anyway.

Incidentally, I didn't apply either magnets or Blu-tak. I just placed the coins on the cone, just like that. And about the percentage decrease in Fs, I got about a 50% decrease with 30g.

Thanks again.
 
jomor said:
i thought Peerless where made in China..... :confused:
The Peerless that most non-Indian DIYers know is a Danish company. Their drivers are made in China today, some other country tomorrow, based on economics. But there's another "Peerless", which is Peerless India, whose Website carries the datasheet I'd pointed to in my earlier post. This is an OEM speaker driver manufacturer. They were born with a tie-up with Peerless of Denmark, to get technology from Denmark and make drivers using cheap labour in India. It seems the Danish tie-up is now gone, but the name remains. Peerless India continues to build drivers in bulk under OEM contracts for a lot of big names in the world. Incidentally, the GR Research M130 series of 5" midbass drivers are made by Peerless India, I believe, and are slightly customised versions of their own S13NG-8 driver or a very close cousin, it is believed. This GR-Research driver is a well respected one in DIY circles: Murphyblaster.com has used it in many projects.

Some of the excess lots of these Peerless India drivers are released in a very unpredictable, unstructured and informal way into the Indian market, and are superior to any other Indian-made drivers available in the market here. The Kevlar drivers I'm working with are made by Peerless India and have been made available in the local market in the same way. They cost the INR equivalent of about $30 each. With this kind of supply channels, you never know when a particular model will disappear from the market's inventory, and whether it'll ever re-appear. Make speakers while they last!

i m not sure which method is more accurate, but i would trust the measurement more than the manufacturer's data. make sure your drivers are runned in before measuring them.
Can you give me some data of exactly what percentage change of Fs, Qts and Vas you have seen before and after breakin? I've heard data from some diyers that break-in is a myth when it comes to impact on measured parameters... hardly any measurable change is seen. I just wanted to verify this fact.
 

wintermute

Administrator
Paid Member
2003-08-03 11:43 am
Sydney
Hi TCPIP,

one thing to be aware of is that as the fs increases the VAS will decrease, therefore since your FS measurements are higher than published specs it is normal that your vas measurements are lower... whether they are in proportion or not I don't know, but I wouldn't worry too much :)

I used delta mass, using a concave washer stuck to the center of the dust cap with some magic tape (the scotch stuff that doesn't leave residue when you peel it off). I did measurements a number of times taking the washer off and putting it back on and got consistent results. I'd previously used bluetack and plasticine, but found that they marked the cone and gave inconsistent results.

I think if the weights were sitting loose they would have been vibrating like crazy and affecting your results... Also if the wheight is significant you might not get accurate results if the cone is horizontal, you probably should mount it vertically.

Tony.
 
wintermute said:
I think if the weights were sitting loose they would have been vibrating like crazy and affecting your results... Also if the wheight is significant you might not get accurate results if the cone is horizontal, you probably should mount it vertically.
The drivers were kept facing up, so I could just place the coin(s) on the cone and do my tests. And the coins must've been vibrating, I guess, but I didn't hear or see any obvious signs. The amplitude of the test signal is quite low, so perhaps the vibrations are practically non-existent?

All this is getting too uncertain. I'll do a sealed box test sometime, before I get into freezing the actual box dimensions. One last question: how much quantitative change does break-in make to the T/S parameters?

Thanks. :)
 

wintermute

Administrator
Paid Member
2003-08-03 11:43 am
Sydney
I'm not at home so don't have my Loudspeaker design cookbook handy, I can't remember the percentage that Vance Dickason says that the parameters can change by (could have been 10% could have been 25% don't remember) (what I do know is that my drivers didn't seem to change at all, but maybe I didn't break them in for long enough)..... but the important thing was that it doesn't effect the design of the cabinet, as the ratios stay the same.

He says that breakin is more important for weeding out faulty drivers because the change in T/S params has little if any effect on the design of the final enclosure.

