Testing 1, 2, 3, testing...

[JonMarsh]

OK, it's a goofy title, but perhaps meant to indicate the tentative nature of any conclusions one might draw-

Got back to doing some driver testing this weekend. Eric Eva brought to my attention some test data published in Germany for a number of roughly 7" midbass drivers, including the Peerless 850439 and the 850467; his point was that based on their measured distortion data, which is lower int he 850467 phase plug version due to the additional of a top side Faraday ring, that I might want to use it instead of the 850439 in the high efficiency high output system Brian Bunge and I have been talking about. This improvement is noted in the midrange, from about 600 Hz on up, and is marked in the range of 1kHz to 1500 Hz.

Some driver manufacturers are noted for their concentration on one aspect of driver performance- Eton, for example, on their honeycomb kevlar and nomex cones- while falling short in other factors- again, Eton, with much higher midrange distortion due to overall mediocre motor design (IMO).

These tests were done only with 2nd and 3rd order harmonic distortion plotted, and are single tone- not a multi-tone test like the Praxis test Mark K has been running on a number of drivers recently.

Eric's comment was a good suggestion, and it got me to thinking some more about some tests I was planning on running - what I'd done the previous weekend was single tone FFT based, looking at all the visible harmonics. But it's very time consuming to cover a wide frequency range for one driver...

Here's an example for the Peerless 850439, at a drive level of 4 VRMS (2 watts at 8 ohms, 90 dB output nominal).



My recent update of CLIOWIN Standard supports harmonic display up to 5 th harmonic. OTOH, though I think my mic (B&K 1/2" condensor) and HP preamp are in good shape, they haven't been NIST calibrated in decades, so I only trust them in the sense of making relative and comparative evaluations, not absolute measurements. Still, I think this is useful, and learned some curious things, yesterday.

First, while I can't measure the absolute sensitivity of the drivers I wanted to test, I can do fairly accurate comparisions, such as between the Extremis 6, RS180, and Peerless 850439, all in the same test cabinet, all with the same 4VRMS verified at the speaker terminals with my Fluke 8922 Wideband True RMS AC voltmeter (and also monitored with my HP 500 MHz scop).

The Peerless is rated at 87.5 dB sensitivity for 1 watt (2.83VRMS), averaged between 100 Hz and 1 kHz. The Dayton is rated at 87.5 dB under the same conditions, and the Extremis 6 is rated at 86 dB.

How do they look? Here are the sweeps, measured at about 6" from the driver front (not my usual extreme nearfield), all in the same cabinet, drive voltage at binding post terminals checked to match within 10 mV.


Peerless Vs Dayton RS180 (Peerless is upper trace)




Peerless Vs. Extremis 6 (Peerless is upper trace).



Doesn't give you warm fuzzies' about manufacturer's sensitivity ratings, does it?


OK, how do they look, overall?

Keep in mind the distortion levels are raised 30 dB to make the graph size reasonable.


Extremis 6, 100Hz - 2 kHz





Peerless HDS 850439 100 Hz to 2 kHz



For reference, blue is 2nd Harmonic, Green is 3rd Harmonic, and Orange is 4th Harmonic. Fifth was genearlly down in the noise floor, and not worth plotting.

NO smoothing at all, here, folks, and their could be some noise contamination (one reason the FFT averaged test is quite good, if quite time consuming).

[JonMarsh]

OK, and here's the Dayton RS180.



The RS180 holds it's own quite well, especially considering it's much, much lower price.

It would be worthwhile to do a similar measurement group in detail just for the lower bass (say, 30 Hz to 100 Hz).

I will get some 850467 in, and see how they stack up- if they're lower in distortion in the upper end, then I'm interested.

Overall, these drivers are fairly close in distortion performance- I wouldn't get in a huff about any of them. I could or should go back and run them at somewhat higher levels, but I thought 90 dB was a reasonable balance to what I was willing to tolerate in the same room- and a moderately decent and respresentative test. 90 dB at 1 kHz is pretty loud...

But, I'm still pretty surprised at the degree of sensitivity difference. Two dB is a lot for nominally idential rated drivers (Peerless and RS180)- remember, 3 dB is twice the power! And the difference with the Extremis is even more marked- this is NOT an 86 dB sensitive driver, IMO... the last published curve on Adire's web makes it look to me like they're 86 dB around 500 Hz, but falling off below that, and maybe 82 dB at 100 Hz. I can't substantiate 86 dB at 500Hz with what I've measured.

