TangBand 75-1558SE on a Waveguide with Measurements

[dlneubec]

FWIW, based only on experience with my DIY waveguide tweeter mountings, if there was one thing I would try first, it would be to keep the depth the same and try a smaller roundover, say 3/4". As several of us have said, what you have is really an improvement, up to a point, and just maybe you can push the dip up a little higher with a smaller radius or alternatively a deeper WG with the same radius. In all cases, I would keep the throat as close to the dome as you can get it.

You really should expect to do a bit of experimentation to find the ideal solution, so you may want to make a baffle, if you have not already, that you can mount your waveguides into, so you just have to fashion each waveguide the same way and mount it into the baffle. I made a bunch of blanks up when I was doing my testing and then was able to quickly route new versions and swap them in and out for testing.

One thing I didn't do and was mentioned above, is take distortion measurements to see what effect the WG has on them. If you have to pad down the mid due to the boost, do you indeed lower distortion levels by some amount over the flush mount version?

[dlr]

One thing I didn't do and was mentioned above, is take distortion measurements to see what effect the WG has on them. If you have to pad down the mid due to the boost, do you indeed lower distortion levels by some amount over the flush mount version?

It will be interesting to see the results of a smaller radius. You'll be trading off low end gain with (possibly) some higher end smoothness.

I would say that it's almost guaranteed that any non-linear distortion is reduced due to the increased sensitivity (the gain) in the low end. However, there is no guarantee that the linear distortion will not change for the worse. That may actually increase in some ways if the response is not as smooth. This would have the caveat that the change is prior to implementation with a crossover.

I'm still in the camp that says that any definitive distortion results require a crossover be in place to know what the real conditions are, both for linear and non-linear distortion. What matters is what results are obtained in a final design.

dlr

[dlneubec]

I agree Dave, especially since the WG will probably require a different crossover implementation than the surface mount. It would certainly have to padded down more, if an equivalent sensitivity were desired and the change in driver AC should require changes to the mid low pass and tweeter high pass as compared to the flush mount. Of course, unless you take the time to develop a crossover and test for both the standard mount and WG mount, we will never know how implemented distortion levels compare.

That said, I think it would be valuable to see how raw distortion measurements change, good or bad. It would be great to see how CSD changes as well. I wish I had done this with my Peerless HDS WG implementation.

[augerpro]

I had asked someone who owned an ATC dome and he said it was maybe 5mm from dome tip to the baffle plane.

I will be trying a smaller roundover also. I just figured that the biggest, most gentle slope would be the least detrimental. But if I think of the dome and roundover as two radiating sources seperated by some distance, with a null at the frequency of twice that distance, than I can see how smaller is better.

I've got two blanks cut right now for my test baffle. My current one is still valuable in that I can change the depth both deeper and shallower and so I can get a handle on that variable. Depending the results I may have a better idea what to do with the new ones.

Dan good question on the CSD. I did not plot it, but when I ran the test teh IR was clearly worse for the WG baffle. There was an additional peak just after the first +/- peaks.

[benchtester]

I will be trying a smaller roundover also. I just figured that the biggest, most gentle slope would be the least detrimental. But if I think of the dome and roundover as two radiating sources seperated by some distance, with a null at the frequency of twice that distance, than I can see how smaller is better.

I think you may be correct. The dip looks like a "waveguide dip to me". I see them a lot at 8 to 10 khz on 1" tweeters. I describe in my blog (click on the 2 on the right) how I used clay to form a waveguide. (Future plans to machine the shape in hardwood.) I ended up with a small radius near the surround and increasing as it went out, looking like the end of a trumpet horn. If I added clay near the outside making it closer to a full radius, the dip deepened.

My $0.02

[benchtester]

than I can see how smaller is better.

Could be, in Zaph's alternate mounting of the D26NC55, he used a .5" radius and the dip was small and over 20 khz.

http://www.zaphaudio.com/tidbits/D26...horn-stock.gif

[Jonathan Anspach]

It's great to see someone doing these types of experiments. Please keep it coming.

Have you read the article here? It might help guide you in what to try next.
http://sound.westhost.com/articles/waveguides1.htm

[Pete Schumacher ®]

I had a thought, and it's probably just a shot in the dark, but the change in response looked familiar to me.

Check out the waveguide response of the TB 3" dome



And now look at the RS28.



Is that dip in the RS28 response waveguide related?

