Impedance testing

[AntonZ]

(not sure what section this should go, so mods feel free to move it to any more appropriate place)

I intend to upgrade my pair of Omni 10.5s from 3 piezo's each to V12 melded arrays. The array will go Omnitop style, in the mouth. Today I have been toying around with a melded array positioned in front of one of the cabs (internal pizzers switched off), comparing to the other. The melded 12 array was louder by a very significant margin, for obvious reasons. But no matter the amount of EQ applied, I could not get the two cabs to sound even remotely similar. Even the DEQ auto-eq could not get one anywhere near the other. After spending quite some time on that, I switched the original pizzers back on, comparing both cabs. They still sounded quite different. That was surprising, they are identical. One of the things that could be wrong is an air leak. This made me think of doing an impedance chart. I have no automated equipment for that, so it would mean hooking up a software tone generator and two digital multimeters: one for current, one for voltage. Then writing down measurements for different frequencies. This can be rather time consuming, but on the other hand I find this somewhat intriguing. So to determine if this is feasible and not too annoying, there are two questions on my mind.

• Any minimum voltage required? If the procedure requires at least 15V, this could become a rather noisy project
• At what frequency steps should I take measurements? Per octave sounds like too big steps. What kind of steps should I take to get meaningful measurements without spending an entire day making funny noises and writing down endless streams of numbers?

I know there is a woofer tester device (from Parts Express I believe), but shipping costs to Europe are prohibitive and I'm not aware of a similarly affordable device available locally.

[Gregory East]

If a crossover is messed up, wouldn't that cause glitchy impedance sweep? I think I'd start with checking the individual peizos, move on to the crossovers and then swap the woofers.

[AntonZ]

My plan is to add the crossover together with the melded arrays. There is no cross-over in there just yet. Just a bank of 3 tweeters, switchable. So currently I can switch off the tweeters and do an impedance sweep of just the woofer section. Still not sure what voltage and what intervals.


Edit:
I just found this article
The Impedance Measuring FAQ

Measurements at 1/3 intervals over the freq range of the woofer section is enough to give a fairly accurate impedance chart. The freq range of the woofer range is some 5 octaves, so that's 15 measurements. Sounds feasible.

I will go with a rather low voltage in order to not loose my hearing as well as upset the neighbours during these sweeps.

[sbe]

You can order the WT3 here:

https://www.intertechnik.de/Lautspreche ... 6175,52567

[AntonZ]

Thanks for the link. Still quite expensive in Europe, $99 or €67 in the US. €99 in Germany, and then it still needs to be shipped to Holland. I will do some manual measurements first, then see if it's worth purchasing such a device. It's not something I need every day.

[Bill Fitzmaurice]

Thanks for the link. Still quite expensive in Europe, $99 or €67 in the US. €99 in Germany, and then it still needs to be shipped to Holland. I will do some manual measurements first, then see if it's worth purchasing such a device. It's not something I need every day.

Getting a useful impedance plot requires at least 100 separate points on the chart, otherwise it's very easy to miss something. Taking a sweep with a WT3 takes about 5 seconds.

[AntonZ]

Doing 100 measurements of voltage and current takes me a tad over 5s (quite a big tad). OK, I should probably look into ways of getting a woofer tester like device. Thanks.


Edit:
It seems readily available software and a handful of passive components can do impedance measurement through a pc soundcard. I may give that a try.

[AntonZ]

So tonight I took a few scrap components, threw together a Q&D jig, and took impedance plots of 6 cabs all in one evening and it cost me nothing but a little time.

Autotuba

Wedgehorn 8

Omni 10.5

[AntonZ]

It got interesting as I got to my OT12 cabs.

OT12 nr 1

OT12 nr 2, very similar

OT12 nr 3, somewhat different

OT12 nr 4, very different

I would think there is something not quite right with the number 4 cab. Could this be a leak of some sorts? Other than the handles/ports it should be pretty much airtight. Can something like a sloppy driver gasket show up like this in the impedance plots?

[Bill Fitzmaurice]

I would think there is something not quite right with the number 4 cab. Could this be a leak of some sorts? Can something like a sloppy driver gasket show up like this in the impedance plots?

Yes. The reduced and shifted lower peak indicates the cab is tuned too high, and either a leak or an over-sized port will do that. But as the upper peak is also shifted higher a driver gasket leak is the most likely source.

[Rune Bivrin]

I think there are a few of us who would be interested in your setup. Not that it would be an insurmountable task to duplicate, but it does look quite polished. That would be a useful addition to the test procedure for the cabs I've built so far (and the ones I will build in the future...)

[AntonZ]

Thanks, will look into the driver gasket on the number 4 cab.

Rune, I will post details about the setup tonight.

[AntonZ]

I don't think I have ever made anything simpler than this. A bit of wire, a few connectors and a resistor. I threw it together in a rather haphazzardish way, so I'm not going to make any pictures of my sloppy solder work and wires hanging loose. Instead, here's the schematic.

Schematic.gif (5.69 KiB) Viewed 2748 times

The program that I used is called LIMP. It comes with Arta (clickable).
Personal license is €79, commercial license €149. Screenshots were made from the evaluation version.

I think the same thing can be done with another program called "Speaker workshop" which is free. I haven't tried that one.

[AntonZ]

As I wrote in the opening post of this thread, I am currently working a bit on my pair of O10.5. So the cover on one of the pair was not tightened down today, because I'm working on the inside. Things like add piezo HP filter, a bit more egg crate foam, get wiring ready for melded array in the horn mouth.

Just to see what effect it has on the impedance curve when the back is not properly sealed, I tested both tonight. Both cabs face down but up from the floor (sitting face down across two other cabs).

One of the pair is properly sealed, screws on the back all tight. The impedance plot looks like this:

On the other one the back is not tightened down. Just gravity pushing down on the back.

As soon as I put a bit of my weight on the back of the last cab, the plot cleans up and starts to look just like its' sibling. Useful stuff, and I will look into the #3 and #4 OT12 cabs. Those will be opened, driver gaskets checked, driver screws either tightened or replaced by bigger size if they have torn from the wood. Going to report back after that (may take a while).

[Dave Non-Zero]

This is an incredibly interesting and useful thread. The ability to detect leaks with impedence plots and the plausibility of using a cheap diy rig to measure it makes me seriously contemplate experimenting myself.