Millenium tweeter issues

[jimangie1973]

I noticed one of my Millenium tweeters had a peak at 12 kHz or so. I was also seeing higher F3 distortion from this unit. I bought new voice coils to hopefully eliminate this difference. Well when I put the new voice coils in and checked the impedance, there were now twin resonant peaks. I've attached plots of the in box frequency response and impedance profile for the two units.

When I changed the voice coils, I also glued down the felt damping pad behind the dome, because it was loose. On one unit, it was put in completely wrong.

Could it be a lack of ferrofluid causing this? Or could the addition of a small amount of glue causing a second resonance? Too much trouble for $200 each tweeters. :roll:

[TacoD]

The voice coils are not properly aligned, this causes the second peak in the impedance curve. The trick is to realign, the thing I do is the following:

measure the tweeter impedance and at the same time move the coil around in the magnet gap. You will hear a resonance when it is not properly aligned, and you will see this second peak. Carefully move the coil assembly until you end up with a as large as possible single peak.

The Seas Millenium is build as a tank, so you have some room for error.

[jimangie1973]

The voice coils are not properly aligned, this causes the second peak in the impedance curve. The trick is to realign, the thing I do is the following:

measure the tweeter impedance and at the same time move the coil around in the magnet gap. You will hear a resonance when it is not properly aligned, and you will see this second peak. Carefully move the coil assembly until you end up with a as large as possible single peak.

The Seas Millenium is build as a tank, so you have some room for error.

Thanks a lot for the advise. I will try rotating the voice coil around relative to the magnet gap and the rear chamber until I get the best impedance response. I believe there are 4 possible positions, but I could be wrong. There's really no 'wiggle' room when the voice coil is installed.

I'll let you know what I find out. Thanks again.

Jim

[Davey]

You can't really rotate the VC since it's keyed to the front plate. I can't quite remember but I think you could turn it 180, but that's it.

The key to this operation is proper sequencing/tightening of the four screws. Assemble the VC onto the guts and attach the front plate with all four screws just short of snug. Play a 500Hz tone at about 1 volt drive level and apply some lateral pressure on the front plate in four different directions and note which way causes distortion and which doesn't. Whichever direction caused the best result tighten the screw on that side first. Then the one opposite and then the two adjacent.

Guaranteed to work well.

Try you impedance plot again and it should look fine.

Cheers,

Dave.

[jimangie1973]

Well I figured it out. The majority of the second peak was because I glued the felt piece. I pulled it out and removed all the glue I could, then reinserted it, now much looser. The second peak was gone. I messed with the positioning of the voice coil and got the 500 Hz resonant peak as high as possible. Now the impedance responses look good and the frequency response is better too. Thanks for the help guys. :T

[Jed]

Can you post your harmonic distortion plots?

[Sakura]

At that price, the build should be pure perfection, anything less is intolerable.

Nice to see that the issue could be resolved though.

[Jonasz]

According to Zaph's and MarkK's measurements the Millenium isn't the distortionchamp it once was. I'm surprised SL is still using them in the Orion. I bet SS6600 or Peerless HDS (and the Peerless to a way better price) would make a better job.

[tktran]

True some other domes beat it in the non-linear distortion stakes. But if you look at it that way I don't think any other driver can match the D6600/81921 in the non-linear stakes.

Certainly the 7" SS Revelator 8531G in John Krutke's ZRT has a hard time.


IMHO a comparison should be made with the other drivers in the system-
ie. midrange, woofers etc. before changing the tweeter.

I'd like to see a full range harmonic distortion sweep to see who's putting out the most distortion? The 2x10" XLS in dipole fashion, the single W22EX or the Millenium.

[bmaupin]

Is the 12k peak gone? I had the same peak in mine. I incorporated a notch in my design. I have quite a few 2.0mH 20ga inductors that I bought cheap that come in handy once in a while

- Brad

[jimangie1973]

Is the 12k peak gone? I had the same peak in mine. I incorporated a notch in my design. I have quite a few 2.0mH 20ga inductors that I bought cheap that come in handy once in a while

- Brad

The peak is mostly gone. I think if you unglue and mess with the positioning of that felt pad behind the dome, you can reduce the peak. I will post images of the new responses when I get a chance.

Also, I will post some non-linear distortion data. I have ARTA only so I don't think I can post a nice swept sine plot, just single frequency snapshots.

Jim

[Dennis H]

I have ARTA only so I don't think I can post a nice swept sine plot, just single frequency snapshots.

Try the companion program Steps.

[jimangie1973]

Try the companion program Steps.

Thanks for the tip. I tried it and it works great.

[jimangie1973]

Here's the distortion result for a 2.83V input. This includes crossover so output is about 87 dB SPL at 1m.

[Jed]

Thanks Jim.

You might want to limit the frequencies to 500hz on up next time if you go much louder.

Jed

[ergo]

Not quite a Millennium topic but similar (do not want to start similar thread).

I recently replaced burned mid and tweeter set on Jamo R909 and got the old burned samples.

Now the tweeter is standard ScanSpeak D2905/970000, so I could get replacement domes from Madisound for example, but I tried to open the tweeter and see the damage etc, but it is not as easy as with Seas tweeters at all. Seems the dome assembly is attached with some amount of glue as well....

Does anyone have an experience changing domes on SS tweeters and could give me some pointers. I do not want to go with brute force on the thing (even though they are damaged already).

Ergo

PS. The mid is M15-CH-JAM so a custom version for Jamo, so I quess there is very little hope to fix those