Assorted driver tests

[augerpro]

Hi everyone. I've just finished some driver tests and put them up here:

Images not available

just look in the albums with the blue names, all the rest on the main page is just my junk.

The FR plots are 1/24 smoothing so they are a little unforgiving. Gate was 7ms for farfield and 200ms for nearfield. The shaded are means the data is invalid.

The THD lower frequencies aren't quite as smooth as Zaph's. I think he uses a higher sampling soundcard though, so that may be why. If you look at the RS150 and 832873 plots they follow Zaph's quite well. It took alot work to get that let me tell you! But I'm pretty confident in the plots now. Please look at the Baffle effect jpg first so you can see what the effect is on the FR. Tweeters aren't really effected here, but the woofers are at lower frequencies. Tweeters were level matched using white noise, and woofers were matched using pink noise at .5 meter. This level is in the file name. There is a room effect at 120-140 Hz. If you guys catch any others let me know, I didn't see anymore though.

I haven't really analyzed anything yet but from a quick glance the Goldwood GW1258 and Selenium didn't do that bad in THD. CSD wasn't too good I don't think and the Z plot is kinda scary. Alot of cone breakups going on I think.

And the Scanspeak 9800 impressed me. Look at that FR! Flat to 1000Hz! Impressive. Now at first glance you may think the lower end distortion is higher on the 9800 vs the Peerless HDS. But these plots are unequalized so the HDS appears lower. If they were equalized i would say the 9800 is slightly higher in 2nd harmonic distortion and lower in the higher orders in the <2000Hz area. Anyone agree? I am reading that graph correctly? One thing you can't see is the higher spectrum noise I see during the sweep. Basically it looks like MarkK's multi tone plots. Here the 9800 was lower in noise than the HDS. Although I don't knwo how important taht is given those frequncies are well above 20k anyway.

Let me know what you guys think of these tests. If you have any advice I'd love to hear it. Also some pro's evaluation of the results would be helpful, especially the proaudio drivers. Over the next couple days I'll be adding some more drivers, particularily some HiVi M and B series drivers, and some small MCM drivers.


EDIT: during the Goldwood and Selenium test the mic level maxed adn clipped above 2000Hz. It's pretty obvious on the plots. So I ran another plot just from 2000-5000Hz with the mic gain down. Also the Goldwood 1258 Q was .48 and the Selenium 10PW3 Q was .71

[augerpro]

Also the SPL level is not absolute, but the GW1258, 10PW3, 9800, and HDS tweeter were all run at the same level so the relative sensitivity is accurate.

[TacoD]

Thanks for posting these results. The scanspeak 9800 is a favourite for me.

[mazurek]

Thanks for the tests. Advice for the future, post charts in a format such as *.png which will not result in hash around the grid and text. Not that it is a big deal, but it will look a little cleaner.

[Gir]

If you zoom in it looks just fine

[augerpro]

I remeasured some drivers so that things were more standardized as far as graph scale and drive levels. The graphs should be more informative and the same for every driver. All drivers should have a THD at 96dB now. I figure that's a good enough level to represent high output-and not fry a SS9800 8O Hopefully I can get some new ones tomarrow.

On second look that HDS tweeter is a superior tweeter. The 9800 goes lower but it just can't compete in distortion at the lowend. I'll be building bookshelves with those tweeters and the RS150 and HDS Nomex this summer to see if I have any preferences for soft/hard domes/cones. Should be fun. Looks like I might have to throw the AC130Mk2 in the mix now too.

Hopefully the HiVi K1 in the waveguide is able to cross low. I want to use it for my PA dipole crossing to the Selenium 10PW3. It just a fun little project but looking at the results for the 10PW3 I don't know if it's even worth using just as an experiment. Looking at the CSD and THD I don't even want to cross it higher than 1k Hz since it gets pretty ugly by 2k Hz. Meh, we'll see...

[Jim Holtz]

I remeasured some drivers so that things were more standardized as far as graph scale and drive levels. The graphs should be more informative and the same for every driver. All drivers should have a THD at 96dB now. I figure that's a good enough level to represent high output-and not fry a SS9800 8O Hopefully I can get some new ones tomarrow.

