NHT Classic Threes

[Beave]

I recently bought a used set of NHT Classic Three bookshelf speakers and the ThreeC center channel from a local seller.

Something's not quite right with the Three bookshelf speakers. Both have a surprisingly 'warmer' than expected sound. Upon careful listening there's a noticeable lack of mid-treble detail. Both bookshelf speakers are like this, but the center channel sounds 'right' to me (with considerably more detail in the treble and otherwise a similar sound).

I checked the drivers and they all work (both midrange domes and both tweeters).

I ran some sweeps and compared the bookshelf speakers with the center channel. Their response is similar up until about 4kHz. From about 4-6kHz there is a big dip in the output of the Three bookshelf speakers that isn't present on the center channel.

That's pretty close to the crossover region between the mid and the tweeter, so I thought maybe they had accidentally wired the tweeter out of phase, thus leading to a cancellation (big hole) in the crossover region.

But upon opening them up, I see that they are pretty foolproof. The + and - wires from the rear panel to the mid-tweeter crossover board have different sized connectors, so they can't be mistakenly reversed. And the wires from the mid-tweeter crossover board to the tweeter are soldered. I found a picture online of the mid-tweeter module (crossover board plus drivers) and it looks to be wired the same.

I've seen three sets of measurements for the Classic Threes. Two are quite flat, while one shows a pronounced dip in the mid treble.

https://www.stereophile.com/content/nht-classic-three-loudspeaker-measurements

https://www.soundstagenetwork.com/measurements/speakers/nht_classic_three/

This one shows a big dip in the mid-to-tweeter crossover range

Does anybody have any experience with these? I don't know much about their history. They're in excellent shape and don't seem tampered with at all (until now!). Were there any revisions/changes made in production? Are there any common build errors? I'm tempted to remove the mid-tweeter module from the center and put it in one of the bookshelf speakers to see what happens.

 

[shadyJ]

Do your own measurements resemble the one in that Eastern European site?
One possible culprit for that strange response is measurement technique. If the mic was too close to the speaker and wasn't pointing toward the tweeter, you could get responses like that where you would otherwise get a flat response with the proper distance and mic positioning.

Maybe another culprit is some component on the crossover isn't working right. You might want to just get in contact with NHT and see what they say.

On another subject, I don't know why that response graph went out to 100,000 Hz when he only measured up to 20 kHz. Makes it a little odd to look at.

 

[Beave]

I don't have a measurement mic, so I just ran some tones and listened. I have the center on a stand and the bookshelf on top of it, with a quick switch a/b box and I've level matched the two. When I play a sweep of tones, the outputs of the center and bookshelf sound about identical in level until 4-6kHz, when the bookshelf output drops very noticeably and the center output stays at a level consistent with the rest of the range. Above about 6kHz it comes right back up in level.

The mid-treble output dip is quite audible all around the room, no matter where I am sitting.

But, without measuring, my *perceived response curve* is most like the third link. Actually the big drop is narrower and deeper, at least it sounds that way to me.

 

[Beave]

Oh, I thought about it being a crossover component gone bad, but the fact that both bookshelf speakers sound the same way (big dip around 4-6kHz) made me think that wouldn't be it. Or at least it would be less likely that both would have the same part go bad. But it's still possible that a cap on each crossover has gone bad in the same way.

 

[Steve81]

Or at least it would be less likely that both would have the same part go bad. But it's still possible that a cap on each crossover has gone bad in the same way.

May not be that unlikely given that they're presumably a matched pair from the same production batch, and they've been used and stored in identical conditions.

 

[Beave]

Good point. I'll look over the crossover boards more closely. I used to have access to a boatload of electronics test gear from work, but I no longer do, so my debugging capabilities are hampered now.

 

[Beave]

I got a schematic for the Classic Threes and compared with mine. All parts are correct values and installed correctly. I see no failed parts. Caps look good.

I might have to send a module to NHT and have them look at it, because I'm about out of ideas of what could be wrong.