Sometime soon I'll try to give the Extremis a workout in the bottom end...

BTW, test amplifier is a Sony TAN9000ES, a mult-channel amplifier capable of 110 watts/ch all channels driven, 125 watts per channel, one channel. Driven at a conservative 4 VRMS, 2 watts @ 8 ohms.

~Jon

[GrahamT]

Thanks for sharing Jon. The Peerless and Dayton have nearly the same FR curve, weird eh? And judging from the tests the Dayton seems to be quite the bargain. Looking forward to the next round.

[Mark K]

Too busy posting and not enough time emptying out your email, Jon.

JonMarsh has exceeded their stored private messages quota and can not accept further messages until they clear some space.

User's mailbox is full: <xxxxxxx@xxxxxxx>

What I was trying to say was.........

What I had suggested is that you consider coming up, not just to drop off the drivers, but to hang around while I/we test them. It shouldn't take more than 4 hours. I'm sure you're busy, like me, but it would get this all out of the way

Some Sunday in Feb might be nice. I should be getting the XXXXXXX 12" sub soon. I can't hold onto them too long.

mark

PS nice work on the distortion plots

[Dennis H]

Thump, thump. Is this thing on?

Jon, just curious about your distortion measurements. Since there is such a difference in sensitivity, wouldn't it make more sense to test distortion at equal SPL rather than equal drive voltage?

[JonMarsh]

Hey Dennis,

You know, I'm basically in agreement with you, EXCEPT that these are supposed to be drivers with very close sensitivity ratings! And also, just to keep things reasonable, I chose a pretty moderate drive level- not too low, not too high. Sort of representative of average moderate listening levels. And, at the moment, I'm more curious about midband distortion- kind of like looking at the difference between a DPL12 and a TC2+, both of which have excellent performance below 50 or 60 Hz, but which deviate ABOVE that... not where people usually get their knickers in a twist, you know?

I've only just recently gotten this CLIOWIN update that allows me to do much of this kind of work - the oiginal CLIOWIN 4.5 that I won as the first place prize in the Audiomatical design Grand Prix contest a few years ago (2000? 2001?) was pretty incomplete and buggy- couldn't even export text data files from MLS mesaurements in a usable format. It was as if someone completely different (from the DOS version) had written it, so mostly I've had to continue using my DOS version until this upgrade.

I don't want to tread too heavily in Mark's or Monte's steps- I've really enjoyed being able to look over their shoulders and glean a lot of useful information- but sometimes I have to do a few comparisons to address my own problems. There were some other interesting measurements Sunday that I'm not ready to discuss, either- more interesting than these. That's still ongoing... we'll publish no data before it's time.

Then it's back to building speakers!

~Jon

[JonMarsh]

Too busy posting and not enough time emptying out your email, Jon.

JonMarsh has exceeded their stored private messages quota and can not accept further messages until they clear some space.

User's mailbox is full: <xxxxxxx@xxxxxxx>

What I was trying to say was.........

What I had suggested is that you consider coming up, not just to drop off the drivers, but to hang around while I/we test them. It shouldn't take more than 4 hours. I'm sure you're busy, like me, but it would get this all out of the way

Some Sunday in Feb might be nice. I should be getting the XXXXXXX 12" sub soon. I can't hold onto them too long.

mark

PS nice work on the distortion plots

You wouldn't believe it, but the last time I downloaded my mail box was this AM before going to work, about 5AM! I got another major data download from Eric, and then a different friend sent me some WMV files that she thought were funny.... sigh. Instant mailbox full.

That's a very good idea (getting together with drivers in Sacramento), and one, as I think you know, I'd been leaning towards, too. It looks like I may be going to Austria on business later in the month (last week), so earlier would be better, or early March (well, after the IEEE APEC conference I'll be at in Austin).

Any idea what day would work for you? Just let me know when you have an idea- I'm fairly open other than the mentioned business travel. We can discuss what to bring further off line, too.

~Jon

[capslock]

Interesting data, Jon. I think I like your unsmoothed graphs better than the graphs we have in our magazines here which leave many questions unanswered, e.g. regarding the frequency resolution and the smooting applied. From the graphs, I have the impression KT uses excessive smooting. The graphs in HH look to me like a spline fit, and occasionally, when something has gone wrong in their typesetting, individual data points are visible which leads me to believe they actually do sine spot testing at discrete frequencies.

