My Tumult sucks (v. Don't believe the model / What to do with a paperweight?) - Page 3

[=][/]

This is one of my original Tumults, No EQ. The only thing is the 80hz XO in my receiver. It is in a sealed 3 ft cubed cabinet... Close mic
Edited by kgveteran - 10/13/12 at 6:55pm

OK, a couple of things:

Here's the latest posted model with the averaged close mic measurement normalized and overlaid:



As in the free air close mic, the low end roll off is inaccurate, pointing to the measurement rig/environment. There is also a bump at 30 Hz in the averaged close mic of the AV and the Tumult. Also, the tune in the close mics is lower than the model predicts, indicating the PRs/net enclosure volume/stuffing, not the drivers. Finally, in all of the close mic measurements, the top end roll off is not accurate. There's no way there is that large an "inductance problem" with the Tumult. That would almost exclusively have to mean there is more VC in your Tumult than any other Tumult, which I'd have to say is just not possible.

If you correct for the 30 Hz anomaly, correct for the excessive top end roll off and low end roll off and adjust the tune a few Hz lower, the model and the result are not too far off.

The difference in top end roll off (inductance-induced roll off) between the AV15H and the Tumult is around 6dB. No way it's double that, IMO.

Did you do a close mic of the AV in free air to compare to your free air of the Tumult?

Dave let me get back to you on the close mic vs groundplane measurements. I believe I have about 6 drivers at least for comparison. I will have to put the pics together.

EDIT: Ok. ..So it is not quite as large a difference as I remembered at 100Hz. After normalizing the measurements to the peaks in each it looks like between 1 and 2.5dB difference by 100Hz on average. Also the low end is a little bit boosted on the close mic. It's probably better to just look at the overall response shape differences though. The bass is more humped up and exaggerated in the close mics.









Edited by Ricci - 10/14/12 at 4:10pm

Glad you figured out where the user error was distorting your results.

Yup, that one resembles my graph, and if you add back in the 80hz rolloff, bingo....

I just read the whole thread and I'm pretty certain we have a combination of issues:

1. The Tumult has relatively high inductance, it is going show a hump and roll-off well beyond the simple Unibox model.
2. The AV15H is one of the lowest inductance subwoofers I know of and therefore should better match the simple Unibox model.
3. It appears your measurements are exaggerating the roll off above 40hz as the AV15H shouldn't be down ~8db by 100hz...could also be an artifact of close mic like Ricci suggested.
4. You should expect some manufacturing variance along with some expectation that parameters can shift over 10 years of storage.
5. You mention a lack of punch and I believe that is easy to explain with your measurements showing a severe roll off likely due to inductance (and possibly combined with some measurement issues).
6. You could simply have a bad driver although I'm guessing that is not the case given that it is not far off what I would expect. Take Bosso's suggestion of a 100L sealed box for testing.

UniBox and WinISD models are only as accurate as the data that is used. In the case of low inductance drivers like the AV15H or many pro drivers you can be fairly confident when using a single inductance number. For long excursion, higher inductance drivers like the Tumult, the simple model will be off and I believe that is mainly what you are seeing.

A higher inductance driver like the tumult will require some amount of EQ to achieve the punch you would get from a lower inductance driver. That is the price you pay for the greater excursion. I'd suggest EQ'ing the Tumult to match the measured response of the AV15H and see how the sound differs. I'm fairly certain you will find the differences to be small assuming you are low passing around 80hz. The AV15H is likely a lower distortion driver above 80hz.

You ask what drivers offer low inductance and high excursion. Well, that is basically what everyone seeks (aside from Bosso, although I'd guess all things being equal he wouldn't mind lower Le). Ultimately, this is expensive to execute and the market is generally small.

If you need the long excursion of the Tumult, I'd suggest EQ'ing. I don't know of a 15" that would replace it to your liking. If you are fine with the AV15H's excursion limits, stick to that.

As far as the reaction you've received, I think your title is at fault. First you trash a product without basis and then you suggest that either the specs or models are faulty. You were asking for assistance but you didn't exactly have an open mind.

Interesting indeed. Man, I appreciate an answer that has actual data as its main body. Since I design and build all of my signal shaping circuits from close mic measurements, it will be interesting to see how the subs measure outdoors, GP.


Good post. I agree with all of it, except #4. My comments:

#4: inductance is the result of the amount of wire on the VC. I'm all but 100% certain that the VCs ordered for any sub of the Tumult caliber has virtually the same amount of wire and that can't change over time. This is the main reason I doubted the original posted measurements.

