12 New Drivers Measured at Zaph Audio

[Amphiprion]

Just an FYI.

ETA: The Usher 8945A looks like the new low distortion champ at $110, better than anything else. There are several cheapies that did quite well (MCM and the like). The new GR was average.

[tktran]

What a great roundup.

The data is clear, but like all research, the results and conclusions is not always straightforward, and make for good discussion. :B

The AA 6.5 Poly would make a great pure midrange. Bargain too.

On brief inspection, the Ushers do well, though I would give the nod to the 8945P.

I recall that Feyz did a nice two way design with the Seas 27TDFC and Usher 9950. Should inject some new interest in the 'old' Usher drivers.

[HMenke]

It's an amazing effort and very well presented. But I need educatin'. What are we looking for on the CSD graphs? It looks like we want to identify resonant peaks that take a long time to decay?

[SteveCallas]

The way that is laid out allowing you to quick switch between the drivers is excellent.

The AA 6.5 Poly would make a great pure midrange. Bargain too

Yeah, that one looks to do surprisingly well even before considering price. How come it doesn't seem to get used very much?

[peterS]

tinsel lead slap
and the guy is untrust worthy

[joecarrow]

So, that driver (the AA 6.5 poly) wouldn't be any good as a replacement for an Extremis for a tiny stealth sub?

[cotdt]

tinsle lead slap
and they guy is untrust worthy

i use these AA6.5's and you can fix the tinsel lead slap problem with a small amount of hot glue. it's very easy to do. and it only occurs at higher output, much higher output than in zaph's tests. even if it occured, harmonic distortion tests won't show noise issues very easily.

with the big xmax (official figure is 11.5mm), most people who use these use them as midbass in cars. surprisingly, zaph's tests showed that it's actually more of a midrange than midbass.

i remember doing blind tests on myself (on good audio equipment) showing that i could only detect 3% harmonic distortion 3rd order and 5% 2nd order on piano music. it's higher for normal music. and this was distortion added across the whole spectrum.

among the better drivers in zaph's tests, harmonic distortion is less than 1% if i'm reading it right. i think zaph tested at too low a level to be useful, since some drivers will scale up better than others. for example, the seas excel W18EX does extremely well in bass and midbass, but only has 3mm of linear xmax despite Seas' claims of higher 5-6mm. i wonder how it would fare at higher output against the other top performers which have more linear motors.

[technimac]

:roll:

tinsle lead slap
and they guy is untrust worthy

Do you have an English translation for that bit of grammatical nonsense?

[cotdt]

:roll:
Do you have an English translation for that bit of grammatical nonsense?

the leads are too long and slap against the cone during high excursion (and this woofer has plenty of it). but there are at least 3 good ways to fix it easily and nothing to complain about given the low price and good build quality. when i got my pair, i noticed that the motors were well-centered unlike other (far more expensive) drivers.

[SteveCallas]

So you hot glue the leads to the back of the cone?

[cotdt]

So you hot glue the leads to the back of the cone?

they are hot glued back to the frame, not the cone! a better way is to just cut and shorten the leads, and solder it back to the terminals. but if you shorten them too much it will no longer support the big excursion.

[SteveCallas]

Lol, ok, that makes much more sense h:

[peterS]

:roll:
Do you have an English translation for that bit of grammatical nonsense?

sorry was writing one handed in the dark while watching tv ops:

[peterS]

it pains me to do buisness with chad again but im tempted-lol

[JoshK]

I might be reading this wrong but it looks like the XG18 would make a better midrange than the AA6.5. Seems problems creep in a little too low to make good transition to most sane tweeters with the AA6.5.

[Scott Simonian]

15 posts and no linky?

[JoshK]

zaphaudio.com most folks here are quite familiar with the site.

[cjd]

Scott, you're admitting publically that you don't know zaphaudio.com? :P

[peterS]

I might be reading this wrong but it looks like the XG18 would make a better midrange than the AA6.5. Seems problems creep in a little too low to make good transition to most sane tweeters with the AA6.5.

i came to that conclusion too but john added some info

http://www.htguide.com/forum/showthread.php4?t=22931

[technimac]

sorry was writing one handed in the dark while watching tv ops:

No problem, but I think there's an Occupational Health Hazard Warning on the bottom of the keyboard about doing "that". :W
Appreciate the insight about the quirks of these drivers.
Cheers, Bruce

[Jim85IROC]

I might be reading this wrong but it looks like the XG18 would make a better midrange than the AA6.5. Seems problems creep in a little too low to make good transition to most sane tweeters with the AA6.5.

