Plane Wave Tube Construction

[1audiohack]

Hi All;

I have decided to build a plane wave tube. For a start I have the AES papers on them. What I would like to do is to duplicate the one used at JBL and or Brush Wellman.

In the transducer library the JBL 2435, 2450 and 2452 were all measured with TEF, the same system I primarily use. The closer I can stay to the JBL set up the more comfortable I will be.

Is there anyone here that has access to this information? Any help will be greatly appreciated!!!

[Mr. Widget]

I recently toured the Brush Wellman factory and testing facility... both were impressive and their attention to QC very impressive, but I do not think they use a plane wave tube. I'll find out and let you know more.


Widget

[yggdrasil]

Here's something I have found somewhere on the internet:

rho = density of air
c = speed of sound
d = diameter
Upper limit = 1,22 * c / d
lower imit = c / 4 / length of tube

Now let's talk about horns driven by compression drivers. One of the handy pieces of information that used to be included with compression drivers (but is not always now) is the plane wave tube response. A plane wave tube is a tube the same size as the exit of the driver (or smaller or larger and including an adaptor to have a smooth transition from the driver's throat to the size of the tube). There is absorbing material in the tube which is designed to provide a certain load to the driver - a load of rho * c * diameter of the tube, where rho is the density of air and c is the speed of sound. This is usually called a "rho c" load; rho * c is the specific impedance of air. This load is supposed to be constant at all frequencies and allow easy comparison of different drivers, but in reality the size of the tube determines the frequency range measurements are useful in. The upper limit of the tube is set by the diameter. 1.22 * c / d is usually given as the upper limit. There will typically be a notch at this frequency and other notches above it. The lower limit is c / 4 / length of the tube. So only the response between these frequencies should really be looked at. With high frequency horns, you can typically build them large enough to obtain a rho * c * some constant input impedance that is relatively flat above the cutoff frequency (depending on the type of horn). This means in theory the frequency response on the horn will be the same as on the plane wave tube (although the sensitivity will be different - compression drivers' sensitivity on a PWT is very high). The only caveat here is that many high frequency horns have a polar response that varies with frequency. In other words, at low frequencies the horn is not big enough to the control the sound radiated by it, so the sound spreads out over a wide angle. At high frequencies, the horn becomes acoustically large compared to the sound waves being radiated, and the sound is confined into a narrower angle. This means that at higher frequencies more power is concentrated over a given area, so the sound pressure level is higher in that area (but lower outside it). This is effectively an acoustic equalizer, so this effect needs to be added to the plane wave tube response to come up with the on-axis response of the horn / driver combination. This also explains why typical horn design programs do not predict the on axis response of high frequency horns very well - they don't typically include this factor.

Attached is a spreadsheet for calculation of PWT's.

[JeffW]

Google Books has a limited preview of John Eargle's "Loudspeaker Handbook" (http://books.google.com/books?id=Twu0oHE1ukgC&pg=PA179&lpg=PA179&dq=plane+wave+tube&source=bl&ots=CmMcs5W-SH&sig=qEXRBdJDJ0xGWUah4pak4EEVaE8&hl=en&ei=mzz6TYz1Fcbr0gHVjP2dAw&sa=X&oi=book_result&ct=result&resnum=15&ved=0CG4Q6AEwDg#v=onepage&q=plane%20wave%20tube&f=false) with formulas and tube sizes/diameters for various compression drivers, pp 180~181

[grumpy]

I've pondered doing this... found clear tube source... bought much cheaper opaque
1.5" and 2" tubes (won't be able to see/monitor absorbent "wedge")... read
about sensitivity of mic port placement and repeatability of mounting and sealing
the drivers under test (precision and stability of mount plate)...
end result: tubes are still in the garage :o:

Hoping you have more drive/desire to see this through :)

[Allanvh5150]

Hi All;

I have decided to build a plane wave tube. For a start I have the AES papers on them. What I would like to do is to duplicate the one used at JBL and or Brush Wellman.

In the transducer library the JBL 2435, 2450 and 2452 were all measured with TEF, the same system I primarily use. The closer I can stay to the JBL set up the more comfortable I will be.

