Frequency discrepancy from measured Fs to REW impedance peak

[4R5]

Hi,

New member here, but not new to audio or electronics.

I would like to thank John Mulcahy for developing REW! It's a far cry from back in the late 70’s when I helped my dad tune a bass reflex with a Heathkit sine/square wave generator, a Heathkit VTVM and the July ‘68 Popular Electronics article “ Tune Up Your Bass Reflex” by David Weems.

I’m seeing a difference in the Fs as I measure a woofer, vs. REW’s impedance peak. Here are 5 different woofers (4 are the same model):

Woofer	My Measurement (Fs)	REW Impedance peak
Seas 25TV-EW (15 ohm) 10” #1	30 Hz	34.9 Hz
Seas 25TV-EW (15 ohm) 10” #2	32 Hz	36.0 Hz
Seas 25TV-EW (15 ohm) 10” #3	30 Hz	36.3 Hz
Seas 25TV-EW (15 ohm) 10” #4	33 Hz	37.8 Hz
Stromberg Carlson RW489 12” 16 ohm	30 Hz	33.4 Hz

My measurement consists of a Wavetek generator with a 50 ohm output impedance, directly driving the woofer, with the DVM measuring the voltage. The generator has sufficient drive to provide 1.5 volts to the woofer.
I’ve verified the frequency with a Fluke DVM and a scope and also verified the waveform is not distorted.

For REW I’m using the Laptop’s internal sound card, stereo outputs and stereo inputs, all “effects” turned off. I’ve checked to see the sample rate is the same for input & output. Calibration went fine, my sense resistor is 105.1 ohms, my reference resistor is 23.1 ohms.

Am stuck – looking for reasons why I see the frequency shift. My readings are close to the published specs (20 - 25 Hz for the Seas, 25 +/- 5 Hz for the Stromberg Carlson).
I’m not concerned about actual impedance values at this point, only the frequencies.

Thanks, Charlie

 

[John Mulcahy]

Seems a bit suspicious that the differences are around 10 %, like 44.1k vs 48k. Probably best to attach some example impedance measurement mdat files to see if there are any clues there. You can also use the REW signal generator and spectrum analyser to compare the frequencies it generates and displays to those you measure with your DVM.

 

[4R5]

Hi John,
Appreciate the response.

Am attaching 2 Z files, one for the Seas #4, the other for the Stromberg Carlson.
I changed the settings on both the PC and REW to 48kHz, no change.
Verified the frequency of the REW signal generator is as indicated (30hz measures 29.99 Hz on my DVM)
Using REWs signal generator with the Z cabling (i.e. 105.1 ohm series resistor) to the woofers and the DVM across the woofer terminals, I measured the peak voltage and recorded the frequency. Using the table I made above, here is the result:

Woofer	My Measurement (Fs)	REW Impedance peak	FS using REW signal generator
Seas 25TV-EW (15 ohm) 10” #1	30 Hz	34.9 Hz	31.5 Hz
Seas 25TV-EW (15 ohm) 10” #2	32 Hz	36.0 Hz	32 Hz
Seas 25TV-EW (15 ohm) 10” #3	30 Hz	36.3 Hz	32.5 Hz
Seas 25TV-EW (15 ohm) 10” #4	33 Hz	37.8 Hz	33 Hz
Stromberg Carlson RW489 12” 16 ohm	30 Hz	33.4 Hz	32.5 Hz

The results are spot on for 2, the other 3 are pretty close to my original measurements.

Thanks, Charlie

 

[John Mulcahy]

You haven't selected the input device or input, best to do that. It is preferable to use the WASAPI exclusive entries for both input and output (name starts with EXCL) to ensure the sample rate used is the one selected in REW rather than have the OS resample, otherwise you need to check the Windows audio properties to make sure the sample rate in the default format is the same as the rate selected in REW. Longer sweeps are better. The driver behaviour will change with drive voltage, how much depends mainly on the surround and its compliance characteristics. TS parameters are for small signal, typically the terminal voltage at the driver would be around 100 mV.

 

[4R5]

Hi John,
I thought I had selected the input device, checked and if I try to select either EXCL input or output, I get errors (attached screen shots).
I made sure both the OS (Win 7) and REW ins & outs were all set to 44.1kHz and 16 bit stereo. No change.
I also slowed the sweep speed down to 23.8 seconds, no change
I'll try 100 mV and see what Fs is.
I did a search and could not find any other posts about this issue; I'm suspecting it's a issue on my end, is it a driver issue on my PC? I'm using an older Dell (M4300) because its the only one that has a stereo input.
Thanks, Charlie

 

[John Mulcahy]

Try setting everything to 48 kHz.

 

[4R5]

Tried that earlier, no change.

 

[4R5]

Hi John,
Problem solved!
It was not the OS, not REW, not the sound card, not the sample rate, not sweep speed.
It was the drive voltage. You touched on that in your 2nd response. I assumed that drive voltage would not matter and figured 1V would still be considered small signal. Not true
Here's a chart of Seas #4 with increasing drive voltages:

Drive Voltage	Fs
100mV	37Hz
200mV	37Hz
300mV	36Hz
400mV	36Hz
500mV	35Hz
600mV	35Hz
700mV	34Hz
800mV	34Hz
900mV	33Hz

1V	33Hz
1.5V	31Hz
2V	30Hz
2.5V	29Hz

The 30Hz Fs at 2V is close to Seas specs on this woofer (26HZ).
Interestingly, Seas, on their web site states they measure Fs at 2 V rms, but they do not state how they measure TS parameters.
I'm also assuming that for drivers like the Stromberg Carlson which was built in 1959, the designers back then probably used a drive voltage in the 1 to 3 (or more) range.

Thanks, Charlie

Not a REW question, but if the Fs changes with drive, and with a bass reflex design, do you design for large signal Fs?

 

[John Mulcahy]

Glad you resolved it.

if the Fs changes with drive, and with a bass reflex design, do you design for large signal Fs?

Can't help with that, sorry.