On second look that HDS tweeter is a superior tweeter. The 9800 goes lower but it just can't compete in distortion at the lowend. I'll be building bookshelves with those tweeters and the RS150 and HDS Nomex this summer to see if I have any preferences for soft/hard domes/cones. Should be fun. Looks like I might have to throw the AC130Mk2 in the mix now too.

Hopefully the HiVi K1 in the waveguide is able to cross low. I want to use it for my PA dipole crossing to the Selenium 10PW3. It just a fun little project but looking at the results for the 10PW3 I don't know if it's even worth using just as an experiment. Looking at the CSD and THD I don't even want to cross it higher than 1k Hz since it gets pretty ugly by 2k Hz. Meh, we'll see...

Very good stuff! I'm enjoying it. Is there any chance you can test the AC130 MK2's? I'd like to see if they test as good as they sound. A line array obviously influences the audible distortion but they strike me as a very well built and thought out driver design. The new drivers Aurum Cantus has listed on their website have the potential to be really good I think. Time will tell.

Thanks!

Jim

[augerpro]

I'll test them for sure. In the "newer" versions AC actually uses in their speakers the difference between the F1 and Mk2 is the use of copper and/or shorting ring in the Mk2. I'm hoping that's true of these of older models as well. Afterall teh F1 did very well in Zaph's testing without the shorting ring, so this Mk2 should be even better.

[Paul W]

...in case you guys didn't recognize the cat, his name is Lucky.

[PoorboyMike]

...in case you guys didn't recognize the cat, his name is Lucky.

That's just his last name. His first name is Extremely!

[augerpro]

Interesting how something can go from lewd to adorable with the addition of kitten :

Added some more drivers: HiVi M8N, Morel MDT22, Goldwood GW-4028S, MCM 55-1500, and MCM 55-1853.

[AJINFLA]

There is a cat in that picture?

[MuaDibb]

Quote: Gir-If you zoom in it looks just fine.


I think this says it.

[augerpro]

Added the Vifa DX25, P Audio WN8S, HiVi M4N and B4N.

Also I redid all measurements due to some noise caused by my mic stand. I decided revamp my methods so the results would be more consistent and accurate. Also the SPL plots are calibrated to reflect the actual sensitivity at 2.83v. The driver tests are only in the subfolders, everything else is my garbage holder. Please read the note in the "please read first" folder for more info.

In the mail: Visiton KE25SC, TB W4-1337, Peerless 830872, RS125S and RS270S :B

[augerpro]

I may add some nearer THD measurements of the woofers over the next couple days. It clears up the plots alot but there is a concern that high frequencies won't be accurate. I'll see how close I can get before the plots start changing.

Or maybe multitone test farfieldish and see if they jive with some nearfield sweeps...

[Preacher]

I didn't see the M4n on the website. I noticed you built an mtm using the m3n's is that right? How do these sound. Right now I have a set of Monitor 30's (polk) with a matching Polk center. I have 6 m4n's sitting here doing nothing so I'm debating on whether or not to build 2 fronts and a center out of the HIVi's (a few months) or sell them and jump for a bigger RS setup (it would take over a year to complete).

[augerpro]

Oops forgot to add the HiVi's.

The little HiVi's are nice little drivers. I love the system I built with them.


EDIT: M4N and B4N added

[Paul Ebert]

I look forward to your assessment of the KE25SC. I've really been curious about that one.

[augerpro]

I look forward to your assessment of the KE25SC. I've really been curious about that one.

You'll need someone more knowledgeable than me for an "assessment". I'm just a monkey with a mic :B

But thanks for your interest!

[augerpro]

Ok ladies and gents, I'd like the opinions of the 50 lb. brains here. The HiVi B4N and M4N are very similar and I'm kind of scratching my head over which is "better". Let me know which you think is better: running as mid in a conventional 3way, running in a 2 way but crossed to sub at say 100 Hz. I'm curious what you all find important.

Images not available

A few things have me wondering.
1)Looking at just the F3 on the THD plots, do you find the M4N's lower level in the lower midrange but a peak around 2k more or less desirable than the B4Ns smoother but higher in level in the midrange but no peak? If that made sense...