The agreement between both magazines is usally ok, sometimes they have good agreement, and when there are differences, they are usually not completely different, i.e. might be attributable to driver tolerances or sloppy level setting. So MLS vs sine testing does not give a fundamentally different picture.

Then, looking at your 100 Hz test of the HDS, -58 dB for second and third order, which is a little over 0.1%, seems a little on the low side, but you may have picked a very good specimen.

Looking at your multi tone measurements, the harmonics again seem a little low. But the thing that is not clear to me is whether those harmonics are relative to the measured FR or relative to some reference level that is constant for all drivers.

Either way, the Extremis was running at the lowest SPL of the pack, hence it's distortion performance at the SPL will be worse than indicated here.

With the published Le curve being that flat (and the large Xmax verified by Klippel as Dan posted) the LF motor linearity is excellent. With the massive copper sleeve and low inductance, the midrange motor linearity should also be blameless (Dan, I would be grateful if you could elaborate on your comment about Le assymmetry).

I then assumed that the suspension was somehow the culprit, but I reread Dan's post and he said Klippel also confirmed excellent suspension linearity.

So what does that leave? Jon may well be right in his assumption that the cone is doing strange things. The other thing might be suspension creep. I suspect in that quick test at CES, Klippel only tested LF suspension linearity. There are, however, some strange effects from suspension hysteresis that can result in strange effects such as frequency dependent assymetry (it's all in Klippel's papers).


I am under the impression that it is easier these days to design and build excellent motors with the software tools now available and cheap Nd magnets and manufacturing facilities in China. We have seen a couple of excellent motors emerge from newcomers (Adire, Aura, Dayton). However, it does not seem quite as easy to get the suspension and cone right, and all these drivers have minor or major weaknesses in these respects.

On the other hand, the traditional manufactors have usually mastered taming cones of all flavors (soft, semi-rigid, rigid) and suspensions. Their motor designs are usually fair (Seas, Audax and Vifa get pretty good distortion performance out of desings w/o any Faraday shielding) to good (Peerless, Excel, Scan), put no effort is taken to really take this to the next level. Scan midbasses and tweeters usually have pretty high second order distortion (probably originating from membrane flexing rather than the motor), and according to one of their designers, their own listening taste is more of a design objective than distortion measurements. Both Scan and Excel use SD-2 type shielding, i.e. shielding above and below the gap, but not inside. No effort was taken in the hexadyme motors to drive the pole piece into saturation, which would have roughly the same effect as full shielding. The high end Excel and Audax tweeters do not even use any shielding, which may again be a question of listening preferences rather vs. commitment to faithful reproduction.

I would really like to see one of the newcomers (such as Adire) join forces with one of the masters of cone/surround engineering (such as Seas).


Eric

[JonMarsh]

Very astute comments, Eric; you sound like someone who's been around and inside this business for a long time. You've made a very observant summary of the design styles of these manufacturers - sometimes I almost find myself pulling my own hair out trying to understand why SS or Seas does one aspect quite well, and seems to ignore something else (I'd love to combine Seas diaghragms with some SS motors, in principle. Maybe that's why I like the SS98000)

I'm curious where you see Focal fitting as regards the design technology, as I think they've put some serious efforts into motor development in their own way, too.

~Jon

[capslock]

Thanks for the kind words, but I am nothing more than a hobbyist making some observations.

I have not worked with any Focal drivers, nor have I particularly closely followed their reviews. Their cone drivers ususally have fair but not excellent distortion AFIR. Also, CSDs were not really good.

Are those 2nd/3rd/4th order distortions relative to measured FR or some reference level?

Greetings,

Eric

[JonMarsh]

The Distortion plots are relative to the measured FR, but with a 30 dB offset.

I haven't been very impressed with the couple of Focal cone drivers I've worked with, but then they were less expensive series- I actually like the polyglass series better than their more expensive sandwich offerings- too many break up modes in the cones.

What I've had an ongoing relationship with, for better or worse, is the TC120 series tweeters- I'm planning on some additional testing for the ones I have on hand.

