DiyMobileAudio.com Car Stereo Forum banner

1 - 20 of 21 Posts

·
Moderator
Joined
·
18,605 Posts
Discussion Starter #1
Driver:
Dyn Audio Esotar2 430 Midrange:
Dynaudio - Authentic Fidelity

This is a test of a 2nd sample of the Dyn 430 midrange. Jerry allowed me to borrow his so I could do further testing with the Klippel mic and new modules. Since I was testing a completely different driver than before, I also tested it with LPM & LSI.
Thanks to Jerry for sending me his driver to test.

This is the first driver tested:
http://www.diymobileaudio.com/forum/klippel-reviews-driver-specs/115960-dyn-esotar2-430-midrange-klippel-testing.html

Picture of the driver from the previous test (it's the same dang driver):



Notes:
  • I've attached reports in pdf format. I have also provided pictures of the more popular type data.
  • FR Testing at 0, 30, and 60 degrees as well as the accompanying HD data in this graph is done at 1m. You can see an approximate sensitivity at 2.83v/1m of 90.5dB.
  • The more strenuous testing was done at 1/2m, 6.5vrms, which gives approximately 101.5dB output at this distance. Note, HD distortion is given as % of fundamental in a standalone. I like this. Makes it a bit easier to decipher.
  • IMD Voice sweep was done at 1/2m at the various voltages presented in the results.

  • Windowing was approximately 9.5ms which provides useful data to just below 200hz.

Let's go...
 

Attachments

·
Moderator
Joined
·
18,605 Posts
Discussion Starter #3 (Edited)
FR @ 0, 30, and 60 Degrees vs HD at 0deg. Taken at 2.83v/1m:





Zoomed In:



FR/HD Taken at ~102dB @ 1/2m:






Multitone Distorion Taken at 10v input, 1/2m distance:





IMD Voice Sweep using described input votlages:
 

·
Moderator
Joined
·
18,605 Posts
Discussion Starter #4
IMD Test, Voice Sweep (~630-8000hz) with fixed bass at 93hz:

Note: PLEASE read the title of the graphs. It tells you what the test is for!
I've tried to put the IMD results together, and then the HD results together for better comparison. If you read the titles, you won't be confused. ;)

















 

·
Premium Member
Joined
·
6,214 Posts
Now I need to really look at the Klippel links you posted in order to fully grasp those plots.

Thanks for the hard work Erin and all the best for the new year...

Kelvin
 

·
Moderator
Joined
·
18,605 Posts
Discussion Starter #9 (Edited)
Alright, slackers, let's talk about this data so I can feel I didn't waste days of my life doing this stuff. ;)


T/S Parameters (highlights; see PDF for full table):

Re = 4.33ohm
Fs = 95hz
Le = 0.137mH
Bl = 5.315 N/A
Qts = 0.289

Note the zoomed in plot for impedance shows a ripple at 2khz. This is likely due to resonance in the basket makeup or spider. I'll have to break out some calipers to do some measuring to know for sure.

Xmax and LSI:
When you compare this test vs the other dyn 430 test you'll see this one has much less Bl offset (while having the same slight assymetry) ultimately contributing to a higher xmax in this regard of about 3.5mm vs the 2.6mm tested last time.
For the suspension limitation, the driver in this particular test has an xmax of 3.9mm vs. the previous tests' 3.1mm. Note, however, this particular driver has a much more severe assymetry in suspension. I haven't yet looked at this driver in this regard but with a nearly full millimeter shift, I imagine it's visible to the naked eye.

Overall, you're Bl limited in xmax at 3.5mm; nearly 1mm higher than with the previous driver and it's seemingly all due to the fact that the coil is more centered in the gap, judging by the offset results.

The one thing I was surprised (saddened, shocked, call it what you will) is the fact there is not a shorting ring used in this driver as evidenced in the assymetry of the Le(x) graph. Given the driver's intended use a a midrange, coupled with the knowledge that Le distortion is very offensive and namely so in the upper midrange, I would have thought it would have been used. Furthermore, at the retail cost of $1600/pair, I don't see how a $0.90 shorting ring would have affected costs. More discussion on this in the IMD distortion section.
*Folks at Dyn, feel free to contact me if you wish to share some information on why you chose not to use this component in your design.*

Make sure to use the LPM data along with the above xmax values to model this driver for low end response.


