Building 2-way array speakers and the tweeter lines don't work

[18Hurts]

Hello all and hope you are having a good weekend,

I'm building some 2-way array speakers for "garage" use--not an audiophile build by any means. My arrays are 6 feet tall and use twelve Sony 5" neodymium woofers and 48 Audax/JBL 10mm tweeters. The crossover is underlapped at 4050 Hz 12dB/Oct for the 5" woofers and 4,300 Hz 24dB/Oct for the tiny tweeters (Fs is 3000 Hz) There is no zobel network on the woofers and they willl get run at rather high power for BBQs, parties and electric guitar use.

Tested the build and the woofers work fine--the crossover rolls off just over 4 KHz as designed. The 48 tweeters measure 11.3 ohms and are wired 8 series and 6 parallel so that is expected. I have a pile of 15 watt power resistors wired before the tweeter crossover and it is set for 16 ohms. The power resistors can be re-wired by changing the wiring harness. The tweeter line is dead--no ouput.

I wired in some surround sound speakers running four 3" Tangband full ranges to the outputs on the crossover. I get all the treble as I would expect from 91 dB--1w/1m full range boxes. The wiring and crossover are fine but the tweeter line is still silent.

Ohming the tweeter line gives me an Re of 11.3 ohms or around 15 ohms or so active. I'm confused, if there was a bad tweeter, one of the 8 tweeter series lines would be out--but they are all silent. Even with the 16 ohm padding resistor inline with the surround sound speakers, they still put out treble but at a much lower volume.

How do you troubleshoot 48 tweeters that are wired 8 series/6 parallel with a 11.3 ohm Re? Can tweeters fail and still give a good ohm reading? When I tested them for impedance, they all worked fine at the 4 KHz test tone. Not a problem but it won't work with the crossover. I connect another speaker to the crossover and it works.

Now I'm confused--am I missing something? Any help is appreciated and enjoy your weekend!

[boydon_lepasci]

A picture would be appreciated in this case. Seeing how the drivers and crossover are wired and how they're connected to each other might help. How hard would it be to isolate each group of drivers? You could bypass the xo and apply a small signal directly to each seperated group.

[18Hurts]

Originally Posted by boydon_lepasci

A picture would be appreciated in this case. Seeing how the drivers and crossover are wired and how they're connected to each other might help. How hard would it be to isolate each group of drivers? You could bypass the xo and apply a small signal directly to each seperated group.

I just find it odd. I tested the 48 tweeter line for it's impedance to get the crossover designed correctly. At 4 KHz, they were at 12.6 ohms and very loud so I installed padding resistors before the crossover inputs. They can be set for multiple resistances and are mounted on an aluminum panel as a heat sink on the rear of the speaker to easily "tune" the tweeter line. The crossover does work as designed, it sends the highs to the at least 10 ohm impedance full range speaker I test with. The highs are there.

The crossover uses 1.5 uF and 3.0 uF caps which test good. They are coupled with 0.3 and 1.0 mH coils which add about .65 ohms to the circuit. Crossover setting is approx 4300 Hz at 24dB/Oct slope and it sees 13.25 ohms at 4,300 Hz. Works fine with other speakers and I ran it for two hours to make sure there was no intermittent issues.

The tweeters are wired 8 series and 6 parallel to raise their impedance to dcr 13 ohms or abouts. Today I will hook a laptop to the amplifier and send a test tone of around 6 KHz to the tweeter line to see if it is still operational. It was a few weeks ago when I did impedance testing.

This is what I find confusing--the crossover works properly when connected to another speaker. The 48 tweeter 8S/6P line used to work without a crossover during testing. It ohms out on the meter at around 13 ohms so an expected load. Even if I had one of the 158 soldered connections on the line fail, I would not expect to loose all 6 segments unless it went open or shorted. The expected resistance reading is there so I'm not getting 24 or 48 ohms that I would expect if I lost segments. I jumpered the padding resistors to run wide open and get no sound.

Can tweeters utterly fail without a change in resistance? I'm used to seeing them with open coils which is simple to spot. They are "rated" for 25 watts each so the 48 tweeter line is "rated" 1,200 watts--it should be very durable.