2)And how about the level of F5? It's higher over a larger area in the B4N than the M4N. But is this audible? Kind of like the Peerless 810921, the first few harmonics are very low, but there is noticeable noise higher in the spectrum. What's worse?

3)Would you prefer a driver with slightly higher distortion if it was very smooth over a driver with very chaotic distortion plots but never quite reached the level at any time of the other driver? Considering the limits of diy measuring this kind of made me wonder.

[JonMarsh]

My nod goes to the B4N; lower THD levels on 300 Hz tone, incl. fewer high order harmonics (I'm looking at the single tone plots), and a less screwy response profile above 3 kHz, probably will be a bit easier to manage the crossover.

Your question #3 is a tough one- I'd try to re-do the measurements under more favorable conditions to resolve matters more accurately.

Oh, and let's not forget listening- why not put them both in a simlar test enclosure, with passive BSC comp, and give them a listen? (I know, radical thought coming from an engineer like me...)

Seriously, when things are this close, I also look closely at the CSD plots- which may be limited in usefulness by measurement conditions, but that B4n still seems smoother up top- hope it's not just unit to unit variation.

[augerpro]

Thanks Jon! About question 3: I first thought about this comparing a RS150 and 832873 at Zaph's. The 832873 had slightly higher distortion in one area but the harmonics were all very smooth and even. The RS150 OTOH was like looking at mountain peaks, even though they never rised higher than the 832873, except I *think* around 2k it was a little higher.

I noticed this same thing comparing the B4N to the RS125 on Zaph's site. Again except for a peak around 2k this is clearly lower distortion than the B4N. But again the plot is just crazy compared to the relatively smooth distortion of the B4N.

Anyhoo I agree the B4N looks like the better driver. Actually looking at the THD, Z, and LeX plots I think the M4N has the marginally better motor (except for the high Q), but the cone just sucks. That breakup is too severe and too low.

Question 1 still stumps me though. Not just in this comparison but in general.

BTW the 830872, RS125, and W4-1337 have arrived. Should have some baffle cut soon and have some measurements up late tonight. Hopefully UPS will be swinging by today with that KE25SC. I'm going to try varying distances for the mic for THD also. I have the formula that *supposed* to determine the farfiled/nearfield transition for a given Hz and cone diameter. The 14" plots seemed to be a good compromise for woofers in that regard, but the plots just aren't that smooth. So I'm going to experiment and plot several distances adn people can pick which means more to them. Coupled with some multitone testing I think you should be able to get a fairly clear picture of what's going on.

[augerpro]

Results are up for the Peerless 830872, Dayton RS125S, TangBand W4-1337S, and Visiton KE25SC. I also added some more single tone distortion tests for the SS 9800, DX25, and 810921.

Also my gram scale is in so I''l be adding TS measurments to the woofers when I get a chance.

[Dennis H]

Wow, thanks for all the hard work, Brandon! :T

[JonMarsh]

Results are up for the Peerless 830872, Dayton RS125S, TangBand W4-1337S, and Visiton KE25SC. I also added some more single tone distortion tests for the SS 9800, DX25, and 810921.

Also my gram scale is in so I''l be adding TS measurments to the woofers when I get a chance.

Well, you're really going to town on this, I'm jealous of all the spare time you must have!! :W

And thanks for sharing your results!


~Jon

[Paul Ebert]

I've never seen a resonance like the 1.7k one on the KE25SC CSD. I wonder if it's caused by the built-in LCR? Any thoughts on what might cause this? Otherwise, the CSD looks pretty clean (to me). Distortion doesn't look all that good for the price, alas.

[augerpro]

I've never seen a resonance like the 1.7k one on the KE25SC CSD. I wonder if it's caused by the built-in LCR? Any thoughts on what might cause this? Otherwise, the CSD looks pretty clean (to me). Distortion doesn't look all that good for the price, alas.

The tall order harmonics are not good in that area. The singletone kind of shows that but when doing the sweep it really shows up. After 2K lower order looks pretty good but there is always some noise higher in th spectrum. I wonder if this is audible though? But the SPL is very smooth and the flange is dead flat, so maybe a good waveguide candidate, which could help with the lower end distortion. I also did some tests with the phase shield off and it was very similar performance.

