The BaSSlines (was High Sensitivity Design)

[Dennis H]

I don't mess with image capture programs. Just size the window how big you want it and hit Alt-PrntScrn. That copies the active window to the clipboard. Then open a paint program, Ctrl-V to paste the image in, and save it as a GIF file. I use Paintshop Pro but MS Paint works fine as well.

[augerpro]

Dan in Photo-bucket's loader window there is an option to resize uploads, I set mine to "800x600 15in screen" and it will downsize everything you upload to that.

Your new plot shows the 1/2" waveguide doing pretty well. Surprisingly even amount of gain. And that was using a 1/2" roundover correct?

[Saurav]

You haven't actually built and listened to that crossover yet, have you? I'm curious about how well the Audax handles a 2nd order 450Hz high-pass on an open baffle. I suspect the B&C will handle it a lot better.

[dlneubec]

Dan in Photo-bucket's loader window there is an option to resize uploads, I set mine to "800x600 15in screen" and it will downsize everything you upload to that.

Your new plot shows the 1/2" waveguide doing pretty well. Surprisingly even amount of gain. And that was using a 1/2" roundover correct?

Correct.

[dlneubec]

You haven't actually built and listened to that crossover yet, have you? I'm curious about how well the Audax handles a 2nd order 450Hz high-pass on an open baffle. I suspect the B&C will handle it a lot better.

No and I likely won't since I don't have enough spare parts to piece all the needed values together. I should have my B&C's in any day now and will move to them. According to Jeff, who modeled the Audax, it will not handle a 450hz crossover in an open baffle without exceeding xmax.

[Saurav]

According to Jeff, who modeled the Audax, it will not handle a 450hz crossover in an open baffle without exceeding xmax.

Thanks, that's good to know. I'm using a 500Hz LR4 crossover, and if I did my math right, I should be within Xmax for the 4W output of my tube amp. Your target SPL levels are quite a bit higher than mine. But if I ever re-do my speakers, the first thing to change would be the midrange, since I had to fit everything else around what the Audax could (or rather, could not) do.

[dlneubec]

For those interested, here are some new and more accurate measurements of the Peerless HDS tweeter in a 1/2" waveguide mount (1/2" deep. 1/2" roundover) and a 3/4" waveguide (3/4" deep, 3/4" roundover) both show spl measurements on axis, 15, 30, and 45 degrees off axis (horizontally). These were taken at 1m, with a 5ms gated window, tweeter at about 80" above the floor (vaulted ceilings).

First 1/2" deep waveguide:



Next 3/4" deep waveguide:

[dlneubec]

This is predominantly for Augerpro, since he seems the most interested. I want to test further versions of simple waveguide mounts. My plan is to test a 5/8" deep and 1/2" deep waveguides, but with a 3/4" roundover. I will do this by starting with a 3/4" roundover on a 1-1/4" dia hole in 3/4" stock and then route out the back to 5/8" depth, test, and then route out another 1/8" to a 1/2" depth and test that. This will change the throat angle from 90º to the tweeter face at a full 3/4" depth to about 100º at 5/8" and about 110º at 1/2".

Hopefully I will find a combination that offers the low end boost I want to flatten the FR of the HDS, with some directivity control, but with a little smoother FR, especially off axis. Below is a drawing of what I will be testing.

So far, my preference is for the 1/2" deep WG with a 1/2" roundover, as posted above. We'll see if these other two offer any improvements.

[Saurav]

As you're changing the throat angle, you're also changing the throat diameter, right? I'd think you'd want to test for that too. It seems to me that you probably want the throat diameter to be as small as possible (i.e. as close to the dome as possible) while changing the angle, which I guess means varying the diameter of the hole that you round over?

[---k---]

Dan,

Is it just me or is the off-axis results worse with every waveguide attempt?

With a lot of the commercial designs that use small waveguides, it seems like the waveguides are much shallower. Looking at the Seas DXT, the waveguide actually goes the other way:
http://www.madisound.com/catalog/pro...oducts_id=8322 Have you given any thought to experimenting with a something like an Oge bit or a raised panel bit that can produce big, wide, shallow waveguides?


Also, did you do any testing on the mid to see what the rear mounting and round-over did to the mid response?