Regarding Le, I have not found that it matters much in the final analysis. The driver designers who know what they're talking about rarely state that they have gone with shorting rings (or similar measures) to flatten the high end response of their systems. Instead, they'll say something like "...to reduce some harmonic distortions".

I did not find the AV15H to be in any way superior to the Tumult. In fact, as I said, it takes a pair of the AV15H to match a single Tumult for clean output at 20 Hz... where it counts.

I would much rather have the available excursion <20 Hz and shape the signal and cross point to arrive at the required top end than to not have to shape the signal or cross point traded for 1/2 the power <20 Hz.

Have a look: Same placement, same input signal, same room, same mic distance and placement, same amplification, same # of drivers, etc., Tumult-based vs AV15H-based subs, same amount of boost in the signal shaper:



Can you tell which is which? Can you see a "hump"? Can you see a 12dB difference in top end? Etc.

I was just suggesting that there could be variances in the specs of drivers. Of course, Le is very unlikely to vary for the reasons you suggest. I was trying to say that don't assume models are incorrect based on unconfirmed specs. A manufacturer could even fudge a number. I've never used Tumults so I wasn't suggesting it necessarily, but that a model that is not matching isn't indicative of a poor model.

I also wasn't suggesting that a lower inductance woofer was altogether better. Would you not take a lower inductance Tumult assuming all else was the same? Of course that is easier said than done. I understand why you choose excursion over lower inductance for greater <30hz headroom for your systems. Ideally, you would have 10000w power handling, 50mm of linear excursion and .1mh of Le. Of course that is not cheap.

Stereodude, it looks like you have it figured out. Now I would suggest using some EQ to bump up the upper response and let us know how you like it.

FWIW (as I'm not in the same league as you guys measurement wise!!) re: Close mic vs ground plane. When I did my review of the SVS SB13-UItra (sealed 13' cube) for the contest at Audioholics, I compared my close mic measurement to the 2M GP FR that Ed sent me from SVS and there was an obvious difference from 100Hz onwards. Ed was also kind enough to send me his close-mic measurements of the SB13 and I overlaid it on mine, and they were almost identical.

Here my close mic ...



Here is SVS's 2M GP ..



Finally, here is my close-mic overlaid over the close-mic from SVS that Ed provided (Ed's is the dotted line, apologies for the quality of my overlaying!)...



Unfortunately the close mic Ed sent me was only up to 300Hz. But as you can see it tracks to mine fairly well. I noticed the same thing when I compared my close mic measurement of the PB13 in sealed mode to Ricci's 2M GP and SVS's 2M GP.

Ricci mentioned creating a "correction file" for his own purposes for this. I was going to try and do the same with the measurements I have but never got around to it.

What I can't wrap my head around is why the differences exist and become so pronounced from around 80Hz or so onwards, but not below that?

In general. I.e., why close-mic measurements seem to be quite different once you get passed 80Hz or so (e.g., differences in Ricci's close mic vs 2M GP measurements at 100Hz onwards by way of example). Why would the measurement be quite close on the lower end using close-mic, but start to fall off as the frequencies increase?

Yeah don't sweat it too much just know that close mic will roll off the top end a little more than what it really is. Just figure on being about 1.5dB less at 100Hz ballooning to about 5dB by 200Hz. If I had to guess I would assume that the longer wavelengths of the bass frequencies take longer to capture a measurement of and in that time they start to get reinforced by the boundaries which are acoustically close at the wavelengths involved so that the mic is getting more and more additional energy from these reflections as the frequency gets lower which boosts the low end more than the top end. The deeper your frequencies the longer they are and the further away you can be and still be considered "close" acoustically. Anyway that's my 2 cent guess on it. Maybe Mark or someone else knows better.

Are you happy with it now? Do you want to sell it?

What about it sounds bad? Just lacking punch? If so, is the low end pretty good?

What mains are you crossing the Tumult with?

I was just about to post that the posted measurement looks exactly like a sub running about 10dB hot.

The steep roll off to 100 Hz, then dead flat to 200 Hz is definitely weird and not like any MKI I've had.



Can you post one with the mains (use an arbitrary LP distance) with the sub calibrated flat level-wise and using a 60 to 80 Hz cross?

Do you have a graph of the LP when listening and saying it sounds bad?

No, I was referring to this one:



Looks exactly like running the sub hot. Either way, it doesn't resemble anything I've ever gotten from 8 Tumults. It just put the thought into my head to have a look at the 2 subs with the mains, as you heard them. Since what we hear, and thus comment on, is FR, it always helps to see what your listening to. Trying 60, 70 and 80 Hz cross points will be interesting as well.

It's your call to go through the exercise or not.

If you're still not happy, do you want to sell it?