[augerpro]

I had figured it was interference due to not having a phase shield as most metal domes do, but it could be some kind of reflection off the faceplate yeah...what's half a wavelength of 15khz?

[ajinfla]

A bit under 1/2". Why?

[benchtester]

I had a thought, and it's probably just a shot in the dark, but the change in response looked familiar to me.

Check out the waveguide response of the TB 3" dome

And now look at the RS28.

Is that dip in the RS28 response waveguide related?

I am proposing that the TB 3" dome dip is waveguide related because the 4 kHz dip was not seen in it's stock FR.

I am assuming the RS28 is stock; and I am not very familar with the construction. I doubt there is a "waveguide" as such, but at such a high frequency it could be a similar reflection from the mounting plate. But I am more inclined to think that the dip shown here is a common effect of metal domes. Often you see a phase plate added to suppress this effect. (I don't think the RS28 has a phase shield, although sometimes they are clear plastic under the screen grille.)

I am a novice understanding the phase shield purpose and would welcome clarification.

As a wild conjecture, what if the waveguide brings out the phase issue with the 3" dome. It does have a pretty big "nipple". But the contrary argument would be that the graph shown is already includes Augerpro's phase shield.

edit: you guys are quick, two posts before I could compose mine (or I could be slow). 1/2" would correlate to a reflection off the mounting plate.

[joe7414]

looking at a cut away of a atc sm75 i estimate the wg is 25 mm deep and meats the dome at an angle of 45' ,there is a 8mm gap betwen the dome and wg
hope this helps v interested to see how it turns out .

[augerpro]

I should be getting more done on this soon, just been so busy with my PA and horn measurements I haven't had a chance to do the woodwork.

[Audio Quest]

Bravo for your effort and special thanks to John Krutke for all the testing and measurment data he supplied recently on the TB 75-1558SE 3" Dome Midrange. I can't say enough good things about John as he has helped so many DIY's by supplying information that has helped so many get past their loudspeaker design and building obstacles.

Horn loading of the TB 75-1558SE was something I had thought about as I would love to have a pair of the ATC SM-75 Dome Midranges. I'm not sure what your next step should be, but I am confident with this group of colleagues that resolving this horn loaded design is within reach.

Thanks again for serving on the front line and taking the bold steps forward. The work that you and John have done may make a horn loaded TB 75-1558SE come into fruition.

[augerpro]

So finally had some time to try out some new waveguides. Made an improvement, but still nothing usable. This is getting to be a tough nut to crack. Just to recap the first waveguide was basically a 1" roundover, but truncated as the baffle was only 3/4" thick, and it terminated quite close to the dome where it meets the surround.

First I tried to add a 3/8" spacer on WG1 just to see what increasing depth would do. Here is the original WG1 (red) versus the deeper WG1 (green):



Interesting but not an improvement. Next is WG2 which is a simple 3/4" roundover, again 3/4" deep (so a full roundover), but a larger throat terminating near the outside of the surround instead of near the inside. Red is a the stock driver and green is waveguide 2:



Finally getting somewhere! Now I thought I'd try a 45 degree chamfer instead of a roundover. I just reused this waveguide which means the chamfer is not full depth since I needed to leave a little edge for the bearing to ride on. Original roundover in red and chamfer in green:



Meh, not much change, maybe a slight improvement. One thing that bothered me about all these is that since the driver flange is not flat, there is an 1/8" gap between the baffle and the area by the surround underneath where the throat terminates. I geeked out in the first waveguide trying to leave that little 1/8" overhang to help this but its effect was swamped by the poor waveguide. For the hell of it just added some tape around the inside of the mouth so the waveguide would sort extend deeper almost touching the surround. Here is what that did:



So this means doing it right either involves making the little PITA overhang, or removing the driver flange so the baffle edge/ throat termination ends as close to the surround as possible. Which I'm a bit hesitant to do since it looks glued and screwed on, and may hold down the diaphragm.

So I have one blank baffle insert to try something new, what do you guys think? Moving the waveguide wall away from the diaphragm has seemed to help. Strange though...I didn't think it would actually raise the dip frequency as the distance from the wall to the dome got longer. I like the chamfer, as it's a small improvement and if would probably be a hair better if it was the full 3/4" deep instead being about 5/8" deep. Maybe an even wider throat? And an overhang to get the termination close to the surround?

