March 2 - TimG

Finally warm enough for glue to dry so fabrication can begin on the test enclosures. The first 2 cabinets are ready to go. These drivers require 3/4" recess for flush mounting so I had to laminate up 1.5" baffles. I will need to add weatherstrip to flush mount the 2234H's in these cabinets because the older 15" drivers only require 5/8" recess for flush mounting. First set of measurements will be done with 1500AL, and 2426H on a 16" x 6" biradial horn.

March 3 - TimG

Budget horn proposal:

I'm thinking that some people may want to build a DIY project similar to the Project May, but will not be able to afford the original drivers or custom CNC machined wood horns (me for instance). As I build cabinets and measure the flagship drivers I will also be measuring other alternative drivers in the cabinets so that I will be able to propose an alternative design that will be more affordable to build.

My plan is to use drivers and horns that are similar enough that someone could start with affordable vintage drivers and then eventually upgrade to the full S9800 driver complement, or at least the less expensive diaphragm versions (2430, 045ti). I don't have access to a 1500FE, but that could be considered for an alternative woofer as well.

I will be measuring a 2226H and 2234H as alternative woofers. For midranges I will be measuring the 2426H on 2 horns that I have on hand. I will also be measuring a BMS 4590P, JBL 2450, and 2446, that a forum member will be sending me, on a 2380A horn.

I found another 1" throat horn and supertweeter that I would like to try, but I can't figure out how to get the faceplate off the supertweeter that I have. Here is the plan:

I want to get this dirt cheap piezo driver tear the horn off and then mount the driver from this combo behind the horn. I already have one of the FT17H drivers.

Here is a mid horn that appears very similar to the S9800 horn and we could add a JBL 1.5" to 1" adapter to allow use of the 2430/435Be as well.

March 4 - TimG

Beginning testing with cabinets below. Started off with 2426H, 2226H, and Sub1500. For stereo testing I have the following amplification available Adcom 555 (300 watts/4 ohms x 2) and 2535 (60 watts/8 ohms x 4). For mono testing both amps can be bridged for 800 watts/4 ohms x 1, 200 watts/8 ohms x 1, 60 watts/8 ohms x 2. Will hook up 1500AL next, now that I know everything is wired properly. The Sub1500 is shaking the concrete with only 300 watts, I'm eager to hear what 2 could do if I had a second 800 watt amp to drive it.

March 9 - TimG

March 18 - TimG

March 19 - TimG

March 21 - TimG

March 22 - Mr. Widget

The initial horn(s) will be something like the illustration below. They will require an 8" tall by 19(?)" wide space to sit. I would expect the final horns to have a lip of some sort to help incorporate them into the design. One possible scenario would be to have the UHF horn incorporated into that lip.

No it is not too late to consider the option of having a recessed UHF horn if the design ends up being TMM.

March 29 - TimG

Everything was looking so good on Thursday, I was thinking I would have measurements posted that night. If only it were so easy. The 4 cubic foot boxes were basically complete, except for front baffle mounting and binding post install, after just 2 Sundays with a helper. I spent the afternoon stopping at 4 stores (Home Depot, Lowes, Radio Shack, and Performance Audio) just to get the supplies I needed to wire up the 4 new boxes, secure the fronts, install t-nuts and remount the front baffles. Now that I have the drivers mounted I can barely lift these things. They have to be 100lb each at this point. On a positive note, they are really solid and don't have a lot of obvious resonance.

Finally I have some measurements we can look at. I'm continually amazed at how powerful SE is. My main computer, with the video capture program, fancy sound card, and SE is downstairs hooked up to the test rig, so I can't post the pictures until I bring it back upstairs. I need to get some more measurement done while everything is in place so I'm not willing to move it yet. What I will accomplish today is getting plots for the MTM off-axis. If I can I will also get plots for the TMM option, but all of this stuff is taking 4 times longer than I previously estimated because I'm doing it all by myself. Surprising, I got my wife to help me haul (4) 100lb cabinets across the yard, onto the porch, and down the stairs without screaming. Unfortunately I had to do 6 hours of yard work in return.

Things will start happening a lot more quickly now that everything is in place. I say we worry about box tuning after we decide on the design format. I will post the plots as soon as I can. Plan on seeing some tonight.

Here are some things I learned:

1) It would be nice to have an indexed volume control on my test receiver so that I could replicate volume levels on different days.
2) Don't try to use a fancy surround sound preamp for testing, it's to much of a PTA to turn off all the features you don't need.
3) Use lots of masking tape to mark the position of your microphone stand on the carpet so you can remember where it was before you had to move it to do the close-miked woofer response.
4) Heavy things move a lot faster when you have 2 people to do it. (I had to take Soma last night, and luckily my back felt undamaged this morning.)
5) Always buy more wire than the amount you were certain would be more than enough.
6) Next time invest in a computer power supply with an ultraquiet fan if you plan on using it for testing, the mic picks up everything
7) 1 watt in a large format compression driver is lots.