Tony.
 

JMB

Member
2004-04-01 8:47 pm
Texas
There are a couple of issues. Not fastening the weight will lead to inaccuracy. It is best to attempt to check T/S in the position in which it will ultimately be played though there are many who check VAS horizontally. According to Dickason, the change in T/S parameters from break in will not be so much as to effect ultimate box size but rather it is a way to assure that the drivers do not have any obvious manufacturing defects. Fs is likely to drop somewhat (expect about 5-10%) as the suspension loosens but VAS will increase proportionately with respect to it's impact on ultimate box size.

It will be better to build a box somewhat larger than you calculate. For one thing, you will want to take into account the space taken up by bracing and of the driver. There is a neat program you can freely download to assist with this called BoxyCAD. In addition, by building a little larger and measuring, non linear or weak responses due to an oversized box can be rectified by adding wood (either baffle re-inforcement or bracing, both good things anyway) to the inside of the cabinet.

Good Luck,

Jay
 
JMB said:
There are a couple of issues. Not fastening the weight will lead to inaccuracy... It is best to attempt to check T/S in the position in which it will ultimately be played though there are many who check VAS horizontally.
I have decided to delay my final box designing and do a sealed-box measurement of impedance, so I guess I'll be able to bypass both these problems.
According to Dickason, the change in T/S parameters from break in will not be so much as to effect ultimate box size but rather it is a way to assure that the drivers do not have any obvious manufacturing defects.
Good. This is a big relief. I have a copy of Dickason 5/ed, but I don't remember this point being mentioned.

It will be better to build a box somewhat larger than you calculate. For one thing, you will want to take into account the space taken up by bracing and of the driver.
Actually, I do precise geometry calculations of each piece of each brace in a spreadsheet (and each piece of every other inclusion too) to calculate the difference between the gross internal volume and the net volume. This way, at least I know my net internal volume is exact to perhaps +/-2%.

Now I come to the next sticky point. I found that Speaker Workshop's box design module is a bit simple (only one flat vented alignment supported, etc), so I looked around and found that people consider Unibox very sophisticated. So I downloaded Unibox, fed in some parameters, and I'm getting two different specs for box size which are hugely different from each other.

These are the parameters I fed in for one of the drivers:
[IMGDEAD]http://www.starcomsoftware.com/tmp/personal/photos/unibox-params.gif[/IMGDEAD]

And here are the box specs I got. Note the two hugely different box sizes: 43.9L and 20L!!!
[IMGDEAD]http://www.starcomsoftware.com/tmp/personal/photos/unibox-vented.gif[/IMGDEAD]

What gives? Which of the two am I supposed to take, and with how much seriousness? And is there any way to specify the alignment, as in SBB4, QB3, and so on?
 

JMB

Member
2004-04-01 8:47 pm
Texas
I'm not certain why the results are so vastly different but if in doubt, I would probably go with Unibox.

Regarding precision measurements, I would still be inclined to build the box larger as formulae are only an estimate and measurements will help you get more precise.

Good luck,

Jay
 

wintermute

Administrator
Paid Member
2003-08-03 11:43 am
Sydney
Hi TCPIP

!st thing take out that Rs 1 Ohm Unless I'm mistaken that is for amplifier output impedance and speaker leads, and will greatly affect the results... probably best to leave it at the default of .2 ohms unless you have measured your amp's output impedance. The first model doesn't take this into account but the second one does, that will be where the big discrepancy is coming from!!

Yeah the fact the unibox doesn't have any alignments built in is a bit of a pain, but if you know the q of the alignment your after you can put that in. The other option is to take what speaker workshop recommends, and then plug that into unibox and tweak it from there.

edit: did you press the optimise for wanted QTC button???

edit2: Actually I just had a play with unibox and I got it backwards.... the Rs does affect the box volume for the standard design, but it doesn't change the second one, it stays static because it allows you to change things to see what happens... best bet is to take the standard design and see how that models, but still I'd get rid of that 1 Ohms!!!! for vented I get something out of speakerworkshop and plug it into the relevant section in unibox :)

Tony.
 