Is there anyone here that has access to this information? Any help will be greatly appreciated!!!

I think this piece of gear was used in the Manhattan project. You really want to split atoms?

[1audiohack]

I think this piece of gear was used in the Manhattan project. You really want to split atoms?

No, I want to split hairs. I don't want to build that thing, well, of course I do but,,,

Fortunately information about PWT's is prevalent, I am hoping to build identical to JBL's units, if the information to do so is not available then I will do the math and build a suitable unit. I am simply hopeful that someone here knows the dimensions of the JBL PWT's.

Thanks all,
Barry.

[JeffW]

The JBL model that 1audiohack wants to replicate is one driver on one tube. He just posted that pic for fun, that's not precisely what he's going to build.

[1audiohack]

Wow someone did some serious house cleaning here! Thank you.

The published JBL measurements are somewhat condensed as are most others I have been able to find for reference. I do however like the fact that the charts shown for the 435Al, 2450 and 2452 are shown with the 2nd and 3rd harmonic traces in actual place as compared to the fundamental tone, I think it gives a better picture to what's real going on.

What is missing is the phase response. Unless the responses are averaged, TEF will provide the phase response of not only the fundamental but the harmonics as well. This is I believe important, a harmonic in a reproducer is of of of course not desirable but a harmonic that is also out of phase (time) is nothing but total noise. Some of the magnitude traces have the second harmonics within 20dB SPL of the fundamental. I have duplicated these measurements very closely and the phase response falls thousands of degrees in reference to the fundamental in the frequency range that is said to be usable. These drivers used this way would not even qualify for use in the Bell Labs Symposium on Auditory perspective published in 1934!

A primer to what I am looking at;

The newer drivers (that I can get my paws on) used in the high-performance audio applications have very good magnitude and phase responses in the minimum phase passband and have markedly low 2nd and 3rd harmonics that also are phase coherent, so the signal to noise ratio is very good but on the ends these drivers loose it big. The trade is improved performance in a more limited band width for a more musical compression driver. The goal here for me is not to design a driver, rather to know how to best use what's available with the fewest compromises in driver integration.

Looking at (measuring) the newer drivers with all of the different diaphragms available has revealed some interesting details to me. I know what I am doing has been done before and is known to some. I fully intend to share when I believe I can defend what I am finding. In short I believe I can now describe many of the subjective differences I hear with objective measurements. Time will tell and some of you will be the first "peer review." Just remember this is a (my) hobby, not my profession and I am just trying to learn.

Now I just have to build the tube assemblies.

I haven't researched or sourced the PWT foam yet, has anyone already done that? Grumpy?

Thanks again all,
Barry

[grumpy]

I was going to use fiberglass, inserted as a long wedge shape. Then measure and rearrange as necessary... Hence my search for clear tubes and a place to put the thing so it would not be disturbed. I expect other materials would hold shape and position better... with less fiddly reassembly.

If I can remember the articles that sort of described several pwt's that were used at JBL, I'll post the links... One was curved for compactness, iirc. You may already have this info.
There was at least one aes paper that described some of the issues involved with building and using pwt for audio testing.

[Allanvh5150]

All interesting stuff. One thing I have never really understood though and no one may have the answer, why did JBL always test drivers into a 1" tube? Surely this would upset the load that the 2" throat driver sees? Please enlighten.

Allan.

[grumpy]

Iirc, for comparison's sake... most mfgs convert their readings to values as though they were taken on a 1" pwt. There are reasons why using just one tube for extended freq measurements might not be adequate (e.g. Hf limits using a large tube).

[Allanvh5150]

Iirc, for comparison's sake... most mfgs convert their readings to values as though they were taken on a 1" pwt. There are reasons why using just one tube for extended freq measurements might not be adequate (e.g. Hf limits using a large tube).

Thanks Grumpy. I thought it would have been a better idea to spec the driver on the horn or horns they were meant for. Real world specs. Do you know of the reason why they tested in this way?

Allan.

[1audiohack]

A plane wave tube is a way to eliminate some of the variables. Without doubt what happens when a horn is installed is most important, that's where the rubber meets the road.

