I just tried to measure the TS parameters of a loudspeaker. Everything was ok, until I had to measure two frequency points (Fl or F1 and Fh or F2, see http://www.epanorama.net/documents/audio/speaker_parameters.html) at which a certain calculated impedance had to occur. While finding these, I noticed strange impedance values, so I decided to put these into a graph (see attachment).
The impedance curve of the measured loudspeaker, in free air, has two maximum impedance points, in stead of one at the resonance frequency. I'm sure I followed the measurement instructions correctly.
The driver has 2 voice coils, they were wired in series when I measured. I already did some measurements some time ago, before the drivers were put in a bassreflex enclosure with only one coil connected. They had a quite low Qes and Qts values, around 0.25
Could someone tell me what the cause of this strange impedance curve could be?
regards
Ewin
The impedance curve of the measured loudspeaker, in free air, has two maximum impedance points, in stead of one at the resonance frequency. I'm sure I followed the measurement instructions correctly.
The driver has 2 voice coils, they were wired in series when I measured. I already did some measurements some time ago, before the drivers were put in a bassreflex enclosure with only one coil connected. They had a quite low Qes and Qts values, around 0.25
Could someone tell me what the cause of this strange impedance curve could be?
regards
Ewin
Attachments
/sreten.The shorting ring would not cause that. The best axiom has always been.."the more difficult the problem to solve..the more basic the mistake." (meaning the root source of the issue at hand is buried in fundamental questions, issues, or 'facts')
I would suggest looking at the whole thing, with fresh eyes, from the start, again.
I would suggest looking at the whole thing, with fresh eyes, from the start, again.
Is the driver mounted in a box or is it in free air? If it is mounted in a BR box the impedance will look similar to what you describe, but typically with a deeper dip between the two peaks.
Is there an extra mass load on the cone (like coins or something for measurement purposes)?
Are there any crossover components still connected to the drive?
What voltage do you use for the measurement, is it possible that the driver is overloaded?
Is there an extra mass load on the cone (like coins or something for measurement purposes)?
Are there any crossover components still connected to the drive?
What voltage do you use for the measurement, is it possible that the driver is overloaded?
sreten said:
I checked the connections several times; I'm sure the speaker was connected correctly. Also, I could feel the cone moving when frequencies were applied. I guess it wouldn't move when the coils were wired in anti-phase...
soongsc said:
As follows:
Amplifier +
Svante said:
KBK said:
sreten said:
I checked the connections several times; I'm sure the speaker was connected correctly. Also, I could feel the cone moving when frequencies were applied. I guess it wouldn't move when the coils were wired in anti-phase...
soongsc said:
As follows:
Amp [+]----- (1K resistor) ----- [+]* (VC1) [- ]------[+] (VC 2) [- ]*------- [- ] Amp
* = connected to multimeter
Svante said:
- This curve also made me think of a BR enclosure, but the measurement was taken in free-air.
- I didn't use any mass load.
- I only used a 1 KOhm resistor between the amplifier and the driver, no other crossover components
- The voltage across the calibration resistor (which is substituted for the speaker coils in the circuit above) was equal to it's resistance: Rcr = 8 Ohms and Vcs = 8 mV. I used this voltage level (voltage equal to resistance) in previous measurements, and the results were accurate. Nevertheless, would it be better if I used a higher voltage level?
KBK said:
I'll definitely do that.
Thank you all for your replies.
regards
Erwin
So, your system setup sure sounds alright. There is a slight possibility that the level was too low, with 1mV/ohm, but I would not think so.
Another explanation that came to my mind. Was the driver lying on a table or something that could flex a bit? It might be that the mass of the driver (magnet system and all) and the springiness of the table happened to resonance at somewhere in the range 30-80 Hz. That would give you an impedance somewhat like the bass-reflex box, I think.
Yet another thing: did you test the multimeter reading with the calibration resistor with different frequencies? I might be that the multimeter has a funny frequency response.
Another explanation that came to my mind. Was the driver lying on a table or something that could flex a bit? It might be that the mass of the driver (magnet system and all) and the springiness of the table happened to resonance at somewhere in the range 30-80 Hz. That would give you an impedance somewhat like the bass-reflex box, I think.
Yet another thing: did you test the multimeter reading with the calibration resistor with different frequencies? I might be that the multimeter has a funny frequency response.
To rule out the meter, substitute a 10 ohm resistor for the driver and see if it measures 10 ohms for all frequencies. Most inexpensive DMM's are quite nonflat at low frequenciies, but usually not complicated like you have there. What sort of Amp is it - many plate amps have a bass boost.....
When I had only a DMM, I measured with a 10 ohm resistor instead of the 1k resistor, taking a voltage measurement across the 10 ohm and then the driver at every point of measurement. Then you can calculate the impedance of the driver based on the voltage ratio. This corrects for meter/amp flatness issues.
Better to just build a jig for Speaker Workshop.
When I had only a DMM, I measured with a 10 ohm resistor instead of the 1k resistor, taking a voltage measurement across the 10 ohm and then the driver at every point of measurement. Then you can calculate the impedance of the driver based on the voltage ratio. This corrects for meter/amp flatness issues.
Better to just build a jig for Speaker Workshop.
Svante said:
- Yes, the driver was positioned between 2 stable tables. But I also did this in previous measurements, with the same multimeter, and the results were accurate.
- I didn't test the multimeter, I assumed it worked OK, as it did before. But the thing is cheap, so I'll check it's accuracy.
soongsc said:
I really don't know. The unit is an old Audax, a previous version of the G8 series I guess.
regards
Erwin
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.