March 29 - TimG

OK, I have the MTM plots. I'm not going to do TMM plots tonight because once I take the MTM apart I'm not lifting that top box 50" off the ground again without help or a hoist. I wish they would have used Neodymium on the motors for these beasts I have on axis, 15 and 30 vertical. I will do the same for horizontal, but MTMs don't suffer comb filtering in the horizontal plane. These plots were taken to confirm the model so I don't think there is a need at this time to do plots every 5 degrees.

There are limitations to these plots because I don't currently have access to an auditorium. I'm limited to far field measurements above about 350Hz. These plots do not include near field response because I'm not sure what 15 degrees off axis looks like when the mic is almost touching the dust cap. Basically adding the near field would raise the level of each of the plots by the same amount.

The procedure for getting the complete far field picture is as follows. Measure far field, move mic adjacent to dust cap, lower output level and measure near field, add baffle gain to nearfield plot, adjust level of near field to far field at 500Hz, merge near field to far field at 500Hz.

To obtain accurate far field response to a lower frequency I would need to move the speakers and measurement system to a larger room. Since my garage is only a 1 car, that won't really help. If I had a crane in my back yard, and it was totally silent outside, I could measure the speakers in my back yard, but neither of those conditions can be met. If I am able to reserve the University Social Science Auditorium in May, for the DIY event, I will be able to take accurate measurements to a lower level, but I'm hoping to have estimates for the crossovers before then, so it would just be for confiratory purposes.

The procedures I am using for measurement were obtained from Joe D'Apollitos book, the SoundEasy manual, the LspCAD manual and the SoundEasy user guide, which is an invaluable time saving resource written by John K.

The image of the setup in my next post shows the woofer cabinets, 4 cu ft internal, with 19" x 21" baffles. The drivers are 1500AL with a reversed cabinet serving as a stand at the bottom. The cabinets are separated by a 8.25" high stack of MDF round cutouts and the horn is a DDS 1" throat 70 by 40 degree horn that I'm borrowing to test a new compression driver. It looks like it would make a great horn for a 12" or 10" midrange crossed over at 1000-1300Hz. I have not yet had a chance to complete the horn baffle/spacer that will house the H9800 horn.

March 31 - TimG

What are those curves telling us??

The curves show some of the comb filtering effects that occur with an MTM with large woofers and a wide center-to-center distance. I previously modeled the comb filtering effects in another program and posted the results. I will do that modeling again now that I have more information about baffle size and correct piston diameter so that we can compare the predicted to the measured results. Unfortunately, I can model more of the comb filtering effect than I can measure. However, the software predictions match reality very closely and are based on vector modeling. Therefore, I feel we can trust the predicted results because they will correspond to the actual results. That is not to say that the final crossovers will not be tweaked by ear.

I have measured TS parameters of the woofers, and I will conduct on an off-axis measures of the 1500ALs in both a MTM and TMM format so that we can decide on the advantages and disadvantages of each design before moving on with the design process.

I have another microphone that I can use, but I don't have a calibration file for it. My Panasonic capsule mic is the same device used in the Mitey Mike. The reference speaker is a good idea. I will shoot some measurements of my reference speaker today. My current reference is a 4-way based on the NHT 3.3. The 3.3 is no longer in production but it is a 4-way based on a 12", 6.5", 4", and 1" dome tweeter. Kind of like a sealed L250ti. Here is a link to a review of its little brother the 2.9.

http://www.stereophile.com/loudspeakerreviews/320/

A friend is also planning to bring over his Focal W-cone version of the Joe D' designed Aria 7 MTM tonight. Here is a link to what they look like. We can measure these too.

http://www.zalytron.com/pictures/Focal6WTC120.jpg

An RTA measurement in this room would be totally colored by room reflections. Gated response measurements are used so room reflections do not contaminate the measurements. In a sense, gated measurements, merged with close mic woofer measurements simulate measurements under anechoic conditions. I can take an ungated measurement of the speaker, but I'm not sure what value it would have. We certainly don't want to design the crossover based on measurements that include room reflections from my small room. However, this issue merits more discussion. Here are just some of the issues involved that we should discuss.

1) You can design the speaker to measure flat at one point in the room, on the listening axis, or use EQ to achieve the same result. However it may sound terrible at every other point in the room, and will not be accurate in any other room.

2) On the other hand, you can design the speaker to have flat frequency and power response under simulated anechoic conditions and then treat your listening room to achieve good response at the listening position, with small tweaks to the crossover as necessary, (i.e., for high frequency rolloff contouring or a BBC dip). This type of speaker has the potential to sound excellent in all rooms, with some work.

I would lean towards the second design philosophy, with the provision for EQ in the bass frequencies (say below 100Hz) since this is the area of response most affected by the room. This is the same design philosophy used in the JBL LSR6300 studio monitor series, and the Infinity Prelude speaker before that. I personally feel that this is the best way to produce a SOTA speaker, and it is also the future of accurate audio reproduction. However, this philosophy would likely involve having to mount the subwoofer separate from the main speakers to produce the best wide range response, not always the most decor or space friendly solution.

Please refer to the white papers located here, particularly the "RABOS" and "Loudspeakers and Rooms" technical papers:

http://www.infinitysystems.com/homeaudio/technology/technology_whitepapers.aspx

and here

http://www.jblpro.com/LSR/PDF/White%20Papers.pdf