What problems would those be? CSD and THD plots look very good to my eyes at frequencies well above normal tweeter crossover regions.

[Drew]

I might be reading this wrong but it looks like the XG18 would make a better midrange than the AA6.5. Seems problems creep in a little too low to make good transition to most sane tweeters with the AA6.5.

i came to that conclusion too
http://www.htguide.com/forum/showthread.php4?t=22931

Well, the tests performed by Jon and John arrive at fairly significantly different results for the XG18. This could be due to different systems, test setup, etc. If you're not doing your own testing, it leaves you in a bit of a difficult place.

I have to admit that I'm a bit skeptical of the AA 6.5 midrange results. I've never seen a poly cone return results anywhere near that good (well, the bass isn't that good).

I don't know enough about driver design to know what mechanisms generate distortion as the frequency changes. I suspect that the suspension and cone are the dominant factors at low frequencies, while the motor dominates at high frequencies. If this is true, it would explain the behavior of the AA 6.5.

Still, it doesn't look that great below about 300-400 Hz. The Peerless 830883 looks better if you are crossing fairly low.

I'm currently in the middle of a search for a "perfect mid" that's not Accuton expensive, and it's a tough call. Even the venerated W18 is looking a bit less compelling compared to some of the new and cheaper competition.

[JonMarsh]

For reference, my tests were done at 1 watt drive level nominal (2.83 VRMS) with Praxis and an Acco Pacific 1/2" condensor mic (think expensive, 40 kHz bandwidth, low distortion, lots of headroom).

Why this drive level? It's moderately loud but not very in a home enviroment, and is a level that might be used for music playback by a lot of people. It's NOT grueling test, but if driver doesn't do well at this level, I don't care what it does elsewhere.

Subwoofer drivers I usually meaure at 4VRMS and 10VRM.

Different drivers have different responses to increasing drive levels. AND different samples of drivers from different production lots may not behave the same, due to minor variations in manufacturing or design changes. For example, the Millenium Excel tweters I have are not as smooth and consistent as I would like to see- but my curves for the RS28a and for the Focal TC120dx2 look just like the factory curves. Measured under the same nearfield conditions with 2' square baffle.

The XT18 isn't bad- it just doesn't seem to me to be as good as some other drivers. It should be easy to work with, and you might make a pleasant sounding speaker for many applications, particularly in HT if you cross to a sub. But alas, it's not a reference level midrange, but it is pretty decent. I wanted it to be better, truthfully- all these drivers I've tested I've paid for myself.

[JonMarsh]

Well, the tests performed by Jon and John arrive at fairly significantly different results for the XG18. This could be due to different systems, test setup, etc. If you're not doing your own testing, it leaves you in a bit of a difficult place.

I have to admit that I'm a bit skeptical of the AA 6.5 midrange results. I've never seen a poly cone return results anywhere near that good (well, the bass isn't that good).

I don't know enough about driver design to know what mechanisms generate distortion as the frequency changes. I suspect that the suspension and cone are the dominant factors at low frequencies, while the motor dominates at high frequencies. If this is true, it would explain the behavior of the AA 6.5.

Still, it doesn't look that great below about 300-400 Hz. The Peerless 830883 looks better if you are crossing fairly low.

I'm currently in the middle of a search for a "perfect mid" that's not Accuton expensive, and it's a tough call. Even the venerated W18 is looking a bit less compelling compared to some of the new and cheaper competition.

For use as a mid, I'd say the 830883 is the bang for the buck champion. You can always do better, but at increasingly disproportionate cost. Sometimes, that's worth doing anyway. Othertimes, not.

YMMV, JMO, etc. All the usual disclaimers.

In the end, we all have to vote with our wallet one way or another, for something that will hopefully please ourselves. Now, if you're in it to sell to other folks, that's a whole other level of issues, as guys like Rick Craig will testify. (I notice Rick is using a lot of Aurasound and Accuton lately, too).

[JoshK]

I had forgotten that our own Jon had tested the XG18....I was looking at only the zaph site comparison thingy (which kicks ass btw :T ).

[JonMarsh]

Yeah, Zaphs full round of tests is pretty cool, he's put a lot of work into it. I'm jealous that he has that much free time!

[TurboFC3S]

I was the guy who sent John the GR M-165x to test ... guess it's obvious that I'm a bit disappointed with the results. Especially disappointed at how OFF the measured T/S params are from GR's published data.

I feel bad for Danny, since his sales will probably drop to zero for that brand new driver. I guess that's what he gets though for not putting any copper in the motor at a $45 price point.

[cjd]

I feel bad for Danny, since his sales will probably drop to zero for that brand new driver.

