Question re near-field low-frequency measurement

[mefisto]

Greetings all, I have rebuild my experimental low frequency enclosure to allow rear driver mounting, and I would like to measure it. However, we are still snowed in, so I was reading on the near-field measurement as proposed by DB Keele Jr in his 1974 paper. However, my driver ALTEC 515 has a hole, protected by wire mesh, in the middle of the dust-cup, so I was wondering if the method will work. Anybody has an experience? Could I tape over the hole for the measurement? Kindest regards, M

[Ian Mackenzie]

Hi,

The vent in the dust cap is not important.

I am not familiar with the change make you your loudspeakers?

However, a near field mic measurement close to the cone will only measure the output of the Altec 515 woofer. Not the output of the bass reflex vent or port. To measure the output of the vent place the mic immediately in front of the vent. Note you will see a broad hump representing the vent output. The output will be different due to the velocity of the air movement in the vent.

If you wish to determine or confirm the vent tuning frequency sweep a test tone from 20 hertz to 50 hertz. Using a finger on the edge of the cone diaphragm the smallest movement of the cone is the true vent frequency. Check for leaks around the driver gasket.

[RMC]

mefisto,

RE Could I tape over the hole for the measurement?

Had this in the making already.

I know Keele's near-field measurement method because its described nicely in J. Eargle Loudspeaker Handbook. Haven't tried it yet, but i'd risk to say try it both ways, with and without tape applied. If the results are the same then its a no brainer.

HOWEVER, be careful in applying the tape don't press it too much in place avoiding it being too sticky so you don't damage the dust cap when removing the tape. In this regard you must remember that some tapes are very sticky and others less so.

For example, the tapes on the pic are all made by 3M. But the Tartan brand sticks a lot, even when lightly pressed, so it shouldn't be anywhere near a dust cap or cone. The other two are painters tape so meant to be removed easier. For delicate jobs i typically use the Blue and don't press it too much in place.

With any tape you plan to use you need to try some on a piece of cardboard or paper first to get familiar with how sticky it may be... Good luck.

Richard

[mefisto]

Hi Ian, thank you for the reply. The enclosure was originally built for a driver to be mounted from the front. This, of course, does not work for the Altec, which has ribs on the back side of the frame, hence the need for rebuild. The enclosure is sealed, so no need to measure the went. I am no expert on fluid dynamics, but the naive intuition would suggest that the pressure at the vent hole will be different than at hole-less dust-cup. Hence my concern. Hi Richard, thank you for the reply. Yes, I had concluded to make two measurements, one with the vent hole as is and second with the went hole covered with a tape. And yes, I am rather worried about the potential damage due to the glue, as well as a different property of the tape vis-a-vis the dust cup. But maybe the latter is a second order issue? Kindest regards, M

[mefisto]

Why do my posts not respect paragraphs? Kindest regards, M

[RMC]

You're not the first i see here with paragraph problems.

I never had an issue with this, but i rarely write posts live. I start a text, save what is done then frequently continue to work on it later since i'm interrupted more than i'd like.

Therefore i write in Windows Word Pad and when i'm ready to post then i copy and paste to LH forum page. Moreover before posting i put double space between para. otherwise its all packed together. Things work well this way, no para. problems for me.

[Mr. Widget]

Why do my posts not respect paragraphs? Kindest regards, M

Last fall Lansing Heritage began a migration process to join AK. It has been a bumpy ride and some things have gotten broken, allowing new members to join for example.

In any event EarlK figured out a hack to allow text formatting: https://www.audioheritage.org/vbulle...l=1#post442784


Widget

[mefisto]

Hi Richard, Mr. Widget,

testing the new settings.

Kindest regards,

M

[sebackman]

Hi mefisto,

Skip the tape idea, it will not be "stiff" enough to resist the air vent pressure.

You can put the mic on the side of the dust cover and there is no difference compared right in front. The reason to use the dust cover is that it is stiff and "small enough" to not flex as the cone can do. With a very stiff cone (light and/or strong) you can measure on the inner edge of the cone.

The only limitation is that you are close enough so that all indirect sounds are much lower than the direct sound = <0,5cm.

