Announcement

Collapse
No announcement yet.

TS Parameter changes during break-in

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • DeZZar
    commented on 's reply
    Oh, sorry to disappoint. I was intending for a sub design - but a fairly different one. Going to be running with two of these in a tower design. A little way off as I've other projects to finish first.

  • Colonel7
    commented on 's reply
    So what are you thinking as far as design? Sub, 2- or 3-way, or something else? Very interested to see a non-sub design

  • Steve Lee
    replied
    Ooooh!

    I'll be watching this project closely - Thanks for doing this.


    Leave a comment:


  • DeZZar
    replied
    Thought I would add in here the Epique E180HE-44. I picked up a couple of these during a recent visit to the US for a future project. I thought I would test out how they measure. I haven't completed a full break in at this stage, and may not take these all the way, as agreement with factory is pretty good after just under an hour running @ 30hz.

    No factory spec to compare with for wiring the voice coils for 4ohm, but 8ohm comparison looking very good.

    Click image for larger version

Name:	E180 Break In.png
Views:	79
Size:	188.8 KB
ID:	1479845

    Based on the measured results a single driver wired for 8ohms will achieve -3db @ 28hz in 30L.

    Factory spec would achieve the same -3db @ 28Hz in 25L by comparison. I'll break these in a little more over the next couple of weeks and see where they land.

    And boy can these little things move. I know they aren't the first small high excursion woofer around or even unique but I still love seeing such a small driver move this much!
    Click image for larger version

Name:	5xynec.gif
Views:	113
Size:	155.3 KB
ID:	1479844

    Leave a comment:


  • DeZZar
    replied
    7 Hours on the SB tweeter now. Because its in the office here with me I had them (two facing each other out of phase) in a small cooler bag, inside a larger cooler, stuffed with a blanket, then the cooler covered in two blankets and by the time I had that it just sounded like plumbing/the shower running in the house - so I could live with it for the day!

    Click image for larger version

Name:	SBMeasurementsComparison.png
Views:	152
Size:	24.7 KB
ID:	1474601

    Interestingly here the Q values are all getting further away from spec. With the Dayton woofer the opposite trend is true - the values get closer to spec. Not sure what to make of it at this point - just reporting what I see.

    Here is the updated frequency response. Because I made the mistake of not setting the voltage with the first measurement I have normalized these measurements to the same relative level @2K - so the changes you see in response are all relative to the level being aligned at 2K.
    Click image for larger version

Name:	SBFrequencyComparison.png
Views:	130
Size:	50.5 KB
ID:	1474602

    Definitely some smoothing apparent in the lower octaves now although some new peaks seem to be creeping in around 2.4K and that 13K region.

    I've taken these measurements a few inches from the tweeter which upon reflection I'm not sure why I did that and I will probably try this again with another set of tweeters measuring right on the driver (near-er field? lol). Might not make much of a difference.

    Distortion.....well I'm no expert in reading distortions. After the initial increase in distortion things have gotten better in some areas and stayed basically the same in others. My noise floor here where I live isn't great so I suspect it might be impacting these measurements more than it should.

    What I notice is that although THD has improved, the 3rd, 4th, 5th harmonics have actually increased quite a lot. Taking 2K as an example. At 1 hour break in the 3rd order was reading 0.05%. After 7 hours it is now reading 0.149% and I notice it is similar for 4th, 5th etc. They've all increased a gotten a lot closer together but the THD at 2K has improved significantly going from 1.48 to 0.96.

    Click image for larger version

Name:	SBDistortion.png
Views:	132
Size:	80.5 KB
ID:	1474603

    Like I said I'm likely dealing with a bad noise floor and I don't think I've taken these measurements close enough so I might wrap this one up here and try again shortly with another set of new tweeters I have on hand.

    I'm also about to run the process with an SB Satori MW16P and an SB23NBACS45 (aluminum 8" woofer). I'll report on these as soon as I can.