[dlneubec]

As you're changing the throat angle, you're also changing the throat diameter, right? I'd think you'd want to test for that too. It seems to me that you probably want the throat diameter to be as small as possible (i.e. as close to the dome as possible) while changing the angle, which I guess means varying the diameter of the hole that you round over?

Based on the 1" dome and it's surround, I figure 1-1/4" is the absolute minimum and that could be a bit too small. Zaph suggested 1-5/16" in his waveguide mount for a 1" dome. The drawing shows the actual dimensions at the different depths. Yes they do change, but not by a lot. For example, the 5/8" deep is 1-9/32", only 1/32" larger than the 1-1/4" minimum. The size at 1/2" depth is 1-11/32", only 1/16" larger in diameter and just 1/32" over Zaphs recommended size. Yes I'm changing two parameters at once. If these show significant enough differences, I may go back and do another comparison with the same throat dia. but I'm not convinced yet that the differences will be that significant. These are pretty small tolerances for mdf. Just sanding them or apply paint might make 1/32" difference.

[Saurav]

I thought Geddes had said something about the optimum WG for a dome tweeter being a conical profile that pretty much starts at the dome, and is tangential to it. But I didn't follow that thread in much detail, so take this with a grain of salt.

These are pretty small tolerances for mdf. Just sanding them or apply paint might make 1/32" difference.

That's a good point.

[dlneubec]

Dan,

Is it just me or is the off-axis results worse with every waveguide attempt?

With a lot of the commercial designs that use small waveguides, it seems like the waveguides are much shallower. Looking at the Seas DXT, the waveguide actually goes the other way:
http://www.madisound.com/catalog/pro...oducts_id=8322 Have you given any thought to experimenting with a something like an Oge bit or a raised panel bit that can produce big, wide, shallow waveguides?


Also, did you do any testing on the mid to see what the rear mounting and round-over did to the mid response?

I agree that shallow and flatter may be better, however, the 1/2"/1/2" combo already does a pretty good job of what I want it to do. I hope to improve on it, but with simple construction techniques and tools that most DIYers might have, like a 3/4" roundover bit.

I agree that it would be worth testing with bits that could do wider, shallower cuts to see if they could produce better results, but not a lot of people will have those. I know I don't. I'm actally trying to get a wider, shallower cut by using the 3/4" roundover, but in a 1/2" depth. If anyone wants to send me a bit to try that I can use on my router table, (1/2" bit size) I'd be more than happy to try it out. I don't have a lot of diameter available through the table (aluminum), so it would be limited by the diameter of the existing hole. I'll check to see what it is, but my guess is it is not much larger than a 2" diameter.

As far as I know, I'm the only one who has implemented this kind of mount (in my Duo's). It was Zaph's idea, but I don't believe he has implemented it in anything yet, so there is not a lot to go on other than my simple experimentation. :E

[augerpro]

Good stuff Dan, keep it coming :T I think the 3/4" roundover was promising. A little more care to eq flat maybe, but the response was much smoother around the typical crossover point. The response for the 1/2" in this area was pretty ragged, at least the on axis response is. I think that would be harder to deal with.

I'd still like to see some 45 degree chamfers. I can send you a couple if you don't have them.

BTW can you do the new measurements under the same conditions so all these plots are directly comparable? And maybe throw in one plot of the tweeter flush mounted so we can see what the "control" is?

Thanks Dan.

[dlneubec]

I can do that. I do have some 45º chamfers. I'll see what I can come up with. Are you thinking 45º from the driver opening? what about the edge at the face of the baffle? How about a 3/4" roundover first and then a 45º chamfer to some depth? Perhaps you can describe what you are thinking.

I think I posted my prelim crossover FR previously (using the Audax mid). I used the 1/2" waveguide measuremet for this and did not find it to a difficult one. I believe the tweeter crossover in the one I posted was only 3parts, 2nd order electrical and padding tweeter, IIRC.

[dlneubec]

Dan,

Also, did you do any testing on the mid to see what the rear mounting and round-over did to the mid response?