An alternative when looking for the resonance frequency is to do as Ian suggest (down-load a signal generator app to the phone and connect to the power amp) but add a 30-60 ohm resistor in series between the driver ad the power amp and use a voltage meter direct over the coil (after the resistor). The voltage meter will give two peaks in the low end, one on each side of the resonance frequency. The frequency where there is resonance is where you get the lowest voltage read between the peaks. Pretty accurate.

[mefisto]

Hi sebackman,

thank you for your reply.

As I have been reading about the measurements, it appears the the requirement is stiffness - as you have also noted - so that the measurement is not contaminated by non-piston modes. Thus, I was thinking that a good point to measure from would be the glue edge of the dust-cup to cone.

Kindest regards,

M

[Ian Mackenzie]

Hi Mefisto,

If you can elaborate on what you are attempting to achieve with the Altec 515 woofer then it would simplify the discussion here via one way posts.

Put simply if we know what your trying to do in the big picture we can give you more complete and concise responses.

What someone published in a book 40 years ago mightn’t be relevant to achieving the outcomes your looking for.

> Describe your project?
> what progress have you made?
> what sort of results are you aiming for?
> Explain what you want from the measurements?

Mid range frequency measurements?
Measurements below 100 hertz?
What measurement system do you have?

As a primer it’s not a simple task to conduct a useful low frequency measurement indoors. This is particularly true if the enclosure is vented or has a port.

In either a vented or an infinite baffle system the key is to understand the T/L parameters of your particular woofer via measurements if possible. Then determine with the help of a simulator program or TL alignments the required enclosure volume. There are some good online programs that allow custom enclosure simulations.

The reason for pointing this out is that with the measurement of your 515 TL parameters you can accurately predict the low frequency performance far more accurately than is possible by a measurement with a microphone indoors. It will also preserve your marriage!

A simple nearfield frequency response measurement is a helpful objective indication of the woofers overall response. But a near field measurement will not detect the effect of the enclosure baffle shape or the effect of the woofer location on the baffle. This can be simulated with some programs.

One simple way of determining if your woofer is going to meet your expectations is to mock up a test enclosure and a simple text book passive crossover network with the high frequency driver - horn and evaluate it subjectively. This assumes the woofer is for inclusion in a full range loudspeaker. Your ears will tell your straight away if the woofer is worth pursuing.

This is exactly the approach described by one of JBL’s most influential and successful engineers Greg Timbers.

Simple on axis measurements are an objective yardstick but they impart very little information about how a loudspeaker actually sounds to your ears in your room. A flat frequency response and a loudspeaker that is subjectively favourable are not mutually exclusive unfortunately. Anyone who thinks otherwise is kidding themselves. This is not generally a problem because they live with their own situation and nobody cares.

Unfortunately magazine curves are the eye candy used to sell subscriptions.

Unfortunately we can be subjectively influenced by what we have seen with these curves before hand. This is why some loudspeaker manufacturers do double blind tests during the development of their designs.

Hopefully none of this will cause too much social disruption or nail biting….Lol.

If reading the above causes deep inner tensions or disagreement go for a walk around the block and it will pass.

[mefisto]

Hi Ian,

thank you for the extensive reply, and no, it did not "cause too much social disruption or nail biting…." But, just to be on the safe side, I have taken a "walk around the" mountain. ROTFLMAO.

To start with the questions under your "big picture":

1. Describe your project.

I would like to make a speaker similar to the one built and documented by Gary Dahl (g3dahl), in the "Beyond the Ariel" thread on diyaudio.com. His project is spread over the ginormous thread, so it is not easy to follow, but I concur with his objectives and would be elated if I could achieved his results.

I understand that Gary used a different driver (Altec 416) in the final implementation, but he did experiment with the Altec 515, and selected the Altec 416 because the Altec 515 had "lean bass" whatever that means, see below, and he did not want to equalize the raising frequency response of the Altec 515. However, in a private communication, he encouraged me to try the Altec 515.

2. What progress have you made?

I have modeled the enclosure and build one to measure. I am currently waiting for some parts to re-build my impedance measurement fixture.

