Measured Fs much higher than spec

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Okay, I finally got around to rebuilding a pair of Vifa P17/D25 speakers that I've had for a while, but never really been too happy with the sound. I originally pulled the drivers out of a second-hand Jaycar JV60 kit.

Measuring the response in Speaker Workshop seems to suggest that the T/S parameters are pretty far out-of-spec, particularly the free-air resonance for the woofer. I know it's typical to see some variation, but I'm getting close to 100 Hz when it should be 37 Hz (strangely it's happening on all three of the woofers - they all seem to be defective by the same amount).

They're probably about 10 years old at this point - is it typical for the parameters to degrade by that much over time? Or even to vary that much out of the factory?

I guess it's possible I'm doing something wrong in speaker workshop, but my impedance jig works for other components and I've tried a few different mounting methods with the same result.

Any suggestions, or are these woofers rubbish?
 

Attachments

  • graph.png
    graph.png
    7.9 KB · Views: 216

GM

Member
Joined 2003
That's way more than I've ever seen. Indeed, the worst cases I'm aware of aren't nearly that bad even with drivers where the surround is just a rolled out/pleated edge of the diaphragm. These can become hard enough to split with minimal motion.

Spiders OTOH once loosened up usually stay fairly flexible once their starch is heated/flexed out of them, although I've seen some so dried out they cracked too.

Anyway, try measuring while heating them up with a hair dryer or better yet, measure them outdoors in a warm/hot noonday sun. If there's no dramatic drop, then WYSIWYG......... Ditto if there's a drop and they rise back up once cooled off.

GM
 
Hmm... strange.

Don't have another driver at the moment but for my next project (once I decide what that is) I'll definitely measure the drivers first. If they're also significantly off, it's probably my method that's off. I could also try to loosen them up, but after years of use I'd assumed they'd reached their final state by now.

Otherwise it's a little puzzling why they'd all be off by the same amount if it's rare for even one to be out by that much to begin with.

Thanks, guys.
 
Or you could be measuring on a different point in the T/S curves than the factory -- althou with that large a difference i'd suspect something is not quite right. With 3 drivers so closely matched i'd tend to suspect the measuring kit 1st.

dave

You mean a different signal level? I'm just using the line-level output and input from my computer's sound card. I wouldn't have thought it would make much of a difference since the T/S parameters are small signal anyway, but I guess I could try it with an amplifier.

For what it's worth, the tweeters are right on the money - Fs is within 10% of the 850 Hz spec for each driver. So if there is a problem with the setup, it's only manifesting itself at lower frequencies.

Edit: I've also tried a frequency sweep with oscilloscope software (rather than what seems like an analysis of white noise with speaker workshop). Same response... although I wasn't expecting any difference.
 
Last edited:
Just another Moderator
Joined 2003
Paid Member
Looking at your graph if anything I'd say your drive level is too high.. all that hash above 1Khz doesn't look great. I'd say either your card is overdriving, or perhaps isn't up to the task.

This is how I would think your impedance plot should look --> http://www.diyaudio.com/forums/showpost.php?p=563617&postcount=14

I use an audigy 2ZX sound card. It took quite a bit of experimenting and tweaking to get good results like this, the levels were very important. I think I got better results driving directly off the soundcard, rather than through an amp, but not all sound cards can drive a speaker with their line level outputs, so you might need to use an amp.

Tony.
 
Im pretty sure this is your soundcards fault.
Ive got similar results before I connected an amp for measuring impedance.
My M-Audio sc couldnt handle low impedance at low frequencies.

It isnt difficult to check, the easiest way is to play sinus tones from 100hz and down and just look for the cone movement, it is highest near the Fs.

Or even better, connect an amp, play the same sinustones while you measure the current through the speaker with a multimeter.
 
Okay, I tried an amplifier and a few different mounting combinations (off the table, held in free space, on top of posts, etc...) and double-checked the results against my subwoofer driver (an Adire Shiva).

Same results for the Vifa's, but the Shiva is very close to spec (Fs is 22 Hz, spec'd at 23).

It's looking likely that the Vifa's actually are out-of-spec. Maybe I just got a bad batch... which is annoying, it would cost about $500 to replace them.
 
OK, could Jaycar have bought some cheap 'B' stock Vifas that didn't meet spec and/or had them whip up some compromised variants to keep cost down?

GM

I suppose it's possible they used 'B' stock units, and seeing as I got them second-hand there's another unknown there (what on earth the original owner could have done to them I don't know)... but they certainly weren't engineered for the off-spec Fs. If anything they were tuned slightly lower than usual.

On the other hand it was my first DIY project and I was never really happy with the speakers... so at least I'm more confident it wasn't completely my fault now.
 
Before giving up on them, try running them in. I apply a low frequency say 30Hz and as much power as to give almost the Xmax for 24 hours, before measuring the T/S parameters
With old speakers they can degrade so that one should be suspicious of 'new old stock' even tweeters.

Thanks, I might try that.

OK, back to square one, what's the average set of complete specs?

GM

Okay, the specs for the four drivers are as follows:

Fs: 96, 108, 84,79
Qts: 0.68, 0.77, 0.58, 0.59
Qms: 1.68, 1.68, 1.47, 1.57
Qes: 1.13, 1.4, 0.95, 0.93
Le: 0.45, 0.43, 0.46, 0.47
L1: 0.47, 0.52, 0.61, 0.54
R1: 3.17, 3.36, 3.29, 2.98
Re: 5.72, 5.93, 5.72, 5.73

Haven't been able to take Vas measurements yet, though.
 
Just another Moderator
Joined 2003
Paid Member
good that you had another driver to do a sanity check against, at least you know now that your measurement setup is working properly! Bit of a bummer about the vifa's though! if you push the cones in does it feel like they are binding? Maybe the coils are scraping the poll peice? The only other thing I can think of is that they have been cooked, perhaps the voice coils have been somewhat melted.. not sure what effect that would have on FS though... probably the most noticable effect would be rather a lot of distortion.

Tony.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.