Could be better if I had an anechoic chamber, but then again, I don't know if they do any good since all speakers go in non-anechoic rooms, lol.
I'm hoping to eventually mate one 10" to a horn-loaded AMT. A 10" midbass/midrange will start beaming above 1kHz IIRC, so the tweeter has to be robust down-low, and even then the trade-off is that it will produce more distortion down low no matter what.
I added a B&C 10PLB76 and an 18 Sound 10NDA610 to this royal rumble.
Same test conditions as before.
B&C 10PLB76 Frequency Response
B&C 10PLB76 Distortion Performance
18 Sound 10NDA610 FR with and without AIC. The addition of AIC results in ~ +5dB hotter output from about 2kHz-upward, with the change starting around 1.5kHz.
Interestingly, the distortion with AIC decreases on the low-end but not on the high-end, but I guess it makes sense since it is not shorted, but rather in parallel to the voice coil.
Without AIC
with AIC
and for the kicker, here it is with the AIC coil shorted to itself
FR comparing without AIC, with AIC, and AIC shorted. A little bit hotter than without, but not as hot as with AIC on the top-end.
18 Sound NDA610 Distortion with AIC shorted
18 Sound NDA610 Distortion Percentage with AIC, without AIC, and AIC shorted.
18 Sound 10NDA610 Impedance/phase & T/S parameters via DATS V2
Without AIC.
With AIC. Top-end impedance and phase shift near resonance definitely flattens out.
Shorted AIC. Impedance Peak and phase shift remains largely unaffected, but the top-end impedance does decrease.
The inductance on my DATS V2 has never checked out compared to mfg spec sheets. 18 Sound claims 0.06mH @ 1kHz w/ AIC, DATS V2 shows ~0.8mH. Oh well, no biggie. I'm sure the DATS V2 has limitations.
Oh almost forgot, BIG thumbs up to Jay @ Loudspeakers Plus for hooking me up with the great deals on all my 18 Sound speakers! Great company to work with.
Typo in some of the graph titles. The test enclosure is the same one as before, 0.5ft³. I don't know why I kept typing "0.4ft³". It's more like 0.6ft³ when you factor in the polyfill.
The latest update is that there is not much of an update, LOL. I moved in a slightly bigger house in September and have been doing so ever since! Amazing how much crap a hoarder like me has stashed away between this stuff, race car parts, metal working fabrication stuff, and BBQ firewood!
Only updates:
Added a Beyma 10G40 into the line-up for testing, and have sold off the Faital Pro 10FH520's, so I no longer have those.
Waiting on the Beyma 10MC700Nd to be released/received here in the States so I can snag a pair for testing.
Oh, yea this house is a toy wonderland, for me at least. Right now I'm still working on setting up the machine shop in the garage. Three welders, plasma cutter, 7x12 bandsaw, Smithy 3-in-1 Lathe/Milling machine/drill, about 1 ton of steel that needs to go up on a rack that I have yet to finish making, shelving with electric and air tools galore, having a 10HP 38 CFM 80gal 2-stage compressor shipped in, re-wiring the electrical, etc. It never ends!
You have certainly done your due diligence on this project.
And that's why it's going to be nothing short of awesome. The best constraint of this project is that I have no constraints! LOL. It's not meant to be a "DIY for others" kind of thing; it's just for me. Heck, all those speakers are laying on the living room floor (aka no wife to tell me otherwise! )
Oh and just wait until you see the bass section. Over-the-freakin-top, 2,600cm² of total piston area for sub-100Hz reproduction. Per speaker. LOL.
Results from my measurements for the Beyma 10G40 and Beyma 10MC700Nd.
Conditions identical as before to keep the measurements as meaningful as possible. Same electronics. Same 0.5 ft³ box stuffed with ½lbs of polyfill. Different room because I moved houses, that's about the only difference.
Beyma 10G40 Frequency response. 2.83V/1meter.
I think on this test one of the screws was loose and there was a good size airleak getting past the gasket, hence why the distortion below 100Hz looks really high. It is of no consequence since I wouldn't be using it below 100Hz anyways. I might end up re-testing it anyways, just to see if there was an airleak.
Beyma 10MC700Nd Frequency Response
Whatever Beyma is doing to that 10MC700Nd near 85Hz is what they need to be doing to the upper-midbass region!
So, it turns out, that ALL of these harmonic distortion tests are a little "off". Turns out, SPL levels should be kept under 120dB for distortion sweeps, but with the mic so close to the speaker it ended up reading approximately 125 dB for the tests. Didn't see that part in the OmniMic Help file, oops, lol. But, I don't feel that it detracts from the validity of the results because they were all done identically. In reality, all of these speakers would have lower harmonic distortion levels that what these graphs would represent. I played around with the microphone distances while testing the 10MC700Nd, and sure enough percentage levels were definitely lower. One day I might re-do all of these with more attention to this part, but none the less the relative levels allow meaningful comparisons, even if the absolute numbers might not be 100% accurate.