Frequency Response:
Measured SPL with 2.83v input at 1m is approximately 90.5dB.

Good usability in all ranges to about 2700hz, which correlates with beaming (Speed of sound/effective driver diameter/2). In the range from 400-2700hz the FR is nearly completely flat - less than 0.5dB delta above 1khz - with more abrupt changes below this value which I contribute more to measurement resolution than the driver itself (ie: data unsmoothed, window is shorter on the zoomed in results). If you want to use this driver higher you'll want to stay on axis, where you have decent response out to about 6.5khz before severe roll off with nearly flat FR to this point sans a dip of about 1.25dB around 4khz. This actually may not be a bad thing depending on your take of the fletcher munson curve.

Overall, IMO, the frequency response of this driver is very, very good given it's intended application as a midrange and not a widebander. Outside of this, and especially in applications where you go off-aix, it's still a fair driver. Use the data provided to draw your own conclusions for your desired install and frequency bandwidth.


Distortion:

Note:
As I've said in previous posts, the room modes contribute to high levels of distortion below my gating frequency (approx 150-200hz, depending on treatment used). Below this, my room takes over and colors the distortion results, so I had to gate out the low end response. However, T/S parameters and xmax measured by LSI are dominant here anyway: use these to model the low end limitations in a program such as WinISD.


Harmonic Distortion of this driver looks good. What's good? At the 102dB level, which was taken at 1/2m - correlating as 96dB at 1m - the THD is under 2% down to 200hz and under 1% above 400hz. However, one thing I find interesting is the rising 3rd order response above 2khz. I'll keep this stored away in my brain for a later time.


Intermodulated Distortion is new to me (us?) in my measurement suite. Klippel was kind enough to lend us this module to do further analysis so I took advantage of it. If you want to read more about it, please visit their website:
klippel.de

On their site you'll find a lot of discussion about IMD testing. Namely, to me, is the way that it brings to light issues standard HD testing does not, such as inductance. In short, inductance rears its head on the higher end of the spectrum and is more readily visible in IMD testing of 2nd order components. Drivers without shorting rings are stated to have a rising 2nd order distortion between (7-20)*fs. (Please view this chart for further information).
Chart Linked said:
Le(x) symptoms:
moderate HD in sound pressure and current for 1.5fs < f < 4fs
high IMD in sound pressure and current (f1 < fs, f2 > 7fs)
You'll recall the mention of this driver not using a shorting ring and you can see in the IMD results this driver clearly exhibits a rising 2nd order IMD trend. This particular test used a fix bass tone at fs with a swept voice tone between approximately 600-8khz. I chose this broad range for the voice sweep as to show the rising 2nd order distortion; otherwise it wouldn't have been as evident. Below 1khz you have an IMD percentage under 10%. However, above this frequency of 1khz out to 8khz, you have a rise to 30% distortion for 2nd order IMD.

Again, note how compare this 2nd order component vs that of the 3rd order component which is much more linear relative to the 2nd order component. This, to me, goes further to agree with what Klippel says about rising 2nd order intermodulated distortion due to inductance.

One thing I find interesting is the correlation (or lack of) between HD and IMD values above the 2khz mark. HD 2nd order drops while HD 3rd order rises. However, in the IMD test, 2nd order rises while 3rd order remains fairly neutral.

The fun stuff, to me, is our ability to now see how increased power (input voltage) affects the distortion values. For example, let's look at IMD 2nd order:


Note that as you increase the voltage you get increased distortion. Duh, right? Well, it's certainly nice to see it all overlayed or simply only having one graph at a given SPL. Now see how, with increasing frequency, you have a larger delta between distortion vs input voltage? For example, at 800hz you have about a 1-2% increase in distortion from 4.33v to 6.00v. Then, at 2khz you have a delta of about 4% (eyeballing it) from 4.33v to 6.00v input. Now at 7khz you have nearly 8%!


Moving down a bit, while still under the IMD results, look at the HD results graphs (the ones with stepped voltage). Look how they resemble the FR/HD test results with some variance (due to amount of points taken in test) but also look at how they change as the input voltage is increased like the IMD example discussed above. This gives you insight in to the driver design and potential trouble areas Cool stuff!