Ryan, I missed this the first time around. My recent measurements with a flush mount versus a rear mount with 1/2" roundover were pretty much identical in the passband. I did not look at them in detail in the stop band area, so there could have been some differences in them there. I saw enough to conclude that the rear mount did not seem to indicate any obvious issues in the Frequency response. I did not look at distortion or CSD, however.

[---k---]

I saw enough to conclude that the rear mount did not seem to indicate any obvious issues in the Frequency response.

Cool.
PSB mounts their drivers similarly in their Synchrony speakers. They claimed in Stereophile that it improved the frequency response. I asked about that and it was called bunk because the frequency range of the driver was low enough that edge diffraction and such weren't in play. But no one had any testing to prove it.

So thanks.

[augerpro]

Dan I'm thinking just a simple 45 degree chamfer the full depth of the wood. So the tweeter is sitting in a cone profile. I think the profile edge should essentially intersect the base of the dome. As far as the edge at the baffle surface, I don't what to expect-that's one reason I'd like to see it. In my gut I think that edge will be less of a problem than placing the dome inside of a tube, which is basically how the round over is for the first 1/4" or so. I guess I'm just more partial to a cone profile, ideally intersecting the dome base at 90 degrees, forgot what the geometry term is for this. Doing a round over first would smooth this edge, but just to compare ideal shapes first I'd like to see the simple chamfer first.

[cjd]

I'd be curious to see if lining the waveguide with wool felt makes any difference in the dip in the top end of response (which is, I think, throat reflections).

Though I'm quietly reading from the sidelines, I'm actually very interested in the results. So thanks for posting.

[Saurav]

I guess I'm just more partial to a cone profile, ideally intersecting the dome base at 90 degrees, forgot what the geometry term is for this.

I think the word you're looking for is tangent. Edit: No it's not, it's the opposite, it's the normal. I think.

Doing a round over first would smooth this edge, but just to compare ideal shapes first I'd like to see the simple chamfer first.

You have the WG geometry, and you have the smoothness of the mouth termination, and with a very shallow WG like this, I'm not sure where one ends and the other begins. IOW, you may gain something by going to a cone profile that has the correct angle w.r.t. the dome curvature (which is what I was trying to describe in my previous post, poorly), but lose something because of the sharper mouth termination, and not know which change caused what effect.

I think Dan should test all possible combinations of all the suggestions he's received so far

[Dennis H]

These are the Peerless HDS right? Seems to me the most important thing is to understand that the tweeter's faceplate forms the beginning of the 'throat' and the best you can do is match your wood to the recess in the faceplate. That's probably causing the HF null with the 3/4" bit. Looks like a chamfer bit at some angle (45 or maybe bigger?) will match up to the 'throat' best and you could sand (or route) a roundover at the front of the baffle so there's no sharp edge. You could use some non-hardening putty to smooth the transition from metal to wood. Maybe kids' modeling clay like Geddes uses or something like Mortite caulk?

[dlneubec]

Guys,

One thing I've been doing which is apparently an issue is using two different tweeters. It looks like one has a dip in the high end response naturally and the other doesn't, not to the same degree anyway. This accounts for some of the differences we have been seeing. I need to start the tests all over again using the same tweeter. That will make comparisons make much more sense.

I have things set up to make a round baffle insert for each mount, so I would be using the same baffle for each test (with the mid installed). I need to come up with a logical way to do the testing. Here are the ones I'm thinking of doing at this point:

1) 1/2" deep, 1/2" roundover
2) 1/2" deep, 3/4" roundover
3) 5/8" deep, 3/4" roundover
4) 3/4" deep, 3/4" roundover
5) 3/4" deep, 45º chamfer
6) 3/4" deep, 3/4" roundover, then chamfered 45º to 1/2" deep
7) ?

Do on and off axis on each. Would 0, 22.5 and 45º off axis be enough to compare? That would save a little more time than every 15º. My goal would be to weed these down into one or two best looking options and then test those two with at least one other starting throat size, say 1-3/8". I would also try felt applications at this time, once the best geometry is discerned.

Any thoughts or suggestions?

BTW, the largest bit that will comfortably fit in my router table would be about 2-1/4" in diameter. The actual size is 2-7/16".

