BillFitzmaurice.info

I've been gathering piezo tweeters for a couple years, throwing them into orders to round up for free shipping etc.

Today I gathered them all up to test them before putting them in a couple melded arrays. Went through about 60-70. I started with the pink noise, got rapidly bored with that. So I brought out my LC meter and checked capacitance on each tweet . My sonic memory isn't as good as sorting by numbers.

Interesting empirical results. I found a continuity from roughly 0.080 uF up to 0.320 uf. The majority ran between 0.130 and 0.150.
The higher the capacitance, the brighter (louder???) the tweeter sounded both with pink noise and music source. The 300 group sounded pretty shrill and raspy (10-12). The 100 or less group were the obvious defectives (about eight). I wound up with a 4th group 200-250 (about 15) that were brighter than the 130-150 group (about 35 tweeters)

I'm going to build my arrays out of the 130-150 group. Just thought I'd throw this out there for comment.

Thank you, sir.

Or is it madam? Lucy?

Interesting, never tried that. Might be faster than the white noise I use. Although I have gotten to the point I identify them in about a second.

LelandCrooks wrote:Interesting, never tried that. Might be faster than the white noise I use. Although I have gotten to the point I identify them in about a second.

As is always the case if can be heard it can be measured. I've never tried measuring the capacitance, but it would be tantamount to measuring the impedance of a voice coil, so it's a logical test method. The only problem with it is that few builders would have a capacitance meter.

They're cheap enough.

True that most won't have an LC meter. Don't know that this is "validated" till I actually build the array. I'm sure that the noise test is certainly workable and fine for most users.

What surprised me was the spread of measured values. In an array, I suspect minor differences will be averaged out. I was hoping for something that was a bit more clear cut, "good/bad" test. Since piezos are basically a capacitive load, the method made sense to try. Again strictly empirical, one set of ears. But higher capacitance values were definitely brighter (or louder) than those with lower values.

As for my name I am definitely a fat old man!
However my avatar/bloodhound, Carmella, is a petite female at 110 lbs. I could see how you might get confused.

Greg Plouvier wrote:They're cheap enough.

These days, quite a lot of DMM-s have a capacitance measurement function. Sadly, very few have inductance measurement.

I build a lot of home audio speakers, and I bought mine several years ago for about $40. Just does Inductance & Capacitance (I have a couple cheapie multimeters for everything else) I typically use it more for inductance. I buy half a dozen big iron core coils then unwind them to whatever I need. I hand wind small coils too. Probably saved $40 in shipping charges over the years in shipping 'cause I didn't need to order a bunch of small parts.

The software trick can be used, with a few cheap and readily available components the only essential being one non-inductive resistor of known value. The rest is just for tying wires together without having to look at schematics.

http://billfitzmaurice.info/forum/viewt ... 93#p170793

There are a bunch of ways to measure components with a computer. I have a Woofer Tester 3, and a complete CAD/Measurement rig.
I like the meter because it's portable and easy.

I was kind of hoping somebody with a practiced ear (Leland, Bill,...?) who has tested a lot of these might confirm the validity of my findings independently. If it works out that this method is viable, it might help some other builders. Trying to go though 40-50 convinced me I didn't want to do it for a living!

LouC wrote:
I was kind of hoping somebody with a practiced ear (Leland, Bill,...?) who has tested a lot of these might confirm the validity of my findings independently.

Right now you're the lead dog. I don't have enough piezos on hand to do a decent job of testing them. Perhaps Leland will get a chance, as he tests his by ear already, and it would be a useful method to weed out the duds.

Getting close to an order. I buy them 500 at a pop, so I can get a pretty good sample to work from to confirm. Not retesting what I have in stock, as you say, it ain't that much fun.

LelandCrooks wrote:Getting close to an order. I buy them 500 at a pop, so I can get a pretty good sample to work from to confirm. Not retesting what I have in stock, as you say, it ain't that much fun.

When I'm able to what few I have I'll do an SPL sweep on and then test capacitance to see if there's a correlation between capacitance and response. Logically there would be.

Did anyone ever close the loop on this? Would it maybe be in another thread. Interested to see what was uncovered.

I didn't, and I believe that Lou C has passed on.