Ultimax UM15-22 break in

[pick]

I recently received my Dayton Ultimax UM15-22 15-inch subwoofer. I connected it to a signal generator and amplifier. I set the generator to the expected resonance frequency of 19.5Hz. I turned the amplifier up until I saw an estimated 1.25 inches of peak to peak cone travel (Xmax should be about 1.5 inch). I let it run this way for 24 hours before testing the free-air characteristics of the woofer. I am measuring with a Dayton WT3 that I have used successfully to test and design several other systems. The results I see are unexpected. I measured DC resistance (Re) of 3.5-ohm with coils wired in series. This seems correct. Free-air resonance is 27.8Hz instead of the specified 19.5Hz. Qts measures 0.67 instead of the specified 0.47. I reconnected the subwoofer to the amplifier and re-started the break-in process. Should I expect the break-in to take this long or should I be concerned about this driver?

[billfitzmaurice]

Whatever break in that will occur has occurred by now. Be sure you actually measure in free air, placed on a table or bench will give inaccurate results.

[dcibel]

Measure VAS and model it in an enclosure. If the model is roughly the same as what you get using manufacturer specs, then there's nothing to worry about even though your measurements differ.

[Chris Roemer]

Yup. If Q and Fs are higher, but Vas is lower, your box model probably won't be much different.
Measure Vas and run the sims, I think you'll like it.

[pick]

Whatever break in that will occur has occurred by now. Be sure you actually measure in free air, placed on a table or bench will give inaccurate results.

It's actually on the floor with the magnet sitting on two 2x4s to keep the vent unobstructed. This is the way I usually do it and I've never seen this kind of error. I'll do it again with the woofer suspended but I don't expect drastically different results. Sent from my SAMSUNG-SGH-I337 using Tapatalk

[pick]

OK I'll check Vas. I'm just concerned that something is wrong based on the magnitude of the error. I've seen around 10% error in Fs before but this is more like 45%. That seems like it should be well outside of manufacturing tolerance.

Measure VAS and model it in an enclosure. If the model is roughly the same as what you get using manufacturer specs, then there's nothing to worry about even though your measurements differ.

Sent from my SAMSUNG-SGH-I337 using Tapatalk

[pick]

Also, Qts is really high. We are getting into infinite baffle driver specs here. I will make new and additional measurements but I'm still a little concerned.

OK I'll check Vas. I'm just concerned that something is wrong based on the magnitude of the error. I've seen around 10% error in Fs before but this is more like 45%. That seems like it should be well outside of manufacturing tolerance. Sent from my SAMSUNG-SGH-I337 using Tapatalk

Sent from my SAMSUNG-SGH-I337 using Tapatalk

[pick]

OK as suggested by some of the contributors to this post (thanks everyone), I took a complete set of measurements, including Vas. I have attached an image of the DLS data sheet and of my measurement results. I suspended the driver from the ceiling so there is nothing within a few feet of it during any tests. The results changed only slightly compared to my previous measurements. I put the new parameters into a simulation. The driver as measured works OK in a huge sealed enclosure and is unusable in vented alignments. This was not surprising considering the measured Qts of almost 0.7. There are some minor cosmetic blemishes on the cone that originally didn't bother me that much. After further inspection and some measurements, I also notice the surround is about 1/8 inch higher on one side of the driver than it is on the other. Considering this driver is so far out of spec and does not seem to be useful for the intended ported alignment, I think this one is going to be sent back.

[Chris Roemer]

Well... those specs DO seem to be pretty far off, but -
(w/Dayton's orig. T/S specs) WinISD likes 15cf, w/an Fb AND F3 in the mid teens.
With your measured specs, ISD still likes 15cf w/an Fb and F3 in the mid teens - very similar model (but with about 1dB of "ripple").

In a nicer looking alignment (8cf tuned to 18Hz) they look even closer w/the orig. specs giving an F3 in the upper teens, and YOUR specs the mid teens.
Also looks really strong in 6cf tuned to 20, w/F3s in the lower 20s (to "spec"), and your driver a few Hz lower (actually better looking, IYAM)

[pick]

Well... those specs DO seem to be pretty far off, but -
(w/Dayton's orig. T/S specs) WinISD likes 15cf, w/an Fb AND F3 in the mid teens.
With your measured specs, ISD still likes 15cf w/an Fb and F3 in the mid teens - very similar model (but with about 1dB of "ripple").

In a nicer looking alignment (8cf tuned to 18Hz) they look even closer w/the orig. specs giving an F3 in the upper teens, and YOUR specs the mid teens.
Also looks really strong in 6cf tuned to 20, w/F3s in the lower 20s (to "spec"), and your driver a few Hz lower (actually better looking, IYAM)

I have attached plots of simulations for the manufacturer's spec and for the unit I received. I also included a simulation of my unit in 6cf tuned to 20Hz. My feeling is, if my measurements had been the published specification, I would not have selected this product for my application. Also, for me personally, I would never design a ported enclosure around a driver with Qts = 0.7. To me, a driver with Q=0.7 is for infinite baffle not ported enclosure. I live in Columbus, so I actually drove to Dayton to exchange this driver yesterday. As usual, Parts Express took care of me without any hassle. Unfortunately, the replacement is even further out of spec. With Fs=29Hz and Qts=0.7. I have to conclude there is a problem with this production lot or with this contract manufacturer. I plan to wait until PE gets the next shipment and try again. If the next one is not significantly better, I will likely go back to the reference series that has treated me well in the past.