It seems lately I keep winding up working with a lot of offbeat or unavailable tweeters, just through sheer luck, and a minor bias against soft dome tweeters. That's how I wound up with the Hales Transcendence tweeter (which is fairly top class based on measured distortion, even if it is basically a really souped up Vifa custom built for them). I've only looked closely at the TC120dx2 above 2.5 kHz in the past; time to do a little more investigation. It's not easy to find an 8 ohm tweeter with 93 dB sensitivity and fairly low distortion.

[Ten 99]

... so earlier would be better, or early March (well, after the IEEE APEC conference I'll be at in Austin).

Hey, don't some of us in Austin owe you a dinner or something?

[capslock]

Looks like we have some more work cut out for us, thanks to Peerless:

http://www.madisound.com/cgi-bin/discuss.cgi?read=334017

Among other goodies, there is a full line of 10 - 20 cm HDS drivers featuring the PP / copper cap motor design in combination with a Nomex/paper cone. Design is plain ugly, though...

In more detail:
- line of Nomex cones with inverted dustcap (i.e. same as traditional non-PP HDS, Al ring, no Cu cap, only different membrane), FR looks quite nice
- line of PP cones with inverted dustcap (who needs them)
- line of Nomex cones with PP and advanced motor design and silver basket
- new 10 cm basket added in all lines except SD-HDS

[JonMarsh]

No kidding! Just when I thought it was safe to venture on to the DST site...




You are very right about the plain ugly part, with regards to the Exclusive series. The inductance specs don't make much sense compared with the curves, but the curves look very promising. I wish the 180's didn't have that little glitch just above 1 kHz; that will bear a little near field investigation.




But I think there might be no problem with my "traditional" lowish crossover frequency. The only remaing issue being the tarrif they see fit to impose for these drivers, and when they're available. I see you have some OEM's on the way to look at?

Even the 200 mm version may be something to look at, as it seems the first cone mode may be above 1 kHz, unlike their earlier products.

[capslock]

The OEMs are Nomex w/o plug and 12 Ohm. Will do measurements once I have SE9 up and running...

The Le-spec didn't make any sense in the traditional HDS182 vs. HDS182pp (1.3 vs. 1.2 mH). I thought it was a typo at the time, but may be due to their fit algorithm. Most measurements I have seen have Le around 0.3 mH for the HDS182PP.

Did you notice how straight the HDS182 Nomex cone is? Funny that the 205 looks less glitchy than the 182.

[capslock]

Interestingly, the 14 and 17 cm ugly silverlings are round frames. Just when I had perfected the art of making templates

[JonMarsh]

I noticed that, too... then again, the rounded edge on the frame makes rabetting and getting a flush fit something of a futile exercise. If one crosses over not much over 1 kHz, I have to wonder if the frame edge will make any real difference.

As nice and flat as these look from 100Hz or so to 1 kHz, it just occured to me that the 8" might be a possible choice for the small line array dipole midbass- 75 Hz to 600 Hz? Not the Xmax I would like, but enough swept area based on Sd; 8 would do, maybe fewer. Depends on how the midrange distortion looks. A fairly open basket, too, which I like.

Something to think about...

[capslock]

Last time I inquired about a direct order from d-s-t (for SLS12), minimum quantity was 47, price was really good.

Maybe we can get a group buy organized :wink:

[taz13]

Jon
Do you have an opinion on Vifa's D25AG3506, are there any mods for it. It just happens to be my fav tweeter and I really don't know why. I love the sound and happen to have 4 more kicking around for a future project or 2. And if they can be made to sound better, I am willing to try almost anything once.

Rick

[JonMarsh]

Last time I inquired about a direct order from d-s-t (for SLS12), minimum quantity was 47, price was really good.

Maybe we can get a group buy organized :wink:

Well, if I went with an interleaved columns of the 6-1/2's, that would be almost half of the 47! Logistics would be a bit of a problem...


~Jon

[JonMarsh]

Jon
Do you have an opinion on Vifa's D25AG3506, are there any mods for it. It just happens to be my fav tweeter and I really don't know why. I love the sound and happen to have 4 more kicking around for a future project or 2. And if they can be made to sound better, I am willing to try almost anything once.