Here is a summary of all midbasses tested:
Obviously, there's many ways to dissect the data. The last chart just uses a made-up formula to compare the values, where the rest are more direct value-derived. Which one am I going to use? That's a tough one. I'm actually torn between a few of them.
Hey, Big O - love your penchant for hard data, and I"m particularly drawn to low distortion drivers these days. That said, I think you are skewing the results a bit, since 2nd harmonic is not objectionable, yet a high 2nd level will skew the results. I think you have flexibility in your analysis pipeline to strip out 2nd harmonic before the final % plot. What do you think? Worth the effort?
But Chahly - Stahkist don't want speakers that look good, Stahkist wants speakers that sound good!
Hey, Big O - love your penchant for hard data, and I"m particularly drawn to low distortion drivers these days. That said, I think you are skewing the results a bit, since 2nd harmonic is not objectionable, yet a high 2nd level will skew the results. I think you have flexibility in your analysis pipeline to strip out 2nd harmonic before the final % plot. What do you think? Worth the effort?
I agree, odd order harmonics should be weighted higher than even order. THD alone doesn't tell much about the sound character of a speaker.
Mathematically, I am not skewing the results, they are what they are. If I mathematically reduce the 2nd order contribution, then that would be more of a subjective thing, as it would depend on your/mine/our decision as to what weight each one has within the THD dB summation. Keep in mind that I do not calculate the THD dB nor % from the higher order harmonics, but rather I use the summation already given by the OmniMic system through the saved data file. I just do the conversion from dB-to-percentage. To strip it out, I'd have to re-work all the data files manually. It would be great if one could specify these things in Omnimic prior to exporting the raw data file.
Not promising anything, but. I can rework the formula to add up the harmonics to not use the OmniMic THD data column, and have a new calculated THD altogether. By how many dB should I subtract/offset the 2nd Order data, and by how many dB should I subtract/offset the 4th Order data?
Mathematically, I am not skewing the results, they are what they are. If I mathematically reduce the 2nd order contribution, then that would be more of a subjective thing, as it would depend on your/mine/our decision as to what weight each one has within the THD dB summation. Keep in mind that I do not calculate the THD dB nor % from the higher order harmonics, but rather I use the summation already given by the OmniMic system through the saved data file. I just do the conversion from dB-to-percentage. To strip it out, I'd have to re-work all the data files manually. It would be great if one could specify these things in Omnimic prior to exporting the raw data file.
OK, I thought you had some scripts to munge the data after export from Omnimic. I think we agree on this, but to be clear, at 1,000 Hz, if 2nd order is at -30dB (not very good) and 3rd order is at -55dB (pretty darn good), the THD will be around -30 dB, due to the high amount of 2nd order, even though it's pretty benign. I think this is why Zaph stopped reporting THD, and just reports 2nd, 3rd, 4th and 5th numbers. In any case, I enjoy your tests a whole lot!
EDIT: Actually, your Excel percentage graphs for each driver has all the HD data anyone would need in a very clear format. That's all I need also. <grin>
But Chahly - Stahkist don't want speakers that look good, Stahkist wants speakers that sound good!
I love the data, but let us know what you picked and why. It's interesting to see what people value in speaker selection. Me any 10 inch I would pick has to do sub 30 Hz
I love the data, but let us know what you picked and why. It's interesting to see what people value in speaker selection. Me any 10 inch I would pick has to do sub 30 Hz
Well these are just the mids, lol. They won't be going much lower than about 100Hz. The bass drivers (four 10" drivers + four 10" PR's per box) will be in an alignment that will not permit sub-30Hz due to enclosure volume. Even so, I don't need sub-30Hz reproduction for heavy metal. I just need LOUD 40Hz-100Hz for fast double-bass drum blast beats and thundering bass guitar riffs.
Someone on DIYAudio.com is sending me their personal pair of JBL 2012HPL mids for testing. First one got here today, and I just finished testing it. I'm gonna wait until I get the second one to obtain averaged results for both FR and HD, but I can already tell some of you (Don) would love to get your hands on one of these! Spectacular 3rd order HD performance for such an "old" unit!
The Beyma 8P300Fe/N? The 8LW30 is cleaner in that range ;-p Also keep in mind, that Beyma publishes their HD tests with 2.83V, so under real power input things aren't as clean, that's why I test with approx 10W input to simulate a bit more realistic scenario, at least for me the way I play my music.
This single 2012HPL averaged 0.4% 3rd order HD from 100-2k, besting even the beast Beyma 10MC700Nd! Second one is not here yet, but now i'm wondering what the newer neo-2251J's have to offer. Technically, the 2012HPL is a bit cleaner, as are all the other ones, but I ended up measuring it identically to put it on the same playing field for a better comparison.
Comments
I'm hoping to eventually mate one 10" to a horn-loaded AMT. A 10" midbass/midrange will start beaming above 1kHz IIRC, so the tweeter has to be robust down-low, and even then the trade-off is that it will produce more distortion down low no matter what.
I added a B&C 10PLB76 and an 18 Sound 10NDA610 to this royal rumble.
Same test conditions as before.