JMB said:
I'm not certain why the results are so vastly different but if in doubt, I would probably go with Unibox.
By "going with Unibox", I presume you mean I should ignore the Standard results (which are reinforced by practically every box design program I've checked with, BTW) and go with Unibox's Universal Box Modelling algorithm? Is there any place where I can find more literature which tells me what Unibox does which is so different?

Regarding precision measurements, I would still be inclined to build the box larger as formulae are only an estimate and measurements will help you get more precise.
.
Yes, I'll have to figure out how to implement this idea, but I think I'll take this suggestion very seriously. My box will be so heavily braced that there may not be any clear, large, floor area, for instance, where I can simply add some extra layers of material and reduce box size. But I'll figure something out.
 
wintermute said:
Yeah the fact the unibox doesn't have any alignments built in is a bit of a pain, but if you know the q of the alignment your after you can put that in.
Is there such a thing as a Q of a vented alignment? Sealed box is hardly a problem, because formulae are simple and there are no issues of choosing from half a dozen alignments.
for vented I get something out of speakerworkshop and plug it into the relevant section in unibox :)
Yes, but if I do that, I still don't understand what Unibox does to arrive at its proprietary calculation.
 

wintermute

Administrator
Paid Member
2003-08-03 11:43 am
Sydney
If you put in the specs for you driver over an existing driver you will find that the box volume tuning etc in the section that you can play with are actually the ones that were there for the previous design! They don't change, the whole point of that section is that you can put in whatever you want! the spreadsheet doesn't initialise it, whatever was there before stays there. If you can remember what the driver was before you put your specs in, go back and take a look and I bet that the figures are the same.

I can only assume that when the database was initially populated that the designer plugged the standard volume into the other section and then clicked on optimize or recalculate to take into acount the stuffing, leakage etc! so that there is something meaningfull in there.

yeah sorry I was thinking closed when I said about the Q of the box.... for vented the best bet is to plug in the volume of the standard design and the port recomendation and then tweak them from there, or as I said before get the alignment you want from speaker workshop and plug it in :)

Tony.
 
hate to bring up an old thread but I'm currently attempting to do the same thing, Vas measurement using added-mass in SW, and am getting REALLY strange results.

I'm using the blue-tac method, so that could have something to do with the problem, but the results I'm getting are far beyond innacuracy into complete nonsense.

Here's the story:

I'm trying to measure one of these: (Dayton DA175)
http://www.parts-express.com/pe/sho... &FTR=da175&CFID=8668781&CFTOKEN=22551141

My free-air impedance looks great, and is within 1hz of the factory.

The "unofficial SW manual version 2" mentions using up to 75% of Mms to change the speakers Fs by upwards of 25% to get a good reading. The datasheet for that woofer lists Mms at 25grams, so I tack 20 grams onto it and measure.

Almost no difference.

I quadruple the mass of the blue-tac and re-test. Now I'm getting an added-mass Fs of about 34hz, with a free-air Fs of 41. When I estimate Vas, it gives me a value of .04 cu ft. The factory specs list .5 cu ft.

I eventually went up to as much as 180 grams of added weight, but the calculated Vas was always the same. .04 cu ft

This is completely out of whack.

Could this be because I'm using entirely blue-tac for my weights, and not coins as well, and the tac is doing far too much damping? For reference, I have the speaker on the bench, facing up, with the tack stuck symmetrically to the back of the cone.

Could it be because I'm not running enough power through the speaker? I'm stricly using my sound card--not an external amp--for impedance measurement, for simplicities sake. The speaker makes plenty of audible noise, this way, and as I mentioned, the free-air measurements are very close to factory specs. (41hz to the factory's 40)