Rick

Well, perhaps I shouldn't tell you how to change the sound too much, (if you like it a lot already), but my suggestion would be to remove the acoustic diffuser/lens, and stuff the center pole piece with lambs wool. This should smooth out the extreme top a fair amount, and refine the rest of the range a bit, too. This is an underhung design; the logical extreme of develpment for this is the Hales tweeter I have a few pairs of; no phase diffuser; larger magnet system and gap, slightly lower Fs, and a Faraday cap for the pole piece. (this tweeter was made by Vifa as a custom)

[Jim85IROC]

I seem to be the only person who feels this way, but I think those HDS Exclusive drivers are downright sexy. For a long time I've always felt that if I were to ever go into the speaker business, I would have a driver produced with a bare (or better yet polished) cast basket and phase plug. While it would require a unique approach to box asthetics, I think overall it could give a fresh, but still very tasteful appearance to the finished product.

[JonMarsh]

I'm just hoping the motor is very linear and the cone is break up free in the region I would want to use it!

I would prefer black, though. The little baby one is pretty cute; the 8 looks a bit odd the way the cone is implemented and the smallish bullet in proportion. Will have to see how they work, but if similar to the 850467, might be very nice. And at a semi-reasonable price.

~Jon

[capslock]

Well, if I went with an interleaved columns of the 6-1/2's, that would be almost half of the 47! Logistics would be a bit of a problem...


~Jon

Well, you come here often enough, you could just lug them back four at a time. I do this all the time (the other way around, actually), as some drivers are available in the US which I cannot get here, and even many European drivers are also way cheaper than what I have to pay here (unless I can get OEM stuff). I sometimes get strange looks when they X-ray them, but when I tell them it's loudspeakers, they usually don't even look at them.

Talking about OEM stuff, I received my Nomex non-plugged HDS last night. Unfortunately, the guy packed 8 of those, rather than the 4x Nomex non-PP and 4x SD PP I had ordered (why isn't there any first time right any more?).

Build qualitity is very nice, so are the looks. The Nomex paper is very rough and thick, and probably stiffer than the 8545 carbon fiber cones. The inverted dustcap is not glossy coated paper as in the traditional HDS series, it's pretty dull. It might be some hard and thin plastic like the material used in the CSC/CSX series. If so, it certainly does not look as tacky as a CSC dustcap.


Interesting that those Hales tweeters have a solid copper cap. The cap on my XT25SC seems to be electroplated.


Greetings,

Eric

[capslock]

Had a closer look at my 830 557 OEM Nomexes. The dustcap might well be paper, but with a matté black coating that you find in expensive catalogue paper.

The magnet is 90 mm rather than 102 mm diameter as in the other HDS 182. The VC diameter is the usual 33 mm, but this being a 12 Ohm version, the winding is quite interesting. They seem to have used the same gauge wire, but the winding in longer by maybe 1.5 mm half way, and it has a third layer that stops maybe 3 mm short of the end of the lower two layers, i.e. a 2 3/3 layer VC! This explains how they have maintained the BxL essentially equal to the 102 mm magnet, regular +/- 5.5 mm excursion version.

The longer winding height will make for longer travel, but the BxL curve sure is going to be interesting!

This tells us another thing (and I confirmed this by taking a guestimating look at regular HDS drivers): if there is room for a third winding, there will also be room to insert an outside copper ring into the gap of two layered drivers. I just wonder how to insert one evently enough without resorting to electoplating...

[JonMarsh]

It sounds pretty tricky to me. It's hard to imagine (for me!) how you'd do it without resorting to electroplaing, for the thinness and the precision.

I remember what an accomplishment I thought it was back in the 70's when I finally got skilled enough to be able to consistently rebuild drivers without botching things, even pro-sound drivers with fairly small VC gaps and big 4" voice coils.

Thanks for the additional scans; I'll try to take a close look at them this evening, after I get finishing testing some sub drivers and doing the documentation. A lot of VC impedance glitches down low in some of those models in the scans, which is usually a clear giveaway of mechanical impedance discontinuity.

BTW, your appellation of "polyfloppy" cones does seem possibliy appropriate for the Extremis. I could show you a 6-1/2" midwoofer (if I still have them around) that Vance Dickason designed for a friend (built in Taiwan) years ago; it's almost unbelievable how soft the cone was, in order to damp resonances, I guess... worse than the old Seas RCY 6's.

[ericpoyer]

Eric

I purchased via Ebay 4 units of this 830557 new nomex drive from peerless.
It's an OEM drive, so technical details and parameter are not available via Tymphany website.
Do you have on your side more details about this drive?