I have swept these things in every configurable way available to me from hanging in space with the rear caps off and nothing bolted to the throat to caps on with the throat stuffed and a plate bolted over it's throat in attempt to discern what features of the impedance trace are connected to the driver it's self and what features are horn related.

It is said that a knowledgable teacher only saves an eager student time. This is surely true, many of the tests I have done most likely serve no real purpose besides narrowing the scope of the search. Getting down to asking the right questions and spending time and energy on the ones that really matter is the process I am still in.

Thanks again for all the help!

[JeffW]

Thanks Grumpy. I thought it would have been a better idea to spec the driver on the horn or horns they were meant for. Real world specs. Do you know of the reason why they tested in this way?

Allan.

Read the link to the Eargle book I posted upthread, it goes into a little detail about using different sized tubes.

[4313B]

JBL uses a smallish curved PWT that fits in the corner of a small room where their Klippel machine is also located. The shelves next to the PWT are stacked with every compression driver JBL ever made. I was not allowed to take a photo. :D At the time, the driver bolted to the PWT was a 275Nd.

The mouth of the vertical portion of the PWT where the compression driver is bolted on is at about waist height and the PWT curves where is meets the floor and extends out for a few feet where it ends. They might have others but the question didn't come up.

[grumpy]

I swear I've seen a picture of that thing somewhere... :dont-know:
or at least a trial version.

1audiohack... in the AES materials (1991 design and practice):
Common materials are 3 lb/ft3 (48 kg/m3) fiberglass, and 70 to 90 pore-per-inch Scott
reticulated foam (available from Scott Paper Foam Division, Essington, PA).
(http://forums.klipsch.com/forums/storage/6/1329418/palnewt.pdf)

[4313B]

I swear I've seen a picture of that thing somewhere... :dont-know:It's certainly possible. Don might have shot one during one of his visits. Heck, I might have taken one too and just don't remember. :p

[Rudy Kleimann]

It's certainly possible. Don might have shot one during one of his visits. Heck, I might have taken one too and just don't remember. :p

I seem to recall a JBL catalog or brochures from the '70's showing a picture of their driver test benches, including the compression driver bench with the PWT. It may (also) be an Altec catalog or brochure.

I want to say I saw it somewhere on the LHS site. It could be on an Altec site. If I had time right now, I'd do some sleuthing for it. Maybe tonight...

Grumpy, thanks for sharing here.

I've met 1audiohack and seen some of his goodies. He can do some really fine work, given the right tools and setup. What I'd give for one week with his TEF analyzer... :hmm:

[grumpy]

Any bonehead can point to what more experienced people have published. I just hope
I point to more good than bad. :)

[1audiohack]

in the AES materials (1991 design and practice):
Common materials are 3 lb/ft3 (48 kg/m3) fiberglass, and 70 to 90 pore-per-inch Scott
reticulated foam (available from Scott Paper Foam Division, Essington, PA).
(http://forums.klipsch.com/forums/storage/6/1329418/palnewt.pdf)

Many thanks.

[1audiohack]

It's alive1 It's alive!! It's alive!!!

I am still tuning the dampening after finally finding a way to pull it through with out tearing it.

I did some driver sweeps and after swapping eight drivers and two mic's, powering down the computer and amp, restarting anew and re-sweeping the first driver I tested a day later the single largest variation on a small peak was 0.4 dB, another small peak was 0.2 dB, the rest was a single line imposed on the original line from 200 to 22000 Hz,,, finally, I have a way to pick the gnat crap out of the pepper, and it's fast!

[JeffW]

Do you have adapters for that pile o' drivers, or just gonna machine something up?

[grumpy]

Nice job. It's that driver/pipe/mic interface, that you've nicely machined, that slowed me to a stop.

[Rudy Kleimann]

What's the input torque rating on that flange and tube?

Looks great! Hardly a "hack" job...:applaud:

Is it a "Plain wave"? :wave:

Seriously though, is the tube I.D. 1.5" or 2"?

It's alive1 It's alive!! It's alive!!!

I am still tuning the dampening after finally finding a way to pull it through with out tearing it.

