Distortion in MTM speaker build - AVS Forum
Forum Jump: 
 
Thread Tools
post #1 of 29 Old 08-01-2013, 04:03 PM - Thread Starter
Member
 
GaryM44's Avatar
 
Join Date: Feb 2012
Location: Spokane, WA
Posts: 17
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 10
I just finished building a pair of MTMs, using Zaph Audio's "Bargain Aluminum MTM" design.

This system uses Dayton DA175-8 woofers and the SEAS 27TBFC/G tweeter.
Crossover is at 1450Hz.

Testing with REW and a calibrated Dayton UMM-6 mic revealed high THD between 100 and 500Hz, tapering off above 500. Both speakers are similar. See plots (I included a shot of a measurement on another speaker I have, a Klipsch RF7-II, for comparison. It shows very low THD).

Notice that above the crossover point THD drops steeply. The tweeters, at least, seem to be on spec. Notice also that most of the distortion is odd-order. 2nd order is actually quite low.

I also tested the left speaker drivers separately, bypassing the crossover, the woofers together w/o the crossover, and the two crossover filters separately. The woofers show very low distortion when connected directly to the amp (< -50 db), and also when the tweeter is in the circuit via the high-pass filter, but the low-pass section is bypassed. It is clearly the low-pass circuit in the crossover causing the problem.

Anyone have any idea how that can happen?

Left speaker:



Right speaker:




Another speaker, using same test procedure, just for comparison:



Left speaker with tweeter connected via high-pass, but low-pass bypassed:



Left speaker woofers only, via low-pass:



Crossover schematic:



Crossover photo:



Suggestions welcome!
GaryM44 is offline  
Sponsored Links
Advertisement
 
post #2 of 29 Old 08-01-2013, 04:38 PM
AVS Special Member
 
Bill Fitzmaurice's Avatar
 
Join Date: Nov 2009
Location: New Hampshire
Posts: 9,283
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 1 Post(s)
Liked: 1195
If the distortion occurs with the LP filter and not without the probable culprit is the LP inductor. Assuming the drivers are parallel wired I'd not use a coil with .47 ohms DCR, 5% of the load impedance is the highest I'd tolerate. That coil's got a lot of wire in it, and that can lead to high THD, as well as a lot of insertion loss.

Bill Fitzmaurice Loudspeaker Design

The Laws of Physics aren't swayed by opinion.
Bill Fitzmaurice is offline  
post #3 of 29 Old 08-01-2013, 05:12 PM - Thread Starter
Member
 
GaryM44's Avatar
 
Join Date: Feb 2012
Location: Spokane, WA
Posts: 17
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 10
Yikes. The only 1.8 mH coil I could find with DCR < 0.2 ohm costs $65!
GaryM44 is offline  
post #4 of 29 Old 08-01-2013, 05:22 PM
AVS Special Member
 
tuxedocivic's Avatar
 
Join Date: Feb 2010
Location: Ladysmith, BC
Posts: 4,361
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 156 Post(s)
Liked: 433
How quite we're the measurement conditions and at what drive level did you drive the DUT.
tuxedocivic is online now  
post #5 of 29 Old 08-01-2013, 05:33 PM
Member
 
Mike-ht's Avatar
 
Join Date: Jun 2013
Posts: 52
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 11
Would a non air-core inductor be acceptable in the LP?
Mike-ht is offline  
post #6 of 29 Old 08-01-2013, 05:48 PM
Senior Member
 
djarchow's Avatar
 
Join Date: Mar 2005
Location: Iowa
Posts: 457
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 1 Post(s)
Liked: 25
Given your inductor placement, could this be an issue of crosstalk between the inductors? Can you run your distortion measurements on just the crossover or the individual components? I have done this before to measure XO transfer functions but have never tried it for distortion.
djarchow is offline  
post #7 of 29 Old 08-01-2013, 07:25 PM
AVS Special Member
 
Bill Fitzmaurice's Avatar
 
Join Date: Nov 2009
Location: New Hampshire
Posts: 9,283
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 1 Post(s)
Liked: 1195
Quote:
Originally Posted by GaryM44 View Post

Yikes. The only 1.8 mH coil I could find with DCR < 0.2 ohm costs $65!
Looking in the wrong places you must be.
http://www.parts-express.com/pe/showdetl.cfm?partnumber=255-106

Bill Fitzmaurice Loudspeaker Design

The Laws of Physics aren't swayed by opinion.
Bill Fitzmaurice is offline  
post #8 of 29 Old 08-01-2013, 07:27 PM
AVS Special Member
 
tuxedocivic's Avatar
 
Join Date: Feb 2010
Location: Ladysmith, BC
Posts: 4,361
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 156 Post(s)
Liked: 433
Seems like an awful lot of added distortion for an inductor to cause. Why not rule out, say, the HVAC coming on?
tuxedocivic is online now  
post #9 of 29 Old 08-02-2013, 09:46 AM - Thread Starter
Member
 
GaryM44's Avatar
 
Join Date: Feb 2012
Location: Spokane, WA
Posts: 17
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 10
Quote:
Originally Posted by Mike-ht View Post

Would a non air-core inductor be acceptable in the LP?

Sure. Are air cores more prone to distortion than other types?

I don't understand, theoretically, how either a cap or coil can introduce harmonic distortion in an AC signal. I'm wondering whether it may be coming from the amp, because that load may be putting some peculiar strain on the amp in that 100-500Hz band.
GaryM44 is offline  
post #10 of 29 Old 08-02-2013, 09:52 AM - Thread Starter
Member
 
GaryM44's Avatar
 
Join Date: Feb 2012
Location: Spokane, WA
Posts: 17
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 10
Quote:
Originally Posted by djarchow View Post

Given your inductor placement, could this be an issue of crosstalk between the inductors? Can you run your distortion measurements on just the crossover or the individual components? I have done this before to measure XO transfer functions but have never tried it for distortion.

Some else suggested that, so I removed the small coil and re-oriented it. Also moved it about 3 cm further from the large coil. THD sweep showed no difference, not even 1 db.
GaryM44 is offline  
post #11 of 29 Old 08-02-2013, 09:56 AM - Thread Starter
Member
 
GaryM44's Avatar
 
Join Date: Feb 2012
Location: Spokane, WA
Posts: 17
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 10
Quote:
Originally Posted by tuxedocivic View Post

Seems like an awful lot of added distortion for an inductor to cause. Why not rule out, say, the HVAC coming on?

It is consistent. The THD is consistently absent when the LP filter is out of the circuit, consistently present when it in the circuit. I do have a window AC unit in another room that might be on the same circuit. I'll make sure its off for next test.

Update: Shut off AC and also large ceiling fan in room. No difference in THD.
GaryM44 is offline  
post #12 of 29 Old 08-02-2013, 09:58 AM - Thread Starter
Member
 
GaryM44's Avatar
 
Join Date: Feb 2012
Location: Spokane, WA
Posts: 17
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 10
Quote:
Originally Posted by tuxedocivic View Post

How quite we're the measurement conditions and at what drive level did you drive the DUT.

As quiet as any average room can be. No noticeable noises. Measured THD at several different levels and mic gain settings.
GaryM44 is offline  
post #13 of 29 Old 08-02-2013, 10:01 AM - Thread Starter
Member
 
GaryM44's Avatar
 
Join Date: Feb 2012
Location: Spokane, WA
Posts: 17
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 10
Quote:
Originally Posted by Bill Fitzmaurice View Post

Looking in the wrong places you must be.
http://www.parts-express.com/pe/showdetl.cfm?partnumber=255-106

Much better! Thanks!

I'll try those if nothing else seems to work. Prob replace the caps too, with polys. But I don't understand, theoretically, how either of those could be responsible for this particular problem.
GaryM44 is offline  
post #14 of 29 Old 08-02-2013, 10:55 AM
Member
 
Mike-ht's Avatar
 
Join Date: Jun 2013
Posts: 52
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 11
Quote:
Originally Posted by GaryM44 View Post

Sure. Are air cores more prone to distortion than other types?

I don't understand, theoretically, how either a cap or coil can introduce harmonic distortion in an AC signal. I'm wondering whether it may be coming from the amp, because that load may be putting some peculiar strain on the amp in that 100-500Hz band.

No. Air core inductors are generally preferred, as they don't have an iron core that can interfere with the sound or saturate. However, the core increases inductance, so less wire is needed for the same value, resulting in a lower cost and lower DCR.
Mike-ht is offline  
post #15 of 29 Old 08-02-2013, 11:24 AM
AVS Special Member
 
Bill Fitzmaurice's Avatar
 
Join Date: Nov 2009
Location: New Hampshire
Posts: 9,283
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 1 Post(s)
Liked: 1195
Quote:
Originally Posted by Mike-ht View Post

Air core inductors are generally preferred, as they don't have an iron core that can interfere with the sound or saturate.
Cored inductors don't have any more tendency to saturate than air core when driven within their specified limits. To the contrary, they tend to have less of a problem than air core. I remember seeing extensive testing of various inductors a few years ago and cored Madisound Sledgehammers came out on top.

Bill Fitzmaurice Loudspeaker Design

The Laws of Physics aren't swayed by opinion.
Bill Fitzmaurice is offline  
post #16 of 29 Old 08-02-2013, 12:02 PM
Member
 
Mike-ht's Avatar
 
Join Date: Jun 2013
Posts: 52
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 11
Quote:
Originally Posted by Bill Fitzmaurice View Post

Cored inductors don't have any more tendency to saturate than air core when driven within their specified limits. To the contrary, they tend to have less of a problem than air core. I remember seeing extensive testing of various inductors a few years ago and cored Madisound Sledgehammers came out on top.

It is the magnetic saturation of the core which I was referring to, which cannot happen in an air core inductor, as air is not magnetic. Properly sized I agree saturation will not be a problem. There are tradeoffs with either, as with just about anything.
Mike-ht is offline  
post #17 of 29 Old 08-02-2013, 03:49 PM - Thread Starter
Member
 
GaryM44's Avatar
 
Join Date: Feb 2012
Location: Spokane, WA
Posts: 17
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 10
Well, just ordered the Jantzen p-core inductors (DCR 0.17 ohm) and 5% poly caps. If those don't work I'll be out of ideas.

Also tried a different amp. Same results.
GaryM44 is offline  
post #18 of 29 Old 08-02-2013, 07:26 PM
AVS Special Member
 
Bill Fitzmaurice's Avatar
 
Join Date: Nov 2009
Location: New Hampshire
Posts: 9,283
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 1 Post(s)
Liked: 1195
I just looked at the schematic of the crossover, and the component values don't compute. The value for C4 seems way too high, and for L8 way too low.

Bill Fitzmaurice Loudspeaker Design

The Laws of Physics aren't swayed by opinion.
Bill Fitzmaurice is offline  
post #19 of 29 Old 08-03-2013, 08:30 AM - Thread Starter
Member
 
GaryM44's Avatar
 
Join Date: Feb 2012
Location: Spokane, WA
Posts: 17
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 10
Quote:
Originally Posted by Bill Fitzmaurice View Post

I just looked at the schematic of the crossover, and the component values don't compute. The value for C4 seems way too high, and for L8 way too low.

But look at the plot above (left speaker, woofers only). The high rolloff begins just where it's supposed to, and the slope is right.

Also, if those values were wrong, it should only affect the pivot point and slope, not create THD. Right?

That's the mystery --- I don't know of *any* error in a crossover that can induce harmonic distortion.
GaryM44 is offline  
post #20 of 29 Old 08-03-2013, 08:51 AM
AVS Special Member
 
Bill Fitzmaurice's Avatar
 
Join Date: Nov 2009
Location: New Hampshire
Posts: 9,283
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 1 Post(s)
Liked: 1195
Quote:
Originally Posted by GaryM44 View Post

But look at the plot above (left speaker, woofers only). The high rolloff begins just where it's supposed to
Or does it? Something is happening much lower than that, and it appears that the tweeter is contributing well below where it should. I'd take that measurement woofer only, then with the coil only in place, then with the cap added, being sure not to change the mic position or drive level. With two woofers paralleled for a nominal impedance of 4 ohms even accounting for the voice coil inductance with a 1450Hz crossover and 1.8mH coil a cap value in the vicinity of 7 to 15mF would be expected, depending on the filter Q. With a 31uF cap the filter Q will be very high, and that could be the source of the distortion. I'd also do an impedance sweep, too large a cap value will also be revealed by the impedance going too low. That is what can cause the distortion, as the amp is being asked to give what it cannot cleanly.
The 0.3 mH coil in the high pass also results in a high Q for that filter, but the 7 ohm resistor keeps that from being obvious via clipping the amp as the load is still high enough.

Bill Fitzmaurice Loudspeaker Design

The Laws of Physics aren't swayed by opinion.
Bill Fitzmaurice is offline  
post #21 of 29 Old 08-03-2013, 10:26 AM - Thread Starter
Member
 
GaryM44's Avatar
 
Join Date: Feb 2012
Location: Spokane, WA
Posts: 17
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 10
Quote:
Originally Posted by Bill Fitzmaurice View Post

Or does it? Something is happening much lower than that, and it appears that the tweeter is contributing well below where it should.

Tweeter is out of the picture on the following sweeps, although the hi-pass filter is live on the last (but not on the first two).
Quote:
I'd take that measurement woofer only, then with the coil only in place, then with the cap added, being sure not to change the mic position or drive level.

Direct to woofers, no crossover:



Via lo-pass coil, no cap:



With both lo-pass coil and cap:



Looks like the coil is causing the prob. Note that the tweeter hi-pass filter is unpowered for the first two plots.
Quote:
I'd also do an impedance sweep, too large a cap value will also be revealed by the impedance going too low. That is what can cause the distortion, as the amp is being asked to give what it cannot cleanly.

That is my best guess also. Wanted to do an impedance sweep. but need a 100 ohm resistor, which I don't have. Have to get one, I suppose.

What could be wrong with that coil to cause that phenomenon? Though all these plots are for the left speaker, the right shows identical behavior. Hard to believe both 1.8mH coils could be bad.
GaryM44 is offline  
post #22 of 29 Old 08-03-2013, 12:29 PM
AVS Special Member
 
Bill Fitzmaurice's Avatar
 
Join Date: Nov 2009
Location: New Hampshire
Posts: 9,283
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 1 Post(s)
Liked: 1195
Quote:
Originally Posted by GaryM44 View Post

What could be wrong with that coil to cause that phenomenon? Though all these plots are for the left speaker, the right shows identical behavior. Hard to believe both 1.8mH coils could be bad.
Accounting for the voice coil inductance you'd expect to see a 1st order filter using a 1.8mH coil to have about a 600Hz. knee. It looks like yours has a 100Hz knee, which would indicate perhaps about 9mH, and that coil sure doesn't look to be that big. Something is major league wrong, that's for sure.

Bill Fitzmaurice Loudspeaker Design

The Laws of Physics aren't swayed by opinion.
Bill Fitzmaurice is offline  
post #23 of 29 Old 08-03-2013, 12:54 PM
Senior Member
 
djarchow's Avatar
 
Join Date: Mar 2005
Location: Iowa
Posts: 457
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 1 Post(s)
Liked: 25
Here is the design page including XO transfer functions etc.

http://zaphaudio.com/BAMTM.html
djarchow is offline  
post #24 of 29 Old 08-03-2013, 05:03 PM - Thread Starter
Member
 
GaryM44's Avatar
 
Join Date: Feb 2012
Location: Spokane, WA
Posts: 17
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 10
Quote:
Originally Posted by Bill Fitzmaurice View Post

Something is major league wrong, that's for sure.

Fixed!!

You won't believe it . . . .

I removed the 1.8mH coil from the board to isolate from everything else. Clipped it between the amp and woofers w/ alligator clips. No distortion.

So then I soldered the cap to the coil, and clipped both into the woofer circuit. Still no distortion.

Soldered them both back to the board, but not tied down, connected the woofers via the terminals on the board and tested again. Still no distortion.

Then I tied the coil down with the wire twist ties I'd been using (see photo in original post). Distortion is back!

Removed the twist ties, distortion is gone again. Removed the twist ties from the coils on the right speaker xover, tested it. No more distortion there, either.

LESSON: Don't secure coils to a PCB board with wire twist ties!

I assume the coil induces a field in the wire tie, which is then induced to the board trace where the tie passes through.

Anyway, THD is now about -47 db, or < 0.5%. And they sound great; very smooth. Plot below has no smoothing:



Thanks for help Bill (and everyone).
GaryM44 is offline  
post #25 of 29 Old 08-03-2013, 05:31 PM
Senior Member
 
NathanJ's Avatar
 
Join Date: Aug 2001
Location: Northern VA
Posts: 210
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 1 Post(s)
Liked: 15
Interesting, is the wire tie a zip tie?
NathanJ is offline  
post #26 of 29 Old 08-03-2013, 06:05 PM
Member
 
Mike-ht's Avatar
 
Join Date: Jun 2013
Posts: 52
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 11
No, it's a twist tie. They have a metal wire inside. Coils of wire are definitely problematic around inductors, as is any ferrite material. Plastic screwdrivers are used for adjustable inductors, for example, otherwise the values are thrown off while making the adjustment.
Mike-ht is offline  
post #27 of 29 Old 08-03-2013, 10:34 PM
AVS Special Member
 
tuxedocivic's Avatar
 
Join Date: Feb 2010
Location: Ladysmith, BC
Posts: 4,361
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 156 Post(s)
Liked: 433
Wow I'm very surprised by these results. I know screws and things can throw off inductor values, but that is really something. Thanks for sharing.

I use hot glue to secure components, most do I think. Consider doing that.
tuxedocivic is online now  
post #28 of 29 Old 08-04-2013, 06:24 AM - Thread Starter
Member
 
GaryM44's Avatar
 
Join Date: Feb 2012
Location: Spokane, WA
Posts: 17
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 10
Quote:
Originally Posted by tuxedocivic View Post

Wow I'm very surprised by these results. I know screws and things can throw off inductor values, but that is really something. Thanks for sharing.

I use hot glue to secure components, most do I think. Consider doing that.

Yah, I know. But that's a pain if you ever want to upgrade or replace components. Think I'll try nylon cable ties first.
GaryM44 is offline  
post #29 of 29 Old 08-04-2013, 07:50 AM
AVS Special Member
 
Bill Fitzmaurice's Avatar
 
Join Date: Nov 2009
Location: New Hampshire
Posts: 9,283
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 1 Post(s)
Liked: 1195
Quote:
Originally Posted by GaryM44 View Post

Yah, I know. But that's a pain if you ever want to upgrade or replace components. Think I'll try nylon cable ties first.
It never occurred to me that the ties might be a problem, as I've never seen anyone use metal core ties, only plastic.

Bill Fitzmaurice Loudspeaker Design

The Laws of Physics aren't swayed by opinion.
Bill Fitzmaurice is offline  
Reply DIY Speakers and Subs

User Tag List

Thread Tools
Show Printable Version Show Printable Version
Email this Page Email this Page


Forum Jump: 

Posting Rules  
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off