[guitar maestro]

Looks like too tight of a spider pack. I wonder if further/more extreme break in would change things for the better.

[danmarx]

This is peculiar for sure. Digging into the T/S parameters a little bit more definitely indicates low compliance (Cms = 0.1 in stead of 0.24), high stiffness (Rms = 24 instead of 14) all which will contribute to a higher Fs as you measured. Mms is also significantly higher (+74 grams) which I'm not sure why the mass of the driver (cone, surround, voice coil, etc) could possibly vary lot to lot that much unless something went terribly wrong somewhere in manufacturing, either a wrong material or excess glue. It's only like 2.5 ounces but for a system that is only supposed to weigh 10 oz. that makes up 25% of the total mass. Higher Mms should result in a LOWER fs, all other things being equal. But the Cms is so low that it counteracts the extra mass still giving you a higher Fs. Vas is also crazy low too, a result of Cms being so low, all akin to the spider and suspension elements being way too stiff, as in more than double the stiffness of a spec'd driver. Also interesting is Qms is just about at spec'd value (2.43 instead of 2.4) which is a function of a bunch of values which in your case given the high fs and low Cms check out, but Qes is way off. Why? Because Qes is calculated using the same parameters as Qms except the impedance at fs is not factored in. Your driver impedance peak at fs is only 13 ohms whereas a normal driver should be around 20 ohms. Again, all signs of low compliance of the suspension components. Above all, the mass has me baffled (that's a speaker joke, haha). Are you using the added mass method to determine Vas? If so what mass are you using and is the mass known/correct? That's the only thing I can think of, but even if we assumed Mms to be actually be correct, the Fs is still too high, Cms would only climb to 0.13 and doing a simple Fs measurement is pretty hard to screw up. Ask PE to confirm your measurements on an independent setup, that is the only way to know for sure. It'd be in their best interest to make sure the drivers are built and operating to spec anyway.

[pick]

Looks like too tight of a spider pack. I wonder if further/more extreme break in would change things for the better.

I have the second driver in break-in right now. Going to let it go for a few days to see what happens.

[pick]

This is peculiar for sure. Digging into the T/S parameters a little bit more definitely indicates low compliance (Cms = 0.1 in stead of 0.24), high stiffness (Rms = 24 instead of 14) all which will contribute to a higher Fs as you measured. Mms is also significantly higher (+74 grams) which I'm not sure why the mass of the driver (cone, surround, voice coil, etc) could possibly vary lot to lot that much unless something went terribly wrong somewhere in manufacturing, either a wrong material or excess glue. It's only like 2.5 ounces but for a system that is only supposed to weigh 10 oz. that makes up 25% of the total mass. Higher Mms should result in a LOWER fs, all other things being equal. But the Cms is so low that it counteracts the extra mass still giving you a higher Fs. Vas is also crazy low too, a result of Cms being so low, all akin to the spider and suspension elements being way too stiff, as in more than double the stiffness of a spec'd driver. Also interesting is Qms is just about at spec'd value (2.43 instead of 2.4) which is a function of a bunch of values which in your case given the high fs and low Cms check out, but Qes is way off. Why? Because Qes is calculated using the same parameters as Qms except the impedance at fs is not factored in. Your driver impedance peak at fs is only 13 ohms whereas a normal driver should be around 20 ohms. Again, all signs of low compliance of the suspension components. Above all, the mass has me baffled (that's a speaker joke, haha). Are you using the added mass method to determine Vas? If so what mass are you using and is the mass known/correct? That's the only thing I can think of, but even if we assumed Mms to be actually be correct, the Fs is still too high, Cms would only climb to 0.13 and doing a simple Fs measurement is pretty hard to screw up. Ask PE to confirm your measurements on an independent setup, that is the only way to know for sure. It'd be in their best interest to make sure the drivers are built and operating to spec anyway.

Yes, I found the results to be very strange. I am measuring Vas using the added mass method. The added mass was measured with a precision scale. I started with about half the specified Mms, resulting in an error message from WT3 that Fs changed less than 25%. I add to the mass in small increments until this error goes away. This process resulted in an additional mass of about 287g. This seemed like too much mass to me because it is almost as much as the specified Mms. The calculated Mms, however, is much higher. The glue under the cone is definitely excessive and a little sloppy. This could result in a little additional mass. Also, the cone is a woven material with a coating, similar to fiberglass. If the contract manufacturer does not have good control of the process, I could see the mass varying more than if they were using something like aluminum.

Ignoring other parameters that are more difficult to measure, Fs and Qts results come from an automated impedance measurement that would be difficult for the user to mess up. I called and asked if someone at PE could verify the unit I returned, but they said they said it was already prepared for return and they could not get to it (strange because I called only a few hours after the return). I did talk to someone at PE who agreed to measure one when their next shipment arrives. If the new shipment is in spec, I will try again.

[guitar maestro]

Same thing happened to me, sorta, with a 10" Ultimax. I returned it because the glue,joint was sloppy, only to get a replacement with the same sloppiness. Then they say that its not a problem/defect to have a sloppy glue joint. Weird.