    Leave a comment:


  • DeZZar
    commented on 's reply
    My pleasure. Happy to share.

  • Chris Roemer
    replied
    Thanks for all your hard work!

    Leave a comment:


  • DeZZar
    replied
    Working with the SB tweeter now. A little more difficult as its so annoying to leave running in the house for such a long time. I 3D printed this little jig to hold the tweeter for measurements and to hold two of them facing each other so you can easily play them out of phase. Amazing how much the volume is cut down by doing this but it doesn't go completely silent. Still need to stick them in a box under a blanket lol. Tomorrow I'm going to try putting them in a cooler.

    Click image for larger version  Name:	IMG20210809174522.jpg Views:	0 Size:	464.4 KB ID:	1474538 Click image for larger version  Name:	IMG20210809174509.jpg Views:	0 Size:	397.8 KB ID:	1474539

    Here are the initial measurements compared with 1 hour of pink noise.
    Click image for larger version  Name:	SBMeasurementsComparison.png Views:	0 Size:	20.2 KB ID:	1474540

    The Q's are all way off at this stage but Re and Fs are pretty well bang on.

    I decided to also see what changes in frequency response and distortion could be seen during break in.

    Here is the frequency, straight out the box and after 1 hour of pink noise.

    Completely ignore the absolute SPL values - its not calibrated. Unfortunately I also forgot to take the initial measurement at 2.83v so this is really just showing relative change at this input level. (2.83 would just make it a little easier for consistent repeatability)

    Click image for larger version  Name:	SBFrequencyComparison.png Views:	0 Size:	49.6 KB ID:	1474541
    I don't think much of a conclusion could be drawn from this yet - I want to say that it appears some areas have smoothed a little - but realistically for everything that's smoother something else is lumpier - so I would just say that response does change.

    I thought a better indicator might be distortion but after one hour of break-in, interestingly distortion has actually increased across the spectrum. Not by much - you can see the percentages recorded in the data above for 1K, 2K, 6K and 10K (arbitrarily chosen points). 6K and 10K saw a respective increase in distortion of...errr....6 and 10% (total co-incidence). Here is the 1 hour distortion results.

    Click image for larger version  Name:	SBDistortion.png Views:	0 Size:	77.5 KB ID:	1474542

    The distortion results surprise me. I expected these to immediately start heading in the opposite direction. I attempted to keep everything in the measurements as consistent as possible (mic distance, levels, mic gain, amp gain, etc etc).

    Tomorrow I'll let these run for an additional 6 hours and retest and see where we land. Will post an update as soon as I can.

    Leave a comment:


  • fpitas
    replied
    Probably if they were in a sturdy cardboard box that would absorb most of the high frequencies.

    Leave a comment:


  • Steve Lee
    replied
    Put them in your car wired as you suggest and put the battery charger on the battery with the ignition in ACC position?

    Leave a comment:


  • DeZZar
    replied
    What's the consensus on tweeter break-in? I'm assuming Pink noise with a high pass? Any other methods?

    I was going to try two tweeters facing each other out of phase in an attempt to reduce the sonic impost of running a tweeter for a prolonged amount of time within earshot. Unfortunately I don't have a shed out the back where I can just leave them running for a while.

    I'll 3D print up a little jig to hold them first.

    Leave a comment:


  • Steve Lee
    replied
    As one can see from DeZZar's excellent work, the compliance of a driver changes based upon its use (Flexibility). Use it frequently or lose it.

    Kind of like the human body as we age . . .

    DeZZar - thanks for posting this stuff and please keep us updated - it is truly educational and fun to see.




    Leave a comment:


  • billfitzmaurice
    replied
    When the voice coil is warm other issues crop up, most notably a rise in DCR. Where home use is concerned it usually doesn't amount to much at normal listening levels. In pro-sound use at high levels the rise in DCR can result in power compression, which can reduce output by 6dB or more.