Compression:
The compression data is given to you to understand just how power compression affects the output of a speaker and to show that it's not something to be taken lightly. The data is given in input voltage, but to make it a bit simpler, let's talk in terms of power (since power is a function of voltage/resistance). We know that doubling of power increases SPL by 3dB. Ideally, this means that as you double your power to the speaker it always increases by 3dB. However, due to thermal issues (heating of the coil) and nonlinearities (which is what the LSI focuses on) this just doesn't happen in high power situations. So, instead, as you increase the power your output doesn't increase in this linear fashion and, instead, your output may only increase a 1dB at very high power (totally hypothetical number; just go with me). This is exactly what compression is. So, how do we determine what is "high power" enough to cause compression? Well, that's where we use the new stuff that klippel has given us.

The way the data in this particular instance is provided, as I understand it, is simply a measurement of how much compression you receive at a given frequency. Everything you see is compared against the baseline (first voltage; in this case 1v). If all the lines are exactly the same, then the system increased in a logarithmic fashion as it should at a given input voltage. When the lines start to separate is where you lose output by compression. Severe cases of this is typically seen at low and high ends of the frequency spectrum and with higher input voltages. With this particular driver, rather than increasing in some logarithmic fashion, you lose about 2dB at the 6v input from the baseline 1v input from about 1.5khz to 8khz.

I know you're wondering, then, WTF is with the discrepancy at ~ 1.3khz. Me, too. I thought it could be some sort of resonance, because at fs you don't have the effect of compression like you do away from resonance. But, honestly, that's all I've got to go on right now. I'll have to do some more digging (and testing) when I get some time.

You'll note that I have not provided low end data for this particular driver. That was because with IMD, I'm not able to test below the gating frequency reliably. However, the compression test is not affected by the room modes so I'll make it a point to go back and provide full bandwidth compression data and at higher input voltage so you can see the severity of power compression.

I encourage you to read up on this here:
Compression of fundamental components
I've tried to give my "in a nutshell" impression.


Summary:
Usable bandwidth as a dedicated midrange is pretty much what I would expect, however I am really impressed with the linearity of the driver itself within its pistonic range; no major FR anomalies to speak of.

In terms of non-linear results, I think it's pretty darn good. Overall HD is down really low -below 2% down to 200hz - at 96dB @ 1m.
The suspension assymetry is pretty bad, though. Though, assymetry lends itself to 2nd order distortion. This could have been a design tweak done on purpose to result in a better sound (based on the theory that 2nd order distortion drives our sense of pleasure in the resultant sound) or simply not. I doubt it was, though, simply because Kms is only applicable to low end (near Fs) response. So, is suspension assymetry bad in this case, then? I don't know the answer. My hunch is no because Bl is still the limiting factor for Xmax.
A shorting ring would certainly have helped limit some high end issues that make themselves readily apparent in the IMD results and (maybe?) lent itself to a more usable bandwidth as well as less compression.

Overall, a good looking driver.

What would really be great is if I could do a full test like this again on the same driver with a shorting ring...




That's it for me... at least for now. I hope that helps you all to understand the data a little more and helps you to draw your own conclusions about tests done in the future. I plan to provide this kind of analysis for all future tests but I ask you to provide your own as well. As a community we learn together. I'm still working to understand the data fully.

To anyone in the industry who's looking, if you want to provide advice or insight in to my test and/or driver design please feel free. This is all about learning and progressing.

- Erin
 

·
Registered
Joined
·
5,234 Posts
Wow Erin, thanks for the summary. While I've been doing some reading lately, the summary helps a lot to understand the numbers.

I have always assumed distortion would go up the lower in frequency you went but it looks like I was wrong. I run these mids from 1.6khz to 8khz. They're almost on axis and it sounds really good like this. Maybe I'm not doing them justice by not running them lower and by running them too high. I think what I have is specific to my car and the kick panel location.

Are you pretty backed up with speakers to test? I would love to see the E650 tested for my own selfish reasons.
 

·
Moderator
Joined
·
18,605 Posts
Discussion Starter #11
I certainly wouldn't run them past 4khz simply due to their directivity. You might, however, be noticing a gain in usefulness beyond that point for the same reason (directivity = reduced reflections). I'm a fan of wide polars but the car makes it a PITA.