Thanks
Eric

[capslock]

Well, the suspension and cone seems to be identical to the regular HDS nomex. I took one apart, and it is a 16 Ohms driver, has no aluminum shorting ring (the regular HDS drivers have one), but it has the same cylindrical pole piece undercut by maybe 0.6 mm.

This was evidently an OEM driver intended for parallelling, hence the high impedance. Also, the OEM cared more about looks than distortion. Bxl linearity will be ok (in spite of the primitive motor design) thanks to the asymmetrical winding. Distortion especially down low in frequency will not be spectacular.

It's an ok driver but it takes a skilled diy modder to make something akin to the HDS exclusive out of it.

If I hadn't 8 of those already sitting on a shelf, I might be tempted to offer to buy them from you

[JonMarsh]

Eric, your motto could be "the only good driver is a modified driver".

Seriously, it's interesting the timing on which this thread floated back up, even if on a slighty OT question, as I'm expecting to revive it soon myself.

My Praxis setup and ACO microphone and preamp arrived this week, as well as the balance of the Seas drivers I wanted to test. I hope to conduct and continue some "fun" shootouts, in which Praxis will play a large part, but CLIO won't be abandoned. CLIO makes it easy to look at some things in a certain way that is useful (i.e., I like FR sweeps with distortion plots as well as multi-tone spectral plots and shaped sine bursts.) This year has been good, but it's going to get better... :B

[capslock]

Eric, your motto could be "the only good driver is a modified driver".

You know how physicists are: they keep telling the engineers that they didn't do their job right and that they can do better...


Seriously: I am beginning to put my workshop back together. Also, I have finally connected the scanner that arrived when we we already beginning to move. If you want to drop me your mailing address, I will send you a DVD one of these days.

[ericpoyer]

Thanks for your answer.

I am starting with soundeasy, so having T/S parameter for these Nomex should have been a great help.
The software package seems very powerful but not so intuitive.
Measurements of T/S parameters is may be not the easiest way to start!

Eric

[ciberproton]

Hi everyone, I'm newest in this foro and I'm really surprised for the diy audio level. I've just bought a pair of 830557 OEM HDS peerlees at ebay if anyone could get the TS parameters it'll be nice.
Regards
Fran

[cobbpa]

I don't have any, but have been wanting to do some work with the nomex drivers. The 830869 (8") version seems to be pretty workable, just that peak at 3 khz. Of course, I haven't seen them or tested them, I don't have the money to just buy them yet...but I spend enough time reading & looking at graphs to know that they seem decent. Not to hijack, but has anyone tested or looked closely at these? I'm curious as to why I haven't seen them implemented or mentioned very much, on any forums I read, really.

[JonMarsh]

The 830884 is of more interest to some of us here because of differences in driver paramters- lower Le, much lower inductance rise at high frquencies (16 ohms at 20 kHz, versus 16 ohms at 3 kHz), expected lower midrange distortion (from those parameters), and Qts better suited to the application's I'm interested in. I probably wouldn't use either one above 1400 Hz. Would probably shoot for a 1300-1400 Hz crossover. Lower in a three way, but I wouldn't use this as a woofer in a 3 way (prefer 10's or larger).

[cobbpa]

This is interesting; will definitely keep an eye on this thread. I had looked at the 830884 too, interesting that they appear to have the same frame & cone, different dustcap / plug ideas though. The little differences must be more of a factor than I recognize. Carry on.

[JonMarsh]

Midwoofer testing

Just a preamble- I'm doing some midwoofer testing, actually in most cases to evaluate for use as a "midrange". By that I mean 250 Hz to 2 kHz.

I'm using techniques that are slightly different from others published on the web, so I wouldn't try to draw direct comparisons- still, might be something useful for you here.

Testing was done not with raw drivers but instead with driver mounted on a 24" X 24" diopole baffle. This pushes the rear wave positive reinforcement to a much lower frequency. I wanted to look at the region from 500 Hz to 4 kHz in some detail, and this should make the results a bit more valid.

Most measurements were done at 6" from the baffle plane, except the "nearfield" measurements which were done flush with the baffle plane, centered on the driver. Microphone is an ACO Pacific 4012 1/2" 200V condensor wtih PS9200 dual channel power supply.