I did some driver sweeps and after swapping eight drivers and two mic's, powering down the computer and amp, restarting anew and re-sweeping the first driver I tested a day later the single largest variation on a small peak was 0.4 dB, another small peak was 0.2 dB, the rest was a single line imposed on the original line from 200 to 22000 Hz,,, finally, I have a way to pick the gnat crap out of the pepper, and it's fast!

[1audiohack]

The I.D. is 1.550". I just made a tapered aluminum sleeve and pressed it in the big end of 2327 for the 1" drivers so the original flare rate is kept until it reaches 1.500". I have a Delerin sleeve for the 2" drivers that slips in the throats up to the screens.

I plan to use a 2451 for critical diaphragm comparisons.

[ivica]

It's alive1 It's alive!! It's alive!!!

I am still tuning the dampening after finally finding a way to pull it through with out tearing it.

I did some driver sweeps and after swapping eight drivers and two mic's, powering down the computer and amp, restarting anew and re-sweeping the first driver I tested a day later the single largest variation on a small peak was 0.4 dB, another small peak was 0.2 dB, the rest was a single line imposed on the original line from 200 to 22000 Hz,,, finally, I have a way to pick the gnat crap out of the pepper, and it's fast!
This is really marvelous work.
Bravo Barry.

[pos]

You did it Barry! :bouncy:

Can't wait to see your measurements!!
Do you have a calibrated measurement mic?

[bstleve]

Great job ! :applaud:

Pos, stop dreaming, that will not fit in your flat ;)

Bertrand

[pos]

Hey Betrand!
Do not forget the balcony ;)
Well, I have no intention of building it: I prefer to rely on Barry's measurements :D

[Mr. Widget]

Very cool... when do we get to see the results.:)

It would be great if you could try to replicate one of JBL's plane wave measurements.


Widget

[1audiohack]

To replicate as closely as possible the JBL tests is the plan. There are at least three JBL sheets that were measured with TEF and the setup parameters are visible on two of the sheets so that parts easy. The tube is currently overdamped and I just don't want to post a graph until it is dialed in. There must be a hundred feet of carefully cut scrap in the shop-vac by now but were getting there.

I do have calibrated mic's.

The only driver I haven't yet made an adapter for is the 435Al.

[ivica]

To replicate as closely as possible the JBL tests is the plan. There are at least three JBL sheets that were measured with TEF and the setup parameters are visible on two of the sheets so that parts easy. The tube is currently overdamped and I just don't want to post a graph until it is dialed in. There must be a hundred feet of carefully cut scrap in the shop-vac by now but were getting there.

I do have calibrated mic's.

The only driver I haven't yet made an adapter for is the 435Al.

I know that using Plane Wave tube is standard method for measurements compression drivers, as I understand mainly because there is not present "interference of the applied horn with the driver", but on the other side such information what would be the LISTENER expectation or experience, is of much less informative, I believe mainly influenced by the driver-horn relationship.
Of course, for the comparison with the producer data, or between the drivers, such method is very good.
In any case Barry's work is great.

[herki the cat]

Wow someone did some serious house cleaning here! Thank you. Now I just have to build the tube assemblies. I haven't researched or sourced the PWT foam yet, has anyone already done that? Grumpy? Thanks again all, Barry

You may be interested for future reference that we had a plane wave tube very similar to yours in RCA Camden. The damping material was commercially available "Cow's Hair" assembled in 6 inch long puffs pulled through the tube by a string. Acoustic calibration was excellent measured with a WECO 640-AA microphone. Cow"s hair with it's complicated mixture of hair components forms a very consistent, firm packing bundle.

The RCA 77 DX microphones, and all the 77' Microphone series ancestors, all used the same "cow's Hair" damping system in a labyrinth tube approx 0.375" diameter by some 60.0 inches long behind the microphone ribbon.

Cheers, herki

[1audiohack]

That's interesting Herki, thanks.