3. What sort of results are you aiming for?

I believe that I have answered this in the 1 above, please let me know if I missed the point of this question. The model predicts that I could achieve the desired result, please see the attached Altec*_515B_120L_0.559.gif. In this regard, I do not quite understand the "lean bass" comment, the prediction shows 98db/30Hz/25W, and since my listening is at average 80 db, . . .

4. Explain what you want from the measurements?

I would like to verify that the prediction correlates with the measurement because:
(i) the simulation is imperfect, e.g., the parameters like leakage, absorption (damping) are only estimates. I have confirmed this when we measured several TAD1601 drivers that a friend acquired for us, the ground plane measurement was correlated to the prediction based on the T/S parameters and the enclosure size, but not as accurate as predicted. Now, it could perhaps be remedied by adjusting the absorbtion, but we were interested in comparing all the drivers. (This is actually is backup idea - to build a TAD Exclusive 2402 replica).
(ii) the simulation does not predict the mid-frequencies for cross-over design (Gary's is at 750 Hz), and
(iii) I want to gain confidence in the measurement, so that I can rely on them when measuring crossover and overall response at the final location of the speakers.

Additional questions:

5. Mid range frequency measurements?
6. Measurements below 100 hertz?

I believe that these are answered above with reference to the predicted response.

7. What measurement system do you have?
UMIK-I microphone, REW and HolmImpulse software.

"A simple nearfield frequency response measurement is a helpful objective indication of the woofers overall response. But a near field measurement will not detect the effect of the enclosure baffle shape or the effect of the woofer location on the baffle. This can be simulated with some programs."

I understand that the near-fields measurement has limitations regarding this issue, but Jeff Bagby wrote an article accompanying his "Blender" spreadsheet enabling "blending" the near-field and mid-frequency measurement. But, that is further down the road. In addition, this may be academic anyway because the snow is melting rather quickly so in a week or two I might be able to do ground-plane measurement outside.

Kindest regards,

M

[Ian Mackenzie]

That all makes sense.

1. The reference to lean bass means an over damped or critically damped close box system.

2. Closed box loudspeaker systems relate closely to high pass filter design.

The Q of a closed box can range from 0.707 for a maximally flat Butterworth design to a Q of 0.5 for a critically damaged design. These lower Q systems have much large enclosure volume. Under damped Chebyshev designs with a Q = 1.00 are also possible. The box volume for a 515 driver is only 20 L for a C 2 under damped design.

So what you have in the diy space is the flexibility to try several different closers box Q’s to determine your preferences. There is no one right answer as long as your chosen design is close in practice.


3. A few questions focussing on T/L measurements.

What system did you use to obtain your T/L parameters? REW or DATS3?
Have you tested your TL data in a software model for errors?

4. T/L Measurements of large drivers can be erroneous and therefore can be difficult to model accurately according to Greg Timbers (JBL engineer).

Models typically don’t account for the motor inductance and real world non linear behaviour of the suspension or the BL curve. The stiffness of suspension compliance of large drivers can make VAS measurements inaccurate.

Small signal T/L data measurements can be different after several test cycles due to the voice coil warming up. I would test the driver for T/L data after running a 25 watt sine wave for 15 minutes and compare the data with a cold driver.

If you are interested in a more in depth understanding of Closed Box T/L parameters have a look at Dr Linkwitz website. There is also a good coverage on models versus practical enclosures in the book Bullock on Boxes.

5. I found the original Altec 515 specs here

http://pdf.textfiles.com/manuals/STA...%20515-16G.pdf

6. The amount of Fill (fibreglass) has the effect of adding virtual volume to the enclosure. You may decrease the physical enclosure volume to account for the internal “Fill”.

Completely Filling the enclosure will reduce the output marginally and lower the closed system.

Try the smart phone enclosure simulator App called Speaker Box Lite.

[mefisto]

Hi Ian,

thank you again for your helpful answers.

Re 1. You are smart as a badger. From my conversation with Gary, he was aiming for a maximally flat system, however, the Q of the Altec 515 is quite lower than that of the Altec 416, which would yield an over-damped Qts, and, hence the "lean bass" comment, correct?

Re 2. IM: "So what you have in the diy space is the flexibility to try several different closers box Q’s to determine your preferences. There is no one right answer as long as your chosen design is close in practice."