Simple swept distortion plots were also made, at nominal SPL of 90 dB, to determine the "best case" distortion performance at nominal/typical SPL levels. I've seen drivers that even when you go easy on them at low levels aren't all that great (sub drivers included), so before detailed and time consuming torture testing, I wanted to weed out the weak players with simpler tests. Tested in this first lot of 7" drivers include the venerable but discontinued Peerless 850439, the Vifa XG18H00, the Dayton RS180-8, and the Peerless 830883.

[JonMarsh]

Peerless 850439

This driver was included because I have a few "in stock", it was very popular in it's day with a number of commercial systems, and it's thought to be fairly easy to work with due to it's smooth upper end response without big peaks.






Here's the basic swept frequency response, at 5 dB/div.




Here's the nearfield response; note that we don't seem to see a cone out of phase notch until almost 4 kHz; not bad performance for a floppy cone.





Next, swept FR and HD2 harmonic distortion, 10 dB/div. Looks pretty nice, I'd say a fairly symmetric motor and suspension at these signal levels.




Swept FR and HD3 looks a bit different, and here the lack of Faraday rings on the pole piece takes it's toll, with steadily rising midrange distortion. Still, between -40 and -50 isn't bad; OTOH, it's not a "loud" level, except perhaps to your great Aunt Tilly.

[JonMarsh]

Vifa XG18H00

The Vifa XG18H00 has attracted a fair amount of attention lately, due to it's reported low midrange distortion, and for that reason I picked up a pair to test. It has a woven fiberglass cone, with resin on the back, NOT a poly and fiberglass contruction as some have reported (on what basis I have no idea) on the Madisound board. It's also thought to be a bit easier to work with, including use of relatively low order crossovers, due to it's softer upper end behavior.





Here's the "standard" FR response on baffle, 5dB/div. The baffle bump at 250 seems higher with the XG18H00 than the others, and it's published data show a similar response as well as irregularities in the impedance curve. Resonance in the double roll suspension? Could be, but I didn't investigate further.





Here's the nearfield response- a rather strong shelving effect is shown.





Here's the FR + HD2 plots, 10 dB/div.






And the FR + HD3 plots.





Interesting that it shows a falling distortion with increasing frequency- probably plenty of copper in the gap, but maybe a bit of inherent motor non-linearity that drops off as cone movement declines?

[Paul H]

Jon,

Your testing format and comments are very user-friendly. :T

Are the speakers mounted centred in the baffle?

Paul

[Evil Twin]

The widenet Imperial monitoring shows that Jon did use drivers centered in the baffle; however, because of the varying distance to the edge depending on whether it's at the side or the corner, the frequency of baffle reinformement doesn't stack up too heavily, and the relatively near field measurement also helps reject this.

[JonMarsh]

Rs180-8

Note I'm making the distinction that this is the 8 ohm sheilded RS180- now that there is a 4 ohm non-shielded version with somewhat different characteristics. I have some of those on the way to test also.





Frequency response at 6" measurement distance from baffle, 5 dB/div.





Here's the nearfield response- note the sharp dip/null at ~2.5 kHz, similar to what I've seen with the RS225 at 1500 Hz.




The FR + HD 2 plot at 90 dB looks pretty good, 2nd harmonic only poking up much above -60 above 1500 Hz.




The FR + HD3 plot doesn't look quite as good, but is still fairly respectable...




Strong performance below 800 Hz, not as good above.

[JonMarsh]

Peerless 830883

The Peerless 830883 is priced similarly to the Vifa XG18H00, which is to say almost double the RS180. While beauty is in the eye of the beholder, and some would argue for a black bullet instead of silver, the overall impression is of a high standard of workmanship, with minimal basket masking of the rear wave.




Here's the 6" measured FR on the same 24" baffle as the others- the somewhat rising top end response is evident. 5 dB/div




This driver reportedly uses a lot of copper in the gap to lower VC inductance, which combined with directivity with increasing frequency could account for the FR curve above. The near field response seems to substantiate that idea, as it looks flatter, being in a fairly tight band to 3 kHz. Again, at 5 dB/div.



Next up is the FR + HD2 distortion at 10 dB/div. Fairly solid performance from 200 Hz up to 2 kHz.





Last, the FR + HD3 distortion, again at 10 dB/div. Nice performance up to 2 kHz.





