Help me fix my broken Revel C12!

[napilopez]

Hey there DIY'ers. I recently ordered an old but "new" Revel C12. It's a 3-way center channel from the aughts.

Unfortunately, it was immediately clear something was wrong with the speaker, and later measurements showed a very deep null around 3.5 kHz.

It seemed like an honest mistake and I've resolved it with the seller, but now I'm wondering if I can salvage the speaker.

As some of you may know, I'm very experienced with measuring finished speakers but know little about actually designing speakers. Thought this could be a fun project to learn from.

The only available measurements for the speaker I could find are the following (both separate measurements from Sound and Vision for some reason):

and

My own quasi-anechoic on-axis measurements, unfortunately, showed the following:

The in room response was a bit better, but clearly still broken:

The bass and mid woofers on their own seem totally fine though (nearfield measurements summed and corrected for baffle step), but in nearfield the tweeter seems like it doesn't reach as low as it should:

I did try physically inverting the polarity of the tweeter, but that only helped a little bit on-axis; the in room response was the same.

Harman customer support has yet to respond to me about whether they have any replacement tweeter's in stock.

Any ideas on how I might go about finding a replacement tweeter? Or do you think it's possible the crossover needs fixing too?

I appreciate your help. I'm a newb in this area .

 

[30 Ounce]

Looks like the tweeter/ midrange polarity is reversed.

 

[napilopez]

Looks like the tweeter/ midrange polarity is reversed.

Thanks, but as noted above, I tried inverting the polarity of the tweeter. I might not have saved the measurement as I couldn't find it, but though it helped with the null on-axis, the overall in-room response didnt change much

 

[Rick Sykora]

Would try changing the polarity only on the midrange. Can you diagram the crossover? Have not found a schematic online.

 

[thewas]

Thanks, but as noted above, I tried inverting the polarity of the tweeter. I might not have saved the measurement as I couldn't find it, but though it helped with the null on-axis, the overall in-room response didnt change much

In this case probably the phase between them is closer to 90° so you just move the lobe by switching the relative polarity but the in-room response doesn't change much due to the reflections and total sound power which doesn't change much.

 

[eddantes]

I am very curious about this. Could a capacitor in the highpass be failing?

 

[617]

I am very curious about this. Could a capacitor in the highpass be failing?

Failed cap would have reduced capacitance and the tweeter would run full range, which would not exhibit the pronounced dip you see here.

It is very curious. If the mid polarity was reversed you'd see two nulls or one big wide one. Mid and tweeter you'd see a null at 300.

 

[617]

Thanks, but as noted above, I tried inverting the polarity of the tweeter. I might not have saved the measurement as I couldn't find it, but though it helped with the null on-axis, the overall in-room response didnt change much

I don't think you're lying but it really looks like an inverted tweeter! 90 degrees cancelation wouldn't give such a deep null. Even a different tweeter or a failed cap wouldn't give such a perfect null.

Provide measurements with the tweeter inverter under identical gated conditions if you can. The reason why, I disagree the tweeter isn't going low enough. It looks like a perfectly executed lr4 with the tweeter flipped, which is what you'd expect from revel.

I'd remove the crossover and take some pictures, noting how it's wired. Might be a pain in the ass but this speaker must have been meddled with. It's a high quality speaker and this is an odd failure.

Worse comes to worse we can find a tweeter that fits and redesign the tweeter network.

 

[Dj7675]

Hey there DIY'ers. I recently ordered an old but "new" Revel C12. It's a 3-way center channel from the aughts.

Unfortunately, it was immediately clear something was wrong with the speaker, and later measurements showed a very deep null around 3.5 kHz.

It seemed like an honest mistake and I've resolved it with the seller, but now I'm wondering if I can salvage the speaker.

As some of you may know, I'm very experienced with measuring finished speakers but know little about actually designing speakers. Thought this could be a fun project to learn from.

The only available measurements for the speaker I could find are the following (both separate measurements from Sound and Vision for some reason):

and

My own quasi-anechoic on-axis measurements, unfortunately, showed the following:

View attachment 359978

The in room response was a bit better, but clearly still broken:

View attachment 359979

The bass and mid woofers on their own seem totally fine though (nearfield measurements summed and corrected for baffle step), but in nearfield the tweeter seems like it doesn't reach as low as it should:

View attachment 359984

I did try physically inverting the polarity of the tweeter, but that only helped a little bit on-axis; the in room response was the same.

Harman customer support has yet to respond to me about whether they have any replacement tweeter's in stock.

Any ideas on how I might go about finding a replacement tweeter? Or do you think it's possible the crossover needs fixing too?

I appreciate your help. I'm a newb in this area .

This might not be of any help. I had 3 C208s and had 1 that appeared to be broken. I compared 2 with a simple 1 point measurement. Had to send one back to get replaced.

 

[Ze Frog]

I'd get a Dayton Audio DATS and Omnimic, measure TS parameters, measure individual driver frequency response on baffle and redo the whole thing from scratch. You probably only need the DATS as you obviously have means to measure. Then use FP GraphTrace(if you can't convert measurements to .FR and .ZMA)and XSIM to design crossover.

And I'll wager you will be able to get a much better response. Crossover design really isn't that daunting at all. I'm amazed actual companies can't actually get it right more of the time.

 

[SDC]

I'll carefully measure the tweeter without passive XO if I was in that situation.

I once had problem with Revel F228be which wasn't fixed with new tweeter, so the dealer replaced the XO and the speaker was fixed.