[Mayhem13]

While it's not going to help driver to driver spacing at all, a large WG is needed to boost the low end response. To get rid of the dip at 4khz, try 1" of spacing bewteen the edge of the guide and the driver surround. I have a hunch there's a relationship between the wavelength of the dip and the circumfrence of the dome.

[dlneubec]

Brandon,

You are making good headway, it looks like to me. Depending on where you crossover, your last implementation could be very usable as is.

My first tip is that you might find that the dips you see on axis disappear off axis, so one thing you should try before anything else, as you do these, is to take a couple off axis shots on each and compare them to the flush mount. for positives, it looks like you have gained a very flat respone out to about 3.5khz and boosted the sensitivity about 2-3db, even some at the low end, which are all good things, IMO. I think you are getting closer.

That said, the problem is that you changed 3 parameters at once with the second implementation and you really should change only one if you want to figure out what is happening. You changed WG depth, throat dia. and roundover radius. You did get a better result, but it is hard to know which of those three were the main reason, which are positive changes and which are negative. You should stick to one throat dia and roundover, for example and try various depths. Then try a different roundover radius at the same depth. Then try a couple wider throat diameters with the other factors remianing the same. Hopefully you will see a trend from each and then have a better idea of which to combine to get the ideal response. It looks to me like you will have fill in that gap in the face plat or remove the faceplate. I wonder if 1/4"w x 1/8" thick weather stripping would work for your testing? I picked up some O-rings to try with my waveguide experiments, but decided it was easier to remove the face plate, though that may not be an option for you. I guess you could try sending an email to Tangband to see if they would tell you if the face plate is removable.

Much depends on what your goals are. My guess is that shallower will get you less boost but a flatter response, so perhaps a 1/2" deep WG with 1/2" roundover would be better, if you don't mind sacrificing the increased sensitivity. If you want to keep the sensitivity, then try the same depth and throat dia. with the larger radius to see how that compares with the chamfer and then keep those the same and try a wider throat dia.

BTW, one trick I employed to do roundovers or chanfers deeper than the baffle was to make a piece for the bearing to ride on and mount it behind the baffle. I simply used old driver cutouts and bored a hole through them of the proper diameter for the throat size I wanted and then stuck it to the baffle using double sided carpet tape. It worked like a charm for me.

There are enough variables that you have to have a lot of patience and a clear methodology to be able to figure out the ideal solution.

[dlneubec]

Going back to the original post, I see that the dip didn't really disappear off axis, so perhaps it won't in these new tests, but I would still do a couple off axis shots to be sure, especially on you last implementation.

[augerpro]

My first tip is that you might find that the dips you see on axis disappear off axis,

Measurements at 20 degrees still showed the dip

That said, the problem is that you changed 3 parameters at once with the second implementation and you really should change only one if you want to figure out what is happening.

I only have so much wood and time to route it man! Yes it's definitely not the scientific method, but I do think we can tease out some observations. The baffle depth was the same, 3/4". The roundover did change from 3/4" to partial 1", but I don't think that had much effect-just look at how little going to the chamfer changed things. I think the real improvement came by making this a larger throated waveguide.

I wonder if 1/4"w x 1/8" thick weather stripping would work for your testing?

I'll try that, good idea. It should work if a thin piece of tape worked.

Much depends on what your goals are. My guess is that shallower will get you less boost but a flatter response, so perhaps a 1/2" deep WG with 1/2" roundover would be better, if you don't mind sacrificing the increased sensitivity. If you want to keep the sensitivity, then try the same depth and throat dia. with the larger radius to see how that compares with the chamfer and then keep those the same and try a wider throat dia.

Right now the priorities are proper depth for a workable AC, smooth response in the passband, and gain, in that order. I may have to try a 1/2" depth to help smooth the response but I wonder how much it's going to hurt where I want the AC? Gain I'm only marginally concerned with compared to the first two. But yeah, I'll have to look into those options you mention.

Another thing Dan mentioned in another forum, which we are on the same wavelength for, is getting rid of the dip. I don't think I really need to get rid of it. I'd just like to push it up in frequency so the typical passband is smooth but with the dip placed over the peak of the natural driver response. An acoustical notch filter.

[Pete Schumacher ®]

I wonder if you could load it with an acoustic bandpass? What if there were a small damped chamber around the dome, with a small opening to the waveguide?

I'll shut up now . . . .