When you look at the FR plots and the basic harmonic distortion plots, this driver is clearly an attractive candidate for my "midrange tasking 250 Hz to 2 kHz; how it stacks up at higher drive levels will be interesting to investigate. I probably should get in some W18s; I'm still waiting on the C90/T6, from which I expect very good performance.

[JonMarsh]

HiVi M8a

Been using these for a few years, it's not the most high tech driver out there, but the price is right. How does it stack up with some of the competition these days?





Since this is a sort of "application specific" test series, I cheesed out and used the Arvo baffle for testing, so don't compare directly with the results above, though the test conditions were otherwise identical.

Here's the basic SPL behavior.




This agrees quite closely with what I've previously measured on the baffle- no surprises here.

Here's the nearfield. You can see why I've used this driver with a 1200-1400 Hz crossover, but steep slope.




This correlates quite well with other nearfield measurements I've done even in closed boxes, including "ultra nearfield" in the center of the cone (at about 1/2"). This is definitely one of the better behaved 8" cones. The best? We'll see.



Here's FR + HD2 at 10 dB/div.



2nd Harmonic distortion isn't too bad, though of course it spikes when we get to the main cone breakup mode. Funny that we don't see more signs down low....


Here's the FR with HD3, at 10 dB/div




Here we see the drawbacks of a motor without copper in the gap to control inductance and modulation distortion. Very similar profile to the Peerless 850439, aside from the extra spikes due to metal cone resonance.

[JonMarsh]

Peerless 830884

This is a driver I've been wanting to take a look at for some time. They've been sitting around Casa de Juan for a while, but just now have had the gear up and running and some time to test.




Here's the SPL plot, also measured on the Arvo test baffle, at 5 dB/div.



Does that bit of a dip at 1 kHz mean anything? Here's the nearfield plot, again at 5 dB/div.



This is not what I was hoping for- what appears to be what I call an "out of phase cone flap" down around 900 Hz. Note that the RS225 has that, but at around 1500 Hz. The RS180 has it, too, at around 2500 Hz.

Here's the FR and HD2 plot, at 10 dB/div.



Fairly decent, but climbing in the upper end. Not as good as it's little Bro, though.

Here's the FR and HD3 plot, at 10 dB/div.




Well, the midrange nonlinear distortion is pretty good for swept tone, though not SOTA; but that notch at 900 Hz bothers me- for me, it's pretty much a deal breaker. Reminds me of the little Carbon Fiber MCM 6-1/2, which has a notch at the same frequency. If I'm going to be true to my school, I have to say, no cigar.

[JonMarsh]

Seas Excel W22

OK folks, this is the big round, the one you've all been waiting for, where we bring out the reigining 8" midrange champion and give him a once over... the Seas W22.



OK, I'm not the best photographer, this driver is expensive and it LOOKS expensive and well made. No surprises there. It's tested under the same conditions as the previous two.

Here's the FR on Arvo baffle, 5 dB/div at 6 inches.




It doesn exhibit the characteristic dip in the 2 kHz region that the W22 is known for- a dip that correlates with some considerable energy storage issues before the main breakup.

Here's the nearfield response at 5 dB/div.





My, my, what do we have here? Third party testing of the W22 often shows ETC curves getting funny arond 1400, and from this plot, I think it's easy to see why- looks like we're going into a bit of a flapper mode around 1700 Hz. Looking at this, I suspect SL is pushing things a bit with a 1400 Hz crossover. I'd shoot for something lower, like I've used with the RS225 in the M8ta. But maybe I'm worrying about nothing- the Orion has a huge following which it certainly deserves, compared to most of everything else out there. What if it could be improved, though?

Next up is FR with HD2 plot, 10 dB/div, now you see what you're spending your big bucks at Seas for....





HD2 is definitely down in the mud, I don't think it gets any better than this for a swept sine (well, haven't seen a C90/T6 yet). Oh yeah, that 830883 was pretty good, wasn't it?

FR with HD3 is next, at 10 dB/div.





Looks good, but doesn't blow us away like the HD2. Hmmmm. Still, as you can see, it's hard to do better than this.

Though I have to wonder if two 830883's wouldn't be a better value proposition. And they don't have the notch at 1700 Hz, either.

Thoughts to ponder this Sunday evening.... :unsure: