Small Syns

Spider maybe? If I understand you are not measuring distortion per se, but the level of the harmonic? I was also thinking something to do with the cabinet?

Yeah, it's a Farina method swept sine, which in effect tracks the harmonics over the sweep. The graph shows the harmonic levels, fundamental level is the black curve at the top. In operation moving the cursor calculates the percent distortion where it points.

Cabinet is still a possibility, but I braced, tightened, stuck microphones around it, couldn't locate anything there related to the peak. The peak seems to drop down when I put some stuffing in the woofer slots, so it does seem to be radiating from there. Putting padding behind the woofers inside the cabinet didn't seem to have much effect.
 
How are the woofers mounted to the box? If stuffing behind the woofer doesn't help, how about a bit of decoupling of the woofer mount? Sorbothane or Neoprane gasket between woofers and box and preferably also decoupling the mounting points/screws? Maybe leave out the screws and mount it using a ring with another decoupling gasket clamping the driver. Just an idea...
 
That might help (and probably isn't a bad thing to do -- where do you get sorbothane gasketing?). But the distortion component does seem to be getting radiated from the driver itself rather than off the box -- if I put some cloth in the woofer's output aperture, the distortion drops fast, covering the cabinet doesn't do much of anything.
 
Founder of XSA-Labs
Joined 2012
Paid Member
It might be turbulent wind noise from the ports. Do a RTA and look at distortion spectrum in real time. If wind noise it will be broad band (but picked up by the peak tracker). Bundled soda straws can act as laminar flow elements and reduce noise while keeping CSA about the same.

Another culprit is sometimes an overtight driver mounting screw. It distorts the frame and HD results. Try loosening the screws holding the mids and woofer in place a bit. A sorbothane gasket on driver bezel can only help.
 
Last edited:
ok, onto the cabinet. Or ports. Or screws. Or bad karma.

Does the woofer have a similar blip in the impedance curve without being in the box?

Yes, it does! On both drivers --
6fe100%20open%20air%20impedances.png


But the 450Hz distortion doesn't happen outside of the box :mad:
barefoot%20woofer%20left%20side%20HD.png

barefoot%20woofer%20r%20side%20HD.png


Time to get out the caulk and put it on all the panel braces and seams, see if that does anything.

The port idea might be something to check (how?), but it doesn't seem like there'd be enough air pressure or velocity happening at that frequency to cause nonlinearity? (the port is all of 1 inch long, too)

Keep the suggestions coming (and thanks)
 
Last edited:
OH FOR CRYING OUT LOUD....!

What an afternoon this has been, even worse than yesterday evening. But I FINALLY figured out what it is. I'll let you wait in suspense while I whine a bit, though.:)

As mentioned above, the woofers were fine when driven by themselves. So, I was about to coat the inside of the cabinet with glue or caulk, but thought it wise to first double check the midrange/tweeter horn. And... (damn, you already guessed...) there was now a 2nd harmonic distortion peak in the lower midrange (350Hz this time). I could even hear it in the sweep, so no "sweeping it under the rug".

Throughout the afternoon, I convinced myself (and then proved myself wrong each time) that it was:
  • The amplifier (which I managed to blow up when some wires shorted... no big loss, it wasn't a very good amp other than having a handy stepped attenuator)
  • Something buzzing in the horn
  • The soundcard
  • The measurement microphone
  • A bad solder connection on a wire (the distortion went away briefly, though it came back
  • The tweeter (which did start distorting until I reseated its back cap)
  • The midrange driver (replaced with new.... same thing)
  • A 100uF NP electrolytic cap, peak seemed to move when I fooled with it. It is in the woofer circuit, but that section of the crossover was still hanging on the amp..

I strapped together enough film caps to replace the 100uF NP cap... same thing! But that led to finally identifying the culprit... I hope. I'm too tired to rebuild everything just now to make sure all plays happy again.

The solution is coming in the next post! (edit: well, maybe the post after that)
 
Last edited:
OK, the secret of a wasted afternoon is now to be revealed.

It's something in the crossover. An ERSE solid-core inductor, to be exact.

When I took a pair of pliers and squeezed its laminations together, the distortion product went away -- completely. Even with the woofers disconnected, it appears that the inductor was able to impress enough 2nd harmonic onto the system (through a longish speaker cable back to the amplifier) to show up on the analyzer plot.

Here's the current schematic --
schematic%207%20June.png


If I had been smart enough to use stuff in my own simulator program, I would have discovered that:
1) With the woofers connected, the current through L1 peaks at ~450Hz.
2) With the woofers disconnected, the current peak moves down to ... 350Hz. And also the system impedance dips below an ohm... oops, rookie move, can't just remove the driver and assume the crossover won't suck current.

So, next move is to either clamp the laminations with something, or get some different inductors (2mH, not a small inductor or something I have a lot of around), put the speaker and crossover back together and make sure there aren't any new surprise distortion peaks.

to be continued....
 
Last edited:
Any ideas on how to fashion a non-ferrous clamp on the inductor laminations? Or maybe dip the whole thing in lacquer?

The alternative is to go with one of their "Super Q" inductors (for near 3x the price). Probably should replace L4, too. Too much inductance needed to go with an air core coil -- by that time, an active crossover starts to look really good.

BTW, it occurs to me that this little adventure shows that there may be something to "bi-wiring". Probably if the mid and woofer circuits weren't sharing the same milliohms cable resistance drop, the inductor distortion wouldn't have been able to get into the other drivers....
 
Great detective work. You heard about my frustrations with months of a wierd 50Hz bass peak that turned out to be a hidden SRS sound processor?

Pretty poor detective work, actually. Should have been more methodical, would've saved time.

I'm wondering whether a secret SRS might be what messed up my "near the wall" measurements earlier. I was playing test tones through a TV system and media player then because I was too lazy to drag my other stuff upstairs! Pretty sure the TV has SRS, would they have been dumb enough to have it affect the line outputs as well? (The EQ and balance controls don't affect the line outputs though, so I just assumed...)
 
Any ideas on how to fashion a non-ferrous clamp on the inductor laminations? Or maybe dip the whole thing in lacquer?
I have had luck quieting noisy inductors with a tightly stretched wrapping of PVC electrical tape around the core.
Of course, the best solution is dipping in wax or varnish.

Thanks for sharing this unexpected source of distortion...something new to keep an eye out for.
It reminded me of the surprisingly detrimental effect a single twist tie can have on an air core inductor.
http://www.diyaudio.com/forums/multi-way/239908-distortion-mtm-build-5.html#post3584890
 
Hmm. I was just thinking about trying plastic cable wire ties to snug all the laminations together (I like the price, hope it works). Tomorrow..

Varnish seems like it should be good, but I think it might be difficult to get it to go between the laminations -- I think I read that high end transformer makers apply varnish in vacuum to pull the air out. Maybe I'll get some varnish and do both varnish dip and cable ties, can't hurt.
 
Last edited:
It reminded me of the surprisingly detrimental effect a single twist tie can have on an air core inductor.
http://www.diyaudio.com/forums/multi-way/239908-distortion-mtm-build-5.html#post3584890

Thanks for that link, I hadn't seen that! Makes sense.

So, if you want 3rd harmonic distortion, use a wire twist-tie around the core. If you want 2nd harmonic, have the laminations be slack.

Maybe this can be the next audiiophile tweak -- think I can get ;philes to buy $100 per pack audiophile cable tie clamps to put around the cores in their speaker crossovers?
 
..Varnish seems like it should be good, but I think it might be difficult to get it to go between the laminations -- I think I read that high end transformer makers apply varnish in vacuum to pull the air out.
Yeah, the wax and varnish used is low viscosity so it will wick up in between the laminations.
Vacuum is used if you need absolutely all air pockets removed; a must for HV applications.

Hmmmm....I wonder if some thin CA glue would work for bonding the laminations together thru wicking action?