I'm always down for another test. ;)
 

·
Registered
Joined
·
5,234 Posts
I certainly wouldn't run them past 4khz simply due to their directivity. You might, however, be noticing a gain in usefulness beyond that point for the same reason (directivity = reduced reflections). I'm a fan of wide polars but the car makes it a PITA.

I'm always down for another test. ;)
Thanks.

I'll get with you in PM when I'm ready to pull the 650s out. I wish someone could contribute an esotec 162 to compare what I percieved as much better midrange from the 650 and more pop to the low end as well. Going purely off of the ts parameters they just don't seem that much different other than xmax.

Plus I like doing weird stuff like bringing my subs up to 200hz and having the 430s take over from there while the 650s are gone just to see how it sounds.
 

·
Registered
Joined
·
4,982 Posts
Nice review! Don't like the rising 3rd order HD above 2,5kHz, quite a bit higher than the 2nd order HD too. Doesn't fully understand the IMD plots but they seem to look pretty good. At least up to 5kHz. The drivers excel in the 250-450Hz area with extremly low HD but I wouldn't run them past 2kHz. Perhaps a 200Hz//2,5Khz 4th order HP//LP. Good drivers but in my opinion the performance don't justify the price even so.

Hard to compare to Zaph's plots since he use a fixed [email protected]/2metre for all drivers but it seems like the Dyn's outperform my Fountek FR88EX below 500Hz, but above Founteks show lower HD, especially after 2kHz. Listened to the Dyn's before, sounded really good but so does the Fountek's... ;)
 

·
Premium Member
Joined
·
6,214 Posts
Crap, I guess I better get out there right now and change my 8khz/12db low-pass.
Don't take this the wrong way, we are just discussing about specs and plots. He clearly stated that HE wouldn't use it above 2kHz and explained why - doesn't mean that you have to do the same thing ;)

Kelvin
 

·
Registered
Joined
·
5,234 Posts
Don't take this the wrong way, we are just discussing about specs and plots. He clearly stated that HE wouldn't use it above 2kHz and explained why - doesn't mean that you have to do the same thing ;)

Kelvin
Of course not. I was being a little sarcastic. If I only had plans to play it up to 2khz I would just run a good 2-way.
 

·
Premium Member
Joined
·
157 Posts
FWIW, if anyone wants to thank me, send some flowers to my wife. Her name is Lindsey. ;)
Moderator bikinpunk -

A Big ole Danke Schon,,,, for the Dynaudio 430 midrange test data effort.

PS. please privately email your (Lindsey) address where flowers can be delivered. The forum's treat!

Cheers, Mike
 

·
Registered
Joined
·
3,684 Posts
Erin,

I want to Thank You for treating my driver well and taking the time to do this test. It seems like we are learning more and more every day and you have gone to the next level with all this.

From reading the data (best I could) and reading your interpretation, it seems like the results back up my subjective midrange test conclusion about these drivers. It will be very interesting to see how they stack up once more midrange drivers are put through this rigorous test.

Again, thanks for all your efforts (and thank your wife as well) and let me know if you need anything else to test. If I have it, I'll send it. (When BuickGN gets the 650's out, I will send a MW162 to compare against)
 

·
Registered
Joined
·
5,234 Posts
Erin,

I want to Thank You for treating my driver well and taking the time to do this test. It seems like we are learning more and more every day and you have gone to the next level with all this.

From reading the data (best I could) and reading your interpretation, it seems like the results back up my subjective midrange test conclusion about these drivers. It will be very interesting to see how they stack up once more midrange drivers are put through this rigorous test.

Again, thanks for all your efforts (and thank your wife as well) and let me know if you need anything else to test. If I have it, I'll send it. (When BuickGN gets the 650's out, I will send a MW162 to compare against)
That would be awesome. With similar TS parameters, I would love to know why/how I can hear a difference in the two. Maybe I should lower the highpass to something a little lower like 50hz to make sure they're broken in well. I really haven't had a whole lot of run time with them.

I need to order those baffles very soon. Ultimately it will be up to Erin but how does early/mid February work for you?
 
1 - 20 of 21 Posts
Top