People buy drivers for all sorts of reasons, and I doubt Danny is going to find orders stop - rather, the same crowd that used to buy from him is likely to continue. Namely, folks that want a driver that is easy to work with and pleasing to the ear. I very much suspect this driver is all these things.

These are the people that might try crossing an RS180 at 2.8kHz

Second order acoustic...

C

[cotdt]

I was the guy who sent John the GR M-165x to test ... guess it's obvious that I'm a bit disappointed with the results. Especially disappointed at how OFF the measured T/S params are from GR's published data.

I feel bad for Danny, since his sales will probably drop to zero for that brand new driver. I guess that's what he gets though for not putting any copper in the motor at a $45 price point.

Harmonic distortion seems to be only 1%, which is not supposed to be audible for music. Sure, the other drivers did better, but when distortion is that low I think it would come down to personal preference. I don't think it would sound muddy, harsh, or grainy because the distortion isn't high enough to warrant that.

[TurboFC3S]

Well, I guess "drop to zero" isn't what I mean ... but sales will certainly be affected by Zaph's tests. For me, especially troubling is the discrepencies in T/S params. The published specs are much 'better' than what Zaph got.

I have my main L/R's done with the M165x in a 2.5 way floorstander. They sound good, but I'm still voicing out the xovers. They do seem to loose a little note separation at higher volumes. They have good bass though, and the crossovers were certainly easy to work with.

[Dennis H]

Harmonic distortion seems to be only 1%, which is not supposed to be audible for music.

Zaph is testing at a pretty low SPL so the distortion will go up as you play them louder. As well, the M-165x has very high levels of 3rd (and 5th) harmonics which sound worse than 2nd order distortion. Not what you'd expect from a paper cone. Must be the motor.

[Scott Simonian]

Scott, you're admitting publically that you don't know zaphaudio.com?

Lol...

No. I just didn't want to use any effort. :P

Busy @ work too!

[JonMarsh]

People buy drivers for all sorts of reasons, and I doubt Danny is going to find orders stop - rather, the same crowd that used to buy from him is likely to continue. Namely, folks that want a driver that is easy to work with and pleasing to the ear. I very much suspect this driver is all these things.

These are the people that might try crossing an RS180 at 2.8kHz

Second order acoustic...

C

Oh CJ, you're so cruel! Sure wouldn't want to hear that! Of course, this is coming from the guy that think's you're tempting fate and reason to try to get any midwoofer up above 2 kHz unless it's just a 5".

[jdybnis]

My Opinion

Well Zaph asked for people to form there own opinions of his measurements. I want to preface this by saying that I haven't used, or even heard any of these drivers with just a few exceptions.

A few overall non-constructive comments first:

1. Just from these measurements, I don't know how to predict the distortion as the signal level increases. If at higher levels non-linear distortion didn't get any worse, or just scaled with the signal then pretty much all these drivers would be excellent non-linear distortion wise.

2. I'm just going to comment on the midrange performance. I don't see how I could tell anything about the low frequency performance from these measurements. The drivers all start rolling off at different points on the low end. The cone excursion and the power delivered would be totally different for each driver in its proper enclosure.

3. I'm not going to comment on the linear distortion. The impact of some anomaly I think comes down to how easy it is to correct in the crossover. I don't have the experience to judge this.

Continued...

[jdybnis]

I split the drivers into 4 groups; from worst harmonic distortion to best (in the midrange).

Group 1: "not worth the time it takes to route a hole in the baffle".
Morel MW166
Fostex FE166ES-R -- I was surprised this did so poorly compared to the Fostex FF165
GR Research M-165X
Dayton DC160
MCM 55-1860
MCM 55-1185
Mach 5 MLI-65
Seas CA18RLY -- borderline; if it had lower 4th hd it wouldn't be that bad

Group 2: "usable but not great"
Hi-Vi F6
Dayton DA175-8
Seas L18
Fostex FF165
Seas P18RNXP
Seas CA18RNX

Group 2.5: "good but not great"
Vifa XT18WH09
Hi-Vi D6.8
Audio Technology 18H520613SD

Group 3: "very good"
Peerless 830883
Seas RW165
Seas W18NX
Dayton RS180

Group 4: "best"
Usher 8945A
Adire Extremis
ScanSpeak 18W8531G
Seas W18EX001
Vifa XG18
Ascendent Audio AA6.5
Usher 8945P

Continued...

[jdybnis]

All metal cone drivers have a high frequency breakup that needs to be handled in the crossover. The distortion products of signals below the breakup frequency excite the breakup too. This is a consideration when using the driver.