As you suggested, there is no "right" in selecting the Qts. When I was researching the choice of Qts, the opinions were ranging from Qts=0.5 to Qts=1.2. Regarding your comments on flexibility, the Vb=120l that I built will be lowered by the volume of the driver, but then increased by the fill. So, the predicted Qts will be about 0.58. I can lower the volume by inserting material into the box, thus increasing the Qts, but I do not know how to decrease the Qts, apart from building another enclosure. I have some notion that Linkwitz transform may change some of the parameters, but I will have to do more research.

Than, there is more overreaching question - how does one know what one prefers if one cannot listen to the different Qts's extensively?

Re 3. IM: "What system did you use to obtain your T/L parameters? REW or DATS3? ?"

Neither, I have measured the Z=f(f) with the cone unweighted and weighted and then used a VituixCAD to calculate the parameters.

IF: "Have you tested your TL data in a software model for errors?"

No, I have no idea that it is even possible. Can you please provide a reference?

Re. 4 IM: "T/L Measurements of large drivers can be erroneous and therefore can be difficult to model accurately according to Greg Timbers (JBL engineer)."

I do understand your point, but, where does the implication lead? If the T/S parameters measurement is so unreliable, what is the point of measuring them in the first place? Just to have a starting point ? And, does it not put into question your assertion form previous answer that the low frequency can be determined accurately from the simulation?

Let me ask you the following: Could one work it out backward? That is, if one had "correct" measurement of the frequency response of a driver in an enclosure, could one derive the T/S parameters thereof?

Re 5. The problem according to people familiar with Altec and 515 in particular is that the parameters vary widely among different samples of the same driver designation. i.e., B, G, etc.

In that regard, I do not know if anything can be told from our measurements of the 12 samples of TAD1601's, but the frequency response measurements were literally on top of one another, the difference being about 1dB. Perhaps the published T/S specifications measured by TAD could be trusted? Maybe I should just build the 2402 replica and be done with it, but I do not like the bass from port assisted speakers.

Re 6. Yes, I know, please see 2 above.

Kindest regards,

M

P.S. Are you taking walks around a block when you read my posts?

M

[Ian Mackenzie]

I am out having a coffee going through your post one step at a time!

Quote

I have put my response in on notes and pasted in the post so you can keep continuity with your questions :-

Hi Ian,

thank you again for your helpful answers.

Re 1. You are smart as a badger. From my conversation with Gary, he was aiming for a critically damped system, however, the Q of the Altec 515 is quite lower than that of the Altec 416, which would yield an over-damped Qts, and, hence the "lean bass" comment, correct?

Yes. You can however manipulate the QES + some series resistance Rs like adding a series 0.2 -.033 ohm resistor and a parallel shunt resister like 80 ohms down to 39 ohms on the woofer.

JbL used this technique to voice some of their bas reflex systems. It’s part of the loudspeaker engineers box of tricks. Many of Greg Timbers systems used a damped woofer to
Improve the low pass crossover filter and hence voicing the overall system response.

The close box is largely mechanically damped with suitable closed box woofers with a relatively high Qts. The the case of the 515 is got a relatively high BL and a low Qts making a sealed system look over damped.

Adding additional fill (fibreglass) will lower the system Q.

You also need to consider how your going to cover the first two octaves. The sealed 515 system will yield a damped mid bass but it has a relatively limited Xmax. Without the support of a bass reflex port or vent the 515 will have subjective limitations on its dynamic output below 100 hertz.

Re 2. IM: "So what you have in the diy space is the flexibility to try several different closers box Q’s to determine your preferences. There is no one right answer as long as your chosen design is close in practice."

As you suggested, there is no "right" in selecting the Qts. When I was researching the choice of Qts, the opinions were ranging from Qts=0.5 to Qts=1.2. Regarding your comments on flexibility, the Vb=120l that I built will be lowered by the volume of the driver, but then increased by the fill. So, the predicted Qts will be about 0.58. I can lower the volume by inserting material into the box, thus increasing the Qts, but I do not know how to decrease the Qts, apart from building another enclosure. I have some notion that Linkwitz transform may change some of the parameters, but I will have to do more research.

Than, there is more overreaching question - how does one know what one likes if one cannot listen to the different Qts's extensively?

That’s a good question.

