Anyone buy the TB W4-656SB driver?

[MarkMcK]

If you bought the Tang Band W4-656SB driver, you did not make a mistake. PPM cone, die cast basket, and shielded magnet structure are all positives in the design. While it does not have the high frequency extension to qualify as a full-range driver and it does need bass augmentation below 95 Hz, it has the potential to be a high quality and high fidelity extended range driver.

In stock condition, it does have some problems. You can, however, improve its performance with a series of modifications simpler than my modifications for the W4-654S paper cone driver. Here is one way of improving the driver.

First, remove the dust cap. The dust cap is a major problem for this driver and I do not understand why TB uses dust caps. Even with their low detail measurement set-up, they have to be able to measure the huge 7 kHz peak in the response.

Next, replace the dust cap with a “phase� plug. These plugs are neither wave-guides nor phase correctors, but they are useful in regulating cone loading and adjusting response. I find the TB phase plug shape to be less than optimal, but with four-inch drivers (and larger) the response is less sensitive to changes in plug shape. I have included a photo of the wooden plug I used for this modification sequence. The plug is 1.5 inches long. The first .75 inches is just a cylinder, the second .75 is tapered. I have made plugs out of aluminum, wood, plastic (both machined and cast), plaster, and hard rubber. I have found that the shape is important; the material of construction is unimportant.

Next, lay down a thin bead of glue (GemTac works well) that is 1 to 1.5 mm in diameter, 6 mm in from the inside diameter of the surround. This stops the origin of a 7 kHz cone vibration mode. Unlike the W4-1052SA modification One, the glue circle does not cause a dip in the response in the 656SB driver. A second cone resonance remains at about 9kHz. I have yet to locate the attractor region of the cone that is the origin of this vibration mode.

Lastly, construct the three-component equalizer circuit. The component values are small and can be obtained inexpensively even when selecting high quality components.

There is a multitude of midrange drivers available. If, however, you wish to design a 2.5 or 2.1 way loudspeaker while pushing the crossover regions as far away from the midrange as possible, this is a desirable driver option.

Good designing,

Mark

[vinnie]

Mighty impressive R&D work on the TB driver Mark.
How is it to put a tweeter cone, like the ones on fostex drivers, on it? will it have any effect in the higher freq. range?

How do you find the "attractor region" for the given freq. is it just try and fail?

How did you calculate the notch filter?

[andrikos]

Fantastic work Mark.
Thanks for making this available to us.
I can see these puppies going in a cheap line array system (full range) maybe a bit equalized.

[rcavictim]

Outstanding Mark! I bought four of these when P.E. started carrying TangBand a couple of years ago. These do not seem to be available anymore. I have been listening to a pair driven fullrange for a couple of years in a rear horn loading cabinet and a Dayton PT-2 tweeter helping fill in the upper registers (2nd order xover @3 KHz). You are right, This TB does not have the top end as supplied and also needs augmentation in the bottom. I ended up using a 6.5" and 8" Dayton woofer run face to face isobaric in the hole in the rear horn originally contemplated for a second TB W4-656. They are run flat out in parallel with the TB no crossover. The net result is very good. I want to try your mod.

I originally wanted to try to use this folded horn cabinet with the W4-656 driver alone as a sort of poor man's Lowther or Fostex. Maybe with your mod I still can.

Can Weldbond (the white universal glue that looks like white wood glue wet and dries clear) be used to damp the cone?

Do you plan to experiment with adding a wizzer to increase the top end capability?

[planet10]

Mark,

Do these plugs look like they would work....they are shorter (developed for FE103s)



dave

[rcavictim]

Dave,

You are probably in violation of the colosally stupid and evil Federal firearms registry. Nice bullets!

Can anyone recommend the safe procedure for removing the dustcap?

[planet10]

Originally posted by rcavictim
You are probably in violation of the colosally stupid and evil Federal firearms registry. Nice bullets!

Thanx -- watch for the advert in the Vendor's Bazaar...

Can anyone recommend the safe procedure for removing the dustcap?

I have a set of pictures taken for just such a page... they will go up on this page which has a preliminary description

http://www.planet10-hifi.com/pp-install.html

dave

[MarkMcK]

Lots of questions.... I will try to address them all in this one post.

First, the glue. I am not familiar with WeldBond glue. The species of glue is called "aliphatic." If it is water soluble, dries clear, and remains flexible it is in the same glue species. I can even give you a detailed chemical recipe for the glue, but that won't help.

The glues vary by water content and adhesion. Some hold to the poly much better than others. The GemTac holds the best of any of the commerical glues I have tried. The WeldBond, if it remains flexible, will work. How easily it can be pulled off the poly I cannot say.

