A loose connection can be madenning


I put together my desktop speakers kind of in a hurry. Low on money and needing to move out of my apartment in the very immediate future. They are 2-way speakers with traditional crossovers. Now one of them has gone mostly silent and troubleshooting it with the right tools reminds me of how hard it must be fore those who don’t have them.

I’m using bot a digital multi-meter (DMM) and a speaker impedance analyzer, DATS V2 from Dayton/Parts Express. I think they are now in V3. Step one, measure the DC resistance. 20 Ohms. Oy. OK. Take off the front panel and the woofer connection is loose. Fix it, reassemble. 9 Ohms. Still too high. Run DATS. Sure enough, the impedance curve from 2 kHz down is around 20 Ohms. Something else is wrong. Disassemble and go through all of the connections in the bass crossover one at a time.

For anyone who has to do any trouble shooting inside a speaker, let alone replace parts, DATS is indispensable. Cheap around $130 far cheaper than some power cables!

Can’t say the process is fun right now though.
erik_squires
Main right woofer was crackling like the mag gap was rubbing. 
Then it went silent.  Pulled it out and the spring loaded lead grabber was lose.  Tightened it up with some thread lock.  All good. 

Fun.
Disassembled the speaker.  The drivers and crossovers are out and on my workbench.  Tested the woofer circuit alone.  It works fine, measures exactly as it should.  The woofer's resonant peak is there, followed by a flat impedance until 1 kHz where the impedance rises, at it should, since this is a low pass filter.

What does this tell me?  The connection problem was undoubtedly in the mating of the crossover to the banana plugs which felt pretty loose, honestly.  I'm remaking them.
Everything has been rewired.  Man, I really was in a hurry, since there are wires in here I don't even know what they were for now. :-)

Tested the impedance curve again from the banana plugs, but front baffle is still not on.  Curve looks perfect, except missing the obligatory 2nd resonant hump in the bass for a bass reflex section.  That will come when I put the baffle back on.  For now I have the correct and expected impedance curve. Now I have to hope nothing comes apart when I put it all in the cabinet.
 Put the baffle back on.  Measured.  Wait what happened to my dual humps?  The peaks and the valley between them are not nearly pronounced enough.  Oh right need to actually tighten the baffle first. :)

There they are!

Why am I taking so many measurements? Because it is easy and it ensures I know when something gets disconnected. Don't want to reassemble the speaker only to find out that I've already broken something else.
Can I just say, yet again, how lucky we hobbyists are to be living in a century where speaker impedance measurement tools are affordable and plug into our USB ports??

Back in the 50’s you would have had to use a hand adjusted frequency source, an amplifier, a scope and resistor and to the math yourself. You didn’t even have a spreadsheet to help you. Then we got to automatic paper plotters which were far too expensive for the average lay person to have lying around. And if you were using a tube amplifier you had to keep an eye on the output voltage too... it is no wonder that companies like Infinity put out speakers with terrible impedance signatures. Optimizing for frequency response and impedance all at once with so many drivers was a real nightmare. It is so much easier for any speaker designer today.  Also, there is absolutely no excuse for designers to do this today.

I also want to say, if you are a hobbyist who wants to do something inside a speaker you 100% need to invest in an impedance analyzer. They are cheaper than a lot of capacitors and coils you are going to try to spend money on, and well worth it.