    The break in source doesn't matter much, but I always use a sine wave well below Fs, for maximum excursion with minimal power, and if the driver isn't mounted it's also pretty much inaudible.

    Break in of speakers by just using them may take up to 40 hours, depending on the material and level. It does not take the hundreds of hours that some manufacturers claim. The main reason for that bit of nonsense is that over time you do get used to the sound of a speaker, so one that sounds bad initially may not seem so bad after a few months. It also may encourage one to keep the speakers beyond the free return window, if there is one.

    By all means don't finalize a design before breaking in and measuring the drivers.

    A spec variance of +/- 10% is considered normal. After break in +/- 5% will have little effect on the result. As I mentioned Eminence does periodic spec checks. They can do that because they build their drivers in house. With sellers of sourced goods, like Dayton, it's not so easy to do. Even if they measured some from the initial batch 3rd party suppliers have been known to substitute everything from cones to voice coils without notice, hoping that no one else will notice.

    Leave a comment:


  • DeZZar
    replied
    28 Hours in now. The final measurement here is with the driver going completely stone cold overnight. The others I had taken after a couple of hours cool down. That might explain some of the creep back up in Fs but its not significant. Re is consistent in any case so voice coil heat seems to be about the same - perhaps just compliance related.
    Click image for larger version  Name:	Measurements Comparison.png Views:	0 Size:	139.8 KB ID:	1474072
    After nearly 30 hours break in I cant see that things will change much from here. They will always continue to change throughout the life of the speaker but for design and modelling purposes I think one can draw the line here.

    Click image for larger version  Name:	FsVasMmsF3.png Views:	0 Size:	190.8 KB ID:	1474070
    Click image for larger version  Name:	Qs.png Views:	0 Size:	191.2 KB ID:	1474071


    I don't know what most manufacturers or Dayton in particular would consider acceptable in terms of variance from specification but as you can see a number of these params are well outside of even a 10% variance.

    In terms of my design it does not make much of a difference. The designed application for these will hit 34Hz f3 in 27L and handle up to 30watts input for content down to 32Hz. If my VAS measurement is unreliable for any reason and it is in fact closer to spec, we'll hit 37Hz, same design but with a little less power handling.

    It's been interesting to document this and I might try it with a few other new woofers I have including some SB Satori's.

    My final thoughts/ponderings:
    1) Should the measurements really be taken with the speaker warm? After all once its in its box playing music it'll be warm....not stone cold?
    2) I cant imagine the type of signal used for break-in would really matter - its current and mechanical movement that achieve break in - not frequency?
    3) The equivalent 'break in' for a consumer set of speakers from a store must take....months?...of regular use...
    4) Should be obvious to most - but definitely don't build your custom designed speakers until you get the drivers, break them in and measure them. In this case if you had just used the manufacturer specs it wouldn't have been a disaster but it wouldn't be optimal. Fortunately I've worked with this driver before and have some already broken-in units on hand so I knew what to expect.
    5) What sort of variance should one expect in advertised specifications? Ironically this is Dayton's own woofers measured with Dayton's own measurement devise...
    Last edited by DeZZar; 08-10-2021, 12:49 AM.

    Leave a comment:


  • billfitzmaurice
    replied
    I'm an AES member, very familiar with AES2-2012. It deals with measuring, not breaking in. Nothing in it says that the specs cannot change after two hours, they just use it as a reasonable period of time to work the driver before taking a measurement. In truth drivers continue to break in almost indefinitely. One of the first drivers that I measured in 1983 was an EVM-12L. Out of the box the Fs was 60Hz. After an overnight break in it came down to the advertised 53Hz. I used that driver in a number of cabs over the years, measuring it again every time I put it in a new box. Every time the Fs had dropped a little more. The last time I did so, around 2001, Fs was down to 46Hz. Bass players are fans of old drivers, because they sound better with age. But that drop in Fs also results in a drop in high frequency extension, so guitar players tend to favor newer drivers, unless they don't care for a bright tone.

    Leave a comment:

Working...
X