 

[napilopez]

Thank you everyone for the help so far! I'll try out some of the suggestions within my means once I get the chance. Have a pretty busy weekend, but I'll let you know once I can measure again.

I genuinely appreciate all the advice. I get the feeling I'm going to learn a lot with this.

 

[napilopez]

Oh, I found the inverted tweeter measurement, just put it in the wrong folder lol. Behold:

Same measurement without gating. Doesn't change much in the upper frequencies as expected for a single measurement and given the distance from the walls.

It's also maybe worth noting that existing measurement of the speaker when working correctly does have a deep null 15 degrees above the tweeter axis, but that null shows up at 2.5 khz rather than the 3.5 khz in mine. Again from sound and vision:

I think I need to spend some time taking more careful measurements as these were all a little haphazard, but it's still clear there's somethign very wrong lol. The inverted tweeter measurement actually does seem a teensy bit better but that scoop shouldn't be there. Also curious that there's a dip centered around 10khz, which isn't present in the two Sound and Vision measurements of the C12.

Lastly, if it gives anyone more ideas, here's the nearfield measurements of all four drivers, uncorrected for baffle step and with the tweeter in its original polarity. In these, the microphone was pointed as close as possible to the drivers without the microphone hitting the drivers during playback, as I would do for quasi-anechoic measurements. The tweeter is protected by the phase shield so I can't get too close.

 

[Blumlein 88]

Just keeps looking to me like a bad crossover issue. Maybe a part of the wrong value installed in the wrong location or some such. Things like that do happen. Don't think it is a bad tweeter or midrange. I'd open it up and look at the crossover itself. Also have you looked at the impedance curve to see if anything looks off around that same dip frequency?

 

[cavedriver]

could be a bad capacitor in a Zobel circuit on the midrange, not sure which component failure in the crossover could drag the tweeter down like that though. If it were me I'd open it up, look for burned out components, and also look for a loose wire. I had a tweeter on one of my Snells drop significantly in output, I thought it was blown, but when I checked I found one of the solder joints had cracked. This can happen over time from vibration.

 

[617]

I would concur with others. Something wrong, most likely with tweeter network. Take out the tweeter network (probably one board for all) and take a good picture. Inspect for damage or poor solder joints. This is an inexpensive speaker with well damped drivers so I would guess the network has a low part count. Take some really good pictures and ASR dorks will let you know a. what components could likely cause this response deviation and 2. which parts are damaged, if any.

In my experience tweeters either work the way they should or are completely destroyed, and I suspect the tweeter is fine in this case; worse case scenario you have to rebuild the tweeter network. I believe you are able to do this.

 

[Looneybomber]

Failed cap would have reduced capacitance and the tweeter would run full range, which would not exhibit the pronounced dip you see here.

As a cap ages, it loses capacitance. A cap with a smaller capacitance filters at a higher frequency. For a tweeter, this means reducing its operating range, not increasing it. It creates the dip that the OP is talking about.

 

[Looneybomber]

Hey there DIY'ers. I recently ordered an old but "new" Revel C12. It's a 3-way center channel from the aughts.

Unfortunately, it was immediately clear something was wrong with the speaker, and later measurements showed a very deep null around 3.5 kHz.

It seemed like an honest mistake and I've resolved it with the seller, but now I'm wondering if I can salvage the speaker.

As some of you may know, I'm very experienced with measuring finished speakers but know little about actually designing speakers. Thought this could be a fun project to learn from.

The only available measurements for the speaker I could find are the following (both separate measurements from Sound and Vision for some reason):

and

My own quasi-anechoic on-axis measurements, unfortunately, showed the following:

View attachment 359978

The in room response was a bit better, but clearly still broken:

View attachment 359979

The bass and mid woofers on their own seem totally fine though (nearfield measurements summed and corrected for baffle step), but in nearfield the tweeter seems like it doesn't reach as low as it should:

View attachment 359984

I did try physically inverting the polarity of the tweeter, but that only helped a little bit on-axis; the in room response was the same.

Harman customer support has yet to respond to me about whether they have any replacement tweeter's in stock.

Any ideas on how I might go about finding a replacement tweeter? Or do you think it's possible the crossover needs fixing too?

I appreciate your help. I'm a newb in this area .

Replace your capacitors. This is a common problem with the Infinity Beta speakers (they use the same tweeters and crossovers are built by the same OEM). It’s fixed by replacing the caps.

As a heads up, the tweeters in these speakers are not very high quality. They blow easily and you’ll start hearing the effects of lost ferrofluid around now or soon depending upon when they were built and how they were stored. I purchased two replacements a little while ago - one to use now and one for the future. Good luck.

 

[617]

As a cap ages, it loses capacitance. A cap with a smaller capacitance filters at a higher frequency. For a tweeter, this means reducing its operating range, not increasing it. It creates the dip that the OP is talking about.

Thanks for reminding me, I was picturing the ladder network in my mind and got it backwards. Hopefully it's something as simple as a cap.

 

[napilopez]

Replace your capacitors. This is a common problem with the Infinity Beta speakers (they use the same tweeters and crossovers are built by the same OEM). It’s fixed by replacing the caps.

As a heads up, the tweeters in these speakers are not very high quality. They blow easily and you’ll start hearing the effects of lost ferrofluid around now or soon depending upon when they were built and how they were stored. I purchased two replacements a little while ago - one to use now and one for the future. Good luck.

Thank you! Where did you get your replacements? I've been too busy to investigate this further but I'll hopefully have some time this weekend.