A driver volume will not have a perceived effect in a 120 L enclosure.

Additional full will lower the system Q, not the woofer Qts as such.

There are ways to manipulate the system Q with some series dc resistance on the woofer which will raise the woofers Qts and additional fill inside the enclosure which will lower the closed box system Q.

But from a technical perspective you need to anchor what the woofer TL parameters are within reason unless there are gross tolerance issues.

Can you quote or attach the Altec quoted Qes and QTs, the Fs and Vas of your 515 version and your findings?

Within the continuum of low to high Qts woofers there are some woofers classes which work in both a bass reflect and a sealed system. But not many hence most woofers these days have a Qts for compact bass reflex designs with a passive radiator. Small enclosures rule these days.

In your situation l think it’s best to try and validate your woofers TL parameters within a model (l can do that for you). If it looks odd then acquire a Dayton DATS3 woofer tester.

It depends if your woofers are noticeably off the manufacturers TL data or not?

Then work with what you have got. See my outline above on raising the system Q.
Lowering the system Q is about more fibreglass insulation or a bigger box. Your 120 L enclosure looks as large as you would want to go. So try removing the fibreglass all together just it see the effect!

You can also also damp the woofer with a parallel resistance of 80 ohms down to 39 ohms.This assumes your power amp has a nominally low output impedance.

It’s also about taste and overall voicing in your listening space. What may look correct on paper does Not always translate into an enjoyable listening experience.

Re 3. IM: "What system did you use to obtain your T/L parameters? REW or DATS3? ?"

Neither, I have measured the Z=f(f) with the cone unweighted and weighted and then used a VituixCAD to calculate the parameters.

IF: "Have you tested your TL data in a software model for errors?"

No, I have no idea that it is even possible. Can you please provide a reference?

My recommendation is to acquire the Dayton DATS3 woofer tester or attempt TL measurements with REW and an audio interface. See the REW help guides online.

FWIW sometimes diy people get too hung up on TL parameters and they end up off the reservation. Allowing for a tolerance in key TL parameters on a run in woofer is a safe strategy. In the scheme of things the group or class in which your 515 belongs on the Qts continuum is more important overall. In other words be mindful of what it’s intended purpose is.

Re. 4 IM: "T/L Measurements of large drivers can be erroneous and therefore can be difficult to model accurately according to Greg Timbers (JBL engineer)."

I do understand your point, but, where does the implication lead? If the T/S parameters measurement is so unreliable, what is the point of measuring them in the first place? Just to have a starting point ? And, does it not put into question your assertion form previous answer that the low frequency can be determined accurately from the simulation?

Let me ask you the following: Could one work it out backward? That is, if one have a "correct" measurement of the frequency response of a driver in an enclosure, can one derived the T/S parameters thereof?

In a Loudsoft loudspeaker driver design problem it is possible to model and predict driver on a particular response, efficiency and enclosure size.

Typically testing of the Qts is done in production for a tolerance range. This is done with rub and buzz tests and sweep tests. The manufacturers know pretty much what a woofer will do after 100 hours of use.

In the case of Altec if their QA was wanting (which is possibly) that’s a different problem. What you need are several of the same woofers to test side by side at the same time properly clamped and so on after break in.

The diy person doing TL testing is okay but it’s un controlled if that makes sense.

Refer to my comments on plugging your results into a model. The DATS3 system does offer sone validation if sone of the parameters are off.

Re 5. The problem according to people familiar with Altec and 515 in particular is that the parameters vary widely among different samples of the same driver designation (B, G, etc).

In that regard, I do not know if anything can be told from our measurements of the 12 samples of TAD1601's, but the frequency response measurements were literally on top of one another, the difference being about 1dB. Perhaps the published T/S specifications measured by TAD could be trusted? Maybe I should just build the 2402 replica and be done with it, but I do not like the bass from port assisted speakers.

See my comments above on QA. I would contact Todd over on the Altec forum and also Great Plains Audio on their experience with original Altec 515s.

I did have a pair of 515 woofers once. I never got much joy out of obtaining a satisfactory outcome to fit my needs. But that’s just me. In the right situation they are excellent.

Re 6. Yes, I know, please see 2 above.

Kindest regards,

M