I don't understand the part of your message about not available anymore? I just checked and the W4-656 is listed on the Parts Express Website and is stated as in stock. I bought mine a little more than a month ago during one of their regular sales.

In other postings I have declared my allegiance to transient coherence. Whizzers are antithetical to transient coherence and resonance control. In my world, the frequency extension provided by whizzers is an illusion and they are all terribly destructive to sound reproduction accuracy. Just as I have taken off the dust cap, if this driver had a whizzer, I would take it off too.

An analogy to a whizzer would be adding a bell to the system. Attach a tiny little clapper to the voice coil and when the voice coil moves the clapper strikes the bell. The system now has more high frequency content. The new high frequency content is created in response to the signal applied to voice coil but it is not a reproduction of the applied signal. It is new and added to the sound.

Second, I would say inexpensive and not cheap. There is nothing cheap about the performance after modification.

Third, Dave, there are so many topics this will apply to, but length and diameter are both important. You can shim your plugs to the 1.5 in spec, but I believe the Fostex uses a 20 mm voice coil and the Tang Band 656SB uses a 25 or 26 mm coil former. It will be impossible to shim the diameter. Your profile seems to be closer to the TB shape than mine. I like a slightly flatter tip with a larger radius near the terminal end. The possibility for innuendo is endless.

Vinnie, thanks for the last two questions. These are great questions for a diy forum. It is classical R&D: empirical research. If my modeling is correct and cone breakup is chaotic, then it will defy simple predictive formulas. Still, there are regions of order within apparent chaos. For example, in four-inch drivers res problems below 10 kHz are always outside edge problems. Above 10 kHz and the triggers are always between half way and inner third. The same corrective techniques do not work for every cone and every material. Aliphatic glue rings won't do a thing on paper (or banana paper). Here it is a matter of matching material mechanical impedances.

I have been working with material resonance problems for so long that I use what academics call an expert system. I have learned about 100 questions to ask when presented with a material resonance problem. I really don't have to think about it very much. I view the data and I make predictions. I found the 7 kHz cone trigger on the second try. My first attempt was at 4mm in and it turned out to be at 6 mm in. The changes in the response to the 4mm ring told me where I had to go. If you go to an analytic chemist to identify a compound, talk to a physicist about the welding properties of an elemental metal, or a medical doctor about diagnosing a disease, the process is the same. You may have to do this a thousand times before you will be as good at it as I am, but it is within anyone's capabilities. You just need to learn the appropriate 100 questions of material science so well that you don't have to think about them anymore. Not only do I know this academically, but also I can testify that this is the way physicists work. I was hanging around a materials joining think tank and sitting in a physicist's office one day when a call came in asking about the oxidation byproducts of welding a particular metal. He just looked at the periodic table on his wall and said, " your metal is in the same column as titanium and so you can expect a similar oxidation spectrum to welding titanium."

The same applies to designing res networks. I have a multi page flip chart of standard component value resonant frequencies. I know the Q limits of the charts and I just flip through them until I find the values I want to use for the shape and frequency of problem I plan to correct. I have the tables in html format and could post them to my Web site if anyone is interested or would find them useful. You could print them and make your own design flip chart. Just let me know.

Lastly, maybe one of you can help me. I have identified a common characteristic of all four-inch cone drivers. Regardless of Fs or enclosure, the output begins to fall at 95 Hz and is minus 3 db by around 72 Hz. I have tried infinite baffle, ported, and sealed enclosures of different sizes and tunings and nothing will change this roll-off. I have not reached expert level yet with transmission line or horn loaded enclosure designs. If this is a physical limitation caused by the falling mechanical impedance of the cone to air coupling, then I do not know if transmission line or horn loading will solve the problem, or if you can fluid couple multiple drivers to make the combined cones act as if they were larger. Would anyone be willing to help me complete this aspect of the R&D? This is relevant to one of you because if my hypothesis is correct, then equalization will not work.

Mark

[MarkMcK]

RCAVICTIM,

Was your question for me and the TB or Dave and the Fostex?

If it was for the TB, don't use a heat gun. The adhesion to the poly is not that great. There are two easy ways to get off the dust cap. The first way is to take the exacto knife and going under the outside edge begin to fold back the dust cap edge. By the time you are half way around the edge you can just pop the cap off.

The second way is to cut the center of the dust cap out, then run the tip under the remaining lip to cut the remaining ring. Then just grab one end and peel off the dustcap ring.

I usually use the first technique only because I like to remove the dust caps undamaged. The second way is easier and pretty much fool proof.

Mark

[mjq]

Thanks for all your work Mark.
I have a pair of 656s in modified Fostex Back horns which work OK.
I haven't measured their response to know if the bass is much better than you have seen in a ported box, but I think they are going deeper than 95hz...

I'll try these mods. and see if I can't them to sound very good...

Jason.