Side chain limiting for excursion and excursion measurment

Member
Joined 2007
Paid Member
Interesting article:
Using Limiters to Help Protect Loudspeakers
the demonstrated idea is to use side chain limiting to ensure the driver is kept in the linear excursion region without using low pass filtering. In a PA application I could see this as useful in lower SPL situations where you might want to EQ in more bass. At higher SPL situations you would want to still use a low pass filter as the limiter would cause an increase in harmonics if hit with a too big low frequency signal that could sound muddy. Speculation on my part though, I haven't tried this yet.

The acelerometer used to generate the excursion-frequency plot is no longer available. I have used Sharp GP2Y0A51SK0F to measure driver excursion before successfully:
measuring excursion - Bass Gear - Data-Bass Forums
The problem with this is the low sampling rate would make measurement of higher frequencies take a long time (operating like a sampling oscilloscope). This causes driver voice coil heating and power compression, making the results inaccurate. Furthermore its output would only be compatible with sine wave testing and no chirp testing.

I also know that the bass horn in the 'real horns' system has real time excursion measurement but this is with an expensive laser based system. Any ideas of a sensor that could work for this application?
 
Interesting article:
Using Limiters to Help Protect Loudspeakers
the demonstrated idea is to use side chain limiting to ensure the driver is kept in the linear excursion region without using low pass filtering. In a PA application I could see this as useful in lower SPL situations where you might want to EQ in more bass. At higher SPL situations you would want to still use a low pass filter as the limiter would cause an increase in harmonics if hit with a too big low frequency signal that could sound muddy. Speculation on my part though, I haven't tried this yet.

Hi kipman, I guess you meant to say high-pass filter. :)

I tried this idea for a 18FH500 in a reflex box.

The box had excessive excursion at 69v around 55Hz, well above f3 tuning at 40Hz.

hornresp excursion 69v.JPG


There was no need to lose the reponse at 40Hz, just a need to tame excursion higher up. So I built this sidechain limiter that responds to level.

sidechain smaart 18fh500 and pl340 Resize.jpg

The voltage chart is the peak voltage a PL340 would output with the limiter in place. The voltages tied to keeping excursion to 9mm across the freq range.
(So 80Vpk = 56Vrms = 9mm @ 55Hz))
 
Last edited:
Member
Joined 2007
Paid Member
Ah yes typo should be high pass.

thanks for showing your results; Low tuned bass reflex looks to be a good use case. The limiter is based off the Hornresp prediction in this case?

Its a bit frustrating that the sharp distance sensors with digital output have a faster update rate Sharp GP2Y0D805Z0F 390Hz! and 0.5 cm to 5cm distance. If this one had analog output it would probably be sufficient with some signal conditioning to plot excursion vs freq transfer functions.
 
Member
Joined 2007
Paid Member
So far the only non very expensive distance sensing option I have possibly found is:
HSDL-9100-021
which is just the IR led and photo diode portion of a detector. The LED and photo diode however are very fast with the photo diode response time listed as 6uS. The companion processing IC is only digital output however so I would need to supply my own processing circuits. These processing circuits would need to pulse the LED, detect the photo diode output and then perform a measurement without pulsing the LED to cancel the ambient light contribution. As the output current from the photo diode is non linear I think it would be best to process the output digitally.
 
Some thoughts for excursion measurement

There are analogue output inductive sensors like this:
https://uk.farnell.com/telemecanique-sensors/xs4p30ab120/inductive-proximity-sensor-15mm/dp/2847730
Would probably need a ferrous target though - could a small piece of steel be attached to the dust-cap?

Another thought would be ultrasonics. You could do this by sending multiple cycles to the speaker and sweep the ultrasonic pulses through the sine wave. So first pulse would be at 0V amplifier output, second pulse would be at 0.1V amplifier output, but on the second sine wave etc.

Or, how about hall effect? Glue a small rare-earth magnet to the dust cap and use a hall effect sensor to measure position. Only trouble is that it's magnetism could be swamped be the speaker magnet

Brian
 
Thinking about the ultrasonic idea - you wouldn't need to synchronise with the sine wave to the speaker (unless you wanted to automate it), you could just choose a slightly lower frequency

So, if you set 20Hz to the speaker and pulsed the ultrasonic at 19Hz, they would meet each other once per second and give you 19 readings of distance for one cycle of the speaker. If you pulsed the ultrasonic at 19.9Hz, it would take ten seconds and give you 199 samples. It's going to take a while to do enough frequencies to get a reasonable curve, but you are only doing it once

I guess that slightly higher frequency would work too

Brian
 
Bag End subs used a variable high pass filter to limit bass excursion. If the bass is too loud, the cutoff points moves up and reduces the bass without affecting higher frequencies. I guess the sidechain has a LPF that tracks the excursion of the woofer.
This is essential for the Bag End subs, which have a 12 dB/octave bass boost to get flat response from a small sealed box (all the way down to 8 Hz... at inaudible SPL). I don't see why it wouldn't be useful for conventional subs (or with a full-range system).