[dlneubec]

These are the Peerless HDS right? Seems to me the most important thing is to understand that the tweeter's faceplate forms the beginning of the 'throat' and the best you can do is match your wood to the recess in the faceplate. That's probably causing the HF null with the 3/4" bit. Looks like a chamfer bit at some angle (45 or maybe bigger?) will match up to the 'throat' best and you could sand (or route) a roundover at the front of the baffle so there's no sharp edge. You could use some non-hardening putty to smooth the transition from metal to wood. Maybe kids' modeling clay like Geddes uses or something like Mortite caulk?

Hi Dennis,

Generally, except for the first 3/4" WG with 1/2" roundover that I posted, all the tests have all been done with the faceplate removed. The driver is then press mounted from the back, so the existing faceplate has been of little consequence for most of the tests I've posted.

That would be the approach I'm planning to continue and the one I used for both tweeters used in my Duo project. It makes sense to me to remove the faceplate as a variable.

[Dennis H]

Hi Dennis,

Generally, except for the first 3/4" WG with 1/2" roundover that I posted, all the tests have all been done with the faceplate removed. The driver is then press mounted from the back, so the existing faceplate has been of little consequence for most of the tests I've posted.

That would be the approach I'm planning to continue and the one I used for both tweeters used in my Duo project. It makes sense to me to remove the faceplate as a variable.

Ah, gotcha, you're way ahead of me. :T

[augerpro]

Maybe throw in a 1/2" deep, 45 degree chamfer, just so all bases are covered? As far as the 3/4" roundover plus 45 degree chamfer, I think the chamfer will cut the entire roundover off, leaving you with the edge at the baffle surface still. You may have to hand sand that if you want it smooth.

I've recently ordered some HDS tweeters from Solen and both of them don't have much of the high frequency rise as the old ones. Which is cool now I don't have to include that notch filter I used before.

[---k---]

BTW, the largest bit that will comfortably fit in my router table would be about 2-1/4" in diameter. The actual size is 2-7/16".

Nah, you just need to make a new router base.

Image not available

Image not available

I can get a 1 1/4" roundover and a big giant Oge in that, no problem. If I need bigger, it can be modified. And it is more stable.

[dlneubec]

Well, I got 4 inserts essentially complete and two more half-way done last night and then broke my jasper jig in half. It has been cracked half-way through for a year or so and I finished it off. I will be able to finish and test the initial 4, since all the circle cutting is done on them. These include the 1/2"deep/3/4" roundover, the 3/4" deep/3/4" roundover, 3/4"deep/full depth 45º chamfer, and 3/4"deep/3/4" roundover, with a 45º chamfer about 1/2" deep. The latter two were hand sanded to provide some roundover at the termimus and throat.

Obviously, I need a replacement for the JJ before I can complete the other two. Does anyone have a recomendation between the Jasper jig, Router Buddy or the Sears circle cutting jig that Zaph has recommended?

[Bill Schneider]

I have both a Sears and a Jasper circle guide. To me, the Sears circle guide is cheaply made (I didn't buy it - it was given to me). If I needed infinite adjustability, I'd go with the Router Buddy instead. It's similar to the Sears, but a WHOLE lot better made. I've seen them (and the Jaspers) at a nearby Woodcraft store.

Personally, I still use the Jasper for my limited speaker making needs. It's very fast to use. With thoughtful use, you'll never have more than a 1/32nd gap between speaker and recess. If that's too much, add an extra layer of paint.

But each to their own...

[cjd]

I make my own of 1/4" HDF or plywood. Drill with a brad-point and you can get extremely accurate. I use a 1/4" pin.

C

[Dennis H]

I replied in your PE thread with a link to Pete Mazz's jig using a template collar. It's cool because you don't have to spin the router and get tangled up in the vacuum hose.

[dlneubec]

Hi guys,

Here are some results for the 4 WaveGuide inserts I tested today. These were all done at 1m, mic about 80" high, 4ms window. The plots are at 0, 15, 30, 45, and 60 degrees horizontal off axis. These use the same tweeter with the same mic, setup, etc. so they are directly comparable. These all were adjusted to have 1-1/4" wide throat diameters. You can see that they all look much more alike than different. I will do the others when I get my circle cutting ability back up.