I'll test them with the XO bypassed with test tones. If they don't work, then it is time to take the parallel segments apart. Many zip ties, heat shrink and mounts for the "trunk" will need to be removed so yes--it is a pain. I assumed since the line worked without a crossover, I could seal it up properly while waiting for the crossover parts to arrive. Murphy's Law came into play with that time saving idea.

Just wondered if there was a "trick". A 1.5V battery is a great trick for woofers but not tweeters!

[Mr. dB]

Have you tested any of the tweeters individually?

[Michael Bean]

So if I understand correctly, the tweeter arrays produce sound when not connected to the crossover, but won't when they are connected to the crossover, and if you connect a test driver to the crossover you do get sound? Can you post a schematic and pictures?

Mike

[Loren42]

Are you measuring DC resistance or AC impedance?

[18Hurts]

OK, did some more testing and jumpers were flying, I did find an error for the second inductor in the forth order filter for the tweeter, it's supposed to be 1.7mH but has a 1.0mH. I'll order the correct inductor this week. (D'oh!) It should push the crossover filter for the tweeters higher, not lower so I won't have to worry about running them too low in frequency.

With only the tweeter line connected, 8 KHz test tone, no crossover in the circuit, the 48 tweeter line is fully operational and has output. All the tweeters have output which I verified with my noise dosimeter. Checked about 20 of them with the ohm meter and they all tested fine. The padding resistors work by lowering the output when connected and raising the output when bypassed with a jumper (shorted)

Connect the crossover and there is no output on the tweeter line. It is dead.

Move jumpers to connect my full range speaker to the outputs of the crossover. I get treble...it works very well. Ran the same 8 KHz test tone through the crossover and the full range connected to the XO had the 8 KHz tone blasting away with no distortion issues. The crossover works with the full range speakers for the highs. Ran music through it and both the woofer and full range "tweeters" work.

DCR readings for the woofer line is 4.9 ohms and 11.3 ohms for the tweeter line. Impedance of the tweeter line at 4 KHz is 12.6 ohms, woofer is 15.7 ohms and and those was the numbers I used to make the crossover. The padding resistors make up for the 10 dB different between the woofers/tweeter levels and are wired in before the tweeter crossover. It does not matter if the padding resistors are connected or bypassed--the tweeter line has no output when connected to the XO.

I think the problem is the crossover. Since I screwed up a calculation on one of the coils, I'm going to change it from 4th order to 3rd. Two caps and a coil instead of two caps and two coils. I'll go for around 4,400 Hz high pass at 18dB/Oct which should be high enough to protect the Fs of the tweeters at 3 KHz. Woofers will be wired for 4250 low pass at 12 dB/Oct.

Confusing? I emailed a guy that used to teach electronics and he replied with this "E = MC Hammer -- Can't Touch This. The only thing I can think of is to start over with the crossover again.

After all the work and hassle of building the monsters--maybe I should of gone with the 12" full range, phase plug mod and Fostex FT-17 super tweeter with just a cap! Oh well, can't learn anything that way. Back to the crossover drawing board I guess.

http://www.apicsllc.com/apics/Misc/filter2.html

I use this calculator for building crossovers, it worked for my center channel so I'm assuming it is correct more or less. Thanks for the help, sometimes pinging the problem off others gets my tired brain awake.

[boydon_lepasci]

If you even think you might make xo design a permanent part of your toolbox, I would buy Joe D'Appolito's book on speaker measuring and start reading over on the Parts Express forums about a free app call PCD(passive crossover designer). It is a powerful bit of code with great support.

[Michael Bean]

Just another thought, I recall you said you mounted the attenuation resistors on an aluminum panel for cooling. Is there any chance that there's a short circuit while the tweeters are connected that disappears when you physically move the leads to disconnected them to test with the other driver? Maybe that is what is causing the confusing situation. It might be worth a look.

Mike

[planet10]

Start simple. Build up the XO a component at a time. Start with just a cap. Then add a choke to ground. Then the 2nd cap. Then the last inductor. Then add the padding resistors.

dave