[augerpro]

I would like to test these: http://www.madisound.com/catalog/pro...a71732079151a7

http://www.madisound.com/catalog/product_info.php?cPath=45_229_236&products_id=1389&osCsid=76202cd8124a3676efa71732079151a7

But I have to get rid of some drivers first so...Visiton KE25SC $90 shipped, HiVi M8N $40 shipped, TangBand W4-1337S $40 shipped, Selenium 10PW3 pair $50 shipped, and if there is any interest I may sell the drivers and XOs from BB project if someone wants to build just the enclosures. I also have a bunch of cheapies like the MCM and goldwood 4 iinchers, Dayton PT2B, and HiVi K1 that I could throw in box for someone who wants to check them out. I think the aluminum MCM is really a standout for the price. PM about any of these sales.

[augerpro]

Well, you're really going to town on this, I'm jealous of all the spare time you must have!! :W

And thanks for sharing your results!


~Jon

Heh, well much more important things ARE being pushed of so I could finish this. But I've found if I don't work on one thing to conclusion before moving on I'll never finish. If I start a bunch of things I'll finish nothing.

PS> any other kinds of tests you'd like to see? Any suggestions for my testing setup/method?

[augerpro]

No takers on any of these drivers? Ok, make an offer...

[augerpro]

Added RS270S

Images not availabe

Room/baffle modes show up in these lower frequencies so keep that in mind when looking at the THD sweep. Especially 110-140 Hz and just above 200 hz.

[dlneubec]

Hi Brandon,

I was looking over your tests and of special interest was your results on the RS150 (I'm using the RS150-4 in my HOSS project). When I look at it's performance for about 3 octaves from 200hz to 1600hz, it looks to be a clear winner when comparing cost, frequency response and distortion. I'm very pleased to see this. You don't happen to have a tone THD test at 1500hz for it do you?

Given the price, it's hard for me to see much reason to use any of the other similar sized drivers tested within this range. Then comparing the RS150-8 versus the RS150-4 on zaphs site, the 4 looks to be just is a little better in all aspects as well. I wonder why we don't see the RS150-4 used more?

[augerpro]

I think the 4 Ohm Dayton's are fairly new so no one has really used them? Then again using those the chances of dropping well below 4 ohm in the final XO are kinda high, so I suppose that keeps people away too. Would be good for an MTM wired in series like yours though.

I'm not so sure the RS150 is clearly better than the Nomex I tested. F2 and F4 are very similar, maybe a slight nod to the Nomex. F3 is *slightly* higher in the Nomex but F5 is significantly lower. You can really see the difference in the tones. Nomex is slightly higher in F3 here and there, but higher orders drop off to almost nothing. The RS150S has bit of higher order stuff going on. Higher orders are more offensive, OTOH what level do they have to be to be audible? That's a tough question that I'm researching. I even have a little program that can add different distortion profiles to a .wav file so you can listen to the difference. The Nomex also has a smoother response and would be much easier to work. For my newb skills that will contribute to my design being successful more than any other single factor I think. OTOH the Dayton motor appears more symmetric and it's half the price.

I have two bookshelf speakers I'm building with both of these to see what appeals to my ears. My end goal is to replace the 5.1 Onix Rocket setup I have now. Since I tested the RS125 and 830872 for center channel duties things become more complex. Here I think the little Nomex is clearly better than the RS125. So unless my ears tell something different with the bookshelves, the deciding factor between the very close RS150 and Nomex may be to match the center channel. Which would be the little Nomex at this point.

Oh ya, no 1500 Hz tone, sorry.

[dlneubec]

I think the 4 Ohm Dayton's are fairly new so no one has really used them? Then again using those the chances of dropping well below 4 ohm in the final XO are kinda high, so I suppose that keeps people away too. Would be good for an MTM wired in series like yours though.

Yes, I was thinking of MTM applications for the 4's, not MT's .

I'm not so sure the RS150 is clearly better than the Nomex I tested. F2 and F4 are very similar, maybe a slight nod to the Nomex. F3 is *slightly* higher in the Nomex but F5 is significantly lower. You can really see the difference in the tones. Nomex is slightly higher in F3 here and there, but higher orders drop off to almost nothing. The RS150S has bit of higher order stuff going on. Higher orders are more offensive, OTOH what level do they have to be to be audible? That's a tough question that I'm researching. I even have a little program that can add different distortion profiles to a .wav file so you can listen to the difference. The Nomex also has a smoother response and would be much easier to work. For newb skills that will help my design become successful more than any other single factor I think. OTOH the Dayton motor appears more symmetric and it's half the price.

