problem leads to intersting discovery

[steelie]

Hey all,

The other day I started having problems hearing my left surround speaker, an RB35. The volume and the high frequencies both appeared way too low to my ears.

I ran an SPL check using the internal test tones in my Denon 2802 and my Radio Shack meter told me all was okay, even though I could hear there was a problem. I put on my Video Essentials DVD, ran the DVD using my digital connection, and again my SPL meter told me all was okay all channels balanced even though I could hear a problem.

Here's the interesting part, at least to me.

I then checked the speaker balance using the internal test tones in my Denon 2200, running the tones through the analog, external inputs. My SPL meter showed my left surround was several decibels lower than the rest of my system. (I had to jack the volume control up by 4 - I think that translates to several decibels) The sound was horrid.

At first I thought there was an issue with the external inputs, which have to be balanced separately from the rest of the system anyway. But it seemed odd to me that there was a fairly large balance problem where there hadn't been one in the past.

The solution was simple. I checked in the back and realized the jumpers weren't making their connections properly. Once I reconnected them, the volume, clarity and quality of the speaker came back to normal. The horn went back to work and all is well.

The quirky thing is why didn't the receiver's internal test tones and the digital connection didn't pick up on that. Even after I reconnected the jumper, the SPLs didn't change at all when I checked the balance through the Denon receiver or the test disc.

Are the 2200's internal mechanisms and analog connections that far advanced compared to the other two options? Does the white noise provided by the receiver not include ANY of the higher frequencies? If they don't, they certainly seem like they should. We're not talking about a small difference in volume here.

Any ideas?

Thanks to all,

steelie

[Scp53]

strange?? hmmm...

[DrWho]

What kind of test tones were each of the devices using? The woofer in the RB-35 is producing sounds up to about 1kHz and is probably capable playing quite higher. When the binding post bridges are disconnected, the crossover is no longer being used and in your case the tweeter wasn't producing sound. If the test tones being used to conduct the test used frequencies that the woofer could play, then you wouldn't measure a difference in volume (even though it'd sound different).

[steelie]

That's what I was thinking, too. But if that's the case, it seems to make the test tones on the receiver totally irrelevant. I can hear a problem and diagnose a several decibel drop with my DVD player running on the external inputs. But if I still can't measure a problem with the receiver's internal test tones, then what's the point of calibrating using the receiver at all? In my case, the balance wasn't even in the ball park.

Just seems too odd.

Thanks for the reply.

[colterphoto1]

just a guess here, but could it be that the 'white noise' test tone although generally described as equal volumes of all frequencies, is actually biased towards the low end of the scale? I don't know all the technicalities of test tones, but the energy needed to create 20HZ at Xdb would be far greater than the electrical energy needed to create 10kHz at Xdb, kind of a logarithmic scale if you will. If that's the case, then the energy output in the woofer's regions would far exceed the energy output in the few higher octaves of the tweeter's regions. Your SPL meter is just measuring total output, not specific frequencies, so it 'misses' the fact that the top 3 octaves or so are totally missing from your test (due to binding post jumpers being disconnected). I've got a Yamaha receiver with test tone, SPL meter and pair of RB3II's, so will try to duplicate your test. Sorry for rambling, but it is an interesting question.

[DrWho]

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

just a guess here, but could it be that the 'white noise' test tone although generally described as equal volumes of all frequencies, is actually biased towards the low end of the scale? I don't know all the technicalities of test tones, but the energy needed to create 20HZ at Xdb would be far greater than the electrical energy needed to create 10kHz at Xdb, kind of a logarithmic scale if you will. If that's the case, then the energy output in the woofer's regions would far exceed the energy output in the few higher octaves of the tweeter's regions.

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

Yes and no...it all depends on the speaker's sensitivity at each frequency. Granted, most speakers don't play 20Hz very well and thus need a lot more power to create 20Hz, but that's because they have a low sensitivity in that region. A pink noise test tone does not weight any of the frequencies (pink noise is the one that had each frequency at the same volume...i forget what white noise is).

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

Your SPL meter is just measuring total output, not specific frequencies, so it 'misses' the fact that the top 3 octaves or so are totally missing from your test (due to binding post jumpers being disconnected).

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

In addition to the SPL meter just measuring total output, when the binding post is removed, the crossover is no longer active and that small woofer is now also playing the higher frequencies that it normally doesn't play. Technically, disengaging the tweeter should reduce the output by about 3dB, but because the crossover gets disabled without the binding post, then the woofer is now "the tweeter" and the difference in output should ideally be 0dB (provided that woofer can play up to 20kHz, which it can't so the difference will be about 1dB which is barely noticeable). The reason it sounds so different though is because that woofer is introducing lots of distortion (if it didn't, then we wouldn't need the tweeter in the first place) In the case of white or pink noise, I've found that distortion will make it seem quieter when it's not.

[artto]

First of all, the Radio Shack SPL meter is not flat in its measured response. It can easily deviate +or- 2dB. Furthermore, its response falls off rather dramatically with increasing frequency. So even though you are able to hear the reduced output from the horns mid/upper frequencies, the meter is still measuring the speakers predominant output in the mid/low frequencies. You would probably need some kind of notch filtering/EQ for a meter to pick it up or use a spectrum analyzer which provides more detail. So, yes, colter is on the right track. Moral of the story.trust your ears!

[steelie]

Got it, at least in general terms. The quirk is still the Denon DVD player, which had internal test tones that were able to identify the reduction in SPLs when run through the analog, external inputs. I'd forgotten to consider a difference between white or pink noise, so that probably explains it. Whatever the DVD player has, it definitely produced specific frequencies in the high end, which, when missing, showed up as a reduction in SPL on the left rear speaker.

Thanks for the input