FWIW, I'm leaning toward the 3/4" deep, 3/4" roundover. as my first choice, so far.

1/2" deep WG, 3/4" roundover



3/4" deep WG, 3/4" roundover



3/4" deep WG, 3/4" roundover, with 1/2" deep 45º chamfer after, sharp edges sanded round by hand



3/4" deep WG, 3/4" deep 45º chamfer, sharp edges sanded round by hand

[Dennis H]

Nice work, Dan! Thanks for sharing your experiments.

[dlneubec]

Ok, here is a little more complete comparison, with some additional results and a repost of some of the graphs from the last post:

HDS mounted Flush: 0-15-30-45-60º



1/2" deep WG, 1/2" roundover: 0-15-30-45-60º



1/2" deep WG, 3/4" roundover: 0-15-30-45-60º



3/4" deep WG, 3/4" roundover: 0-15-30-45-60º



3/4" deep WG, 3/4" roundover, with 1/2" deep 45º chamfer after, sharp edges sanded round by hand: 0-15-30-45-60º



3/4" deep WG, 3/4" deep 45º chamfer, sharp edges sanded round by hand: 0-15-30-45-60º



Here are the flush mount (red), 1/2" deep/1/2" roundover(pink), 1/2" deep/3/4" roundover (blue) and 3/4" deep/3/4" roundover (black), all on axis.



My conclusion is: it depends on what you are looking for. The deeper wavegides have more boost, but all peak at about 4kHz on axis. The 1/2" deep waveguide, with 1/2" roundover has the flattest FR on axis from 4kHz and up on axis. The 1/2" deep/3/4" roundover WG and the 3/4" deep/3/4" roundover WG have the best off axis response when you get out to 30º and beyond and the most directivity. I'm now leaning toward either the 1/2" deep/3/4" roundover of the 1/2" deep/1/2" roundover WG's from a raw response point of view. I will see how they look in the crossover design.

I hope these results will encourage some of you to experiment with this concept. I think it has a lot of merit and is not at all hard to do with just a router and standard bits. It helps if you have a drill press and a router table as well.

[dlneubec]

I've started separate thread on the waveguide testing so folks not following this project can pick up on the info:

https://www.htguide.com/forum/showthread.php?t=31478

[dlneubec]

Here are some preliminary crossover results:

These are measured results, taken at 6' with the speaker up on stands so the tweeter and mic is about 80" high. A 6ms gated window was used, so the woofer output looks low, which is due only to the window length. You should pretty much ignore the data below about 500-600hz. I'd have to combine a longer gated farfield measurement on the woofer, with a nearfield woofer and port measurment to give an accurate picture of the low end.

The actual crossover points for the woofer and mid are about 450hz, the mid to tweeter, about 2.8kHz. I used the 1/2" deep waveguide with 3/4" roundover for these measurments. The 1/2" deep/1/2" roundover would give a flatter response in the 4-10k area, but is not as good off axis in terms of directivity. I will measure with it as well, so I can make a comparison. I will also listen to both to compare them.

At this point the crossover indcludes a notch filter to suppress the 6md38's peak at 5kHz, but it is mildly implemented at this time. When it comes to listening and voicing, I will try it with more suppression of the peak and with the notch filter out completely, to see what my ears prefer. I like to start with a pretty flat FR and then adjust by ear. I did keep the tweeter a little hot to start out with so that at 15º off axis it is about flat (see graph 2). Impedance looks like a nominal 8ohm, with a low of about 6 ohms. I've chopped off the graphs in order to keep them as small as possible while still easily readable.

Next up is lots of listening, which includes trying at least two different slot port lengths to determine what sounds best at the low end as well.

SPL with individual drivers shown (woofer not accurate below 600hz or so)



SPL on axis (black) and 15º off axis (blue)



Impedance/Impedance phase

[ttan98]

Hi,

I thought you may like to know J. Salk is testing a new 6.5" mid range from AE on his new open baffle speakers. If it is good enough he will use it in lieu of the PHY.

[Landroval]

Hi,

I thought you may like to know J. Salk is testing a new 6.5" mid range from AE on his new open baffle speakers. If it is good enough he will use it in lieu of the PHY.