It looks like your single tone THD figures are lower throughout for the RS150 as compared to the Nomex. I did say considering price, so at 2x the cost, I would expect the Nomex to provide more distance in its performance and the RS150. :B

[augerpro]

Considering the price the Dayton's are tough to beat.

[chrismercurio]

KE25SC - Tests

I have seen tests of this tweeter in two other places on the web.

http://www.hifi-selbstbau.de/text.php?id=111&s=read

http://www.audiocircle.com/gallery/displayimage.php?album=983&pos=7

I'm wondering if anyone can comment on the results AugerPro has vs the links above?

Thanks,

C

[Jed]

I have seen tests of this tweeter in two other places on the web.

http://www.hifi-selbstbau.de/text.php?id=111&s=read

http://www.audiocircle.com/gallery/displayimage.php?album=983&pos=7

I'm wondering if anyone can comment on the results AugerPro has vs the links above?

Thanks,

C

Well, in those links the distortion appears lower, especially for odd order harmonics. That's why it is important to test more than one tweeter, as the results appear dependent on which sample you get. That to me is also concerning, especially if one is under the assumption that a tweeter is very low distortion and it ends up being just an average performer in some areas. I like how Zaph gives a consistency rating as well, something we don't see all that often with the magazine tests.

[Florian]

According to their webpage, Hifi-Selbstbau test at least two drivers. The pictures in the second link are scans from the Hobby Hifi magazine, but I don't know if they measure more than one driver.
The Hobby Hifi test was measured at 85dB so it's not really possible to compare it to the other two tests. In the Hifi-Selbstbau distortion measurement, K3 is at the same level as in augerpro's test, K2 is even higher for low frequencies.
In general, I think the agreement of the measurements is quite good considering the different measurement conditions and software.

[augerpro]

Well there's not much I can add to this discussion. I agree the results are somewhat close and sometimes not so much. I tend to beleive my Z plot since this does look like there is an integrated LCR at work. I don't see that in the other plot. The Hifi mag response and mine are pretty close. So is the other one really although the response seems to drop off, which I don't think is the case with a harddome tweeter. That said I do have a peak that always happens just before cutoff. It's not real. I haven't done anything about since it's over 20k and until I get a higher sampling soundcard to extend measurments up to 48k I won't do anything about. Just being able to sample up to 48k may fix it anyway.

In the THD the scale is % HD,not a dB level, so I don't know how it compares really. If you compare it to my %THD (green line at the bottom of the graph) it's pretty close though. And really F2-F4 drop off to very respectable levels after 2k. But there is always some higher order noise as you can see in the single tone plots. I don't know how audible these are though. It's on my list of things to experiment with. Under 2k I think the F5 is accurate, the spectrum has alot of noise in this area. I took off the grill thinking it or the phase shield was buzzing. But plots are nearly identical. The good thing is the faceplate is flat so it's easy to use a waveguide. Which is good since with a waveguide and crossed highish it's probably a really a good tweeter. The SPL response is very nice.

Anyway I'll keep improving my method but at this point I'm as confident of my measurments as I am of a some random magazine. And when it comes to THD and CSD different methods and environment can make a siginificant difference. In fact I can't believe my measuremts agree as well they do with Zaph's, and from the couple of drivers I've compared, MarkK. But none of these should be seen as directly comparable. You must take a "preponderence of the evidence" approach when comparing things I think.

[Jed]

Another note about the audiocircle link, which is a scan I uploaded that I'll be deleting soon (copyrights etc), is that it is a test from mid 2005 Klang and Ton magazine. The Ke25 is now made in Taiwan, not that that makes a difference, but I'm not sure how tight the tolerances are etc versus when they were made in Germany. Only a handful of the Visaton drivers are still made in Germany now. Used 2.5K and above, I bet this little Ke25 is VERY good, which the graphs suggest from all these sources.