B&C 10PLB76 Frequency Response
B&C 10PLB76 Distortion Performance
18 Sound 10NDA610 FR with and without AIC. The addition of AIC results in ~ +5dB hotter output from about 2kHz-upward, with the change starting around 1.5kHz.
Interestingly, the distortion with AIC decreases on the low-end but not on the high-end, but I guess it makes sense since it is not shorted, but rather in parallel to the voice coil.
Without AIC
with AIC
and for the kicker, here it is with the AIC coil shorted to itself
FR comparing without AIC, with AIC, and AIC shorted. A little bit hotter than without, but not as hot as with AIC on the top-end.
18 Sound NDA610 Distortion with AIC shorted
18 Sound NDA610 Distortion Percentage with AIC, without AIC, and AIC shorted.
18 Sound 10NDA610 Impedance/phase & T/S parameters via DATS V2
Without AIC.
With AIC. Top-end impedance and phase shift near resonance definitely flattens out.
Shorted AIC. Impedance Peak and phase shift remains largely unaffected, but the top-end impedance does decrease.
The inductance on my DATS V2 has never checked out compared to mfg spec sheets. 18 Sound claims 0.06mH @ 1kHz w/ AIC, DATS V2 shows ~0.8mH. Oh well, no biggie. I'm sure the DATS V2 has limitations.
18 Sound 10NDA610 Impedance/Phase combined.
Only updates:
And that's why it's going to be nothing short of awesome. The best constraint of this project is that I have no constraints! LOL. It's not meant to be a "DIY for others" kind of thing; it's just for me. Heck, all those speakers are laying on the living room floor (aka no wife to tell me otherwise!
Oh and just wait until you see the bass section. Over-the-freakin-top, 2,600cm² of total piston area for sub-100Hz reproduction. Per speaker. LOL.
Conditions identical as before to keep the measurements as meaningful as possible. Same electronics. Same 0.5 ft³ box stuffed with ½lbs of polyfill. Different room because I moved houses, that's about the only difference.
Beyma 10G40 Frequency response. 2.83V/1meter.
I think on this test one of the screws was loose and there was a good size airleak getting past the gasket, hence why the distortion below 100Hz looks really high. It is of no consequence since I wouldn't be using it below 100Hz anyways. I might end up re-testing it anyways, just to see if there was an airleak.
Beyma 10MC700Nd Frequency Response
Whatever Beyma is doing to that 10MC700Nd near 85Hz is what they need to be doing to the upper-midbass region!
So, it turns out, that ALL of these harmonic distortion tests are a little "off". Turns out, SPL levels should be kept under 120dB for distortion sweeps, but with the mic so close to the speaker it ended up reading approximately 125 dB for the tests. Didn't see that part in the OmniMic Help file, oops, lol. But, I don't feel that it detracts from the validity of the results because they were all done identically. In reality, all of these speakers would have lower harmonic distortion levels that what these graphs would represent. I played around with the microphone distances while testing the 10MC700Nd, and sure enough percentage levels were definitely lower. One day I might re-do all of these with more attention to this part, but none the less the relative levels allow meaningful comparisons, even if the absolute numbers might not be 100% accurate.
Here is a summary of all midbasses tested:
Obviously, there's many ways to dissect the data. The last chart just uses a made-up formula to compare the values, where the rest are more direct value-derived. Which one am I going to use? That's a tough one. I'm actually torn between a few of them.
Pic of my baby...
That said, I think you are skewing the results a bit, since 2nd harmonic is not objectionable, yet a high 2nd level will skew the results. I think you have flexibility in your analysis pipeline to strip out 2nd harmonic before the final % plot.
What do you think? Worth the effort?
InDIYana Event Website
Wolfy's Photobucket
Not promising anything, but. I can rework the formula to add up the harmonics to not use the OmniMic THD data column, and have a new calculated THD altogether. By how many dB should I subtract/offset the 2nd Order data, and by how many dB should I subtract/offset the 4th Order data?
I think we agree on this, but to be clear, at 1,000 Hz, if 2nd order is at -30dB (not very good) and 3rd order is at -55dB (pretty darn good), the THD will be around -30 dB, due to the high amount of 2nd order, even though it's pretty benign.
I think this is why Zaph stopped reporting THD, and just reports 2nd, 3rd, 4th and 5th numbers.
In any case, I enjoy your tests a whole lot!
EDIT: Actually, your Excel percentage graphs for each driver has all the HD data anyone would need in a very clear format.
That's all I need also. <grin>
Well these are just the mids, lol. They won't be going much lower than about 100Hz. The bass drivers (four 10" drivers + four 10" PR's per box) will be in an alignment that will not permit sub-30Hz due to enclosure volume. Even so, I don't need sub-30Hz reproduction for heavy metal. I just need LOUD 40Hz-100Hz for fast double-bass drum blast beats and thundering bass guitar riffs.
This single 2012HPL averaged 0.4% 3rd order HD from 100-2k, besting even the beast Beyma 10MC700Nd! Second one is not here yet, but now i'm wondering what the newer neo-2251J's have to offer.