Seas RW165 and W18EX001 -- within the group, these two drivers have lowest frequency breakups. They have to be crossed lower than the others to prevent the 3rd harmonic distortion products from audibly exciting the breakup. To cut out the 3rd harmonic the crossover needs to be below 1.6kHz (1.5kHz for the RW165). On the plus side the higher order distortion products (4th order and up) are so low that they probably will not cause any problems.

Dayton DA175-8 -- has the highest frequency breakup of the bunch, unfortunately it is also the broadest bandwidth breakup and the probably the most audible. I would still try to crossover below 1.6kHz on this one too, although not such a steep filter is necessary as with the W18EX001.

Seas L18 -- this is nice because the breakup is very high in frequency and very narrow in bandwidth. Unfortunately the overall distortion products are the highest in this group. The 5th harmonic at 1.3kHz is most likely going to be audible.

Dayton RS180 -- The easiest to work with. The breakup is pretty high up there in frequency, and the higher order distortion is low enough that the distortion products are probably not an issue. With a very low crossover (1.4kHz or lower) it's a toss up between this and W18EX001 for lowest harmonic distortion.

Continued...

[jdybnis]

Lowest Midrange Harmonic Distortion: Usher 8945P
2nd Lowest (and best value): Ascendant Audio AA6.5

[cjd]

All metal cone drivers have a high frequency breakup. The distortion products of lower frequencies excite the breakup. This is a consideration when using the driver.

ALL drivers will have a breakup of some sort due to the same reasons - the question is how well self-damped they are. Which then makes me wonder if that's a good thing, because if such often dramatic breakups are damped, how much are sounds we *want* also damped?



I didn't see the RS180 in your initial 4 categories - am I just blind?

C

[cotdt]

ALL drivers will have a breakup of some sort due to the same reasons - the question is how well self-damped they are. Which then makes me wonder if that's a good thing, because if such often dramatic breakups are damped, how much are sounds we *want* also damped?

C

I would like to add to this. In my experience the higher dampened cones using more dampened materials and high-loss surrounds create a much bigger subjective difference than harmonic distortion. Specifically, the transients are less sharp. It may be less fatiguing this way while maintaining the resolution of the driver, but imaging and spaciousness is weakened. For example, Seas Excel W18 has very sharp transients while the Peerless Exclusive has soft transients, so the sound is very different though they both have low distortion and good resolution in their passband.

[WillyD]

I didn't see the RS180 in your initial 4 categories - am I just blind?

Looks like he called it "Seas RS180" .

[Jim85IROC]

I would like to add to this. In my experience the higher dampened cones using more dampened materials and high-loss surrounds create a much bigger subjective difference than harmonic distortion. Specifically, the transients are less sharp. It may be less fatiguing this way while maintaining the resolution of the driver, but imaging and spaciousness is weakened. For example, Seas Excel W18 has very sharp transients while the Peerless Exclusive has soft transients, so the sound is very different though they both have low distortion and good resolution in their passband.

wouldn't this lack of transient response show up as slower decay in the CSD? After all, if the speaker isn't "fast" enough to produce the transient, then it shouldn't be "fast" enough to show quick decay either.

[cjd]

It might show up in a CSD but perhaps not as slower decay. A transient may simply be sucked up - damped to softness. Do the softer cones that don't exhibit strong breakup nodes always decay slower than the metal cones?

C

[Jim85IROC]

It might show up in a CSD but perhaps not as slower decay. A transient may simply be sucked up - damped to softness. Do the softer cones that don't exhibit strong breakup nodes always decay slower than the metal cones?

C

If they're storing the energy that they absorbed during the transient, I would think they would, since that stored energy is being dissipated somehow. But... I'm just an amature when it comes to the mechanical properties of cones. :

[Chris7]

I would like to add to this. In my experience the higher dampened cones using more dampened materials and high-loss surrounds create a much bigger subjective difference than harmonic distortion. Specifically, the transients are less sharp. It may be less fatiguing this way while maintaining the resolution of the driver, but imaging and spaciousness is weakened. For example, Seas Excel W18 has very sharp transients while the Peerless Exclusive has soft transients, so the sound is very different though they both have low distortion and good resolution in their passband.

Usually, the "speed" of transients comes from the tweeter, not from the midrange driver. That's because transients have a Fourier decomposition that is dominated by higher frequency components.

That said, I do agree with your perceptual generalization, but I don't agree that it has anything to do with cone damping. It may have more to do with the distortion profile of rigid cones (tilted towards odd harmonics).

[peterS]

does someone else want to test the aa 6.5" ?