The AES paper refers to The Scott Paper Company's Foam Division as the source of the 80 PPI reticulated foam used as dampening material for the tube. As it turns out that division was sold off, and has been at least two separate companies since. I chased them down to find that they have large minimum orders and a whole loaf of foam at $1100+ USD barely qualifies. :^( I found places that sell in smaller quantities but have no serious cutting or fabrication capacities. In the end to have 1.550" diameter round bars cut 6 feet in length of 80PPI zapped reticulated urethane foam the cost was approaching $2000 and I would still have to cut it in wedges. I had high hopes of finding a shop that could cut conical shapes with some measure of precision but I found none. I even seriously considered buying Foam Trends when it was for sale but could not make voice contact with them. So, to make a long story short, TOO LATE!!! I am using fiberglass.

I have experimented with density and shapes for many nights and I have a shop vac with hundreds of feet of fiberglass snakes in it and have found that a fill from very light, just a whisp of loosely teased glass to packed tight enough to have to push it out in pieces into the waiting shop vac with a stick that the curve is only shifted from by a maximum of 7+ dB on the top end, what really changes is the impedance swept values. I currently have a reasonable duplicate of the impedance values as compared to what the JBL graphs show but the HF is attenuated about 10+ dB.

I am still testing with the 2426 on the modified 2327 adapter while waiting for the 2451 to arrive so I haven't firmly decided that this is as close as I am going to get but it may be.

Since this is a reference tool for comparison and not a listening device maybe the discrepancy is OK?

If anyone is still reading:p and has any experience working with one of these things, what do you think?
I believe I have chased the tuning with fiberglass to an end.

Thanks,
Barry.

[grumpy]

Port placement for the mic can have a significant effect on the HF response, if I recall right, being a few inches from the diaphragm with the 2327 adapter might be contributing to what you're seeing. Unfortunately this is more of an arm waving recollection than a specific recommendation. Hoping the throat-less driver response is closer to what you've set out to achieve.

[1audiohack]

I am hoping that is the case. Looking at the beast I could easily machine the flange down 0.500" or more to get the driver closer to the mic, or worst case weld up the hole and remachine it closer.

I suppose we'll know soon enough.

Thanks Grumpy.

[1audiohack]

Finally I think it's pretty close comparing these to the 1.5" PWT plots found here in the transducer library. Does anyone see anything amiss besides the SPL level is low?

Thank you,
Barry.

[ivica]

Finally I think it's pretty close comparing these to the 1.5" PWT plots found here in the transducer library. Does anyone see anything amiss besides the SPL level is low?

Thank you,
Barry.

I always wonder are those measurements (that producer has presented) are done under:
(i) "CONSTANT VOLTAGE AMPLITUDE" for example U=2.83Veff (or 4Veff)
or
(ii)"CONSTANT INPUT POWER" , what is much different then above, because
P=(U*U)/Z (= 1W), and if Z is (as usual) very much varying with frequency, applied voltage "U" has to be changed in order to satisfy P = 1W, depending of
the applied frequency.

[1audiohack]

Constant voltage I'm quite certain, here's what I have taken from the JBL test sheets from the transducer library:

2452, frequency curve driven at 400mV for H and 566mV for J, the harmonics test was at 5.5V.

2450H, 4V.

2435H, frequency response at 1V, harmonic test at 7.8V

They are the only ones I have written down.

[ivica]

Constant voltage I'm quite certain, here's what I have taken from the JBL test sheets from the transducer library:

2452, frequency curve driven at 400mV for H and 566mV for J, the harmonics test was at 5.5V.

2450H, 4V.

2435H, frequency response at 1V, harmonic test at 7.8V

They are the only ones I have written down.

I am amazed that the data for 2452H, it is only 20mW power, while for 2450H it is 2W.
May be some other things make a differences between your tube and original JBL.

[grumpy]

averaged 4-turns (90º rotation)?
mic depth?
same resolution as JBL tds file?

Above 10KHz I expect reproduce-ability for comparison sake (vs. JBL curves) to be
tough, if not impossible.

There might also be some presumed corrections introduced to make the curves represent
both a standard drive level and tube diameter... I'd guess those might account for
part of your mid-band amplitude differences. Calibrated mic and pre-amp gain?

At any rate, unless the drivers are severely broken, you could probably add in
your own fudge factor to match the JBL mid-band curve and then compare.