More info about the 6.5" can be found here:

http://www.aespeakers.com/news.php?start=0&news_id=16

http://www.aespeakers.com/phpbb2/viewtopic.php?f=1&t=1705

[ttan98]

More info about the 6.5" can be found here:

http://www.aespeakers.com/news.php?start=0&news_id=16

http://www.aespeakers.com/phpbb2/vie...php?f=1&t=1705

Not cheap, normal price $195 , introductory offer is $150. The B&C midrange is $92.

[dlneubec]

I've been talking with John about the Lambda 6.5". I may end up with one for testing. Actually, Jeff Bagby is the one who is looking at the Lambda for Jim Salk.

[ttan98]

I've been talking with John about the Lambda 6.5". I may end up with one for testing. Actually, Jeff Bagby is the one who is looking at the Lambda for Jim Salk.

Please report back on yr comparison between the B&C and AE midranges I am interested in buying one of them.

Thanks.

[JoshK]

Hi,

I thought you may like to know J. Salk is testing a new 6.5" mid range from AE on his new open baffle speakers. If it is good enough he will use it in lieu of the PHY.

It looks like a cool driver. BTW, I think you meant PHL.

[ttan98]

It looks like a cool driver. BTW, I think you meant PHL.

PHY is a very good driver at a reasonable price less than $100 each.
AE 6.5" midrange at $199 retail and $150 introductory price is by no means cheap. It has to to a clear winner(compared with PHY) for someone to pick it.

[John_E_Janowitz]

PHY is a very good driver at a reasonable price less than $100 each.
AE 6.5" midrange at $199 retail and $150 introductory price is by no means cheap. It has to to a clear winner(compared with PHY) for someone to pick it.

I'm sure which driver you're thinking of that could be less than $100. The driver Jim and Jeff Bagby are comparing to is the PHL 1120, not a PHY driver. PHY makes some ultra high cost widerange Alnico drivers. The PHL1120 sells for $189.50 on the e-speakers site. The TD6.5M we are doing now is less than $10 more. I am confident it will have lower distortion than the PHL and be much smoother on the top end as well.

Yes, it is not a cheap driver, but it is fully made here in the US. We buy US made steel and hand machine each T-yoke. The copper tubing for the shorting rings is made here in the US. Phase plugs are hand machined from solid aluminum bar made in the US. Cones, spiders, VC's, surrounds again are all made in the US. The only parts that come originally from outside the US are the frames and magnets, which we just have no option to get from here or we would. Cones, spiders, and surrounds are all hand treated and damped. Each driver is fully broken in and tested before shipping. We verify parameters, and if someone orders a pair of drivers we match DCR on all the coils in that batch for the absolute best consistency possible.


John

[A9X]

PHY is a very good driver at a reasonable price less than $100 each.
AE 6.5" midrange at $199 retail and $150 introductory price is by no means cheap. It has to to a clear winner(compared with PHY) for someone to pick it.

Audiomarketplace.com.au has them for $A260 + postage
PHL1120

However, I still think I'll buy some of the AE's to experiment with as I have a mate with the PHL and I'm very impressed with the other AE drivers I have.

[ttan98]

I'm sure which driver you're thinking of that could be less than $100. The driver Jim and Jeff Bagby are comparing to is the PHL 1120, not a PHY driver. PHY makes some ultra high cost widerange Alnico drivers. The PHL1120 sells for $189.50 on the e-speakers site. The TD6.5M we are doing now is less than $10 more. I am confident it will have lower distortion than the PHL and be much smoother on the top end as well.

Yes, it is not a cheap driver, but it is fully made here in the US. We buy US made steel and hand machine each T-yoke. The copper tubing for the shorting rings is made here in the US. Phase plugs are hand machined from solid aluminum bar made in the US. Cones, spiders, VC's, surrounds again are all made in the US. The only parts that come originally from outside the US are the frames and magnets, which we just have no option to get from here or we would. Cones, spiders, and surrounds are all hand treated and damped. Each driver is fully broken in and tested before shipping. We verify parameters, and if someone orders a pair of drivers we match DCR on all the coils in that batch for the absolute best consistency possible.

John

Try this site:

http://www.zalytron.com/

PHL 1120 at $120.