[chrismercurio]

I hope it is.

I bought AugerPro's and need to pick up an additional to make a pair.

I had lunch with Ken Kantor of Tymphany a few months back and he is more than confident about the moving of production from Europe to China. He gave several reasons that I need not list and ultimately feels that a superior product can be built for far less money.

C

[augerpro]

If you're at all interested I'd like to measure another sample. Let me know.

[Jed]

I decided to see if I could translate that article on the KE25 using google translator. Here it is:

"Our opinion:
The outside impression:
The KE25SC is Understatement purely: the empfindlche ceramic(s) diaphragm is hidden behind a screen, which is sturdy metal front plate very dezent painted. On the back it shows an unusual appearance: the magnet is not to be recognized, but it is shapeless like an embossment whale! The mystery solution is an integrated Entzerrungsnetzwerk that under a metal cover hides itself. Thus the KE25SC is also magnetically protected.

The TSP:
The KE25SC comes in series with a Entzerrungsnetzwerk that the relatively high resonant frequency of approx. 1250 cycles per second nice-proves so well “smoothly” that the production of a passive frequency switch functions also with small separation frequency after text book. Unfortunately the Impedanzentzerrung does not fold completely perfectly, but remaining impedance “Spitzchen” is harmless. Who wants can the circuit also simply “deactivate”, as the led out cables are simply entlötet. Of it however use should be made only with employment into active boxes or highly separated 3-Wege-Boxen, or if one can be provided individually adapted Entzerrung. That is not to be managed however without measuring technique, in addition one should the Hochtöner then only a bringing in time gönnen after the slogan: if already, because already!

The frequency response:
Starting from 1,4 kHz the KE25SC runs as pulled with the ruler. Also outside of the axle the frequency response drops only small and evenly, which is positively noticeable also when listening to the angledependent sound pressure. But a diffuser is responsible in form of a transparent foil ring beside the rigid diaphragm material, which is appropriate within the screen. The step response does not show high frequency resonances. The chassis swings on the mechanical periodic resonance after. The effect degree is sufficient with 90.7 dB/2.83V/m considering the small DC voltage resistance of only 5,26 ohms rather small, in the very most cases however perfectly.

The distortion factor:
The distortion factor essentially consists of the “harmless” K2 and the “inharmonious” K3: K4 to K8 are above 1 kHz with 95 railways < 0,035%! The distortion factor k2 rises below 4000 cycles per second, k3 only below 1250 cycles per second. Even with 95 railways k3 is < 0,15% there! K2 exceeds the 1%-Marke with 2 kHz and 95 railways only scarcely and is inaudible thereby how the following MP3-Datei beweisst. There first 1 second comes. long 2 kHz without distortion factor, then 1 second. with 1% k2 and then additionally 0,1% k3. Occasionally one fades out always briefly, so that one white which one straight hears:


MP3-Datei (23 KB)



--------------------------------------------------------------------------------

HiFi Selbstbau result:


--------------------------------------------------------------------------------

The VISA CLAY/TONE KE25SC looks inconspicuous, has it however fist-thickly behind the covering grid plate: also with high volumes and hard impulses compressibility effects hardly show up. By the broad radiation a stronger meaning comes to the damping of the hearing area: in “cool” furnished designer dwellings with tile and/or Parkettboden and many glass areas the KE25SC will sound simply only brittle and hard. With a reverberation time which is because of the lower edge of the ideal range it might accumulate to the maximum form. In the halation in gel churches it gave anyhow to the activated Trio a very good conception.



The KE25SC is not due to the high price of 125 €/Stück a straight chassis which one lies down times in such a way. If however its qualities (linear frequency response, deeply usable, small rattling/clinking actuator) are needed are it each euro worth! All in all a wide-band applicable Hochtöner for the more impulsive moments of the life!"

[TacoD]

The translation is harder to read than the German article . You can try babelfish, maybe you get a better translation.

[Jed]

The translation is harder to read than the German article . .

:lol: Yeah, I get the idea though. I'm going to spend some time comparing Jon's measurements of the C13 to the KE25. Looking primarily for <3k, and now that Meniscus said no-go for the C13s, I'm looking around again. Maybe its a sign to go SEAS Magnum or Ke25?